HANDBOOK OF LANGUAGE CONTACT. 9781119485025, 1119485029

Citation preview

The Handbook of Language Contact

Blackwell Handbooks in Linguistics This outstanding multi‐volume series covers all the major subdisciplines within linguistics today and, when complete, will offer a comprehensive survey of linguistics as a whole. The most recent publications in the series can be found below. To see the full list of titles available in the series, please visit www.wiley.com/go/linguistics-handbooks The Handbook of Pidgin and Creole Studies Edited by Silvia Kouwenberg & John Victor Singler

The Handbook of Language Emergence Edited by Brian MacWhinney & William O’Grady

The Handbook of Language Teaching Edited by Michael H. Long & Catherine J. Doughty

The Handbook of Discourse Analysis, Second Edition Edited by Deborah Tannen, Heidi E. Hamilton, & Deborah Schiffrin

The Handbook of Phonetic Sciences, Second Edition Edited by William J. Hardcastle & John Laver The Handbook of Language and Speech Disorders Edited by Jack S. Damico, Nicole Muller, & Martin J. Ball The Handbook of Computational Linguistics and Natural Language Processing Edited by Alexander Clark, Chris Fox, & Shalom Lappin

The Handbook of Korean Linguistics Edited by Lucien Brown & Jaehoon Yeon The Handbook of Speech Production Edited by Melissa A. Redford The Handbook of Classroom Discourse and Interaction Edited by Numa Markee

The Handbook of Language and Globalization Edited by Nikolas Coupland

The Handbook of Narrative Analysis Edited by Anna De Fina & Alexandra Georgakopoulou

The Handbook of Hispanic Sociolinguistics Edited by Manuel Diaz‐Campos

The Handbook of English Pronunciation Edited by Marnie Reed & John M. Levis

The Handbook of Language Socialization Edited by Alessandro Duranti, Elinor Ochs, & Bambi B. Schieffelin

The Handbook of Bilingual and Multilingual Education Edited by Wayne E. Wright, Sovicheth Boun, & Ofelia Garcia

The Handbook of Phonological Theory, Second Edition Edited by John A. Goldsmith, Jason Riggle, & Alan C. L. Yu

The Handbook of Contemporary Semantic Theory, Second Edition Edited by Shalom Lappin & Chris Fox

The Handbook of Intercultural Discourse and Communication Edited by Christina Bratt Paulston, Scott F. Kiesling, & Elizabeth S. Rangel

The Handbook of Portuguese Linguistics Edited by W. Leo Wetzels, Joao Costa, & Sergio Menuzzi

The Handbook of Historical Sociolinguistics Edited by Juan Manuel Hernandez‐Campoy & Juan Camilo Conde‐Silvestre The Handbook of Hispanic Linguistics Edited by Jose Ignacio Hualde, Antxon Olarrea, & Erin O’Rourke The Handbook of Conversation Analysis Edited by Jack Sidnell & Tanya Stivers The Handbook of English for Specific Purposes Edited by Brian Paltridge & Sue Starfield The Handbook of Bilingualism and Multilingualism, Second Edition Edited by Tej K. Bhatia & William C. Ritchie The Handbook of Language Variation and Change, Second Edition Edited by J. K. Chambers & Natalie Schilling

The Handbook of Linguistics, Second Edition Edited by Mark Aronoff & Janie Rees‐Miller The Handbook of Translation and Cognition Edited by John W. Schwieter & Aline Ferreira The Handbook of Technology and Second Language Teaching and Learning Edited by Carol A. Chapelle & Shannon Sauro The Handbook of Psycholinguistics Edited by Eva M. Fernandez & Helen Smith Cairns The Handbook of Dialectology Edited by Charles Boberg, John Nerbonne, & Dominic Watt The Handbook of Advanced Proficiency in Second Language Acquisition Edited by Paul A. Malovrh & Alessandro G. Benati The Handbook of the Neuroscience of Multilingualism Edited by John W. Schwieter

The Handbook of Spanish Second Language Acquisition Edited by Kimberly L. Geeslin

The Handbook of TESOL in K-12 Edited by Luciana C. de Oliveira

The Handbook of Chinese Linguistics Edited by C.‐T. James Huang, Y.‐H. Audrey Li, & Andrew Simpson

The Handbook of World Englishes, Second Edition Edited by Cecil L. Nelson, Zoya G. Proshina, & Daniel R. Davis

The Handbook of Language, Gender, and Sexuality, Second Edition Edited by Susan Ehrlich, Miriam Meyerhoff, & Janet Holmes

The Handbook of Informal Langauge Learning Edited by Mark Dressman & Randall William Sadler The Handbook of Language Contact, Second Edition Edited by Raymond Hickey

The Handbook of Language Contact Second Edition

Edited by

Raymond Hickey

This edition first published 2020 © 2020 John Wiley & Sons Ltd John Wiley & Sons Ltd (1e, 2010) All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Raymond Hickey to be identified as the author of this editorial material in this work has been asserted in accordance with law. Registered Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging‐in‐Publication Data Name: Hickey, Raymond, 1954– editor. Title: The handbook of language contact / Raymond Hickey. Description: Second edition. | Hoboken, NJ : Wiley-Blackwell, [2020] |   Series: Blackwell handbooks in linguistics | Includes bibliographical   references and index. Identifiers: LCCN 2020007524 (print) | LCCN 2020007525 (ebook) | ISBN   9781119485025 (hardback) | ISBN   9781119485063 (pdf) | ISBN   9781119485056 (epub) Subjects: LCSH: Languages in contact. Classification: LCC P130.5 .H36 2020 (print) | LCC P130.5 (ebook) | DDC  306.44–dc23 LC record available at https://lccn.loc.gov/2020007524 LC ebook record available at https://lccn.loc.gov/2020007525 Cover Design: Wiley Cover Image: © FRANZ MARC/AKG Images Set in 9.5/11.5pt Palatino by SPi Global, Pondicherry, India

10 9 8 7 6 5 4 3 2 1

Contents

Notes on Contributors ix Preface xvii

Language Contact and Linguistic Research Raymond Hickey

1

Part I – Contact, Contact Studies, and Linguistics

31

1 Contact Explanations in Linguistics Sarah Thomason

33

2 Contact, Bilingualism, and Diglossia Lotfi Sayahi

51

3 Language Contact and Change through Child First Language Acquisition Carmel O’Shannessy and Lucinda Davidson

67

4 Contact and Grammaticalization Bernd Heine and Tania Kuteva

93

5 Contact and Language Convergence Anthony P. Grant

113

6 Contact and Linguistic Typology Oliver Bond, Helen Sims‐Williams, and Matthew Baerman

129

7 Contact and Language Shift Raymond Hickey

149

8 Contact and Lexical Borrowing Philip Durkin

169

9 Contact and Code‐Switching Penelope Gardner‐Chloros

181

10 Contact and Mixed Languages Peter Bakker

201

11 Contact and Sociolinguistic Variation Maya Ravindranath Abtahian and Jonathan Kasstan

221

12 Contact and New Varieties Paul Kerswill

241

vi Contents 13 Contact in the City Heike Wiese

261

14 Linguistic Landscapes and Language Contact Kingsley Bolton, Werner Botha, and Siu‐Lun Lee

281

Part II – Case Studies of Contact

301

15 Contact and Early Indo‐European in Europe Bridget Drinka

303

16 Contact and the History of Germanic Languages Paul Roberge

323

17 Contact in the History of English Robert McColl Millar

345

18 Contact and the Development of American English Joseph C. Salmons and Thomas Purnell

361

19 Contact and African Englishes Rajend Mesthrie

385

20 Contact and Caribbean Creoles Edgar W. Schneider and Raymond Hickey

403

21 Contact and the Romance Languages John Charles Smith

425

22 Contact and Spanish in the Pacific Eeva Sippola

453

23 Contact and Portuguese‐Lexified Creoles Hugo C. Cardoso

469

24 Contact and the Celtic Languages Joseph F. Eska

489

25 Contact and the Slavic Languages Lenore A. Grenoble

501

26 Contact and the Finno‐Ugric Languages Johanna Laakso

519

27 Language Contact in the Balkans Brian D. Joseph

537

28 Turkic Language Contacts Lars Johanson, Éva Á. Csató, and Birsel Karakoç

551

29 Contact and Afroasiatic Languages Zygmunt Frajzyngier and Erin Shay

571

30 Contact and North American Languages Marianne Mithun

593

31 Contact and Mayan Languages Danny Law

613

Contents  vii 32 Contact and South American Languages Lyle Campbell, Thiago Chacon, and John Elliott

625

33 Contact among African Languages Klaus Beyer

649

34 Contact and Siberian Languages Brigitte Pakendorf

669

35 Language Contact: Sino‐Russian Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko

689

36 Language Contact and Australian Languages Jill Vaughan and Debbie Loakes

717

37 Contact Languages of the Pacific Jeff Siegel

741

Index 763

Notes on Contributors

Maya Ravindranath Abtahian is a sociolinguist who studies language variation and change in the context of language contact, maintenance, and shift. She is an Assistant Professor of Linguistics at the University of Rochester, USA, and received her Ph.D. in Linguistics from the University of Pennsylvania, where her dissertation focused on variation and contact in Garifuna (Belize). She also studies language shift in Indonesia, and language variation and change in the northern United States. Matthew Baerman is Principal Research Fellow in the Surrey Morphology Group at the University of Surrey, UK. He works on the typology of inflectional morphology, with a particular focus on structures that are theoretically or descriptively challenging. He is the co‐author (with Dunstan Brown and Greville Corbett) of Morphological Complexity (Cambridge University Press, 2017) and the editor of The Oxford Handbook of Inflection (Oxford University Press, 2015). Peter Bakker is Associate Professor in Linguistics at Aarhus University, Denmark. He specializes in new languages, including pidgins, creoles, and mixed languages, as well as twins who create their own languages. In his research, he combines typology, anthropology, and history in order to explain the genesis and structural properties of these languages. He is best known for his work on Michif (Cree‐French mixed language), intertwined varieties of Romani, and the typological characteristics of creoles and mixed languages. Klaus Beyer is a Research Associate in the Department of African Studies at Goethe University Frankfurt‐am‐Main, Germany. Trained in African linguistics, history, and anthropology, he worked on historical linguistics and language documentation in West Africa. His interest then broadened to the fields of multilingualism, language contact, and sociolinguistics. His current focus lies on the interplay of language repertoires and choices and their relations with speakers’ characteristics and social network factors in non‐­ standardized languages. Kingsley Bolton is Professor of English Linguistics at Nanyang Technological University, Singapore, and Professor Emeritus at Stockholm University, Sweden. He has published widely on English in the Asian region, language and globalization, sociolinguistics, and world Englishes. He is co‐editor (with Daniel R. Davis) of the journal World Englishes. Oliver Bond is Reader in Linguistics in the Surrey Morphology Group at the University of Surrey, UK. His research investigates the ways in which cross‐linguistic variation can form the empirical base for developing linguistic theory. He is the co‐editor, with Greville G. Corbett, Marina Chumakina, and Dunstan Brown, of Archi: Complexities of Agreement in

x  Notes on Contributors Cross‐Theoretical Perspective (Oxford University Press, 2016) and, with András Bárány and Irina Nikolaeva, of Prominent Internal Possessors (Oxford University Press, 2019). Werner Botha is Assistant Professor at Nanyang Technological University, Singapore. His  academic interests include the use of English in Asian higher education, educational ­linguistics, multilingualism, and language variation and sociolinguistics, with particular ­reference to the Asian region. Lyle Campbell is Professor (Emeritus) of the University of Hawai‘i, Manoa, USA. His specializations are: historical linguistics, American Indian languages, language documentation, and typology. He has published 23 books and about 200 articles, held appointments in Linguistics, Anthropology, Spanish, and Latin American Studies, and has taught in universities in nine different countries. He won the Linguistic Society of America’s “Leonard Bloomfield Book Award” twice. Hugo C. Cardoso is Assistant Professor at the University of Lisbon, Portugal. His research focuses on language contact involving Portuguese, having worked on Saramaccan, a ­creole of Suriname, but especially on the Portuguese‐based creoles of South Asia – those of India (in particular, Diu and the Malabar Coast) and Sri Lanka. He has also undertaken extensive documentation on endangered languages. His work combines a synchronic descriptive perspective with an interest in the history of these languages and in comparative approaches. Thiago Chacon is Assistant Professor at the University of Brasília, Brazil. He has been a research fellow at the Collegium de Lyon, the University of California in Santa Barbara, and has a Ph.D. from the University of Hawai‘i with a dissertation on the phonology and morphology of Kubeo, an Eastern Tukanoan language. He has done extensive comparative work with Tukanoan and Northwestern Amazonian languages, as well as other Amazonian languages, including the linguistic isolate Arutani and the Yanomaman language Ninam. His main theoretical interests are historical linguistics, phonology, typology, and language description and documentation. Éva Á. Csató is Professor Emeritus in Turkic Languages at the Department of Linguistics and Philology, Uppsala University, Sweden. She has worked on the typology of Turkic languages as well as the documentation of less‐studied and endangered Turkic languages. Together with Lars Johanson she edited the volume The Turkic Languages (1998). She has co‐edited several books on Turkic language contacts. She is a member of the editorial board of the journal Turkic Languages. Lucinda Davidson is a Postdoctoral Research Fellow in the School of Languages and Linguistics at the University of Melbourne, Australia, whose primary focus is children’s communicative competence. Working in two remote Aboriginal communities, her current research explores children’s development in the Australian languages, Murrinhpatha and Pitjantjatjara. Bridget Drinka is Professor of Linguistics, at the University of Texas at San Antonio. Her research focuses on such issues as the sociolinguistic motivations for language change, the role of contact in linguistic innovation, and the importance of geographical contiguity in the diffusion of changes across the Indo‐European languages. Her 2017 book, Language Contact in Europe: The Periphrastic Perfect through History (Cambridge University Press), explores the

Notes on Contributors  xi complex development of a grammatical category as it spread across the map of Europe. It was awarded the Leonard Bloomfield Book Award in 2019 by the Linguistic Society of America. As President of the International Society for Historical Linguistics, Drinka organized the International Conference on Historical Linguistics in 2017. Philip Durkin is a lexicographer, etymologist, and historian of English, with a particular interest in the history of lexical borrowing in English. For the past 25 years he has been on the staff of the Oxford English Dictionary, where he is currently Deputy Chief Editor and Principal Etymologist. His publications include The Oxford Guide to Etymology (Oxford University Press, 2009), Borrowed Words: A History of Loanwords in English (Oxford University Press, 2013), and (as editor) The Oxford Handbook of Lexicography (Oxford University Press, 2016). John Elliott is a Ph.D. candidate at the University of Hawai‘i at Manoa, USA, whose research focuses on the documentation and description of Enxet Sur, an Enlhet‐Enenlhet language of the Paraguayan Chaco region, including work on the historical linguistics of the Enlhet‐ Enenlhet family and the evidence of language contact between different indigenous groups in the Chaco region. Elliott has also researched and published work on indigenous language policy and endangerment in the Chaco. Joseph F. Eska is Professor of Linguistics at Virginia Polytechnic Institute and State University, USA. His research focuses upon the diachrony of the ancient and medieval Celtic languages. He is the editor of the North American Journal of Celtic Studies and co‐editor of Indo‐European Linguistics. Zygmunt Frajzyngier is Professor of Linguistics at the University of Colorado in Boulder, USA. His interests include the foundations of syntax and semantics in cross‐linguistic perspective; typological explanations in grammar; emergence of forms and functions in language; grammaticalization; Chadic and Afroasiatic linguistics; relations between grammatical structures and the lexicon; descriptive grammars and dictionaries. He is the author, co‐ author, and editor of 23 books, and over 130 papers. His current projects include: study of the emergence of functions; languages in the Afroasiatic phylum (with Erin Shay and Marielle Butters, University of Colorado), and a study of Sino‐Russian idiolects, with Natalia Gurian and Sergei Karpenko, of the Far Eastern Federal University, Vladivostok, Russian Federation. Penelope Gardner‐Chloros was an interpreter for the EU before studying linguistics and taking a Ph.D. at Strasbourg University (published as Language Selection and Switching in Strasbourg, Oxford University Press, 1991). She became a Professor at Birkbeck, University of London, UK, specializing in bilingualism and is the author of Code‐Switching (Cambridge University Press, 2009). She has also done research on youth language in Paris with Jenny Cheshire, published as a special issue of the Journal of French Language Studies in 2018. She is currently a Research Fellow at Wolfson College Oxford, investigating bilingual artists, notably El Greco. Anthony P. Grant has been Professor of Historical Linguistics and Language Contact at Edge Hill University, UK, since 2008. A native of Bradford, UK he was educated at the Universities of York and Bradford and has also worked at the Universities of Manchester, St  Andrews, Southampton, Sheffield, and Duisburg‐Essen. His specialist interests are Romani, creolistics, Austronesian, and Native American languages. He has published over 150 articles, chapters and reviews, and is editor of The Oxford Handbook of Language Contact and co‐editor (with Bettina Migge) of The Journal of Pidgin and Creole Languages.

xii  Notes on Contributors Lenore A. Grenoble holds a joint appointment in the Department of Linguistics and the Department of Slavic Languages and Literatures at the University of Chicago, USA. Her research interests include Slavic, Tungusic, and other languages of the North, discourse and conversation analysis, deixis, contact linguistics, and language endangerment, attrition, and revitalization. Her fieldwork focuses on languages in Siberia and she is currently engaged in research on the interrelations between language shift, cultural change and the environment in the North. Her recent publications include Saving Languages, co‐authored with Lindsay J. Whaley (2006); Endangered Languages: Current Issues and Future Prospects, co‐edited with Lindsay J. Whaley (1998); Language Policy in the Former Soviet Union (2003); Evenki, co‐authored with Nadezhda Ja. Bulatova (1999); and Deixis and Information Packaging (1998). Natalia Gurian is Professor at the Far Eastern Federal University, Vladivostok, Russian Federation. Her main interests are Chinese language and linguistics; traditional Chinese philology and lexicography; early Chinese dictionaries; history of Chinese characters; Russian‐ Chinese language contact and contact languages. She is the author of one book, and more than 50 papers. Her current projects include a study of early Chinese, semantically organized dictionaries and character books from the third century bce to the third century ce, and a study of Sino‐Russian varieties, with Zygmunt Frajzyngier (University of Colorado) and Sergei Karpenko (Far Eastern Federal University, Vladivostok, Russian Federation). Bernd Heine is Emeritus Professor at the Institute of African Studies, University of Cologne, Germany. He has held visiting professorships in Europe, South Korea, Japan, China, Australia, Kenya, South Africa, the United States, and Brazil. He has been a fellow of the Center for Advanced Study in the Behavioral Sciences, Stanford, USA, of the Netherlands Institute for Advanced Studies, and the Tokyo University of Foreign Studies. His main research areas are presently grammaticalization theory, endangered languages in Africa, and discourse grammar. He has published more than 40 books and over 200 articles. Raymond Hickey is Professor of English Linguistics at the University of Duisburg and Essen, Germany. He has written several books on varieties of English including Dublin English. Evolution and Change (John Benjamins, 2005), Irish English, History and Present‐day Forms (Cambridge University Press, 2007), A Dictionary of Varieties of English (Wiley‐Blackwell, 2014), Sociolinguistics in Ireland (Palgrave‐Macmillan, 2016), Listening to the Past, Audio Records of Accents of English (Cambridge University Press, 2017), The Cambridge Handbook of Areal Linguistics (Cambridge University Press, 2017), and English in Multilingual South Africa (Cambridge University Press, 2020). On the Irish language he has published The Dialects of Irish (Mouton, 2011), The Sound Structure of Modern Irish (Mouton, 2014). Lars Johanson is a linguist and Turcologist. For many years he was Professor of Turcology at the University of Mainz, Germany. He has been instrumental in transforming the field of Turcology, which was traditionally more philologically oriented, into a linguistic discipline. Apart from his contributions to Turcology, Lars Johanson has made a number of pioneering contributions to general linguistics and language typology, in particular to the typology of viewpoint‐aspect systems and the theory of language contact, see Structural Factors in Turkic Language Contacts (2002). He is the editor of the journal Turkic Languages (Harrassowitz) and the monograph series Turcologica (Harrassowitz). Brian D. Joseph is Distinguished University Professor of Linguistics, and the Kenneth E. Naylor Professor of South Slavic Linguistics, at the Ohio State University, USA, where he has

Notes on Contributors  xiii taught since 1979. His research focus is historical linguistics, especially sound change and morphological change and the interaction between them, and especially regarding Greek, from Ancient through Modern, in its genealogical context as an Indo‐European language and its geographic, and thus language‐contact, context within the Balkans. Birsel Karakoç is Professor of Turkic languages at the Department of Linguistics and Philology, Uppsala University, Sweden. She obtained her Ph.D. from the University of Mainz (Germany) in 2001, writing a dissertation with the title Das finite Verbalsystem im Nogaischen. Between 1999 and 2006 she worked as a research fellow at the University of Hamburg (Germany). Her research focuses on questions concerning the typology of Turkic languages (especially Noghay, a minor Turkic language spoken in the Northern Caucasus, and other Kipchak Turkic languages, Turkish, and Turkic varieties spoken in Iran), the bilingual acquisition of Turkish in Western Europe, comparative Turkic linguistics, and language contact. Sergei Karpenko is a Professor at the Far Eastern Federal University, Vladivostok, Russian Federation. His main interests are cross‐linguistic studies of syntax and semantics in English and Russian; description of forms and functions from a cognitive linguistic perspective. He is the author/co‐author of three books and over 30 papers. His current projects include: a study of Sino‐Russian varieties, with Zygmunt Frajzyngier (University of Colorado) and Natalia Gurian (the Far Eastern Federal University); the reference system of Russian (with reference to Russian dialects) with Zygmunt Frajzyngier (University of Colorado). Jonathan Kasstan is a Lecturer (Assistant Professor) in French and Linguistics at the University of Westminster, UK. His research focuses on language variation and change, with a particular focus on endangered and heritage languages. He is currently completing a Leverhulme‐funded project on grammatical change in language obsolescence. Paul Kerswill is Professor of Sociolinguistics at the University of York, UK; he has previously held positions at the Universities of Durham, Cambridge, Reading, and Lancaster. His doctoral research in Bergen in Norway initiated his interest in the linguistic outcomes of migration. This was followed by a study of Milton Keynes in Britain and Multicultural London English in the capital. His collaborators include Jenny Cheshire, Sue Fox. Eivind Torgersen, and Ann Williams. A parallel track has been his interest in language and development in West Africa. He was elected a Fellow of the British Academy in 2017. Tania Kuteva is a Full Professor of English Linguistics at Heinrich Heine University Düsseldorf and a Professorial Research Associate of the School of Oriental and African Studies (SOAS), University of London, UK. She has held Visiting Professorship positions at the Max Planck Institute for Evolutionary Anthropology, Leipzig, the Netherlands Institute for Advanced Studies (NIAS), Wassenaar, the University of Texas at San Antonio, the Jawaharlal Nehru University, New Delhi, the Radboud University of Nijmegen, University College London, and the School of Oriental and African Studies, University of London. Her main interests include grammaticalization, linguistic typology, language evolution, discourse grammar. She is the author and the co‐author with Bernd Heine of several books published by Cambridge University Press and Oxford University Press as well as a number of articles. Johanna Laakso studied at the University of Helsinki, Finland, where she completed her Ph.D. Since 2000, she has held the chair of Finno‐Ugric studies at the University of Vienna, Austria. Her research interests include historical linguistics (in particular, the history of

xiv  Notes on Contributors Uralic derivation), language contact and contact‐induced language change, multilingualism and minority languages, and gender linguistics. Danny Law is Associate Professor of Linguistics at the University of Texas at Austin, USA. His research explores the history and development of Mayan languages, particularly contact‐ induced language change. He also works on the documentation and description of several contemporary and ancient Mayan languages, including the language of ancient Mayan hieroglyphics. Siu‐Lun Lee is Senior Lecturer in the Yale‐China Chinese Language Centre at the Chinese University of Hong Kong. His research interests include applied linguistics, Cantonese studies, Chinese linguistics, sociolinguistics and issues of language teaching. Debbie Loakes is a phonetician at the University of Melbourne, Australia. Her postdoctoral project is supported by the Australian Research Council Centre of Excellence for the Dynamics of Language, and focuses on speech production and perception by L1 Aboriginal English people in Victoria (Australia). She is also affiliated with the Research Unit for Indigenous Language at the University of Melbourne. Robert McColl Millar is Professor of Linguistics and Scottish Language at the University of Aberdeen, Scotland. His books include Modern Scots: An Analytical Survey (2018), Contact: the interaction of closely related varieties and the history of English (2016) and English Historical Sociolinguistics (2012). Rajend Mesthrie is Professor of Linguistics at the University of Cape Town (UCT) , South Africa, where he holds a National Research Foundation (NRF) chair in Migration, Language, and Social Change. His main interests are in general linguistics and sociolinguistics. He was head of the Linguistics Section at UCT (1998–2009), President of the Linguistics Society of Southern Africa (2001–2009), co‐editor of English Today (2007–2012), and President of the International Congress of Linguists (2013–2018). Amongst his publications are Language in South Africa (ed., Cambridge University Press, 2002), World Englishes (with Rakesh Bhatt, Cambridge University Press, 2008) and A Dictionary of South African Indian English (UCT Press, 2010). Marianne Mithun is Professor of Linguistics at the University of California, Santa Barbara, USA. Her work ranges over morphology, syntax, discourse, prosody, and their interrelations; language contact and language change; typology and universals; language documentation; revitalization work with communities; and the languages indigenous to North America and Austronesia. Carmel O’Shannessy is a Lecturer in the School of Literature, Languages and Linguistics at the Australian National University, Australia. Her research is on language contact and acquisition, including the emergence of Light Warlpiri, a new Australian mixed language, and children’s development of Light Warlpiri and Warlpiri. She has been involved with languages and education in remote Indigenous communities in Australia since 1996, in the areas of bilingual education and language contact and acquisition. Brigitte Pakendorf, who holds Ph.D. degrees in both Molecular Anthropology and Linguistics, was a Max Planck Research Group leader at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, from 2007 to 2011. Since 2012 she has been

Notes on Contributors  xv a senior scientist at the CNRS research unit “Dynamique du Langage” in Lyon, France. Her  current research focuses on the documentation of the Tungusic languages Even and Negidal and on language contact in Siberia (in particular involving Even and the Turkic ­language Sakha). Thomas Purnell is Professor of English (Language and Linguistics) at the University of Wisconsin–Madison, USA. He received his Ph.D. in Linguistics from the University of Delaware. His research interests focus on variation within sound systems, particularly in Upper Midwestern US English. His research has appeared in American Speech, Journal of Dialect Geography, Journal of English Linguistics, Journal of Language and Social Psychology. He is the editor of American Speech. Paul Roberge earned his Ph.D. at the University of Michigan and is currently Professor of Germanic Languages and Joint Professor of Linguistics at the University of North Carolina at Chapel Hill, USA. His research and teaching interests are Germanic linguistics, Old Norse language and literature, socio‐historical linguistics, pidgins and creoles, Afrikaans, and the origin and evolution of human language. Joseph C. Salmons is the Lester W. J. “Smoky” Seifert Professor in Language Sciences at the University of Wisconsin–Madison, USA. With Jim Leary, he co‐founded the Center for the Study of Upper Midwestern Cultures and has edited Diachronica: International Journal for Historical Linguistics from 2002 until 2019. He is the author of A History of German: What the Past Reveals about Today’s Language (Oxford University Press, 2nd edition, 2018). His research focuses on language change and linguistic theory, especially sound systems. Lotfi Sayahi is Professor of Linguistics in the Department of Languages, Literatures, and Cultures at the University at Albany, State University of New York, USA. His research focuses on language variation and change in situations of bilingualism and language contact. He is the author of Diglossia and Language Contact: Language Variation and Change in North Africa (Cambridge University Press, 2014). He has published more than 30 articles and book chapters that have appeared in the Journal of Sociolinguistics, Journal of Language Contact, International Journal of the Sociology of Language, Journal of Language Sciences, among others. Edgar W. Schneider is full Professor of English Linguistics at the University of Regensburg, Germany, after previous appointments in Bamberg, Georgia, and Berlin. He has written and edited several books (including American Earlier Black English (1989); Introduction to Quantitative Analysis of Linguistic Survey Data (1996); Focus on the USA (1996); Englishes Around the World (1997); Degrees of Restructuring in Creole Languages (2000); Handbook of Varieties of English, 2nd edition (2020); Postcolonial English (2007)) and has published and lectured widely on the dialectology, sociolinguistics, history, semantics and varieties of English, and was the editor of the scholarly journal English World‐Wide along with an associated book series. Erin Shay is a member of the Linguistics Department of the University of Colorado, Boulder, USA, and has spent over 25 years researching, recording, analyzing, and describing ­languages of the Chadic branch of the Afroasiatic family. She has authored or co‐authored grammars of five Chadic languages as well as many books and papers concerning the description and analysis of languages within and outside the Chadic family and the application of such analyses to wider areas of linguistics, such as linguistic forms and functions, language history, language change, and written preservation of endangered languages. Her latest volume is A Grammar of Pévé [Chadic], (2020).

xvi  Notes on Contributors Jeff Siegel is Emeritus Professor in Linguistics at the University of New England, Australia. His main area of research has been on language contact, concentrating on the origins of pidgins, creoles, and new dialects, and on the use of these varieties in formal education. Recently, he has changed his focus to language documentation, working on Nama, a Papuan language of southern New Guinea. Helen Sims‐Williams is a Research Fellow in the Surrey Morphology Group at the University of Surrey, UK. Her research interests lie in historical linguistics, morphology, and typology, focusing particularly on tendencies of language change and the evidence they contribute to linguistic theory. Her work to date has investigated the role of analogy in morphological change and the loss of inflection cross‐linguistically. Eeva Sippola is Associate Professor of Ibero‐American languages and cultures at the University of Helsinki, Finland. She has previously held positions at the University of Bremen (Germany) and Aarhus University (Denmark). Her research interests have a broad focus on contact linguistics and critical sociolinguistics in the Hispanic world. Sippola has published on descriptive and comparative creolistics and sociolinguistic issues, including language endangerment, ideologies, and folk perceptions. John Charles Smith is an Emeritus Fellow of St Catherine’s College, Oxford, UK, where he was Official Fellow and Tutor in French Linguistics for 20 years, and is Deputy Director Emeritus of the University of Oxford Research Centre for Romance Linguistics. Before returning to Oxford, where he was an undergraduate and graduate student at the Queen’s College, he held posts at the Universities of Surrey, Bath, and Manchester; he has also held visiting appointments in Limoges, Paris, Berlin, Melbourne, and Philadelphia. His main research interests are in refunctionalization, morphology, and deixis, with special reference to the Romance languages. Sarah Thomason is a Professor of Linguistics at the University of Michigan, USA, and is a historical linguist specializing in language contact. Among her publications are Language Contact, Creolization, and Genetic Linguistics (with Terrence Kaufman, University of California Press, 1988), Language Contact: An Introduction (University of Edinburgh Press, 2001), and Endangered Languages: An Introduction (Cambridge University Press, 2015). She has also worked with the Seliš‐Q’lispe Culture Committee since 1981 to document the Montana Salish language. Jill Vaughan is a Research Fellow with the Research Unit for Indigenous Language at the University of Melbourne, Australia. Her work in sociolinguistics and linguistic anthropology is concerned with multilingualism, contact and variation in Indigenous languages of northern Australia, and language practices in the context of the Irish diaspora. Heike Wiese is Professor of German in Multilingual Contexts and speaker for the Centre “Language in Urban Diversity” at the Humboldt University, Berlin, Germany. She is interested in the dynamics of multilingual settings, especially in urban areas, with foci on grammatical‐pragmatic interfaces, linguistic architecture, and language perception and monolingual ideologies. In her research, she investigates urban contact dialects, grammatical developments within heritage speakers’ repertoires, urban markets as metrolingual sites, and the dynamics of German in the multilingual context of Namibia. In transfer and outreach activities, she cooperates with educational institutions, museums, and speaker communities.

Preface

The past decade since the publication of the first edition of this handbook has seen a sizable increase in research concerning language contact. It is perhaps significant that at least four major publishers (Oxford University Press, Cambridge University Press, Routledge, and Mouton) are preparing or have just produced handbooks of language contact, testifying to the relevance of the subject to linguistics as a whole. The position of language contact within linguistic research has steadily become more central given the increasing number of studies which see contact as an essential element of virtually all language change. During the first two decades of the present century contact research has broadened to interface with other fields, such as areal linguistics, code‐switching, grammaticalization, creole studies, complexity theory and linguistic typology, to mention just a few. In recent years an increasing focus has been on the role of individual agents and the link between contact and linguistic identity, especially in the context of mixed languages and high‐contact forms of language, often in modern situations of urban multilingualism and transnational communication. The philosophy behind the present handbook has been to provide two broad sections, the first considering the interface between language contact and other closely related areas of linguistic research, such as typology or grammaticalization, with the second part presenting overview chapters of language contact from the standpoint of major languages and language families or areas. In this second part many of the theoretical issues surrounding contact research are discussed in detail and supported by relevant linguistic data. It is hoped that the contributions in the present edition of the handbook will improve our understanding of typical contact scenarios attested across the world. For my part, I would like to thank all my colleagues who readily agreed to contribute to this volume and so made it possible from the very beginning. Hopefully, they and the readers of this handbook will be pleased with the result of their work and find the volume useful and insightful. My thanks also go to the staff at Wiley‐Blackwell for their continuing support during the compilation and production of this second edition, especially to Tanya McMullin and Merryl Le Roux who were always ready to answer questions and offer useful advice.

Münster October 2019

Language Contact and Linguistic Research RAYMOND HICKEY 1 Introduction As a field of linguistic study, research into language contact has long since established itself. Its relevance to investigations of sociolinguistic variation and language change has gained increased recognition during the present century as the number and scope of contact‐based studies have extended and scholars have come to see contact between languages and/or varieties as default and zero‐contact scenarios as the exception. The linguistic publications on language contact in the past few decades cover articles, monographs, edited volumes, special issues of journals (see the references at the end of each chapter).1 While the classic study of language contact by Uriel Weinreich was published in 1953, the following two decades were years which saw not just the heyday of early generative linguistics but also the rise of sociolinguistics (dealing initially with monolingual contexts), and it was those two directions in linguistics which were to dominate the research activity of scholars for a number of decades, certainly in the Anglophone world and frequently elsewhere as well. But the late 1980s witnessed the publication of Sandra Thomason and Terrence Kaufman’s large‐scale study of various contact scenarios with many generalizations about the nature of contact and the range of its possible effects (Thomason and Kaufman 1988). Due to the carefully mounted cases and several stringent analyses, this study led to the re‐ invigorization of language contact studies and the re‐valorization of language contact as a research area. As well as highlighting the field of language contact within linguistics, the study also allowed for virtually any type of change as a result of language contact, given appropriate circumstances to trigger this. The impetus provided to contact studies was felt throughout the 1990s and into the twenty‐first century yielding the situation now (2020) with contact considerations center‐stage in many areas of linguistics. Previously language contact characterized the work of scholars somewhat outside the mainstream. Smaller departments at universities, dealing with non‐Indo‐European languages or Indo‐European ones apart from the Germanic and Romance languages, often produced research in which contact was pivotal. But for scholars in the English‐speaking world, or dealing with varieties of English, language contact was not a primary concern during the 1960s and 1970s. Apart from the dominance of other approaches to linguistics at this time (as just mentioned), there were further reasons for the relative neglect of language contact. Older literature which looked at contact tended to assume uncritically that contact was always the source of new features registered in particular languages, assuming the presence of at least

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

2  Raymond Hickey two in any given scenario. Furthermore, early studies did not necessarily provide rigorous taxonomies for the various types of language contact and their effects (though Weinreich is a laudable exception in this respect). Nor did they usually distinguish individual tokens of language contact, manifest usually in lexical borrowings, from the contact of language systems and the indirect effects which the latter situation could have on the further development of languages. Overviews of aspects of language, which also touched on contact, did of course have relevant chapters, e.g. that by Moravcsik (1978) in the Greenberg volumes on language universals. And the early 1980s did see studies of language contact, e.g. Heath (1984), but other suggestions for the triggers of language change were preferred, at least in mainstream language studies, such as varieties of English, see Harris (1984), an influential article arguing against the role of contact in the rise of varieties of English in Ireland or Lass and Wright (1986) offering similar arguments against contact with Afrikaans as a source for specific ­features of (White) South African English.

2  Recent Studies of Language Contact The studies of language contact during the 1990s and into the 2000s varied. Some of these are in a more traditional style, e.g. Ureland, and Broderick (1991), but others show a linguistically nuanced analysis of the effects of contact, see the contributions in Fisiak (1995) and Thomason (1997a), along with the typological overview in Thomason (1997b). Indeed these publications often contain a blend of contact studies and a further approach in linguistics, consider the sociolinguistically based investigation of language contact in Japan by Loveday (1996) or the large‐scale typological studies in Dutton and Tryon (1994). The 2000s opened with a number of analyses of different contact scenarios. There is the general overview of language contact and change by Frans van Coetsem (van Coetsem 2000) along with the overview article by Thomason (2000), the study of contact within the context of the Slavic languages2 by Gilbers, Nerbonne, and Schaeken (2000) and the investigation of lexical change due to contact in King (2000),3 to mention just three of the publications from that year. 2001 saw the publication of Sarah Thomason’s introduction to language contact (Thomason 2001) and of a volume on language contact and the history of English (Kastovsky and Mettinger 2001), as well as the overview of features in English‐lexicon contact languages (pidgins and creoles) by Baker and Huber (2001). The latter type of investigation characterizes volumes such as that by McWhorter (2000), the full‐length study by Migge (2003), the edited volume by Escure and Schwegler (2004), as well as the special journal issue by Clements and Gooden (2011). Clyne (2003) is a monograph which examined language contact between English and immigrant languages in Australia. This type of contact is grounded in bilingualism, an avenue of research which has been pursued intensively in the past few decades; see Myers‐ Scotton (2002) as a representative example; see also Field (2002). Further studies concern other kinds of contact‐based varieties of English far from the European context, e.g. Chinese Englishes, see Bolton (2003).

2.1  Overviews of Language Contact The increase in the data on language contact has led to more general reflections on the nature of contact and its effects. This is something which can be observed in other fields as well. Once most of the groundwork has been done and bodies of data have been collected, scholars begin to reflect on the status of the field as a scholarly endeavor. It is in this light that one can view publications like those by Winford (2003, 2005, 2008) and Matras (2009).

Language Contact and Linguistic Research  3 A further sign of the maturity of a field is the publication of handbooks dedicated to it. This shows that it has become sufficiently mainstream for it to appear in dedicated courses at universities and hence to be worthy of treatment in this form. Handbooks dealing with language contact began with the comprehensive, two‐volume work by Goebl, Nelde, Stary, and Wölck (eds., 1996) and continued with the first edition of the present handbook (Hickey ed., 2010). Both these works are now in their second editions. The earlier work by Goebl et al. is in the process of appearing in revised form, the first volume of which has been published as Darquennes, Salmons, and Vandenbussche (2019) with the second volume projected for 2023. At the end of the second decade of the present century three further handbooks have been published or are being prepared: Grant (ed., 2019), Mufwene and Escobar (forthcoming), and Adamou and Matras (forthcoming). Dedicated volumes on contact often have a typical range of languages, often recognizable in the subtitle, e.g. Ansaldo (2009) which looks at language contact and change in a South Asian context, Norde, de Jonge, and Hasselblatt (eds., 2010) which deals largely with North Germanic languages, Léglise and Chamoreau, (eds., 2012) which concentrates on French‐lexified creoles, and Law (2014) which looks at scenarios in the Mayan languages; Velupillai (2015) examines pidgins, creoles, and mixed languages, as does Grant (2019: 27–30), while McColl Millar (2016) is a study concerning the historical development of varieties of English. Hundt and Schreier (eds. 2013) deals explicitly with contact scenarios for English around the world. Yet another sign of the advances made in contact studies is the comprehensive guide to contact languages by Bakker and Matras (2013) and is found in the series on contact maintained by publishers. For instance, Cambridge University Press has a series entitled Cambridge Approaches to Language Contact, De Gruyter Mouton has Language Contact and Bilingualism as well as Sociolinguistics and Language Contact while Benjamins has the series Contact Language Library, a continuation of the Creole Language Library. One can also mention the center‐stage treatment of language contact accorded in handbooks of historical linguistics, such as McMahon (1994), McColl Millar (2007), and Campbell (2013).

2.2  What Does Contact‐induced Language Change Encompass? The term “contact‐induced change” has been invoked in a wide range of studies (including those in the present volume) and a number of authors have pointed to the necessity of determining just what is meant by this. Hence the question of what kinds of change can be traced to a contact source has been revisited in a number of recent studies, such as Siemund (2008), Poplack, Zentz, and Dion (2012), Poplack (2018), Lucas (2012) and Seifart (2019). This issue is closely connected to examining the possible effects of contact; see the collection by Wiemer, Wälchli, and Hansen (eds., 2012) and the comprehensive introduction (Grant, 2019) to the Oxford Handbook of Language Contact. In the past two decades sociolinguistic studies have increasingly highlighted the role of the individual in variation and possible change. This standpoint may well have been the impetus4 for the consideration of agency and identity in contact scenarios, see the edited volume by Schwägerl (ed., 2010) and that by Hinrichs and Farquharson (eds., 2011) for the creole context. In volume 1 of the new De Gruyter Mouton handbook on language contact there is an entire section on language contact and the individual (Darquennes, Salmons, and Vandenbussche 2019: 136–256).

2.3  Code‐switching and Contact Code‐switching has been a staple of language contact studies and has been examined from different perspectives, with recent work concentrating on the grammatical change which it

4  Raymond Hickey can engender, see Torres Cacoullos (ed., 2015) and Torres Cacoullos and Travis (2016) A key question, but one surprisingly seldom addressed head‐on except as regards the debate concerning loanwords versus single‐word code switches, is how we determine that a word has been borrowed into another language – this issue is addressed by Durkin (this volume). He furthermore makes the point that words which are copied from a source language may, in their morphological transparency, lead to analogical formations within the receiving language, see ‐ment in Late Middle English. Borrowing is an issue addressed repeatedly in the current volume. Contemporary analyses tend to favor a different terminology, referring to copying and/or replication. Nonetheless, the established discourse around borrowing and loans has been retained by many authors, see Durkin (2014) and the comprehensive guide, Haspelmath and Tadmor (eds., 2009).

2.4  Language Contact, Linguistic Areas, and Typology Research into language families and linguistic areas received considerable impetus during the 2000s. The native languages of northern South America were scrutinized in Aikhenvald (2002a, 2002b). This vein of investigation was continued with Aikhenvald and Dixon (2001, 2003, 2006). Johanson (2002) looked at structural change in the Turkic languages which can be traced to contact (see also Johanson, Csató, and Karakoç, this volume). Similar studies from the early 2000s, e.g. Haspelmath (2001), attest to this revitalized interest in the study of linguistic areas (Matras, McMahon, and Vincent 2006; Muysken ed., 2008). A comprehensive overview can be found in Hickey (ed., 2017). Language typology and its connection with language contact is a theme in studies which congregate around families and areas; see the contributions in Haspelmath et al. (2001), Dahl and Koptjevskaja‐Tamm (2001), Aikhenvald and Dixon (2006). Furthermore, there are languages whose entire development and history is dominated by contact with other languages: Romani and Yiddish are good examples of this situation, see Matras (1995, 2002) and Jacobs (2005) on these two languages respectively; for the genesis of Hebrew, see Doron, Rappaport Hovav, Reshef, and Taube (eds., 2019). The investigation of languages which have virtually no written records presents a special set of problems. This is particularly true of native American languages (Mithun, this volume), of African languages (Beyer, this volume), of Australian languages (Vaughan and Loakes, this volume) and of creole languages in the Pacific arena (Siegel, this volume). Recent book‐ length studies have been dedicated to investigating language in such scenarios; see Dakin, Parodi, and Operstein (eds., 2017) and Berez‐Kroeker, Hintz, and Jany (eds., 2016).

2.5  Language Contact and Creole Studies A central discussion in creole research has concerned the putative exceptionalism of these forms of language, i.e. whether they represent a typological class5 on their own, see McWhorter (ed., 2000) with opposing views put forward which give greater weight to substrate influences, see DeGraff (2003) as a representative example (Schneider and Hickey this volume offer an assessment of this issue). Most of the discussions have been of English‐lexified creoles, but recently studies of Afro‐Hispanic varieties and Ibero‐Asian creoles have been published, see Sessarego (2019) and Cardoso, Baxter, and Pinharanda Nunes (eds., 2012) respectively. Languages whose existence is regarded as deriving from contact between other languages are labeled ‘contact languages’ and have been the subject of various investigations; see Siemund and Kintana (eds., 2008) and Huber and Velupillai (eds., 2007).

Language Contact and Linguistic Research  5

2.6  Language Contact and Mixed Languages Not unrelated to this type of situation is that of mixed languages, the result not just of contact but of fusion, to which the attention of the scholarly community was drawn by a number of seminal publications, among the earliest of which was Muysken (1981) which presented the case of Media Lengua, a mixture of Quechua and Spanish (see Muysken 1997 for a later overview). A broader perspective was provided by the collection of studies on a number of mixed languages to be found in Bakker and Mous (eds., 1994). Cases of mixed languages have also been reported in language endangerment situations (Comrie 2008), e.g. that of light Warlpiri in Northern Australia (O’Shannessy 2005). An instance of a mixed language from the Slavic area would be Surzhyk, a blend of Russian and Ukrainian; see Grenoble (this volume). A further example is Trasianka (a blend of Belarusian and Russian). The Romance languages also have similar mixtures which arose due to contact (Smith, this volume).

2.7  Language Contact, Obsolescence, and Death Language obsolescence (Dorian 1989) and language death (Nettle and Romaine 2000; Harrison 2007) are further issues closely related to language contact. After all, the endangerment of a language always goes hand in hand with contact with one or more dominant languages, the latter threatening the continuing existence of the minority language, or indeed in many cases leading to its disappearance.

2.8  Language Contact and Grammaticalization The study of grammaticalization received significant impulses from the research of Elizabeth Traugott, Bernd Heine, and Paul Hopper in a number of landmark publications, such as Traugott and Heine (2001), as well as the accessible textbook, Hopper and Traugott (2003 [1993]). In the context of the present volume the focus on grammaticalization and language contact6 was made in the programmatic article by Heine and Kuteva (2003) which was followed up by the full‐length study Heine and Kuteva (2005); see Heine and Kuteva (this volume) as well as Heine (2008) on contact‐induced word‐order change.

2.9  Language Contact and Complexity Complexity in language is a recurrent theme in linguistic research over the past decade or so and its relationship to questions of contact and typology has been addressed in a number of publications; see the chapters in Miestamo, Sinnemäki, and Karlsson (2008) and the study by Mufwene (2008). Works in this vein have been continued in the past decade or so; see Sampson, Gil, and Trudgill (eds., 2009) Kortmann and Szmrecsanyi (eds., 2012), Trudgill (2011), McWhorter (2011), Aboh (2015, ed., 2017), Mufwene, Coupé, and Pellegrino (eds., 2019) as representative examples as well as the individual studies by Bastardas‐Boada (2019) and Maitz and Németh (2014). By and large the view is that language contact only necessarily leads to simplification if the contact is largely between adults (Trudgill 2011: 55) whereas situations of longer duration result in stable bilingualism, with children acquiring languages naturally, and hence complex morphosyntactic structures and distributions are not lost in such contact scenarios.

2.10  Early Language Contact and Older Hypotheses The assessment of language contact in the history of established languages is a matter which has varied in the relevant scholarship. For the history of English it is clear that the influence of other languages – bar Latin, Old Norse, and Anglo‐Norman – has been played down by

6  Raymond Hickey the majority of scholars in the field.7 But in recent years, a reexamination and reassessment of the role of contact in the development of the Germanic dialects in the period subsequent to the transportation to England has taken place. Specifically, the role of British Celtic in this context has been highlighted by publications such as Hickey (2012a, 2012b), Filppula, Klemola, and Pitkänen (2002), Filppula, Klemola, and Paulasto (2008), reconnecting to an older hypothesis put forward by German and Scandinavian scholars in the first half of the twentieth century; see Preußler (1938), Dal (1952), and Braaten (1967). Contact as a source of change has been further extended to encompass later, nonstandard features of English such as the so‐called Northern Subject Rule, see Klemola (2000). For details on the “Celtic hypothesis” in the history of English, see Filppula, Klemola, and Paulasto (2008), Filppula and Klemola (eds., 2009), Filppula and Klemola (2014). Several studies of contact have stretched backwards to reach greater time depth using the tools of contemporary research.8 Andersen (ed., 2003), Askedal, and Nielsen (eds., 2015), Boas and Höder (eds., 2018), Braunmüller, Höder, and Kühl (eds., 2014), Drinka (2017), and Ross (2003) are examples of this in their investigations of prehistoric language contact. Salmons and Joseph (1998) look at the evidence for and against Nostratic, an undertaking in which contact is center‐stage. For contact and early Finno‐Ugric, see Laakso (this volume) and for contact and Arabic, see Sayahi (this volume).

2.11  Language and/or Dialect Contact It is obvious that the difference between language contact and dialect contact is more one of degree than of kind. The interaction of dialects with one another is a topic which received considerable impetus from Peter Trudgill’s 1986 study Dialects in Contact after which the treatment of this subject was seen as on a par with that of languages in contact. Given the great diversity of varieties of English, this approach proved to be fruitful in the Anglophone world and has been adopted by many scholars since, especially by considering the notion of accommodation together with existing data not hitherto analyzed from this perspective. Dialects in contact are treated in this volume in the contributions by Kerswill (in the context of new varieties) as well as Salmons and Purnell (in the context of American English).

2.12  Language Contact in Pluricentric Languages Major European languages, above all English and Spanish, are characterized by having several centers across the world due to their spread during the colonial period (c. 1500– 1900). These centers are inherently locations of contact; see Sessarego and González‐ Rivera (eds., 2015), King and Sessarego (eds., 2017), Ortiz López, Guzzardo Tamargo, and González‐Rivera (eds., 2020), Clements and Gooden (eds., 2011) and Mackenzie (2017). With pluricentric languages different locations have been contrasted with each other, see Orozco (2018). Pluricentric languages are often characterized by diglossic situations with a standard form of language deriving from an original source alongside local vernaculars in various countries. This situation has been investigated recently for Arabic, see Sayahi (2014, this volume) and Manfredi and Tosco (eds., 2017) on contact in Arabic in general.

2.13  Language Contact and Diasporic Varieties Languages which are characterized by geographical spread frequently have diasporic communities, cf. Hungarian treated in Fenyvesi (ed. 2005). In the Anglophone world, the largest diaspora is that of Indians which is the subject of the volume by Hundt and Sharma (eds., 2014).

Language Contact and Linguistic Research  7

2.14  Language Contact in English Studies In English studies the significance of contact in the rise of nonstandard vernaculars was given increasing recognition during the 1980s. Rickford (1986) is a well‐known example of work in this vein, here with specific reference to dialect transportation and contact at overseas locations. However, not all scholars saw contact as a prime source of new features in varieties, some put more emphasis on the continuation of vernacular traits at new locations. This stance forms the so‐called retentionist hypothesis which enjoyed greatest favor among Anglicists. However, by the late 1980s and into the early 1990s, the considered case for contact in certain scenarios regained acceptance and was underlined by key publications such as Mesthrie (1992) which showed clearly the role contact played in the rise of South African Indian English. The dichotomy of contact versus retention continued to occupy scholars into the 2000s; see Filppula (2003) which provides a fresh look at the arguments. The role of contact in the formation of different varieties of English at various geographical locations has been considered, e.g. Bao (2005, 2010) which examines substratist influence on the aspectual system of English in Singapore. For contact and African Englishes, see Mesthrie (this volume).

2.15  Vernacular Universals and Contact The notion of vernacular universals is something which has been dealt with by Anglicists in recent years, above all by Jack Chambers (see Chambers 2004). It refers to features found across varieties of English in different parts of the world and postulates that the occurrence of such features is due to universals of language development, specifically in the context of new dialect formation (see Gold 2009, for example). The issue has spawned a number of publications the most comprehensive of which is the volume by Filppula, Klemola, and Paulasto (2009b) in which vernacular universals are viewed within the framework of language contact; see the introduction to that volume (Filppula, Klemola, and Paulasto 2009a) and also the contribution by Donald Winford (Winford 2009).

2.16  Sociolinguistic Perspectives on Language Contact An emphasis on the social setting in which language contact can take place is found in many publications, e.g. those in Potowski and Cameron (2007) on Spanish‐based contact and in particular in studies of pidgins and creoles (Deumert and Durlemann 2006; Holm 2004, 2010; Schneider and Hickey, this volume). Studies like Siegel (1987), where the plantation environment of the Fiji Islands in the nineteenth century is investigated, implicitly adopt this stance. The role of substrate in the rise of these contact languages has also been pursued in other publications by Siegel (1999, 2000a, 2000b, 2008, this volume). In a far‐eastern context this issue has also been broached; see the discussion in Matthews (2010). In handbooks on sociolinguistics and models of socially‐determined language change, chapters on contact can also be found, indeed in the second edition of the Handbook of Language Variation and Change there is a four‐chapter section devoted to contact (Wolfram and Schilling 2013: 469–554), see also Sankoff (2002). A broader view than just the social setting can be found in considerations of a language’s ecology; see Mufwene (2001, 2007) and more recently, Ludwig, Pagel, and Mühlhäusler (2018).

2.17  Contact in Urban Environments In the past, contact studies did not usually deal with the rural–urban dichotomy, probably because at the time of contact this division was assumed not to have been relevant for the communities in question. However, contemporary investigations of contact,

8  Raymond Hickey either interlinguistic or intralinguistic, are frequently of urban scenarios, e.g. Silva‐ Corvalán’s 1994 study of Spanish and English in Los Angeles or Hickey’s 2005 study of language variation and change in Dublin, where dissociation (Hickey 2013a), triggered by internal contact between differing varieties in the city, has been the driving factor. Other urban environments have provided further examples of change and development through contact, e.g. the creative language mixture found in the Sheng and Engsh codes in urban Kenya (Abdulaziz and Osinde 1997). Triggered not least by discussions of superdiversity, research into urban contact scenarios has increased; see Duarte and Gogolin (eds., 2013) and Wiese (this volume).

2.18  Language Contact and Globalization Connected with an augmented focus on language in cities is the issue of globalization in linguistics. The main author in this field, Jan Blommaert, has discussed situations of ­language contact across national boundaries, indeed in international settings; see the discussions in Blommaert (2010), but generally without speaker contact. Communication across the internet and mobile networks, the subject of Deumert (2014), is dominated by English which exercises a transnational influence (Schneider 2014) on virtually all languages.

3  Generalizations Concerning Contact It would seem that language contact always induces change. History does not provide instances of speech communities which adjoined one another, still less which intermingled, and where the languages of each community remained unaffected by the contact.9 However, there may well be a difference in the degree to which languages in contact influence each other, that is a cline of contact is often observable and it can be bidirectional or largely unidirectional. Furthermore, influence may vary by level of language and depend on the nature of the contact, especially on whether bilingualism exists or not and to what degree and for what duration. These remarks refer to communities of speakers as contact is understood here as between groups. As contact is a cline, any community can exhibit low or high contact; the extreme end for the former situation would be a situation of complete monolingualism which may have held and still holds for isolated or indeed uncontacted communities, though zero contact may be more a theoretical scenario rather than an actual reality. High‐contact situations have and still do abound and the extreme end of this type is one of bi‐ and/or multilingualism where all speakers in a community are fluent in two or more languages. Community‐wide bi‐/multilingualism, if it lasts for some time, usually leads to the formation of a linguistic area as the mutual permeation of the languages involved is insured through the acquisition of two or more languages by child learners. The Balkans area (Joseph, this volume) and India (Schiffman 2010) provide good evidence for this. Low‐contact communities also show certain features which are characteristic of them in general. Because of the lack of contact with other groups, such communities are usually more closely knit and certain cognitive distinctions may come to be expressed in their languages. A good example is evidentiality, the confirmation to one’s interlocutor of evidence for an event or state. This can be morphologically marked, as an inflection or a clitic, in low‐contact languages, but in high‐contact, diffuse languages, evidentially is usually expressed periphrastically, e.g. by means of modal verbs like should, ought to. For example, Eastern Pomo (a nearly extinct Pomoan native American language in northern California) has inflectional evidential marking (McLendon 2003; Joseph 2003).

Language Contact and Linguistic Research  9 speakers (1) / \ internal external – – – communities (2) / \ internal external – – – outside groups type: internally motivated externally motivated group: first language acquirers adult speakers

Figure 0.1  Differentiation of the internal‐external distinction and typical time spans for types of language change.

3.1  Internal vs External Reasons It is scholarly practice to distinguish between internal and external reasons for language change (Hickey 2002a, 2002b). However, these terms are used in two distinct ways (see Figure 0.1). The first concerns a motivation (internal) for change which is determined by the structure of a language and which appears during first language acquisition while change motivated by contact with other members of one’s group and ultimately sociolinguistically motivated is seen as external. The second use of the dichotomy “internal ~ external” is where internal change is seen as that which occurs within a speech community, generally among monolingual speakers, and external change is viewed as that which is induced by contact with speakers of a different language, usually from a different community (Hickey 2012c). Opinions are divided on when to assume contact as the source of change. Some authors insist on the primacy of internal factors (e.g. Lass and Wright 1986) and so favor these when the scales of probability are not biased in either an internal or external direction for any instance of change. Other scholars view external reasons more favorably (Vennemann 2001, 2002b) while still others would like to see a less dichotomous view of internal versus external factors in change (Dorian 1993; Jones and Esch 2002). The role of contact in the diversification of languages is also a theme in the seminal monograph by Johanna Nichols (1992); see also Nichols (2010).

3.2  Substrate and Superstrate A lot of attention has been paid in the literature to the relative social status of two languages in contact situations. Two established terms are used to label the language with less status and that with more, namely, “substrate” and “superstrate” respectively. The superstrate is regarded as having, or having had, more prestige in the society in which it is spoken, though just precisely what “prestige” refers to is something which linguists like James Milroy have questioned. Nonetheless, there would seem to be a valid sense in which one of two languages has, or had, more power in a contact situation. Asymmetrical levels of power in a contact situation play a definite role in the results of contact (Durkin, this volume).

3.3  Relative Status and Direction of Influence The standard wisdom has traditionally been that the language with more status influences that with less, i.e. borrowing is from the superstrate by the substrate. This is, however, a ­simplistic view of the possibilities of influence in a contact scenario. Vocabulary, as an open class with a high degree of awareness by speakers, is the primary source of borrowing from the superstrate; French and Latin in the history of English are standard examples.

10  Raymond Hickey However, if contact persists over many generations, then the substrate can have a gradual and imperceptible influence on the superstrate, leading in some cases to systemic change at a later time. This type of contact can be termed “delayed effect contact” (Hickey 2001) and may well be the source of syntactic features in English which the latter has in common with Celtic (Poussa 1990; Vennemann 2002a; Isaac 2003; Hickey 2012a). This line of thought is pursued by Filppula (2010), who presents the arguments for Celtic influence on English. In addition to structural parallels there is further evidence here. Consider the fact that in Old English wealh was the word for ‘foreigner’ but also for ‘Celt’. The word came to be used in the sense of ‘servant, slave’ (cf. wielen ‘female slave, servant’ with the same root, Holthausen 1974: 393), which would appear to be an indication of the status of the Celts vis‐à‐vis the Germanic settlers.10 Not only that, the meaning of ‘servant’ implies that the Germanic settlers put the subjugated Celts to work for them; this in turn meant that there would have been considerable face‐to‐face contact between Celts and Germanic settlers, in particular between the children of both groups. As the latter context was one of first language acquisition it provided an osmotic interface for structural features of Celtic to diffuse into Old English. Given that written Old English was dominated by the West Saxon standard, it is only in the Middle English period that the syntactic influence of Celtic becomes apparent in the written record, e.g. in the appearance of possessive pronouns in cases of inalienable possession.

3.4  Where Does it Start? The Locus of Contact It is a convenient shorthand to claim, for example, that language A borrowed from language B. However, this is already an abstraction as the appearance of borrowings in a speech community can only be the result of actions by individual members of this community. If one puts aside cases of “cultural” borrowings, e.g. from Latin or Greek into later European languages or from English into other modern languages, then it is probably true that the borrowing of “systemic” material – inflections, grammatical forms, sentence structures – can only occur via bilinguals. This view has a considerable tradition. Weinreich (1953) saw the true locus of contact‐induced change in the bilingual individual who moves between two linguistic systems. Some scholars go further and consider bilinguals as having a single system, e.g. Matras (2010) who contends that bilinguals “do not, in fact, organize their communication in the form of two ‘languages’ or ‘linguistic systems’.” The awareness of linguistic systems on the part of speakers is a difficult issue to resolve. It may well be that in prehistory and in nonliterate societies today the awareness of the separateness of languages was/is less than in present‐day literate societies. If one of the languages a bilingual uses is the sole language of a country then the bilingual’s awareness of switching between languages increases. Matras (2010) maintains that bilinguals “operate on the basis of established associations between a subset of structures and a set of interaction contexts.” The communicative competence of the bilingual then includes making the appropriate choices of structures for communication in given contexts. Whatever the degree of awareness by bilinguals of the separateness of their linguistic (sub)systems, the presence of competence in two languages fulfills the precondition for the adoption of material from one language into another. The next, and crucial, question is how borrowings, made on an individual level, spread throughout a community and are accepted by it. This step is essential for borrowings/ items of transfer to become part of a language/variety as a whole and hence be passed on to later generations as established features (see Hickey, Chapter 7).

3.5  What Can be Attributed to Language Contact? The current volume is dedicated to analyses of language contact, the situations in which it is or was to be found, and the results it engenders or has engendered. This focus should not imply a neglect of changes, indeed types of change, which are not due to language contact.

Language Contact and Linguistic Research  11 Consider for instance, reanalysis by language learners. A specific instance of this is provided in the prehistory of Irish. The precursors of all the Celtic languages inherited complex suffixal inflections from still earlier stages of Indo‐European when these were central to morphology. Sometime before the Celtic languages appeared in writing (in the first centuries bce) the languages changed their typology. They began to abandon suffixal inflections as a means of indicating grammatical categories and adopted a new system whereby these categories were indicated by changes to the initial segments of lexical words, so‐called initial mutation. This typological shift came about by children reanalyzing phonetic changes at the beginnings of words (external sandhi) as having systemic status (for a fuller discussion, see Hickey 1995a, 1995b, 2003a). This is an entirely language‐internal change, though the original trigger for the phonetic changes, which were later reanalyzed, may have been due to contact.

3.6  Pushing the Question Back Contact treatments tend to push the question of origin back a step but do not necessarily explain how a phenomenon arose in the first place. For instance, if one believes that the VSO word order of Insular Celtic (Eska, this volume) is due to contact with a Semitic language (Pokorny 1949; Vennemann 2011) present in the British Isles before the arrival of the Celts, one still has not accounted for the rise of VSO in the source language.11 Thus contact differs from explanatory models of language in that it offers more or less plausible accounts for the appearance of linguistic features. However, the analysis of contact mechanisms and speaker strategies in contact situations can indeed have explanatory value.

3.7  The History of Contact Phenomena It can be salutary to bear the attested history of contact phenomena in mind. The paths of contact may be multiplex and varied. Take, for instance, the immediate perfective of Irish English which is (rightly) regarded as a calque on Irish. (1)

Tá sé tar éis an ghloine a briseadh. [is he after the glass COMP break‐NONFINITE] ‘He is after breaking the glass.’, i.e. ‘He has just broken the glass.’

Both the Irish and the Irish English structure have gone through historical developments while in contact. Originally, the structure could be used in both Irish and Irish English with future, i.e. prospective, reference and it is attested from the seventeenth and eighteenth centuries in this sense. However in both languages, the prospective use declined and an exclusively past, i.e. retrospective, use came to the fore, gradually replacing the former one in the latter half of the nineteenth century (McCafferty 2004).

3.8  Contact in Hindsight If centuries lie between the period of contact and the present it may be difficult to reconstruct the social circumstances of the contact. However, the nature of the contact can often be gleaned from the results it engendered. To illustrate this consider lexical changes in the period immediately after the coming of the Anglo‐Normans to Ireland in the late twelfth century. Many loans from Anglo‐Norman appear and not a few of them are “core” vocabulary items like the words for ‘boy’ (garsún < Anglo‐Norman garçon) and ‘child’ (páiste < Anglo‐Norman page). Given that Anglo‐Norman was the superstrate language in the late Middle Irish period, why should the Irish have borrowed such “noncultural” core items as

12  Raymond Hickey ‘boy’ or ‘child’? The answer would seem to lie in the manner in which these words entered Irish. Assume that they were not borrowed by the native Irish directly, but rather that the Anglo‐Normans used them in their variety of Irish. It is a historical fact that the Anglo‐ Normans lived in the countryside among the Irish and gradually shifted to their language. During the shift period an intermediate variety was spoken by the Anglo‐Normans in which they used words from their own language like garçon and page. Because of the power the Anglo‐Normans had in Irish society, the native Irish adopted core vocabulary items of this Anglo‐Norman variety of Irish and, for example, the negation structure Níl puinn Gaeilge agam [is‐not point Irish at‐me] ‘I cannot speak Irish’, which shows the negative use of French point (Rockel 1989: 59). The likelihood of this scenario is strengthened by considering that the Anglo‐Norman loans in Irish did not necessarily replace the native Irish words. For instance, the Anglo‐Norman loan páiste exists side by side with the original Irish word leanbh ‘child’. For ‘boy’ Irish has two words: garsún and the original buachaill. To use a metaphorical term, the Anglo‐Normans “imposed” parts of their second language variety of the Irish language on those who surrounded them in late medieval rural Irish society (see discussion of terminology pp. 18–20).

3.9  Category and Exponence In borrowing one must distinguish between systemic and nonsystemic elements. The latter are typically individual words or phrases, pragmatic markers, sentence adverbials, or other free‐ floating elements which are not part of the grammatical structure of a language. Such elements travel well because they do not require integration into the system of the borrowing language and can be picked up by adults in a contact situation. Indeed they often migrate from one language to another via code‐switching (Muysken, 2000; Gardner‐Chloros, this volume). In Irish, for instance, English well, just, really have been borrowed and are used continually by native speakers, although the grammatical structures of Irish and English are very different. When looking at systemic elements, one must bear an essential distinction in mind, namely, that between a grammatical category and the exponence of this category. The reason the distinction is essential in contact studies is that some languages borrow a grammatical category but not the manner of expressing it in the source language. Indeed it may be true that adopting a category rather than its exponence is not so much a feature of borrowing but of transfer in language shift (see the discussions of the borrowing of structural elements in Winford 2010 and Hickey, Chapter 7 this volume).

3.10  Transfer in Language Shift The term “borrowing” implies that speakers, adopting some element or category from a source language, do not switch to the latter. The Middle English borrowings from French may be an example of this (for a detailed discussion, see Durkin 2009 and this volume). However, many contact situations involve language shift. When viewing the past few centuries in Ireland one can see that the majority of the population was originally Irish‐speaking and that they gradually transferred to English, particularly during the nineteenth century. There was no general schooling for the Irish before the 1830s so that the native Irish learned English by picking it up – in adulthood – from others who had a somewhat better knowledge of the language. This is a situation of unguided adult second language acquisition. Here the transfer of features from the outset language (Irish) to the new language (English) was at a premium. An obvious example of this is accent: initially, adult learners of a second language use the phonetic realizations of phonological units from their first language when speaking the second one.12 This can still be recognized in rural western forms of Irish English where the phonetic realization of /ai, au, oː, uː, ʌ/ is the same in English as it is in Irish.

Language Contact and Linguistic Research  13

3.11  The Search for Categorial Equivalence On a grammatical level one finds similar behavior: speakers involved in language shift transfer categories from their native language to the new one they are moving to. But what if the new language does not have an automatic equivalent to a grammatical category in the outset language? This triggers a search for categorial equivalence. Take, for instance, the habitual aspect of Irish. This has no formal parallel in English so that speakers during language shift would not have found a ready equivalent to it. Instead the nonlexical verb do was co‐opted to express a habitual in Irish English, as can be seen in the following example: (2)

Bíonn sí ag déanamh imní faoi na leanaí. [is‐HABITUAL she at doing worry about the children] ‘She does be worrying about the children.’ (vernacular Irish English)

What one can say here is that the category “habitual aspect” was transferred during the language shift process from Irish to Irish English. The exponence which was chosen derives, however, from the co‐option of English do to express this category. The verb do is suitable for the expression of habitual aspect as it denotes the carrying out of an action. This means that a construction of the kind “X does be Y‐ing” had a high probability of diffusion and acceptance in the community of new English speakers which arose in Ireland during the language shift period. Recall, furthermore, that for the vast majority of Irish speakers, the language shift took place in a nonprescriptive environment, one in which creativity was not restricted by notions of correctness. This meant that a degree of restructuring occurred in English which is normally only found in creolization scenarios (Hickey 1997a and Chapter 7).

3.12  Neglect of Distinctions in Language Shift Just as speakers search for equivalents in the target language to categories of their native language during language shift, so they also neglect distinctions in the target language which are not found in their native language. This neglect may become established if the transfer variety of the target language stabilizes and becomes focused, as has been the case with Irish English. The relative infrequency of the present perfect in Irish English has been remarked on by many authors since the nineteenth century (Hickey 2007: 142–5). As this verbal category does not exist in Irish one can surmise that it was neglected by learners of English in the period of language shift. In Irish English actions which began in the past and continue into the present, or which are relevant to the present, are expressed by the simple present or past, whichever is appropriate. (3)

a. b. c. d.

I know her since a long time now. Tá aithne agam uirthi le tamall fada anois. [is knowledge at‐me on‐her with time long now] He’s married for ten years. Tá sé pósta le deich mbliana anuas. [is he married with ten years down]

3.13  What Does not get Transferred in Language Contact? Not only is there a neglect of grammatical distinctions during contact with shift, as just ­outlined, but there are certain categories which appear never to be transferred. Consider the following list

14  Raymond Hickey which shows features which do not occur in contact varieties of Irish English where one would expect the greatest influence of Irish grammatical structures (Hickey 2007: 142). 1. 2. 3. 4. 5. 6.

VSO word order Dhíol mé mo theach. ‘I sold my house.’, lit. ‘sold I my house’ Post‐posed adjectives an fear bocht ‘the poor man’, lit. ‘the man poor’ Post‐posed genitives teach Sheáin ‘John’s house’, lit. ‘house John‐GEN’ Absence of personal pronoun in present tense (pro‐drop) Ní thuigim tada. ‘I don’t understand anything.’, lit. ‘not understand‐1ST_PERS_SG anything’ Autonomous verb form Rinneadh an obair. ‘The work was done.’, lit. ‘done‐was the work’ Rugadh mac di. ‘She bore a son.’, lit. ‘born‐was a son to‐her’ Possessive pronoun and “verbal noun” Bhí sé á bagairt. ‘He was threatening her.’, lit. ‘was he at‐her threatening’

The absence of some of these features can be accounted for by the lack of typological fit between the two languages. For instance, item (5) was rendered via the English passive and item (6) was expressed using a simple direct object. But items (1)–(4) seem to be of a different nature. They would appear to derive in Irish from parameter settings: (1)–(3) are dependent on the setting for direction of modification and (4) for that of the pro‐drop parameter. Irish has post‐modification (VSO, N + Adj, N + Gen) and a positive setting for pro‐drop (in the first person singular and plural). In English the reverse is the case. It would appear that Irish speakers switching to English intuitively recognized this and never used the Irish values in their shift variety of English.

3.14  Nonbinary Categories in Contact The results of contact are not always obvious at first sight. Where something is borrowed – a sound, a word, or a structure – which hitherto did not exist in the borrowing language, the matter is fairly clear. But what of the following situation? A language has a certain grammatical device X and so does another language with which it is in contact. However, X in the second language has a much greater range of applications than in the first language. In the course of time, the first language expands the range of contexts in which it can use X. Here is a concrete example. Irish allows the fronting of elements of a sentence in order to topicalize them. The process involves clefting, that is a dummy verb forms a main clause with the topicalized element and the remainder of the sentence is contained in a following relative clause. (4)

a. b.

Is go Gaillimh atá sé imithe. [is to Galway that‐is he gone] ‘It’s to Galway he’s gone.’ Is mór le Máire atá a mac. [is great with Mary that‐is her son] ‘It’s friendly with Mary her son is.’

This example shows the difficulty of deciding whether contact is the source of a nonbinary category, i.e. one of degree and not of presence versus absence. Indeed such instances highlight the nature of linguistic argumentation in contact scenarios where one operates with statements of probability.

Language Contact and Linguistic Research  15

3.15  Permeability of Linguistic Systems There is nothing in the structure of a language which is excluded from borrowing/ transfer through contact. Given sufficient intensity and duration as well as appropriate social conditions and attitudes, all linguistic subsystems can be affected, even the core morphology (Seifart 2015, 2017). Nonetheless, there are areas of language which show much greater movement in a contact situation. Single words and phrases as well as pragmatic markers and sentence adverbials are borrowed easily (Matras 1998). The reason is clear: such elements do not require integration into the grammatical system of the borrowing language and can be accommodated without restructuring. In a language shift situation syntactic variation can occur as a result of transfer during the shift phase, often due to the development of alternative strategies to reach equivalents to grammatical categories and structures of the outset language, e.g. relative clauses in South African Indian English (Mesthrie and Dunne 1990) whose speakers have South Asian heritage languages. Among the other motivations for borrowing/transfer are (1) the resolution of ambiguity and (2) the filling of gaps in paradigms for which the following examples can be given.

3.16  Borrowing to Resolve Ambiguity A well‐known instance where a borrowing helped to resolve internal ambiguity in the morphology of a language – or at the very least potential ambiguity – is provided by the third person plural pronominal forms in th‐ (they, them, their) which derive from Old Norse during the Old English period, initially in the north of England. The position before the borrowing was such that both the third person singular and plural pronominal forms began in h‐ and the difference between the third person masculine singular and the plural was that between /eː/ and /iː/ (hē, hı̄). Borrowing the Old Norse forms thus increased the phonetic distinctiveness of the singular and plural forms, though it also produced suppletion in this morphological paradigm (Werner 1991).

3.17  Borrowing to Fill Gaps in Paradigms There have been many cases where a gap in a morphological paradigm has been filled by a borrowing from a further language. The borrowing can take the form of a direct loan from a language or of a lexical creation made by scholars, e.g. the Latinate adjectives of English such as marine (noun: sea), aquatic (noun: water), equestrian (noun: horse), which appeared in the early modern period. Borrowing, or creation on the basis of external models, does not have to be the path taken. Consider, for instance, the paradigmatic pressure which arose in English with the demise of a clear distinction between the singular and plural of second person pronouns. This issue was resolved language‐internally for the majority of vernacular varieties which now show a pronominal distinction between the singular and plural second person, e.g. y’all, y’uns, youse, yez for ‘you’‐PLURAL. But in the case of Caribbean Englishes, the form unu – a borrowing from input West African languages to the area – filled the gap (Hickey 2003b).

3.18  Convergence Scenarios There are cases where a certain instance of change could derive from both an internal development, often “tension” in the language system, and the influence of another language (Hickey 1997b). An example is provided by the development of stress patterns in the dialects of Irish. Briefly, the situation is as follows: Old Irish (600–900) had lexical root stress but in Middle Irish (900–1200) long vowels developed through vocalization of voiced fricatives in

16  Raymond Hickey non‐initial position. This led to tension with initial short vowels and long vowels in later ­syllables. The three major dialect areas reacted differently to this tension. (5)

Long vowels in unstressed syllables North (Ulster) Post‐initial vowels are shortened West (Connacht) Maintains the original stress pattern, possibly with syncope of the initial short vowel in some cases South (Munster) Stress is shifted to post‐initial long vowels scadán ‘herring’ North ’V.VV > ’V.V /’skadan/ /’skʊdaːn/ West ’V.VV > ’V.VV /skǝ’daːn/ South ’V.VV > V.’VV

The southern shift of stress to non‐initial long vowels might look like a purely internal solution to the tension between quantity and stress which existed prior to this. However, the south is the region where the influence of Anglo‐Norman, which had non‐initial stress for long vowels, was greatest. In addition it is known that Anglo‐Norman affected varieties of English in the south‐east of Ireland, inducing stress on non‐initial long vowels. One can say of the development in the south of Ireland that it is a typical convergence scenario: there are cogent arguments for both an internal and external motivation for the stress change. In the absence of any clinching evidence either way, one maintains that both sources are possible, i.e. both “converged” to produce the observed output (see further the discussion of convergence in Matras 2010).

3.19  Internal Developments which Favor Borrowing A nuanced view of the convergence scenario would ask whether internal changes can render a language more susceptible to borrowing. Recall that English is unique among the Germanic languages (Roberge, this volume) in requiring possessive pronouns in instances of inalienable possession (internal possessor construction). The inherited Old English type involved a personal dative (external possessor construction) as seen still in modern German, e.g. Er legte ihm die Krone aufs Haupt [he laid him‐DATIVE the crown on the head] ‘He laid the crown on his head.’ This type of structure disappears in early Middle English and is replaced by one in which the possessive pronoun is used (see gloss). Now the Celtic languages were, and still are, remarkable in demanding the use of a possessive pronoun for inalienable possession (internal possessor construction). Could British Celtic of the Old English period have provided the model for the English marking? Yes, but the demise of morphological marking of the dative (contrasting with the accusative) in the late Old English and early Middle English periods provided an impetus for alternative marking of possession. In this situation the likelihood of the adoption of a strategy from Celtic with which Old English was in contact was increased.

3.20  Contact and Geographical Spread Features which cluster in geographically confined areas and are found in languages which are not necessarily genetically related (Noonan 2010), or only distantly so, are said to characterize a linguistic area, such as the Balkans (Joseph, this volume), or, in a larger framework, South Asia (Schiffman 2010). For this to occur, many centuries of prolonged contact and population interaction would seem to be required, especially as the common features of such areas belong to the closed classes of the languages involved, typically to the morphology and syntax. Furthermore, the languages of a linguistic area show not only internal coherence but

Language Contact and Linguistic Research  17 Table 0.1  Features with a considerable geographical spread. Feature

Geographical spread

Front rounded vowels, /y/ and /ø/ Uvular /r/ [ʁ] Initial voicing of fricatives Vowel epenthesis in syllable‐ coda clusters High mid to front realization of /u/ [u̶]

The north and center of Europe, excluding the British Isles and the Iberian Peninsula A wide band from northern France to southern Sweden Flemish, Dutch, and southern dialects of English The Netherlands and the north Rhenish dialects of German Scotland and Northern Ireland (all varieties of English, Scottish Gaelic, and Irish)

also recognizable external boundaries with languages immediately outside the area (Haspelmath 2004: 211). There are also cases where features show a considerable geographical spread without the preconditions of linguistic areas being met. Usually, single features are involved and often these are features which involve the sound systems of the languages in question. Table 0.1 shows a small selection of such features (with differing distribution sizes) taken from European languages.13 Such features are often prosodic or realizational (uvular [ʁ], for instance, Bergs 2006) and while their geographical diffusion may be due to low‐level copying of speech habits, they can in time achieve systemic status as is postulated for the spread of prosodic factors in the South Asian context which resulted in systemic tone contrasts for many languages (Matisoff 2001).

3.21  Regularization and Language Contact The outcome of language contact, either through borrowing or transfer, is of interest compared to internal developments, especially with regard to whether there is a difference in kind between the types of change resulting from these sources. One of the differences which may apply to internal but not external sources concerns the regularization of a language’s grammar. Regularization is a common phenomenon in internal change when this can be traced to phenomena in first language acquisition which are carried through into adulthood, leading in some cases to community‐wide change. It is known that children overgeneralize patterns and constructions and in some cases these can survive beyond childhood, especially if the social environment is nonprescriptive, e.g. nonliterate or pre‐literate. An appropriate example is the regularization of gender in Middle High German. Here disyllabic words in final ‐e became feminine because the majority were feminine anyway.14 Thus the word for ‘flower’ was formerly masculine but became feminine (> die Blume, Kluge and Seebold 2002: 134), making it conform to the established pattern of words like die Sonne ‘sun’, die Decke ‘cover, ceiling’, die Kerze ‘the candle’, etc. Returning to language contact, one can say that this type of development is not typical of contact, unless it is prolonged and intensive, i.e. continues over several generations. If that is the case then the same type of regularizing tendencies, observed for language learners in monolingual situations, can in theory occur in contact communities (see the discussion along these lines in Trudgill 2010). Furthermore, with long and close contact a regular system or subsystem could be adopted from another speech community, perhaps replacing an existing subsystem showing more irregularity than the new one.

18  Raymond Hickey Here it is worth considering what developments can be “beneficial” for a linguistic system. If a language borrows elements of a morphological paradigm then this will result in suppletion (paradigmatic irregularity) but the increase in formal distinctiveness between the elements of a paradigm improves communication, especially if these elements often stand in contrast to each other. Clear cases of “beneficial suppletion” include the borrowing of th‐ pronominal forms from Old Norse into Old English (see discussion pp. 5–6). The increase of formal distinctiveness is no doubt behind the development of English she (of disputed origin) which contrasts with he in its initial consonant.

4  Terminology in Contact Studies In the discussion so far various terms have been used, e.g. borrowing, transfer, imposition. These are not always used with the same meanings by all authors, so it is advisable to offer some clarification. 1. Borrowing Items/structures are copied from language X to language Y, but without speakers of Y shifting to X. In this simple form, borrowing is characteristic of “cultural” contact, e.g. Latin and English in the history of the latter, or English and other European languages today. Such borrowings are almost exclusively confined to words and phrases. In terms of agency, it is the target language users who copy elements from the source language. 2. Transfer During language shift, when speakers of language X are switching to language Y, they transfer features of their original native language X to Y. Where these features are already present in Y the transfer is imperceptible but when they do not already exist in Y the transferred features represent an innovation. For grammatical structures one can distinguish (i) categories and (ii) their exponence (means of realization). Both (i) and (ii) can be transferred, or in some cases, only (i), see the discussion of habitual aspect in Irish English p. 13. a. Supportive transfer: A feature in language X is also found in language Y ensuring its continuation in the shift variety of language Y. Example: Irish has a distinction between second person singular and plural for pronouns. This fact meant that in Irish English this distinction, available in vernacular varieties, was supported and continues to this day. b. Innovative transfer: A feature in language X is not found in language Y so that its transfer constitutes an innovation in Y. Example: The immediate perfective of Irish was transferred to Irish English, e.g. They’re after sinking the boat, representing an addition to the aspectual distinctions already available. This type of transfer is also referred to as interference, especially in the context of guided second language acquisition where the evaluative connotation of “interference” is often intentional. 3. Imposition A community contains two groups: a minority with high status and a majority with lower status. The minority acquire the language of the majority, eventually relinquishing their original native language (a shift scenario). In the process the majority adopt features of the shift variety of the high‐status minority. Example: Anglo‐Norman and Irish in late medieval Ireland where the Anglo‐Normans “imposed” features of their shift variety of Irish on the majority native population (see discussion pp. 11–12). This situation may well have held for the influence of Anglo‐Norman on English during the early Middle English period, though the demographic distribution of the Anglo‐Normans among the local population was not as clear as it was in Ireland. 4. Metatypy Discussing a Melanesian case, Malcolm Ross (1996, 2001) coined the term “metatypy” to denote the sharing of organizational structures across languages in a

Language Contact and Linguistic Research  19 situation where social attitudes disfavor the replication of concrete word forms whose origin in another language is easily identifiable. Metatypy is what gives rise to a Sprachbund or language alliance, where two or more languages are in contact over a lengthy period and become structurally more and more similar, as has happened with diverse Indo‐European languages in the Balkans (Joseph 1983) and with Indo‐Aryan and Dravidian languages in India (Emeneau 1980). But what happens to languages during this growth in similarity and how this process occurs are less well known. It is often assumed that languages simply grow more ­similar to each other, converging on some kind of mean. However, almost all case studies show a one‐sided process: one language (the primary lect) adapts morphosyntactically to the constructions of another (the secondary lect), with no change occurring in the latter. (Ross 2003: 183)

5. Convergence A feature in language X has an internal source, i.e. there is a systemic motivation for the feature within language X, and the feature is present in a further language Y with which X is in contact. Both internal and external sources “converge” to produce the same result. Example: The progressive form in English. The two main views on this are: (a) it was an independent development in English (Visser 1963–73; Mitchell 1985) or (b) it results from contact with Celtic (Dal 1952; Preußler 1938; Wagner 1959; Braaten 1967). A type of progressive structure in which a gerund was governed by a preposition existed in Old English: ic wæs on huntunge ‘I was hunting’ (Braaten 1967: 173). The step from this structure to I was hunting is small, involving only the deletion of the preposition. The fully developed progressive form appears in Middle English, but the apparent time delay between the contact with Celtic in the Old English period and the surfacing of the progressive later can be accounted for by the strong tradition of the written standard in Old English (Dal 1952: 113). The progressive is found in all the Celtic languages and can be clearly recognized today in Irish, for example, Tá mé ag caint léi [is me at talking with‐her] ‘I am talking to her’. This in itself is a good example of a locative expression for progressive aspect and is typologically parallel to Old English ic wæs on huntunge. In both Old English and Celtic there was a progressive aspect, realized by means of a locative expression and with a similar functional range (Mittendorf and Poppe 2000: 139). Both languages maintained this aspect and English lost the locative preposition, increasing the syntactic flexibility and range of the structure, perhaps under the supportive influence of contact with Celtic.

4.1  Variation in Use of Terms Across the literature on language contact there is variation in terminology, a situation which is not unique to this field by any means. Some of this variation arises from the imprecision of original terms and some from differing uses of the same term. Consider the term “borrowing”: this is imprecise, i.e. nothing is “borrowed” from A to B, but the term is established in the field and its use ensures continuity with existing literature. On the other hand, the term “copying” (or “replication”) is in fact more accurate: speakers of language A copy features found in language B into their own language (the replicate the features). This usage is found in Johanson (2002, 2008) and in Johanson, Csató, and Karakoç as well as Pakendorf (both this volume). The term “convergence” is found to refer to the coming together of community internal and external factors to produce the same output, but the term can also be used to mean that two languages become more similar in structure, usually by one language approximating to the other (Ross 2001: 139). This latter development can occur without any internal motivation, i.e. change in one language results entirely from contact with another.

20  Raymond Hickey There is a use of “imposition” which goes back to van Coetsem (1988), continued by van Coetsem (2000), and taken up, most notably, by Donald Winford (2005, 2010) which is essentially the same as “transfer” as defined in (2) above, i.e. innovative transfer of features from language A to language B by speakers in a shift scenario (see the discussion in Salmons and Purnell, this volume, section 1.1). Van Coetsem’s use of the term “imposition” has two disadvantages: it denies the use of the intuitively more obvious and neutral term “transfer” and precludes the use of “imposition” in the sense under (3) above, for which there is strong sociolinguistic motivation. However, in other respects the approach initiated by van Coetsem has distinct merits, such as his highlighting of the relative linguistic dominance of languages and the direction of agency in contact situations. This can best be illustrated by an example. Consider the position of English and Spanish in the south‐west/west of the United States, especially in the large urban centers such as Los Angeles (Silva‐Corvalán 1994). First, Spanish influenced the English of the Chicanos, then after some time, the type of English they developed had a reverse influence back on their Spanish because for many their English has become more dominant (see the discussion in Fought 2010). The same is true for other large immigrant groups, especially in later generations, e.g. the Turks in Germany whose Turkish is now influenced by German (Kern 2015; Auer 2004) although for the first generation of immigrants in the 1950s and 1960s Turkish was the dominant language and this influenced the kind of German they spoke.

5 Conclusion There can be little doubt that the value of contact explanations in linguistics has increased with the greater level of differentiation and nuance provided in the many analyses presented in the last few decades (Thomason, this volume) and in the new compendia which have appeared, and in the analyses to be expected from those in preparation. It has ensured that, where linguists investigate language change and the rise of new features, the possibility of contact is scrutinized objectively. There should be no a priori preference for contact accounts nor should there be for purely language‐internal explanations either, and indeed the combination of external and internal developments should also be considered; in the latter scenario one can speak of a cline from contact‐facilitated to contact‐induced change. The amount of change deriving from contact can vary in the development of a language. The period of contact, its intensity and duration, and the social setting are all factors which need to be weighed up carefully. Furthermore, languages naturally continue to develop after the dust of contact has settled. Some contact features establish themselves, especially in varieties which derive from earlier language shift, but others recede if not accepted by the speech community in later generations. A balanced consideration of all these factors is essential in determining the effects of contact on a speech community and, in the long term, the language change which it engenders.

NOTES 1 There is also a dedicated journal published by Brill (Leiden): Journal of Language Contact, URL: www.//brill.com/view/journals/jlc 2 On contact and the Slavic languages, see also Grenoble (this volume). 3 See Zuckermann (2003) for another lexical investigation of contact, this time in the context of modern Hebrew. Drechsel (2014) deals with contact in the Pacific arena as do Makihara and Schieffelin (2007).

Language Contact and Linguistic Research  21 4 5 6 7 8 9

10 11 12 13 14

It should be said that it was preempted by some studies, e.g. Myers‐Scotton (ed., 1998). On the possible status of language shift varieties as a typological class, see Hickey (2020). This approach had been anticipated to a degree by Bisang (1998). Some scholars did consider contact with Celtic along with other possibilities when examining particular instances of change in the history of English. See Ellegård (1953). This work can involve using evidence from archaeology as well (Fortescue 1998). In the following sections the majority of examples are taken from languages in Ireland. The history of English and Irish is characterized by permanent contact and mutual influence so that many examples of different phenomena are attested (Hickey 2012b). Given that I know this material best, I have chosen to use it for illustrations. However, similar situations and influences can be found in other scenarios of contact for other languages. Compare the similar use in modern Greek of the word filipineza (female immigrant from the Philippines) for ‘maid, domestic servant’. The appearance of the word, first as a dictionary entry, created a storm of protest in 1998. Proto‐Semitic is generally regarded as having VSO (Hetzron 1987: 662). How this arose is unaccounted for. If fossilization sets in, this situation can in fact become permanent. For discussions of features found exclusively in varieties of English, see the contributions in Hickey (ed., 2012). There are only a few of the old masculines left, e.g. der Friede ‘peace’, der Same ‘seed’, der Name ’name’.

REFERENCES Abdulaziz, Mohamed H. and Ken Osinde 1997. Sheng and Engsh: development of mixed codes among the urban youth in Kenya. International Journal of the Sociology of Language 125: 43–63. Aboh, Enoch O. 2015. The Emergence of Hybrid Grammars. Language Contact and Change. Cambridge: Cambridge University Press. Aboh, Enoch O. (ed.) 2017. Complexity in human languages: a multifaceted approach. Special issue of Language Sciences, Vol. 60. Adamou, Evangelia 2016. A Corpus‐Driven Approach to Language Contact. Endangered Languages in a Comparative Perspective. Berlin: de Gruyter Mouton. Adamou, Evangelina and Yaron Matras (eds.) forthcoming. The Routledge Handbook of Language Contact. London/New York: Routledge. Aikhenvald, Alexandra Y. 2002a. Language Contact in Amazonia. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. 2002b. Mechanisms of change in areal diffusion: new morphology and language contact. Journal of Linguistics 39: 1–29. Aikhenvald, Alexandra Y. and Robert M. W. Dixon (eds.) 2001. Areal Diffusion and Genetic Inheritance. Oxford: Oxford University Press.

Aikhenvald, Alexandra Y. and Robert M. W. Dixon (eds.) 2003. Studies in Evidentiality. Amsterdam: Benjamins. Aikhenvald, Alexandra Y. and Robert M. W. Dixon (eds.) 2006. Grammars in Contact: A Cross‐linguistic Typology. Oxford: Oxford University Press. Anchimbe, Eric A. (ed.) 2012. Language Contact in a Postcolonial Setting. The Linguistic and Social Context of English and Pidgin in Cameroon. Berlin: de Gruyter Mouton. Andersen, Henning (ed.) 2003. Language Contacts in Prehistory. Studies in Stratigraphy. Amsterdam: Benjamins. Ansaldo, Umberto 2009. Contact Languages: Ecology and Evolution in Asia. Cambridge: Cambridge University Press. Askedal, John Ole and Hans Frede Nielsen (eds.) 2015. Early Germanic Languages in Contact. Amsterdam: Benjamins. Auer, Peter 2004. Türkisch sprechen nicht nur die Türken: über die Unschärfebeziehung zwischen Sprache und Ethnie in Deutschland. Berlin: de Gruyter Mouton. Baker, Philip and Magnus Huber 2001. Atlantic, Pacific, and world‐wide features in English‐lexicon contact languages. English World‐Wide 22: 157–208. Bakker, Peter and Yaron Matras (eds.) 2013. Contact Languages. A Comprehensive Guide. Berlin: de Gruyter Mouton.

22  Raymond Hickey Bakker, Peter and Maarten Mous (eds.) 1994. Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam: Institute for Functional Research into Language and Language Use (IFOTT). Bao, Zhiming 2005. The aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics 41: 237–67. Bao, Zhiming 2010. A usage‐based approach to substratum transfer: The case of four unproductive features in Singaporean Englsh. Language 86.4: 792–820. Bastardas‐Boada, Albert 2019. Complexity and language contact: a socio‐cognitive framework. In Salikoko Mufwene, Christophe Coupé, and François Pellegrino (eds.), Complexity in Language Developmental and Evolutionary Perspectives. Cambridge: Cambridge University Press, pp. 218–44. Berez‐Kroeker, Andrea L., Diane Hintz, and Carmen Jany (eds) 2016. Language Contact and Change in the Americas: Studies in Honor of Marianne Mithun. Amsterdam: Benjamins. Bergs, Alexander T. 2006. Spreading the word: patterns of diffusion in historical dialectology. In Markku Filppula, Juhani Klemola, Marjatta Palander, and Esa Penttilä (eds.), Topics in Dialect Variation. Joensuu: University Press, pp. 5–30. Bisang, Walter 1998. Grammaticalisation and language contact, constructions and positions. In Anna Giacalone Ramat and Paul J. Hopper (eds.), The Limits of Grammaticalization. Amsterdam: Benjamins, pp. 13–58. Blommaert, Jan 2010. The Sociolinguistics of Globalization. Cambridge University Press. Boas, Hans C. and Steffen Höder (eds.) 2018. Constructions in Contact. Constructional Perspectives on Contact Phenomena in Germanic Languages. Amsterdam: Benjamins. Bolton, Kingsley 2003. Chinese Englishes. Cambridge: Cambridge University Press. Braaten, Björn 1967. Notes on continuous tenses in English. Norsk Tidskrift for Sprogvidenskap 21: 167–80. Braunmüller, Kurt, Steffen Höder, and Karoline Kühl (eds.) 2014. Stability and Divergence in Language Contact. Factors and Mechanisms. Amsterdam: Benjamins. Campbell, Lyle 2013. Historical Linguistics: An Introduction, Third edition. Edinburgh: Edinburgh University Press. Cardoso, Hugo C., Alan N. Baxter, and Mário Pinharanda Nunes (eds.) 2012. Ibero‐Asian Creoles. Comparative Perspectives. Amsterdam: Benjamins.

Chambers, J. K. 2004. Dynamic typology and vernacular universals. In Bernd Kortmann (ed.), Dialectology Meets Typology. Dialect Grammar from a Cross‐Linguistic Perspective. Berlin: de Gruyter Mouton, pp. 127–45. Chamoreau, Claudine and Isabelle Léglise (eds.) 2012. Dynamics of Contact‐Induced Language Change. Berlin: de Gruyter Mouton. Clements, J. Clancy and Shelome Gooden (eds.) 2011. Language Change in Contact Languages. Grammatical and Prosodic Considerations. Amsterdam: Benjamins. Clyne, Michael 2003. Dynamics of Language Contact: English and Immigrant Languages. Cambridge: Cambridge University Press. Comrie, Bernard 2008. Inflectional morphology and language contact, with special reference to mixed languages, In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages. Amsterdam: Benjamins, pp. 15–32. Dahl, Östen and Maria Koptjevskaja‐Tamm (eds.) 2001. The Circum‐Baltic Languages: Typology and Contact. Amsterdam: Benjamins. Dakin, Karen, Claudia Parodi and Natalie Operstein (eds) 2017. Language Contact and Change in Mesoamerica and Beyond. Amsterdam: Benjamins. Dal, Ingerid 1952. Zur Entstehung des Englischen Participium Praesentis auf ‐ing. Norsk Tidskrift for Sprogvidenskap 16: 5–116. Darquennes, Jeroen, Joseph C. Salmons, and Wim Vandenbussche 2019. In Jeroen Darquennes, Joseph C. Salmons, and Wim Vandenbussche (eds.), Language Contact. An International Handbook. Vol 1. Berlin: de Gruyter Mouton, pp. 1–12. DeGraff, Michel 2003. Against creole exceptionalism. Language 79: 391– 410. Deumert, Ana 2014. Sociolinguistics and Mobile Communication. Edinburgh: Edinburgh University Press. Deumert, Ana and Stephanie Durlemann (eds.) 2006. Structure and Variation in Language Contact. Amsterdam: Benjamins. Doron, Edit, Malka Rappaport Hovav, Yael Reshef, Moshe Taube (eds) 2019. Language Contact, Continuity and Change in the Genesis of Modern Hebrew. Amsterdam: John Benjamins. Dorian, Nancy C. (ed.) 1989. Investigating Obsolescence: Studies in Language Contraction and Death. Cambridge: Cambridge University Press. Dorian, Nancy C. 1993. Internally and externally motivated change in contact situations: doubts about dichotomy. In Charles Jones (ed.),

Language Contact and Linguistic Research  23 Historical Linguistics: Problems and Perspectives. London: Longman, pp. 131–55. Drechsel, Emanuel J. 2014. Language Contact in the Early Colonial Pacific. Maritime Polynesian Pidgin before Pidgin English. Cambridge: Cambridge University Press. Drinka, Bridget 2017. Language Contact in Europe. The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Duarte, Joana and Ingrid Gogolin (eds.) 2013. Linguistic Superdiversity in Urban Areas. Research Approaches. Amsterdam: Benjamins. Durkin, Philip 2014. Borrowed Words: A History of Loanwords in English. Oxford: Oxford University Press. Dutton, Thomas E. and Darrell T. Tryon (eds.) 1994. Language Contact and Change in the Austronesian World. Berlin: de Gruyter Mouton. Ellegård, Alvar 1953. The Auxiliary Do: The Establishment and Regulation of Its Use in English. Gothenburg Studies in English 2. Stockholm: Almqvist and Wiksell. Emeneau, Murray B. 1980. Language and Linguistic Areas, edited by Anwar S. Dil. Stanford: Stanford University Press. Escure, Genevieve and Armin Schwegler (eds.) 2004. Creoles, Contact and Language Change: Linguistic and Social Implications. Amsterdam: Benjamins. Fenyvesi, Anna (ed.) 2005. Hungarian Language Contact Outside Hungary. Studies on Hungarian as a Minority Language. Amsterdam: Benjamins. Field, Frederic W. 2002. Linguistic Borrowing in Bilingual Contexts. Amsterdam: Benjamins. Filppula, Markku 2003. The quest for the most “parsimonious” explanations: endogeny vs. contact revisited. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 161–73. Filppula, Markku 2010. Contact and the early history of English. In Raymond Hickey (ed.) Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 432–53. Filppula, Markku and Juhani Klemola (eds.) 2009. Re‐Evaluating the Celtic Hypothesis. Special issue of English Language and Linguistics, 13.2. Filppula, Markku and Juhani Klemola 2014. Celtic influences in English: a re‐evaluation Neuphilologische Mitteilungen 115.1: 33–53. Filppula, Markku, Juhani Klemola, and Heli Paulasto 2008. English and Celtic in Contact. London: Routledge. Filppula, Markku, Juhani Klemola, and Heli Paulasto 2009a. Digging for roots: universals and contact in regional varieties of English. In

Markku Filppula, Juhani Klemola, and Heli Paulasto (eds.), Vernacular Universals and Language Contacts: Evidence from Varieties of English and Beyond. London: Routledge, pp. 231–63. Filppula, Markku, Juhani Klemola, and Heli Pitkänen (eds.) 2002. The Celtic Roots of English. Studies in Language 37. University of Joensuu: Faculty of Humanities. Fisiak, Jacek (ed.) 1995. Language Change under Contact Conditions. Berlin: de Gruyter Mouton de Gruyter. Fortescue, Michael 1998. Language Relations across Bering Strait: Reappraising the Archaeological and Linguistic Evidence. London and New York: Cassell. Fought, Carmen 2010. Ethnic identity and linguistic contact. In Raymond Hickey (ed.) Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 282–98. Gilbers, Dicky G., John Nerbonne, and Jos Schaeken (eds.) 2000. Languages in Contact. Amsterdam: Rodopi. Goebl, Hans, Peter H. Nelde, Zdenek Stary, and Wolfgang Wölck (eds.) 1996. Kontaktlinguistik/ Contact Linguistics/Linguistique de Contact, 2 vols. Berlin: de Gruyter Mouton. Gold, Elaine 2009. The case of Bungi: evidence for vernacular universals. In Markku Filppula, Juhani Klemola, and Heli Paulasto (eds.), Vernacular Universals and Language Contacts: Evidence from Varieties of English and Beyond. London: Routledge, pp. 163–76. Grant, Anthony 2019. Contact‐induced language change: an introduction. In Anthony Grant (ed.), The Oxford Handbook of Language Contact. Oxford: Oxford University Press, pp. 1–50. Grant, Anthony (ed.) 2019. The Oxford Handbook of Language Contact. Oxford: Oxford University Press. Harris, John 1984. Syntactic variation and dialect divergence. Journal of Linguistics 20: 303–27. Harrison, K. David 2007. When Languages Die: The Extinction of the World’s Languages and the Erosion of Human Knowledge. Oxford: Oxford University Press. Haspelmath, Martin 2001. The European linguistic area: Standard Average European. In Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals. Berlin: de Gruyter Mouton, pp. 1492–510. Haspelmath, Martin 2004. How hopeless is genealogical linguistics, and how advanced is areal linguistics? (Review article of Alexandra

24  Raymond Hickey Y. Aikhenvald and Robert M. W. Dixon (eds.) 2001. Areal Diffusion and Genetic Inheritance. Oxford: Oxford University Press.) Studies in Language 28.1: 209–23. Haspelmath, Martin, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.) 2001. Language Typology and Language Universals. Berlin: de Gruyter Mouton. Haspelmath, Martin and Tadmor, Uri (eds.) 2009. Loanwords in the World’s Languages : A Comparative Handbook. Berlin: de Gruyter Mouton. Heath, Jeffrey 1984. Language contact and language change. Annual Review of Anthropology 13: 367–84. Heine, Bernd 2008. Contact‐induced word order change without word order change. In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages. Amsterdam: Benjamins, pp. 33–60. Heine, Bernd and Tania Kuteva 2003. On contact‐induced grammaticalization. Studies in Language 27: 529–72. Heine, Bernd and Tania Kuteva 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Hetzron, Robert 1987. Semitic languages. In Bernard Comrie (ed.), The World’s Major Languages. London: Croom Helm, pp. 645–63. Hickey, Raymond 1995a. An assessment of language contact in the development of Irish English. In Jacek Fisiak (ed.), Language Change under Contact Conditions. Berlin: de Gruyter Mouton, pp. 109–30. Hickey, Raymond 1995b. Sound change and typological shift: initial mutation in Celtic. In Jacek Fisiak (ed.), Linguistic Typology and Reconstruction. Berlin: de Gruyter Mouton, pp. 133–82. Hickey, Raymond 1997a. Arguments for creolisation in Irish English. In Raymond Hickey and Stanislaw Puppel (eds.), Language History and Linguistic Modelling: A Festschrift for Jacek Fisiak on his 60th Birthday. Berlin: de Gruyter Mouton, pp. 969–1038. Hickey, Raymond 1997b. Assessing the relative status of languages in medieval Ireland. In Jacek Fisiak (ed.), Studies in Middle English Linguistics. Berlin: de Gruyter Mouton, pp. 181–205. Hickey, Raymond 2001. Language contact and typological difference: transfer between Irish and Irish English. In Dieter Kastovsky and Arthur Mettinger (eds.), Language Contact and the History of English. Frankfurt‐am‐Main: Peter Lang, pp. 131–69.

Hickey, Raymond 2002a. Internal and external factors again: word order change in Old English and Old Irish. In Raymond Hickey (ed.), Collecting Views on Language Change. Special issue of Language Sciences 24.3–4: 261–83. Hickey, Raymond 2002b. Language change in early Britain: the convergence account. In David Restle and Dietmar Zaefferer (eds.), Sounds and Systems: Studies in Structure and Change. A Festschrift for Theo Vennemann. Berlin: de Gruyter Mouton, pp. 185–203. Hickey, Raymond 2003a. Reanalysis and typological change. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 258–78. Hickey, Raymond 2003b. Rectifying a standard deficiency: pronominal distinctions in varieties of English. In Irma Taavitsainen and Andreas H. Jucker (eds.), Diachronic Perspectives on Address Term Systems. Amsterdam: Benjamins, pp. 345–74. Hickey, Raymond 2005. Dublin English: Evolution and Change. Amsterdam: Benjamins. Hickey, Raymond 2007. Irish English: History and Present‐Day Forms. Cambridge: Cambridge University Press. Hickey, Raymond (ed.) 2010. Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell. Hickey, Raymond 2012a. Early English and the Celtic hypothesis. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 497–507. Hickey, Raymond 2012b. Assessing the role of contact in the history of English. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 485–96. Hickey, Raymond 2012c. Internally and externally motivated language change. In Juan Manuel Hernández‐Compoy and Juan Camilo Conde‐Silvestre (eds.), The Handbook of Historical Sociolinguistics. Malden, MA: Wiley‐ Blackwell, pp. 401–21. Hickey, Raymond (ed.) 2012. Areal Features of the Anglophone World. Berlin: de Gruyter Mouton. Hickey, Raymond 2013a. Supraregionalisation and dissociation. In J. K. Chambers and Natalie Schilling (eds.), Handbook of Language Variation and Change, 2nd edition. Malden, MA: Wiley‐Blackwell, pp. 537–54. Hickey, Raymond 2013b. English as a contact language in Ireland and Scotland. In Daniel Schreier and Marianne Hundt (eds.) English as

Language Contact and Linguistic Research  25 a Contact Language. Cambridge: Cambridge University Press, pp. 88–105. Hickey, Raymond 2017. Areas, areal features and areality. In Raymond Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 1–15. Hickey, Raymond (ed.) 2017. The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press. Hickey, Raymond 2020. Shift varieties as a typological class? A consideration of South African Indian English. In Raymond Hickey (ed.), English in Multilingual South Africa. The Linguistics of Contact and Change. Cambridge: Cambridge University Press, pp. 265–87. Hinrichs, Lars and Joseph T. Farquharson (eds.) 2011. Variation in the Caribbean: From Creole Continua to Individual Agency. Amsterdam: John Benjamins. Holm, John 2004. Languages in Contact: The Partial Restructuring of Vernaculars. Cambridge: Cambridge University Press. Holm, John 2010. Contact and change: pidgins and creoles. In Raymond Hickey, Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell , pp. 252–61. Holthausen, Friedrich 1974. Altenglisches etymologisches Wörterbuch, 2nd edition. Heidelberg: Winter. Hopper, Paul J. and Elizabeth C. Traugott 2003 [1993]. Grammaticalization, 2nd edition. Cambridge: Cambridge University Press. Huber, Magnus and Viveka Velupillai (eds.) 2007. Synchronic and Diachronic Perspectives on Contact Languages. Amsterdam: Benjamins. Hundt, Marianne and Daniel Schreier (eds.) 2013. English as a Contact Language. Cambridge: Cambridge University Press. Hundt, Marianne and Devyani Sharma (eds.) 2014. English in the India Diaspora. Amsterdam: Benjamins. Isaac, Graham 2003. Diagnosing the symptoms of contact: some Celtic‐English case histories. In Hildegard L. C. Tristram (ed.), The Celtic Englishes III. Heidelberg: Carl Winter, pp. 46–64. Jacobs, Neil G. 2005. Yiddish: A Linguistic Introduction. Cambridge: Cambridge University Press. Johanson, Lars 2002. Structural Factors in Turkic Language Contacts. London: Curzon. Johanson, Lars 2008. Remodeling grammar: copying, conventionalization, grammaticalization. In Peter Siemund and Noemi Kintana (eds.), Language Contact and

Contact Languages. Amsterdam: Benjamins, pp. 61–79. Jones, Mari C. and Edith Esch (eds.) 2002. Language Change: The Interplay of Internal, External and Extra‐linguistic Factors. Berlin: de Gruyter Mouton. Joseph, Brian D. 1983. The Balkan Infinitive. Cambridge: Cambridge University Press. Joseph, Brian D. 2003. Evidentials: summation, questions, prospects. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Studies in Evidentiality. Amsterdam: Benjamins, pp. 307–27. Kastovsky, Dieter and Arthur Mettinger (eds.) 2001. Language Contact in the History of English. Berlin: de Gruyter Mouton. Kern, Friederike 2015. Turkish German. Language and Linguistics Compass. https://doi. org/10.1111/lnc3.12136 King, Jeremy and Sandro Sessarego (eds.) 2017. Language Variation and Contact‐Induced Change. Spanish Across Space and Time. Amsterdam: Benjamins. King, Ruth 2000. The Lexical Basis of Grammatical Borrowing: A Prince Edward Island Case Study. Amsterdam: Benjamins. Klemola, Juhani 2000. The origins of the Northern Subject Rule: a case of early contact? In Hildegard L. C. Tristram (ed.), The Celtic Englishes II. Heidelberg: Carl Winter, pp. 329–46. Kluge, Friedrich and Elmar Seebold 2002. Etymologisches Wörterbuch der deutschen Sprache, 24th edn. Berlin: de Gruyter. Kortmann, Bernd and Benedikt Szmrecsanyi (eds.) 2012. Linguistic Complexity: Second Language Acquisition, Indigenization, Contact. Berlin and New York: de Gruyter de Gruyter Mouton. Lass, Roger and Susan Wright 1986. Endogeny vs. contact: Afrikaans influence on South African English. English World‐Wide 7: 201–23. Law, Danny 2014. Language Contact, Inherited Similarity and Social Difference. Amsterdam: Benjamins. Léglise, Isabelle and Claudine Chamoreau. 2013. Variation and change in contact settings. In Isabelle Léglise and Claudine Chamoreau (eds.), The Interplay of Variation and Change in Contact Settings. Amsterdam: Benjamins, pp.1–20. Loveday, Leo J. 1996. Language Contact in Japan: A Socio‐linguistic History. Oxford: Clarendon Press. Lucas, Christopher 2012. Contact‐induced grammatical change. Towards an explicit account. Diachronica 29.3: 275–300.

26  Raymond Hickey Ludwig, Ralph, Steve Pagel, and Peter Mühlhäusler 2018. Linguistic Ecology and Language Contact. Cambridge: Cambridge University Press. Mackenzie, Ian 2017. Language Contact and the Future of English. London: Routledge. Maitz, Péter and Attila Németh 2014. Language contact and morphosyntactic complexity: evidence from German, Journal of Germanic Linguistics 26.1: 1–29. Makihara, Miki and Bambi B. Schieffelin 2007. Consequences of Contact. Language Ideologies and Sociocultural Transformations in Pacific Societies. Oxford: Oxford University Press. Manfredi, Stefano and Mauro Tosco (eds.) 2017. Arabic in Contact. Amsterdam: Benjamins. Matisoff, James. 2001. Genetic versus contact relationship: prosodic diffusibility in South‐ East Asian languages. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance. Oxford: Oxford University Press, pp. 291–327. Matras, Yaron (ed.) 1995. Romani in Contact: The History, Structure and Sociology of a Language. Amsterdam: Benjamins. Matras, Yaron 1998. Utterance modifiers and universals of grammatical borrowing. Linguistics 36.2: 281–331. Matras, Yaron 2002. Romani: A Linguistic Introduction. Cambridge: Cambridge University Press. Matras, Yaron 2009. Language Contact. Cambridge: Cambridge University Press. Matras, Yaron 2010. Contact, convergence, and typology. In Raymond Hickey, (ed.), Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 66–85. Matras, Yaron, April McMahon, and Nigel Vincent (eds.) 2006. Linguistic Areas: Convergence in Historical and Typological Perspective. Basingstoke: Palgrave Macmillan. Matthews, Stephen 2010. Language contact and Chinese. In Raymond Hickey (ed.) Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 757–69. McCafferty, Kevin 2004. Innovation in language contact: be after V‐ing as a future gram in Irish English, 1670 to the present. Diachronica 21.1: 113–60. McColl Millar, Robert 2007. Trask’s Historical Linguistics, revised 2nd edition. Oxford: Oxford University Press. McColl Millar, Robert 2016. Contact. The Integration of Closely Related Linguistic Varieties and the History of English. Edinburgh: Edinburgh University Press.

McLendon, Sally 2003. Evidentials in Eastern Pomo with a comparative survey of the category in other Pomoan languages. In Alexandra Y. Aikhenvald, and Robert M. W. Dixon (eds.), Studies in Evidentiality. Amsterdam: Benjamins, pp. 101–29. McMahon, April 1994. Understanding Language Change. Cambridge: Cambridge University Press. McWhorter, John H. (ed.) 2000. Language Change and Language Contact in Pidgins and Creoles. Amsterdam: Benjamins. McWhorter, John H. 2011. Linguistic Simplicity and Complexity: Why do Languages Undress? Berlin: de Gruyter Mouton. Mesthrie, Rajend 1992. English in Language Shift: The History, Structure and Sociolinguistics of South African Indian English. Cambridge: Cambridge University Press. Mesthrie, Rajend and Timothy T. Dunne 1990. Syntactic variation in language shift: the relative clause in South African Indian English. Language Variation and Change 2.1: 31–56. Miestamo, Matti, Kaius Sinnemäki, and Fred Karlsson (eds.) 2008. Language Complexity: Typology, Contact, Change. Amsterdam: Benjamins. Migge, Bettina 2003. Creole Formation as Language Contact. Amsterdam: Benjamins. Mitchell, Bruce 1985. Old English Syntax, 2 vols. Oxford: Clarendon Press. Mittendorf, Ingo and Erich Poppe 2000. Celtic contacts of the English progressive? In Hildegard L. C. Tristram (ed.), The Celtic Englishes II. Heidelberg: Carl Winter, pp. 117–45. Moravcsik, Edith A. 1978. Language contact. In Joseph H. Greenberg (ed.), Universals of Human Language, vol. 1: Method and Theory. Stanford: Stanford University Press, pp. 94–122. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Mufwene, Salikoko S. 2007. Population movements and contacts: competition, selection, and language evolution. Journal of Language Contact 1: 63–91. Mufwene, Salikoko. S. 2008. Language Evolution: Contact, Competition and Change. London: Continuum. Mufwene, Salikoko and Anaa María Escobar (eds.) forthcoming. The Cambridge Handbook of Language Contact. Cambridge: Cambridge University Press. Mufwene, Salikoko, Christophe Coupé, and François Pellegrino (eds.) 2019. Complexity in

Language Contact and Linguistic Research  27 Language Developmental and Evolutionary Perspectives. Cambridge: Cambridge University Press. Muysken, Pieter 1981. Halfway between Quechua and Spanish: the case for relexification. In Arnold Highfield and Albert Valdman (eds.), Historicity and Variation in Creole Studies. Ann Arbor: Karoma, pp. 52–78. Muysken, Pieter 1997. Media lengua. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 365–426. Muysken, Pieter 2000. Bilingual Speech: A Typology of Code‐Mixing. Cambridge: Cambridge University Press. Muysken, Pieter (ed.) 2008. From Linguistic Areas to Areal Linguistics. Amsterdam: Benjamins. Myers‐Scotton, Carol (ed.) 1998. Codes and Consequences. Choosing Linguistic Varieties. Oxford: Oxford University Press. Myers‐Scotton, Carol 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Nettle, Daniel and Suzanne Romaine 2000. Vanishing Voice: The Extinction of the World’s Languages. New York: Oxford University Press. Nichols, Johanna 1992. Linguistic Diversity in Space and Time. Chicago: University of Chicago Press. Nichols, Johanna 2010. Macrofamilies, macroareas, and contact. In Raymond Hickey (ed.), Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 361–79. Noonan, Michael 2010. Genetic classification and language contact. In Raymond Hickey (ed.), Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 48–65. Norde, Muriel, Bob de Jonge, and Cornelius Hasselblatt (eds.) 2010. Language Contact. New perspectives. Amsterdam: Benjamins. Orozco, Rafael 2018. Spanish in Colombia and New York City. Language Contact Meets Dialectal Convergence. Amsterdam: John Benjamins. Ortiz López, Luis A., Rosa E. Guzzardo Tamargo and Melvin González‐Rivera (eds.) 2020. Hispanic Contact Linguistics. Theoretical, Methodological and Empirical Perspectives. Amsterdam: Benjamins. O’Shannessy, Carmel 2005. Light Warlpiri: a new language. Australian Journal of Linguistics 25.1: 31–57. Otheguy, Ricardo and Ana Celia Zentella 2012. Spanish in New York. Language Contact, Dialectal Leveling, and Structural Continuity. Oxford: Oxford University Press. Pokorny, Julius 1949. Zum nicht‐ indogermanischen Substrat im Inselkeltischen. Die Sprache 1: 235–45.

Poplack, Shana, Lauren Zentz, and Nathalie Dion 2012. What counts as (contact‐induced) change. Bilingualism: Language and Cognition 15: 247–54. Poplack, Shana 2018. Borrowing. Loanwords in the speech community and in the grammar. New York: Oxford University Press. Potowski, Kim and Richard Cameron (eds.) 2007. Spanish in Contact: Policy, Social and Linguistic Inquiries. Amsterdam: Benjamins. Poussa, Patricia 1990. A contact‐universal origin for periphrastic do with special consideration of Old English‐Celtic contact. In Sylvia Adamson et al. (eds.), Papers from the Fifth International Conference on English Historical Linguistics. Amsterdam: Benjamins, pp. 407–34. Preußler, Walter 1938. Keltischer Einfluß im Englischen. Indogermanische Forschungen 56: 178–91. Restle, David and Dietmar Zaefferer (eds.) 2002. Sounds and Systems: Studies in Structure and Change. A Festschrift for Theo Vennemann. Berlin: de Gruyter Mouton. Rickford, John R. 1986. Social contact and linguistic diffusion: Hiberno‐English and New World Black English. Language 62: 245–90. Rockel, Martin 1989. Grundzüge einer Geschichte der irischen Sprache. Vienna: Austrian Academy of Sciences. Ross, Malcolm D. 1996. Contact‐induced change and the comparative method: cases from Papua New Guinea. In Mark Durie and Malcolm D. Ross (eds.), The Comparative Method Reviewed: Regularity and Irregularity in Language Change. Oxford: Oxford University Press, pp. 180–217. Ross, Malcolm D. 2001. Contact‐induced change in Oceanic languages in North‐West Melanesia. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance. Oxford: Oxford University Press, pp. 134–66. Ross, Malcolm D. 2003. Diagnosing prehistoric language contact. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 174–98. Salmons, Joseph C. and Brian D. Joseph (eds.) 1998. Nostratic: Sifting the Evidence. Amsterdam: Benjamins. Sampson, Geoffrey, David Gil, and Peter Trudgill (eds.) 2009. Language Complexity as an Evolving Variable. Oxford: Oxford University Press. Sankoff, Gillian. 2002. Linguistic outcomes of language contact. In J. K. Chambers, Peter

28  Raymond Hickey Trudgill, and Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. Oxford: Blackwell, pp. 638–68. Sayahi, Lotfi 2014. Diglossia and Language Contact. Language Variation and Change in North Africa. Cambridge: Cambridge University Press. Schiffman, Harold 2010. Language contact in South Asia. In Raymond Hickey, (ed.), Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 738–56. Schneider, Edgar W. 2014. New reflections on the evolutionary dynamics of world Englishes. World Englishes 33.1: 9–32. Schreier, Daniel and Marianne Hundt (eds.) 2013. English as a Contact Language. Cambridge: Cambridge University Press. Schwägerl, Christian 2010. Language Contact and Displays of Social Identity. Tübingen: Narr. Seifart, Frank 2015. Direct and indirect affix borrowing. Language 91.3: 511–31. Seifart, Frank 2017. Patterns of affix borrowing in a sample of 100 languages. Journal of Historical Linguistics 7.3: 389–431. Seifart, Frank 2019. Contact‐induced change. In Jeroen Darquennes, Joseph C. Salmons, and Wim Vandenbussche (eds.), Language Contact. An International Handbook. Vol 1. Berlin: de Gruyter Mouton, pp. 13–23. Sessarego, Sandro 2019. Language Contact and the Making of an Afro‐Hispanic Vernacular. Variation and Change in the Colombian Chocó. Cambridge: Cambridge University Press. Sessarego, Sandro and Melvin González‐Rivera (eds.) 2015. New Perspectives on Hispanic Contact Linguistics in the Americas. Madrid: Iberoamericana Vervuert. Siegel, Jeff 1987. Language Contact in a Plantation Environment. Cambridge: Cambridge University Press. Siegel, Jeff 1999. Transfer constraints and substrate influence in Melanesian Pidgin. Journal of Pidgin and Creole Languages 14: 1–44. Siegel, Jeff 2000a. Substrate influence in Hawai‘i Creole English. Language in Society 29: 197–236. Siegel, Jeff (ed.) 2000b. Processes of Language Contact: Studies from Australia and the South Pacific. Quebec: Fides. Siegel, Jeff 2008. The Emergence of Pidgin and Creole Languages. Oxford: Oxford University Press. Siemund, Peter 2008. Language contact: constraints and common paths of contact‐ induced language change. In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages. Amsterdam: Benjamins, pp. 3–11.

Siemund, Peter and Noemi Kintana (eds.) 2008. Language Contact and Contact Languages. Amsterdam: Benjamins. Silva‐Corvalán, Carmen 1994. Language Contact and Change: Spanish in Los Angeles. Oxford: Clarendon Press. Thomason, Sarah G. 1997a. A typology of contact languages. In Arthur K. Spears and Donald Winford (eds.), The Structure and Status of Pidgins and Creoles. Amsterdam: Benjamins, pp. 71–88. Thomason, Sarah G. (ed.) 1997b. Contact Languages: A Wider Perspective. Amsterdam: Benjamins. Thomason, Sarah G. 2000. Contact as a source of language change. In Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics. Oxford: Blackwell, pp. 687–712. Thomason, Sarah G. 2001. Language Contact: An Introduction. Edinburgh: Edinburgh University Press. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Torres Cacoullos, Rena (ed.) 2015. Gauging Convergence on the Ground: Code‐switching in the Community. Special Issue of the International Journal of Bilingualism 19.4. Torres Cacoullos, Rena and Catherine E. Travis. 2016. Two languages, one effect: structural priming in spontaneous code‐switching. Bilingualism: Language and Cognition 19: 733–53. Torres Cacoullos, Rena and Catherine E. Travis 2018. Bilingualism in the Community. Code‐ switching and Grammars in Contact. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs and Bernd Heine (eds.) 2001. Approaches to Grammaticalization, 2 vols. Amsterdam: Benjamins. Trudgill, Peter 1986. Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 2010. Contact and sociolinguistic typology. In Raymond Hickey (ed.), Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 299–319. Trudgill, Peter 2011a. Sociolinguistic Typology. Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. Trudgill, Peter 2011b. A tale of two copulas: language speculations on first‐millennium England. Language and Literacy in Early Scandinavia and Beyond, vol. 62/63. Odense: University Press of Southern Denmark, pp. 285–320.

Language Contact and Linguistic Research  29 Ureland, P. Sture and George Broderick (eds.) 1991. Language Contact in the British Isles: Proceedings of the Eighth International Symposium on Language Contact in Europe. Tübingen: Niemeyer. van Coetsem, Frans 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. van Coetsem, Frans 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Winter. Vennemann, Theo 2001. Atlantis Semitica. Structural contact features in Celtic and English. In Laurel Brinton (ed.), Historical Linguistics 1999. Amsterdam: John Benjamins, pp. 351–69. Vennemann, Theo 2002a. On the rise of “Celtic” syntax in Middle English. In Peter J. and Angela M. Lucas (eds.), Middle English: From tongue to text. Selected papers from the Third International Conference on Middle English: Language and Text held at Dublin, Ireland, 1–4 July 1999. Frankfurt: Lang. Vennemann, Theo 2002b. Semitic > Celtic > English: the transitivity of language contact. In Markku Filppula, Juhani Klemola, and Heli Pitkänen (eds.), The Celtic Roots of English. Studies in Language, 37. University of Joensuu: Faculty of Humanities, pp. 295–330. Vennemann, Theo 2011. English as a contact language. Typology and comparison. Anglia 129: 217–57. Velupillai, Viveka 2015. Pidgins, Creoles and Mixed Languages. An Introduction. Amsterdam: Benjamins. Visser, Federicus T. 1963–73. An Historical Syntax of the English Language, 4 vols. Leiden: Brill. Wagner, Heinrich 1959. Das Verbum in den Sprachen der britischen Inseln. Tübingen: Niemeyer.

Weinreich, Uriel 1953. Languages in Contact: Findings and Problems. The Hague: de Gruyter Mouton. Werner, Ottmar 1991. The incorporation of Old Norse pronouns into Middle English: suppletion by loan. In P. Sture Ureland and George Broderick (eds.), Language Contact in the British Isles. Tübingen: Niemeyer, pp. 369–401. Wiemer, Björn, Bernhard Wälchli, and Björn Hansen (eds.) 2012. Grammatical Replication and Borrowability in Language Contact. Berlin: de Gruyter Mouton. Winford, Donald. 2003. An Introduction to Contact Linguistics. Oxford: Blackwell. Winford, Donald. 2005. Contact‐induced changes: classification and processes. Diachronica 22.2: 373–427. Winford, Donald 2008. Processes of creole formation and related contact‐induced language change. Journal of Language Contact. Available at www.jlc‐journal.org. Winford, Donald 2009. The interplay of “universals” and contact‐induced change in the emergence of New Englishes. In Markku Filppula, Juhani Klemola, and Heli Paulasto (eds.), Vernacular Universals and Language Contacts: Evidence from Varieties of English and Beyond. London: Routledge, pp. 206–30. Winford, Donald 2010. Contact and borrowing. In Raymond Hickey (ed.), Handbook of Language Contact, 1st edition. Malden, MA: Wiley‐Blackwell, pp. 170–87. Wolfram, Walt and Natalie Schilling 2013. American English. Dialects and Variation, Third edition. Malden, MA: Wiley‐Blackwell. Zuckermann, Ghil’ad 2003. Language Contact and Lexical Enrichment in Israeli Hebrew. Basingstoke: Palgrave Macmillan.

Part I – Contact, Contact Studies, and Linguistics

1 Contact Explanations in Linguistics SARAH THOMASON

Language contact has been invoked with increasing frequency over the past two or three decades as a, or the, cause of a wide range of linguistic changes. Historical linguists have (of course) mainly addressed these changes from a diachronic perspective – that is, analyzing ways in which language contact has influenced lexical and/or structural developments over time. But sociolinguists, and many or most of the scholars who would characterize their ­specialty as contact linguistics, have focused on processes involving contact in analyzing synchronic variation. A few scholars have even argued that contact is the sole source of lan­ guage variation and change; this extreme position is a neat counterpoint to an older position in historical linguistics, namely, that language contact is responsible only for lexical changes and quite minor structural changes. (This long‐superseded older position is echoed, surpris­ ingly, in Labov’s recent claim that ‘structural borrowing is rare’ and superficial (Labov 2007: 349, 383).) In this chapter I will argue that neither extreme position is viable. This argument will be developed through a survey of general types of contact explanations, especially explanations for changes over time, juxtaposed with a comparative survey of major causal factors in internally‐motivated language change. My goal is to show that both internal and external motivations are needed in any full account of language history (Hickey 2012) and, by implication, of synchronic variation. Progress in contact linguistics depends, in my opinion, on recognizing the complexity of change processes – on resisting the urge to offer a single simple explanation for all types of structural change. The structure of the chapter is as follows. Section 1 provides some background concepts and definitions, and sections 2 and 3 compare and contrast contact explanations with internal explanations of change. Section 4 discusses a much‐cited typology‐based theory about contact vs. internal explanations. Section  5, finally, is a brief conclusion that includes a warning about the need to be cautious in making claims about the causes of change – both because in most cases no cause can be firmly established and because of the real possibility that multiple causes are responsible for a particular change.

1  Some Background Concepts First, what counts as language contact? The mere juxtaposition of two speakers of different languages, or two texts in different languages, is too trivial to count: unless the speakers or the texts interact in some way, there can be no transfer of linguistic features in either direction. Only when there is some interaction does the possibility of a contact explanation for

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

34  Sarah Thomason synchronic variation or diachronic change arise. Throughout human history, most language contacts have been face to face, and most often the people involved have a nontrivial degree of fluency in both languages. There are other possibilities, especially in the modern world with novel means of worldwide travel and mass communication: many contacts now occur through written language only. Second, what is not language contact? That is, under what circumstances is an internal explanation for variation and/or change the only possible explanation? This question is trickier than it might seem to be, because it can rarely or never be answered by ascertaining that there was no language contact when a variant entered a language. For one thing, it is difficult, and maybe impossible, to find a language anywhere that is not in contact with one or more other languages at any given time. For another, in a sense all linguistic variation and every linguistic change must necessarily involve language contact. This is true because there are always two steps in the establishment of an innovation in a speech community: the initi­ ation of variation and change must begin with an innovation in one or more speakers’ speech, and the spread of that variant through a speech community is always a matter of transfer from speaker to speaker – i.e. via language, or at least dialect, contact. An innovation that remains confined to a single speaker cannot affect the language as a whole. In spite of this complication, it has long been traditional to posit contact explanations for linguistic var­ iation and change only when two or more different languages are concerned. Third, what is contact‐induced change? Contact is a source of linguistic change if it is less likely that a particular change would have happened outside a specific contact situation. (Note that this is a definition, not a criterion for establishing contact‐induced change: there is no way to measure ‘less likely’ precisely for any past linguistic change.) The definition has several parts. It specifies ‘a source’ rather than ‘the source’ because a growing body of evi­ dence suggests that multiple causation – often a combination of an external and one or more internal causes – is responsible for a sizable number of changes. It specifies ‘less likely’ rather than ‘impossible’ because historical linguists are wary, with good reason, about making strong claims about impossible changes. And it specifies ‘a specific contact situation’ because efforts to argue for contact‐induced change without identifying a contact situation in which it could have occurred are doomed to failure. One encounters occasional arguments to the effect that (for instance) change x must have been due to some kind of substrate influence because it could not have happened through strictly internal means. This generally means that the linguist making the claim cannot find an internal route to innovation; but appealing to an unidentified and unidentifiable substrate language cannot possibly explain anything – it just adds an extra layer of mystery to the historical puzzle. The motivation for appealing to a mystery language and inferring part of its structure from the structure of the putative receiving language appears to lie in a belief that every linguistic change can be explained. This brings me to a fourth background point: in spite of dramatic progress toward explaining linguistic changes made in recent decades by historical linguists, variationists, and experimental linguists, it remains true that we have no adequate explanation for the vast majority of all linguistic changes that have been discovered. Worse, it may reasonably be said that we have no full explanation for ANY linguistic change, or for the emergence and spread of any linguistic variant. The reason is that, although it is often easy to find a motivation for an innovation, the combinations of social and linguistic factors that favor the success of one innovation and the failure of another are so complex that we can never (in my opinion) hope to achieve deterministic predictions in this area. Tendencies, yes; probabilities, yes; but we still will not know why an innovation that becomes part of one language fails to establish itself in another language (or dialect) under apparently parallel circumstances. There is an element of chance in many or most changes; and there is an element of more or less conscious choice in many changes. Even if we could pin down macro‐ and micro‐social features and detailed linguistic features to a precise account (which we

Contact Explanations in Linguistics  35 cannot), accident and the possibility of deliberate change would derail our efforts to make strong predictions. This assessment should not be taken as a defeatist stance, or discourage efforts to find causes of change: recent advances in establishing causes of changes after they have occurred and in tracking the spread of innovations through a speech community have produced notable successes. More successes will surely follow; seeking causes of change is a lively and fruitful area of research. But the realistic goal is a deeper understanding of processes of change, not an ultimate means of predicting change. The point about not being able to explain most changes will perhaps become clearer if we consider the fate of most innovations. Speakers who are tired, or drunk, or non‐native, or three years old, or even just verbally inept frequently produce innovative forms. To take one type of example, collecting speech errors from weather reports read by radio broadcasters turns up examples like dreezle (in the sentence We could have some freezing dreezle) and frog (in the noun phrase patchy frog, a combination of frost and fog). Neither dreezle nor frog in this sense is likely to enter the language as a permanent feature (although smog, a combination of smoke and fog exactly parallel to frog, has certainly done so); but the forms were uttered by native speakers, and that makes them linguistically possible changes in the language. Their subsequent history – whether they are ever used again by the same speakers and, if they are, whether they are adopted by other speakers – is a matter of linguistic and social probability, not possibility. It is easy to think of linguistic reasons why (for instance) frog might never turn up in an English dictionary (its homophony with the semantically unconnected pre‐existing word frog would probably hurt its chances); but offering such an explanation for its ephem­ erality is a Just So story, mere speculation. There is no evidence to support it. And the same is true of most or all lexical and structural innovations that do not achieve ultimate success. Fifth, how can internal and external causes of change be established? Traditional ways of identifying internal causes of change have emphasized learnability, in the form of pattern pressures, or structural imbalances, that make a particular set of forms or syntactic structures relatively hard to learn. In phonological systems, especially in the past 30 years or so, phonetic causes of change have been identified. There is now experimental research sup­ porting certain claims of ease of learning, but to date these comprise only a very small fraction of past and ongoing changes. In another line of research, over the past 40 years variationists have studied the spread of variants within speech communities, identifying (for instance) social networks (Milroy 1987) that contribute to the success of a given innovation. Establishing external causes of change has proved especially difficult, in part because most historical linguists used to be reluctant to admit any contact explanation for a structural change. Even now, there is a strong tendency to consider the possibility of external causation for a change only when the search for an internal cause has failed to produce a plausible result. This tendency has weakened in the past 20 years or so, but it has not vanished. It is therefore necessary to be quite explicit, and rigorous, about criteria for establishing contact‐ induced change. Here is a list of conditions that must be met (see Thomason 2001: 93–4 for further discussion). These conditions are usually needed only to support claims that structural fea­ tures have been transferred without the morphemes that express them in the source lan­ guage; given a reasonable amount of luck, borrowed lexical and grammatical morphemes will declare their origin, in which case no further effort is needed. Attaching an external cause to an innovative structural feature is much more difficult. The first requisite is to con­ sider the proposed receiving language (let us call it B) as a whole, not a single piece at a time: the chances that just one structural feature traveled from one language to another are vanish­ ingly small. Second, identify a source language (call it A). This means identifying a lan­ guage – or, if all speakers of A shifted to B, one or more closely related languages – that is, or was, in sufficiently intimate contact with B to permit the transfer of structural features. Third, find some shared features in A and B. They need not be identical in the two languages,

36  Sarah Thomason and very often they will not be, because transferred features often do not match in the source and receiving languages. They should, however, belong to a range of linguistic subsystems, e.g. both phonology and syntax, so as to rule out the possibility of structurally linked internal innovations. Fourth, prove that the features are not innovations in A. And fifth, prove that the features ARE innovations in B, that is, that they did not exist in B before B came into close contact with A. For the sake of completeness, the search for causes should not end here, even if an external cause has been solidly established, because the influence of internal causal factors must also be considered. The best explanation for any linguistic change will take all discoverable causal factors into account. The rather extensive literature that attempts to decide between an internal and an external cause of a particular change is a waste of effort – the dichotomy is false, and the best historical explanation might well have to appeal to both causes. If all five of the prerequisites can be met, then the case for contact‐induced change is solid. But if one or more cannot be met, then any claim of external causation must be tentative at best. The most common weakness in a case for external causation concerns requisites four and/or five. Here are two examples to illustrate an all‐too‐common situation. In the well‐ known Pacific Northwest linguistic area (Oregon, Washington, British Columbia), a number of striking structural features are shared widely in all languages of the region. These include, among other things, typologically marked features such as velar vs. uvular obstruents, labial­ ized dorsal consonants, lateral obstruents, a weak noun/verb distinction, verb‐initial sen­ tence structure, optionality of plural inflection, and numeral classifiers. All specialists agree that the region is a true linguistic area. But it turns out that all of the most widespread areal features must be reconstructed for all three of the major language families in the region – Proto‐ Salishan, Proto‐Wakashan, and Proto‐Chimakuan. This means that we have no actual evi­ dence for diffusion of any of the features. They could, in principle, be either inheritances from a common ancestor for the three families or due to diffusion among the three proto‐languages. But just as we have no evidence of diffusion, because we cannot satisfy the fourth and fifth conditions for establishing contact‐induced change, we also have no evidence that two or all three of the language families are changed later forms of a single parent language. (Such a relationship has been proposed, most prominently in recent times in Greenberg 1987, but Greenberg’s methods are considered by virtually all specialists to be fatally flawed, and his results have not been accepted by historical linguists who specialize in Native American lan­ guages.) At least we can rule out accident as the source of the shared area‐wide features, because the package is too specific and too unusual to make accidental similarity an attractive hypothesis. But this does not help us to choose between the remaining two possibilities. This last point suggests that there may be no plausible historical explanation at all for a particular change or set of changes. This inference is correct. Our goal in analyzing linguistic changes, always, is to arrive at the best available historical explanation for a change. But in many cases, as noted on p. 34, no historical explanation is available; and in others, like the question of the origin of the shared area‐wide features in the Pacific Northwest, two expla­ nations are in principle available, but there is no supporting evidence for either of them.

2 Contact Explanations and Internal Explanations of Change: Social Predictors Not surprisingly, contact‐induced change can be viewed from a variety of perspectives. The focus in this section and in section  3 is on factors that predict the outcome of change processes – that is, on social and linguistic explanatory factors. The ones presented here are discussed in more detail in Thomason and Kaufman (1988) and Thomason (2001, ch. 4), though without the present emphasis on comparing and contrasting external and internal

Contact Explanations in Linguistics  37 causation. As we will see, certain factors are unique to contact‐induced change, while others are shared with internally‐motivated change; but there do not appear to be any factors that are unique to internally‐motivated change. I will argue (following Dorian 1993) that there is often no clear‐cut dichotomy between internally‐ and externally‐motivated change, and that this makes the search for causal explanations even harder than it might otherwise be. Let us turn now to a consideration of some major predictors of linguistic change – not predictors in an absolute sense of predicting that change WILL occur, but predictors in the sense of conditions under which certain kinds of change CAN take place. The major (gen­ eral) social factors that are relevant for predicting the effects of contact‐induced change are the presence vs. the absence of imperfect learning, intensity of contact, and speakers’ atti­ tudes. Of these, the first will not be relevant for internal explanations of change, because the agents of internally‐motivated change are native speakers of the changing language, or non‐ native speakers who have native‐like fluency in the receiving language, or incipient native speakers (if children produce innovations during the process of first language acquisition). This is true both of the original innovator(s) of a novel linguistic feature and of those who participate in the spread of the innovation through a speech community. The presence or absence of imperfect learning by a group of people is, however, a major predictor of the outcome of contact‐induced change (Thomason and Kaufman 1988; Thomason 2001: 66–76). Here is a brief characterization of the contrasting expectations. When the agents of change are fluent speakers of the receiving language, the first and pre­ dominant interference features are lexical items belonging to the nonbasic vocabulary; later, under increasingly intense contact conditions, structural features and basic vocabulary may also be transferred to the receiving language. The only major type of exception to this predic­ tion is found in communities where lexical borrowing is avoided for cultural reasons; in such communities, structural interference may occur with little or no lexical transfer. A striking example is reported from the Vaupés region of Amazonia (Aikhenvald 2002: 187 et passim). The prediction for the outcome of contact situations in which one group of speakers shifts to another language, and fails to learn it fully, stands in sharp contrast to a situation in which imperfect learning plays no role: in shift‐induced interference, the first and predominant interference features are phonological and syntactic; lexical interference lags behind, and in some cases few or no lexical items are transferred from the shifting group’s original language to their version of the target language (see Hickey, Chapter 7, this volume). Here too there is a class of potential exceptions. If the shifting group is a superstrate rather than a substrate population, then there may be a large number of transferred lexical items. (But it is doubtful that the famous case of the shift by Norman French speakers to English in England ­during the thirteenth century, which is often cited as a prime example of superstrate shift, belongs in this category: by the time of the shift, the Norman French speakers in England were largely bilingual, so that imperfect learning in fact probably played no role in the ­process of shift.) The second social factor that affects contact‐induced change, intensity of contact, is rele­ vant to both of these general types of contact‐induced change. Where imperfect learning plays no role, intensity of contact is typically derived from things like the duration of contact and the level of bilingualism in the receiving‐language speech community. The longer the contact period and the greater the level of bilingualism, the more likely it is that structural features will be transferred along with lexical items. In shift‐induced interference, intensity has to do with such factors as the relative sizes of the populations speaking the source and receiving languages, the degree of access to the target language by the shifting group, and the length of time over which the shift occurs: if there are many more shifting speakers than original target‐language speakers, if the shifting group has only limited access to the target language, and if the shift occurs abruptly, a relatively large amount of shift‐induced interfer­ ence is likely. In the opposite situation, little or no shift‐induced interference can be expected.

38  Sarah Thomason A typical example of the former type is the variety of English spoken in a community of Yiddish speakers in the United States. In this community, Yiddish‐speaking immigrants learned English, but Yiddish was their first language and remained their main language. Their contact with native English speakers was somewhat limited; within their community, they were certainly the majority; and they learned English as a second (or third, or...) lan­ guage. As a result, their English displayed moderate lexical interference but strong phono­ logical and morphosyntactic interference from Yiddish, their first language (Rayfield 1970: 85). (These immigrants did not in fact shift from Yiddish to English; they remained Yiddish‐ dominant bilinguals throughout their lifetimes.) Ironically, given its reputation as an extreme case of (superstrate) shift‐induced interference, the structural influence of Norman French on English illustrates the opposite set of conditions: between 1066 and the thirteenth century, when the majority of Normans finally shifted to English in England, they were always vastly outnumbered by English speakers, and their access to English as spoken by native speakers was apparently unlimited. Nor was their shift to English abrupt; as noted on p. 37, by the time it occurred, the Norman population in England had apparently long been bilingual. Structural interference of French on English is quite modest, especially compared to the huge number of French loanwords. One complication here is that a significant amount of shift‐induced interference is some­ times found in a long‐term contact situation in which imperfect learning was important early on, but then bilingualism was established and maintained for a considerable period of time: the shifting group’s version of the target language was fixed at a time when the level of bilingualism was low among members of the shifting group, and (possibly for attitudinal reasons; see below) it never converged toward the target language as spoken by the original target‐language speech community. An example is the French spoken in the (originally) Breton community of Ile de Groix, France: as of 1970, only the oldest inhabitants were fluent speakers of Breton, but the French spoken in the community was heavily influenced by Breton in both structure and vocabulary (Fowkes 1973: 195, in a review of Ternes 1970). Intensity of contact is a factor in internally‐motivated change if (and only if) we include under “contact” processes of person‐to‐person transmission of variants within a single speech community – that is, the spread of an innovation. The concept of social networks, as developed in e.g. Milroy (1987), is of obvious relevance here: the idea is that variants spread through networks, and close‐knit social networks characterized by intense contact among the participants can facilitate the spread of innovations. Intensity of contact of course plays no role in the initial innovation of an internally‐motivated change; it can be relevant only for the spread of a change. The third general social predictor, speakers’ attitudes, is certainly relevant both in contact‐ induced change and in internally‐motivated change. This is admittedly a very vague notion, but it is difficult to make it precise – and equally difficult, unfortunately, to prove that it has affected the course of language history. Speakers may or may not be aware of the attitudinal factors that help to shape their linguistic choices, and historical linguists are (of course) unable to establish speakers’ attitudes except in the tiny handful of instances in which meta­ linguistic comments on innovations are found in old documents. To take just one of many frustrating examples, a striking areal feature of at least some parts of the Pacific Northwest Sprachbund is an avoidance of loanwords from French and English, the European languages with which the indigenous peoples first came into contact. In Montana Salish, for instance, there are hardly any English (or French) loanwords, in spite of massive cultural assimilation to European‐derived culture; a similar situation obtains in Nez Perce, an unrelated language whose speakers have long had close contact with Salishan tribes. The two tribes have instead constructed descriptive words from native morphemes to name items imported from Anglo culture, as in the Montana Salish word for ‘automobile’, p’ip’úyšn, literally ‘it has wrinkled feet’ (so named because of the appearance of tire tracks), and the Nez Perce word for

Contact Explanations in Linguistics  39 ‘telephone’, cewcew’in’es, literally ‘a thing for whispering’. Modern elders, when asked by young tribal members how they would say (for example) ‘television set’ in Salish, make up comparable words on the spot; but when asked why they do not just use the English word with a Salish pronunciation, they do not know – they merely shrug and say that’s how it is. In other words, the reason for this culturally‐determined pattern of lexical innovation is unknown to current tribal members, and was possibly never a conscious avoidance pattern. The clearest examples of speakers’ attitudes as a cause of change therefore come from cases of deliberate change (though even here it is often impossible to prove that a change was made with full conscious intent). Some of these are internally motivated, at least in the sense that no direct language contact was involved; others are externally motivated, according to the definition of contact‐induced change given in Section  1. Language planning is one prominent source of internally‐motivated deliberate changes. For instance, the twentieth‐ century Estonian language reformer Johannes Aavik introduced hundreds of new words and a sizable number of morphosyntactic features into Estonian, and many of these became permanently fixed in the language (Saagpakk 1982; Thomason 2007). Several striking exam­ ples of contact‐induced changes that must be explained by speakers’ attitudes come from situations in which a speech community wishes to distinguish its language, or more likely its dialect, more sharply from its neighbors’ speech. Changes range from lexical substitution in certain languages of Papua New Guinea (Kulick 1992: 2), to reversal of gender assignments, also in Papua New Guinea (Laycock 1982; Thomason 2007), to phonological distortion of words in Lambayeque Quechua in Peru (David Weber, p.c. 1999, citing research by Dwight Shaver), and to a combination of the first and third of these types of change in Mōkkı ̄, a lan­ guage of Baluchistan that was created by lexical substitutions and distortions (Bray 1913: 139–40). Lexical substitutions are also found, of course, in teenage slang, which is usually (at least in the United States) a product of internally‐motivated lexical innovation. A consideration of social predictors of externally‐ and internally‐motivated language change could be extended to more fine‐grained social factors, but the very general factors discussed here provide a good overall comparative picture of external and internal explana­ tory social factors. One of the three categories, speakers’ attitudes, seems to be equally effec­ tive in internal and external causation. A second, intensity of contact, is very important to both innovation and spread of an innovation in contact‐induced change, but it is relevant only to the spread of an innovation in internally‐motivated change. The third category, the differential effects of interference depending on whether or not imperfect learning played a role, is relevant only for contact‐induced change.

3 Contact Explanations and Internal Explanations of Change: Linguistic Predictors As with social predictors, the discussion here of linguistic predictors is confined to quite general categories; here too the analysis could be extended to cover more specific linguistic factors, but the broader categories will serve to develop an overall comparative picture. For contact‐induced change, the most important linguistic predictors are typological distance, universal markedness (with its ultimate appeal to ease of learning), and degree of integration within a linguistic system. The first of these is relevant only to contact‐induced change, but the other two are equally important for internally‐ and externally‐motivated change. Typological distance is a (very informal) measure of structural differences between two linguistic systems; it cannot be relevant for making predictions about internally‐motivated change because internal motivations arise within the same overall system. If different ­dialects are involved in motivating a change, then that is contact‐induced change, not internally‐ motivated change. If differences exist in two parts of the same system – say, in two different

40  Sarah Thomason inflectional classes in a language’s system of noun declension – then analogical changes that bring about leveling of the two classes are indeed akin to contact‐induced change if they are interpreted as interference between two different (sub)systems. But in such a case the typo­ logical distance between the two (sub)systems is minimal or zero: with minor variations, noun classes in a single language will share the same morphological and syntactic properties. In contact‐induced change, the degree of typological distance between specific subsys­ tems of a source language and a receiving language helps to predict the kinds of interference that may occur under differing degrees of contact intensity. Where typological distance is small, linguistic subsystems in which contact‐induced change is in general rare may undergo such change. This principle is most obviously illustrated by changes in the inflectional mor­ phology, which tends to lag behind phonology and syntax and even derivational mor­ phology in a catalogue of interference features, in both shift‐induced interference and interference in which imperfect learning plays no role. Specifically, minimal typological distance is in part responsible for the frequency of inter­ dialectal interference involving inflectional features that are rarely transferred in cases of foreign interference. An example is the realignment of syncretic categories in the plural oblique cases of masculine and neuter nouns in certain nonstandard dialects of (the lan­ guage formerly known as) Serbo‐Croatian. Where Standard Serbo‐Croatian has syncretism in the Dative‐Instrumental‐Locative plural, whose suffix ‐ima is opposed to the Genitive plural suffix ‐a: , some nonstandard dialects of the Čakavian dialect group had instead a Genitive‐Locative suffix –i:h vs. Dative plural ‐o:n and Instrumental plural ‐i. But in the major study of one of these Čakavian dialects, that of the Adriatic island Hvar, the analyst found that only older speakers still retained the inherited Čakavian system; by contrast, younger speakers were using ‐i:h only for the Genitive plural, and they had borrowed the standard dialect’s suffix ‐ima for the Dative, Instrumental, and Locative plural cases (Hraste 1935: 17–25). It is important to emphasize here that the role of typological distance in this and other cases is not to motivate the change itself; social factors – in this instance an expanded educational system that heavily exposed younger speakers to the standard dialect, together with the prestige of that dialect – were surely the primary cause of the change. (When I con­ ducted dialect research in Yugoslavia in 1965 it was still fairly easy to find villagers as young as 60 who had never attended school. Ten years later this would almost certainly have been impossible.) The role of typological distance, in a case like this one, is to make possible a type of change that would be a great deal less likely if the source and receiving systems differed, at the relevant structure points, in some or all of their inflectional categories (gender, number, case, and/or noun class). Minimal typological distance sometimes does facilitate otherwise rare types of contact‐ induced change in different languages. A well‐known example is the borrowing of Bulgarian person/number verbal suffixes in Megleno‐Rumanian, yielding double‐marked verb forms: original Megleno‐Rumanian forms like aflu ‘I find’ and afli ‘you find’ were already marked for 1sg and 2sg subject, respectively, but the borrowed Bulgarian 1sg and 2sg suffixes ‐m and ‐š were nevertheless added, producing aflum and afliš (Sandfeld 1938: 59). Bulgarian and Megleno‐Rumanian are only distantly related to each other, but their verbal systems have the same person/number combinations, so that these redundant suffixes could be added without any adjustment in the categories  –  there was already a one‐to‐one correspondence in the expression of 1sg and 2sg in verb inflection. In other words, the typo­ logical distance between the two languages at this structure point was zero. It is far more common, however, for different languages to differ significantly in inflec­ tional and other categories, and in these cases typological distance is likely to prove a barrier to contact‐induced change. Not an absolute barrier: under circumstances of intense contact, any linguistic feature can be transferred to any other language. But the effect of typological

Contact Explanations in Linguistics  41 distance when typologically dissimilar languages are in intimate contact is reflected in the various borrowing scales that have been proposed in the literature (e.g. in Thomason and Kaufman 1988: 74–6): with casual contact, only easy‐to‐borrow items can be transferred from source to receiving language, primarily or only nonbasic vocabulary; with increasingly intense contact, harder‐to‐borrow features may be transferred, ultimately including even inflectional morphology, basic vocabulary, and typologically disruptive features (e.g. pre­ fixes introduced into a language that was previously exclusively suffixing). Examples can be found in a recent collection of papers on morphological borrowing, Gardani et  al. 2015, which contains five case studies of inflectional borrowing. Importantly, no borrowing scale can be valid for cases of shift‐induced interference, because in shift‐induced interference the primary agents of contact‐induced change are speakers of the source language who have learned the receiving language imperfectly. In particular, neither introducing features from one’s first language into a second language nor failing to learn aspects of the second language (the target language) requires any great knowledge of the second language – by definition, in the case of learning failure. The validity of borrowing scales therefore rests on their appropriateness for cases of interference in which imperfect learning plays no role. Certain features that appear at the hard‐to‐borrow end of a borrowing scale are hard to borrow because they are relatively hard to learn (typically, because they are universally marked): you cannot borrow what you do not know, so only the most fluent bilinguals can introduce hard‐to‐learn features into their second language. But part of the reason features are hard to borrow has to do with the degree to which they are integrated into the linguistic system. This is why inflectional features are so much less likely than other structural features to be transferred from one language to another: inflectional systems tend to be very tightly inte­ grated in a system of interlocking structural relationships, and it is thus likely to be difficult both to extract a single form or set of forms from one system and to insert foreign forms into another inflectional system, especially one that is typologically incompatible. (See pp. 42–44 for a more detailed discussion of the role of markedness and of systemic integration.) In fairly intense contact situations where imperfect learning plays no role and where the languages involved are very different typologically, then, interference features are most likely to include much nonbasic vocabulary, some function words and derivational mor­ phology, borrowed phonetic and phonological features confined to loanwords, and bor­ rowed syntactic features that do not cause major typological change. But toward the extreme end of the borrowing scale, where contact is very intense, typologically significant contact‐ induced changes may occur: borrowed basic vocabulary, borrowed phonology and pho­ netics in native vocabulary, borrowed syntactic features that do alter the receiving language’s syntactic typology, and even borrowed inflectional categories and patterns. The Iranian lan­ guage Ossetic, which has borrowed extensively from neighboring Caucasian languages, pro­ vides an example from the middle of the scale, with moderate structural interference. In addition to many loanwords from Georgian (a South Caucasian language), Ossetic has bor­ rowed ejective stops, which appear even in native Ossetic vocabulary, a declensional system with more cases than Ossetic inherited from Proto‐Indo‐European via Proto‐Iranian, agglu­ tination (replacing the largely flexional morphology typology typical of Indo‐European lan­ guages), and a more rigid SOV word order, with more postpositions, than one finds in Iranian languages that have not undergone interference from Georgian or other Caucasian languages. In sum, the popularity of borrowing scales, and the fact that they seem to be largely valid (in the absence of powerful social factors that skew the results, such as a culturally motivated disinclination to borrow words), indicate the need for ideal social conditions – in the form of very strong cultural pressure from a dominant language – in order to override typological barriers to the transfer of hard‐to‐borrow structural features. It may reasonably be surmised

42  Sarah Thomason that typological barriers could also be broken down in situations where attitudinal factors favor the adoption of typologically disruptive innovations. Here are two examples that illus­ trate this possibility, though both are admittedly minor in terms of their current and probable ultimate effects on the receiving languages’ systems (see Thomason 2007 for further discussion). First, the remaining speakers of Ma’a, a mixed language of northeastern Tanzania, are bilingual in Shambala and in fact form part of the Shambala speech community. Ma’a com­ bines Cushitic lexicon and a few residual Cushitic structural features with Bantu lexicon and structure, plus a sizable component of Maasai words; its complex history involves heavy Bantuization of an originally non‐Bantu (presumably Cushitic) language, and also a near‐ complete shift from Ma’a to Shambala, with retention only of some non‐Bantu lexicon – and one phonological feature that is foreign to the nearby Bantu languages, a voiceless lateral fricative /ɬ/. Nowadays Ma’a speakers emphasize the distinctness of their heritage lan­ guage by introducing this fricative into their Bantu discourse, using it not only in Ma’a words but also in Bantu words (Mous 1994: 199). The effect is to make their Bantu speech less Bantu‐like, and apparently also to impress their non‐Ma’a‐speaking Bantu interlocutors with their ability to pronounce such an exotic sound. The introduction of this fricative into Shambala is both phonetically and typologically novel. The second example is syntactic accommodation to English structure in Nisgha, an indig­ enous language spoken in northwestern British Columbia. This example does not involve actual language change; but the deliberate innovation attests to the possibility of a future contact‐induced syntactic change that would be typologically disruptive in a major way. In working with speakers of Nisgha, in which objects are deleted under identity with the object of a previous clause (as in pseudo‐English They heard him but couldn’t see) except in emphatic contexts, Tarpent found that Nisgha/English bilinguals tended to insert object pronouns into their Nisgha speech when they were trying “to approximate the English utterance” (1987, 157–8). She also observed that “some bilingual speakers...tend to stick very close to English surface structure” when they translate English sentences, resulting in very strange‐ sounding Nisgha. Let us turn now to the other two major linguistic predictors of contact‐induced change, the two that are also relevant for internal explanations of language change: universal mark­ edness and degree of integration within a linguistic system. Markedness, as noted on p. 41, is ultimately connected with ease of acquisition. Universally marked features are believed to be those that are harder to acquire; unmarked features should be easier to acquire. The amount of evidence available to support this connection is unfortunately limited, but two main lines of evidence – frequency of cross‐linguistic occurrence (expected to be greater for unmarked than for marked features) and the age at which children learn them in acquiring their first language (earlier for unmarked features, later for marked features) do converge on certain features, especially phonological ones. To take a rather trivial example, a phoneme /t/ is close to universal in the languages of the world, while a phoneme /θ/ is quite rare; and /t/ is acquired earlier than /θ/ by children learning English as a first language. So, arguably, /t/ is less marked than /θ/. Markedness plays an important role in shift‐induced interference, both because the shift­ ing group is more likely to learn unmarked target‐language features than marked ones and because, if the two speaker groups eventually merge into a single speech community, original target‐language speakers are less likely to adopt marked features than unmarked ones from the learners’ version of the target language. But in contact‐induced changes that do not involve imperfect learning, markedness is likely to be of considerably less importance, and may not play any role at all. The reason is that the agents of change in this type of interfer­ ence are fluent speakers of the receiving language and of (at least the relevant aspects of) the source language as well, so ease of learning does not enter the picture; if they borrow

Contact Explanations in Linguistics  43 structural features, they are as likely to incorporate marked as unmarked features. The Ma’a example on p. 42 is a case in point: the reason for introducing the lateral fricative into Ma’a speakers’ Bantu speech was precisely its exoticness, its status as a hard‐to‐pronounce phoneme. In internally‐motivated change, ease of learning is the major component of the phenomenon known as drift, the only internal factor traditionally recognized as a cause of language change. (The other two traditional causal factors are external: dialect borrowing and foreign interference. These external factors are of course sometimes difficult or impos­ sible to distinguish from one another.) The idea behind the concept of drift (which was first proposed, though with now controversial trappings, in Sapir 1921) is that pattern pressures, or structural imbalances, motivate language change because they are structure points that cause learning problems. Irregularities are harder to learn than regular patterns, for instance; and marked features, which by definition impose a burden on learning, are relatively hard to learn and are therefore likely to be diachronically unstable. A well‐known characteristic of drift is that it often brings about similar or identical changes in related languages, especially during the period immediately following the split of two or more sister languages from their common parent – because all the daughter languages will have inherited the same pattern pressures from the parent, and efforts to ease the learning burden may well take the same route in each. In the Indo‐European language family, for instance, the history of all extant branches displays a tendency toward inflectional simplification (with resulting complication in the syntax); in noun declension, for example, even those languages that best retain many aspects of Proto‐Indo‐European noun declension have collapsed several inherited noun classes, which were semantically opaque, into just three based on biological gender, with a few residual forms from other classes. It is hardly surprising, given that markedness plays an important role both in shift‐ induced interference and in internally‐motivated structural change, that different (unre­ lated) languages can undergo the same change for quite different reasons. It therefore makes no sense to ask (for instance) whether the loss of certain plural case distinctions in the Serbo‐ Croatian dialect of Hvar is due to contact‐induced change or to the same long‐term process of drift that has eroded Indo‐European inflectional distinctions over many centuries. Of course in that case there is no doubt about the contact influence  –  the borrowed Dative‐ Instrumental‐Locative plural suffix ‐ima attests to its standard‐dialect origin – but drift may well also have played a role in motivating the change. And if (as is often the case) it had been merely the standard‐language Dative = Instrumental = Locative plural pattern that was involved, and not the actual suffix, it would still make no sense to assume that a choice bet­ ween an internal cause and an external cause was necessary. The third major linguistic causal factor, degree of integration into the system, also has its place in explaining both internally‐ and externally‐motivated changes. Contact‐induced changes are much more common in loosely‐integrated linguistic subsystems than in subsys­ tems which – like the inflectional morphology – are characterized by sets of interconnected forms organized into paradigms. In fact, as noted on p. 41, inflectional morphology is the linguistic subsystem that is least likely to be affected by contact‐induced change, owing in part to typological distance and in part to the fact that the inflectional morphology typically displays the highest degree of integration of elements, of interlocking structure. Transfer of inflectional features from one language to another is therefore likely to happen only under very intense contact conditions, including processes of language shift as well as changes in which fluent bilinguals are the agents. And most such changes fit into the receiving lan­ guage’s inflectional patterns without significant typological disruption. So, for instance, Lithuanian acquired three new noun cases when speakers of Finnic languages, which are rich in case distinctions, shifted to Lithuanian; but the cases fit naturally into the already case‐rich Lithuanian system of noun declension and caused no typological change. As we

44  Sarah Thomason saw on p. 40, the same is true of the Bulgarian person/number suffixes added to Megleno‐ Rumanian verbs. In general, however, less close‐knit subsystems more readily admit new elements; that is why nonbasic vocabulary is the predominant category of interference fea­ tures in contact situations where imperfect learning plays no role. It is far from obvious that tightly‐integrated inflectional systems tend to resist internally‐ motivated change, or that they change more slowly than other linguistic subsystems (although claims of super‐stable morphology have certainly been made; see Thomason 1980 and sources cited there for a discussion). Analogical changes within inflectional paradigms (as well as in other subsystems) are routine occurrences in languages all over the world. One might argue, however, that the analogical processes themselves arise out of tightly‐integrated systems, where partial regularities can be found on multiple axes and thus motivate analog­ ical changes. In this sense, the degree of integration can be said to serve as a predictor of internally‐motivated change. In sum, like the effects of social predictors of change, the effects of linguistic causal factors are unevenly distributed between externally‐ and internally‐motivated change. Of all the predictors surveyed here, two – presence or absence of imperfect learning and typological distance – are restricted to contact explanations. Intensity of contact is highly important for contact explanations but less so for internal explanations. The remaining three factors, degree of integration within a system, speakers’ attitudes, and markedness, are significant in both types of explanation, although in contact‐induced change the effects of markedness seem to be largely restricted to shift‐induced interference, where imperfect learning is important.

4 On a Current Theory of Internal vs. Contact Explanations of Language Change Many authors have claimed over the years that language contact simplifies grammars. For instance, Kusters has argued that “Type 1 communities [relatively isolated, mainly L1 speakers] permit complexities, while Type 2 communities [many L2 learners] restrict them” (2003: 9); Karlsson et al. propose that “Contact‐induced grammatical change is likely to pro­ duce outcomes simpler (in some sense) than the original ones...” (2008: viii); and McWhorter has claimed that a “crucial factor in language change is adults abbreviating the machinery of a language...” (2005: 265). Perhaps the most carefully elaborated theory along these lines is Peter Trudgill’s, and I will focus on his theory here. In an early formulation of his typology‐ based approach he claimed (2004: 306) that Small, isolated, low‐contact communities with tight social network structures are more likely to be able to maintain linguistic norms and ensure the transmission of linguistic complexity from one generation to another....On the other hand, communities involved in large amounts of language contact, to the extent that this is contact between adolescents and adults who are beyond the critical threshold for language acquisition, are likely to demonstrate linguistic pidginisation, including simplification, as a result of imperfect language learning.

More recently, Trudgill has refined his theory to cover the many instances of language con­ tacts that (he argues) can add complexity to a language’s structure. He distinguishes bet­ ween contact situations that feature child bilingualism, where he expects what he calls additive complexification, and contact situations in which “adult second‐language learning... dominates,” where he expects structural simplification (2010: 313). He believes that the con­ ditions for the latter type of contact situation are essentially modern, suggesting that “wide­ spread adult‐only language contact is mainly post‐neolithic and indeed a mainly modern phenomenon, associated with the last 2000 years” (2011: 20, quoting Trudgill 2009: 109).

Contact Explanations in Linguistics  45 Trudgill’s more recent formulation is reminiscent of the distinction described on p. 37 ­ etween contact‐induced changes in which imperfect learning plays a role and contact‐ b induced changes in which imperfect learning plays no role (Thomason and Kaufman 1988; a very similar dichotomy is proposed in van Coetsem 1988). A major difference is that Thomason and Kaufman found a robust correlation between the kinds of transferred fea­ tures and the presence or absence of imperfect learning: in shift situations with imperfect learning, the dominant interference features were phonological and syntactic, while in bor­ rowing situations with no imperfect learning the dominant interference features were non­ basic vocabulary, followed (with increasing intensity of contact) by phonological and syntactic features of more to less superficial types. Inflectional morphology lagged behind in both types of contact situation. The relevance of this difference between Trudgill’s approach and mine is that shift by a speech community to the language of another speech community will involve everyone, from young children to adults – and there is no way to tell, for us or for Trudgill, which age groups, if any, predominated in a long‐past shift situation. The point here is that I found the same types of transferred features in all shift situations that resulted in significant amounts of shift‐induced interference; so if different age groups dominated in different shift situa­ tions, the linguistic evidence indicates that age differences did not affect the outcome. I there­ fore see no evidence to support Trudgill’s proposed distinction between contacts involving adult learners only and other kinds of language shift situations. I do, of course, agree with his basic dichotomy between contact‐induced changes with imperfect learning and contact‐induced changes without imperfect learning. But again, the evidence that the latter contact situations crucially involved child bilingualism seems to me to be thin. A great many adults in a great many speech communities are fluent bilinguals and multilinguals. Of course most such communities will also have child bilinguals, but there are documented contact situations with considerable amounts of interference in which children almost certainly played no role in the transfer of structural features. A fairly typical scenario is one in which the men in a village leave home to work in a nearby town where a different language is dominant, while the women and children stay home in a largely monolingual environment. The men become bilingual and are the agents through which interference features travel from the town’s language to the village’s lan­ guage. A probable example is Asia Minor Greek (before World War I): in at least some Greek communities in Turkey, the men apparently became bilingual, while the women and chil­ dren stayed home and were less likely to speak Turkish. The most‐affected dialects of Asia Minor Greek acquired a large number of Turkish structural features along with much Turkish lexicon (see Thomason and Kaufman 1988: 215–22 for a list of some of the borrowed Turkish features). Less typical situations in which active participation by very young children is excluded are those in which the few well‐documented bilingual mixed languages developed. These languages were definitely created by fully fluent bilinguals, not by very young children, because the components show too little distortion to be a plausible result of either second language learning or early first language acquisition; moreover, the social motivations for creating the bilingual mixtures could not have been present in the very young children (Thomason 2013). Michif, with its French noun phrases and its Cree verb phrases, is one example; another is Media Lengua, with its almost entirely Spanish lexicon and its Quechua grammar; and a third is Mednyj Aleut, whose finite verb morphology (but not its elaborate native Aleut noun morphology) was imported wholesale from Russian. Michif and Mednyj Aleut arose in mixed‐blood populations that had distinct legal and social status in contrast to speakers of their input languages; Media Lengua was created by men who had left their Quechua villages to work in Spanish‐dominant cities and returned home partly acculturated and fluent in Spanish. Although stable mixed languages of this general type are rare, as far

46  Sarah Thomason as we know, their existence provides solid evidence that contact‐induced language changes carried out by bilinguals cannot be assumed to be the result of child bilingualism. I am also skeptical about Trudgill’s suggestion that language shift by adults is a (relatively) modern phenomenon. There are numerous documented modern instances of language shift by hunter‐gatherer communities to the language of a more powerful hunter‐gatherer or agri­ cultural community; these have nothing to do with modern nation‐states, and they may well reflect conditions that have existed for several thousand years. Herodotus, for instance, reports an example of an ancient language shift: “The Sauromatae speak the language of Scythia, but have never talked it correctly, because the Amazons learned it imperfectly at the first” (The Persian Wars, Book IV, chapter 117). Admittedly, Herodotus cannot be treated as a totally reliable scholarly source; but his description of language shift is modern enough to carry conviction, whether or not it describes a specific shift situation accurately. My major reservation about Trudgill’s theory, however, is that there are too many clear counterexamples to his predictions to justify maintaining it. Both kinds of contact situations, with and without imperfect learning, result in changes that simplify the structure, changes that complicate the structure, and changes that have neither overall effect. This can be established easily in spite of the great difficulty (acknowledged by Trudgill) of pinning down the notion of structural complexity with any precision. Trudgill argues that simplification is not the norm in internally‐motivated language change – that it instead happens when a language is the target of many adult learners. He notes correctly that internally‐motivated changes that accumulate over many centuries very often lead to irregularities and lack of transparency, especially in the inflectional morphology; but he does not allow for another powerful force in internally‐motivated change, namely, analogical leveling, which very often erases morphological irregularities and leads to more transparency. (Analogical extension often has this effect too, but it also sometimes causes actual complications.) He also does not allow for the complex interplay between sound change and morphosyntactic change, the effects of which sometimes simplify structure. Internally‐motivated phonological changes also do not predictably lead to either complica­ tion or simplification. Mergers like the Balto‐Slavic collapse of the Proto‐Indo‐European voiced and murmured stop series simplify the inventory; splits like the phonemicization of the formerly allophonic English pair [k] and [č] complicate the inventory. Such examples can be multiplied endlessly. The same is true of contact‐induced changes with and without imperfect learning. Without imperfect learning, for instance, there are simplifying changes like the loss of syl­ lable‐initial consonant clusters in the Salishan language Comox under Wakashan influence (Thomason and Kaufman 1988: 88), the loss of gender in Latvian under Finnic influence (Comrie 1981: 147), and the partial loss of gender distinctions and adjective‐noun agreement under Turkish influence in some dialects of Asia Minor Greek (Thomason and Kaufman 1988: 220, citing Dawkins 1916: 115, 119, 125). And there are also complicating changes, e.g. the development of ejective stops in the Iranian language Ossetic and of extra locative cases under Georgian influence (Dawkins 1916: 84). Again, such examples can easily be multi­ plied many times over. With imperfect learning, a few simplifying examples are the merger in Czech of the plain and palatalized lateral phonemes under the influence of shifting German superstrate speakers (Jakobson 1962 [1938]: 240) and the loss of the dual number category in Ethiopic Semitic under the influence of shifting Cushitic speakers (Thomason 2001: 112–13, citing Leslau 1945, 1952). Two complicating examples are the development of labialized dorsal phonemes and of a future tense category in Ethiopic Semitic and the development of three new morphological case distinctions in Lithuanian under the influence of shifting Finnic speakers (Senn 1966: 92; cf. Fairbanks 1977: 117). Here, too, examples of both types are easy to find.

Contact Explanations in Linguistics  47 But by far the largest category of contact‐induced changes both with and without imperfect learning comprises interference features that are neither clearly simplifying nor clearly complicating. The most obvious examples are the very common changes in basic sentential word order, most often from SOV to SVO or vice versa. For these and many other replacement features, arguing for a simplifying or complicating effect ranges from difficult to impossible. Because this is the largest category, and because it is so easy to find both sim­ plifying and complicating changes in contact situations both with and without imperfect learning, Trudgill’s proposed sociolinguistic typology lacks predictive force. What makes this unsurprising is the fact that shift‐induced interference involves at least two different (but not mutually exclusive) processes. First, L2 learners may fail to learn certain features of the target language (the TL); many of those failures will simplify the learners’ version of the TL. Second, L2 learners may carry over certain features of their L1 into the TL; many of those features will complicate the learners’ version of the TL. (Neither the learning failures nor the carryovers of L1 features, of course, will necessarily be total: partial feature transfer is extremely common in all kinds of contact‐induced change.)

5 Conclusion What, then, is the value of contact explanations in linguistics? This very general question has not been fully answered in this chapter, but by surveying causal factors in both externally‐ and internally‐motivated change, I hope to have sketched the beginning of an answer. In  analyzing the linguistic effects of language contact, the linguist needs to consider all ­possible causal factors. If any factors can be ruled out, the task of explaining the changes should become easier. Only some predictors discussed in sections 2 and 3 have relevance for internally‐motivated change, so the range of internal explanations for a given set of changes in an intense contact situation is, to begin with, narrower than the range of potential contact explanations. The search for internal and external causation should proceed in parallel because of the strong possibility, an actuality in many cases, that both internal and external causes influenced the linguistic outcome of change. We have also seen that certain social factors – most obviously the presence or absence of imperfect learning and intensity of contact – set the stage for different linguistic outcomes. For example, if contact is intense enough, especially if no imperfect learning is involved, then typological distance is no barrier to extensive structural borrowing; to take another example, speakers’ attitudes can trump expectations for types and degree of both externally‐ and internally‐motivated change. In other words, in this domain social factors rule. This of course does not mean that linguistic predictors are necessarily less important or less significant in a given case than social predictors. It only means that, in cases where linguistic and social factors point to different outcomes, the social factors will be more effective. A caveat is needed here: although it is easy to find clear examples of causation, with all the causal factors discussed here, it remains true that for the vast majority of known linguistic changes there is no adequate explanation. Some of the reasons are obvious: often, in past contact situations, too little is known of the social and linguistic circumstances to satisfy the requisites for establishing contact‐induced change, and the same is unfortunately true for internal causation. This should not discourage the search for causes. The search itself is illu­ minating, and concentrating on changes for which we can build a solid explanatory account helps to advance our understanding of the probabilities of change. This leads to a final point. Historical linguists and sociolinguists employ such different methods that they sometimes misunderstand each other’s results; and, more ominously, specialists in each subfield are sometimes inclined to dismiss the results achieved in the other subfield. This inclination, if indulged, diminishes our chances of arriving at a single unified account of contact‐induced

48  Sarah Thomason change, and of contact explanations more generally. There can surely be no doubt that a sat­ isfactory theoretical understanding of contact phenomena must be compatible with the results of both subfields – and must, for that reason, help each set of specialists in arriving at a better understanding of the processes they study.

REFERENCES Aikhenvald, Alexandra Y. 2002. Language Contact in Amazonia. Oxford: Oxford University Press. Bray, Denys de S. 1913. Census of India, 1911, vol. IV: Baluchistan. Calcutta: Superintendent Government Printing, India. Comrie, Bernard 1981. The Languages of the Soviet Union. Cambridge: Cambridge University Press. Dawkins, R. M. 1916. Modern Greek in Asia Minor. Cambridge: Cambridge University Press. Dorian, Nancy C. 1993. Internally and externally motivated change in contact situations: doubts about dichotomy. In Charles Jones (ed.), Historical Linguistics: Problems and Perspectives. London: Longman, pp. 131–55. Everett, Daniel L. 1982. Phonetic rarities in Pirahã. Journal of the International Phonetic Association 12.2: 94–6. Fairbanks, Gordon 1977. Case inflections in Indo‐European. Journal of Indo‐European Studies 5: 101–31. Fowkes, Robert A. 1973. Review of Ternes 1970. Language 49: 195–98. Gardani, Francesco, Peter Arkadiev, and Nino Amiridze (eds.) 2015. Borrowed Morphology. Berlin: de Gruyter Mouton. Greenberg, Joseph H. 1987. Language in the Americas. Stanford: Stanford University Press. Hickey, Raymond 2012. Internally and externally motivated language change. In Juan Manuel Hernández‐Compoy and Juan Camilo Conde‐ Silvestre (eds,) The Handbook of Historical Sociolinguistics. Malden, MA: Wiley‐Blackwell, pp. 401–21. Hraste, Mate 1935. Č akavski dijalekat ostrva Hvara. Južnoslovenski Filolog 14: 1–59. Jakobson, Roman 1962 [1938]. Sur la théorie des affinités phonologiques entre des langues. In his Selected Writings, vol. 1: 234–46. The Hague: de Gruyter Mouton; reprinted from Actes du Quatrième Congrès International de Linguistes, Copenhagen: Munksgaard, pp. 48–59, 1938). Karlsson, Fred, Matti Miestamo, and Kauis Sinnemaki 2008. Introduction to Matti

Miestamo, Kauis Sinnemaki, and Fred Karlsson (eds.), Language Complexity: Typology, Contact, Change. Amsterdam: Benjamins, pp. i–xi. Kulick, Don 1992. Language Shift and Cultural Reproduction: Socialization, Self, and Syncretism in a Papua New Guinean Village. Cambridge: Cambridge University Press. Labov, William 2007. Transmission and diffusion. Language 83: 344–87. Laycock, Donald C. 1982. Melanesian linguistic diversity: a Melanesian choice? In R. J. May and Hank Nelson (eds.), Melanesia: Beyond Diversity. Canberra: Australian National University Press, pp. 33–8. Leslau, Wolf 1945. The influence of Cushitic on the Semitic languages of Ethiopia: a problem of substratum. Word 1: 59–82. Leslau, Wolf 1952. The influence of Sidamo on the Ethiopic languages of the Gurage. Language 28: 63–81. McWhorter, John H. 2005. Defining Creole. Oxford: Oxford University Press. Milroy, Lesley 1987. Language and Social Networks, 2nd edition. Oxford: Blackwell. Mous, Maarten 1994. Ma’a or Mbugu. In Peter Bakker and Maarten Mous (eds) Mixed Languages. Amsterdam: Institute for Functional Research into Language and Language Use, pp. 175–200. Rayfield, Joan Rachel 1970. The Languages of a Bilingual Community. The Hague: de Gruyter Mouton. Saagpakk, Paul F. 1982. Estonian‐English Dictionary (Introduction, section “Linguistic innovation in Estonian”). New Haven: Yale University Press, pp. lxxix–lxxxvii. Sandfeld, Kristian.1938. Problèmes d’interférences linguistiques. Actes du Quatrième Congrès de Linguistes. Copenhagen: Einar Munksgaard, pp. 59–61. Sapir, Edward 1921. Language: An Introduction to the Study of Speech. New York: Harcourt, Brace and World.

Contact Explanations in Linguistics  49 Senn, Alfred 1966. Handbuch der litauischen Sprache. Heidelberg: Carl Winter. Tarpent, Marie‐Lucie 1987. Between ergative and accusative syntax: lessons from Nisgha/ English syntactic interference. In John Dunn (ed.) Papers from the 22nd International Conference on Salish and Neighboring Languages, 149–71. Ternes, Elmar 1970. Grammaire structurale du Breton de l’Île de Groix. Heidelberg: Winter. Thomason, Sarah G. 1980. Morphological instability, with and without language contact. In Jacek Fisiak (ed.), Historical Morphology. Berlin: de Gruyter Mouton, pp. 359–72. Thomason, Sarah G. 2001. Language Contact: An Introduction. Edinburgh and Washington, DC: Edinburgh University Press and Georgetown University Press. Thomason, Sarah G. 2007. Language contact and deliberate change. Journal of Language Contact 1: 41–62. Thomason, Sarah G. 2013. Innovation and contact: the role of adults (and children). In Daniel Schreier and Marianne Hundt (eds.),

English as a Contact Language. Cambridge: Cambridge University Press, pp. 283–97. Thomason, Sarah G. and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Trudgill, Peter 2004. Linguistic and social typology: the Austronesian migrations and phoneme inventories. Linguistic Typology 8: 305–20. Trudgill, Peter 2009. Sociolinguistic typology and complexification. In Geoffrey Sampson, David Gil, and Peter Trudgill (eds.), Language Complexity as an Evolving Variable. Oxford: Oxford University Press, pp. 98–109. Trudgill, Peter 2010. Contact and sociolinguistic typology. In Raymond Hickey, ed., The Handbook of Language Contact. Oxford: Wiley‐ Blackwell, pp. 299–319. Trudgill, Peter 2011. Sociolinguistic Typology: Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. van Coetsem, Frans 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris.

2 Contact, Bilingualism, and Diglossia LOTFI SAYAHI 1 Introduction This chapter presents the differences between language contact in the context of classical diglossia and language contact in the context of bilingualism. In the first case, the genetic relationship between the two varieties involved is a very close one (e.g. the relationship between Standard Arabic and vernacular Arabic varieties), as emphasized in the work of Charles Ferguson (1959, 1991). In the second case, the two varieties in contact are clearly distinguishable languages that may even belong to separate families (e.g. Tunisian Arabic and French). The latter situation has been referred to in the literature as “extended diglossia,” following Joshua Fishman’s (1967, 1980) proposal that separate languages can also be in a diglossic situation. It will be argued here, however, that extending the term “diglossia” to contexts where a functional compartmentalization between any two linguistic systems obtains, regardless of their genetic relatedness, has rendered the concept of diglossia less productive. Because of internal and external factors, contact between varieties of the same language that are in a diglossic situation takes a different path and may lead to different outcomes than contact between separate languages. The conflation of the two types of situations has led to the relative marginalization of the concept of diglossia within contact linguistics. Often, diglossia is only used to describe the functional predominance of one language over another in official domains in the form of an assumed clear‐cut compartmentalization. But, to reclaim the concept of diglossia as a useful one for contact linguistics, questions need to be asked about the nature of contact in situations that involve different varieties of the same language and situations that involve clearly distinguishable languages. Furthermore, it is useful to explore contact situations that involve varieties of the same language and other separate languages, as is the case today in many parts of North Africa, the Middle East, and South Asia. These include questions about the direction of lexical and structural influence between all varieties involved and the implications for language maintenance, language shift, and the emergence of new varieties.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

52  Lotfi Sayahi

1.1 Definitions 1.1.1  Classical Diglossia The term “diglossia” in its early uses (Krumbacher 1902; Psichari 1928; Marçais 1930) and in the definition presented by Ferguson (1959), which brought the concept to the forefront of sociolinguistic theory, refers to the existence of two varieties of the same historical language. One variety is standardized and often used in written form and in formal domains, the High variety (H), and another variety that is not standardized and is not used in education or formal domains, the Low variety (L). This type of situation where two varieties of the same language are in a functional distribution within the same community is referred to in the field as classical, genetic or endoglossic diglossia. Ferguson (1959) presents a theory of diglossia that lists several features of situations that can be described as diglossic. The most salient feature is the functional distribution that limits the L variety’s use to informal contexts and extemporaneous communication, while the H variety is reserved for formal domains and written communication. A second feature describes a positive attitude by the speakers towards the H variety and a much less positive attitude toward the L variety, which is often described as a “deteriorated” version of the “original” language. The third feature refers to the existence of an established body of literature in the H variety that harkens back to its early history, while the literature produced in the L variety is often transmitted orally and rarely recorded in written form. In this last case, the situation has changed rapidly with the spread of literacy and digital communication. For example, it is becoming increasingly common to find electronic publications and digital private and quasi‐private communication recorded in vernacular Arabic varieties (Høigilt and Mejdell 2017; Sayahi 2019). A fourth characteristic of a classical diglossic situation has to do with the natural cross‐generational transmission of the L variety as a first language in contrast to the need for formal instruction to acquire the H variety. Ferguson points out as a fifth feature the fact that the H variety is a highly codified system, while references that prescribe the use of the L variety are non‐existent. The sixth feature established by Ferguson is that diglossic situations tend to be stable, i.e. this type of situation takes time to emerge and time to come to an end. At the linguistic level, Ferguson analyzed three additional features typical of a classical diglossic situation. Feature seven refers to the idea that the grammar of the H variety is more complex than that of the L variety. For example, the number of subject pronouns in Arabic vernaculars is usually half that of subject pronouns in Standard Arabic. In Tunisian Arabic, there are no dual pronouns and there is no difference between masculine and feminine third person plural pronouns while in Standard Arabic there is. The eighth feature establishes that, in general terms, both varieties share the same phonological inventory. This also varies although it is true that the differences are often an issue of frequency, as opposed to the categorical presence or absence of a given sound. For example, in some Arabic dialects although /g/ dominates where Standard Arabic uses /q/, we can still find words where /q/ is used in all dialects as in the case of the word /qurʔa:n/ ‘Quran.’ Finally, feature nine captures the idea that, despite the obvious differences often signaled by the speakers themselves, the greater part of the lexicon is shared between both varieties. This also varies according to the history of the L varieties and the contact situations in which they have been involved. These nine features as described by Ferguson, although often in need of adjustment according to each situation, remain central to any description of the processes and outcomes of language contact under classical diglossia. In his original article, Ferguson (1959) references the situation of four different languages as cases of classical diglossia: Arabic, Modern Greek, Swiss German and Haitian Creole. Although the case of Haitian Creole has been debated and theories about its formation make its situation less of a typical classical diglossic one, research on the other defining languages

Contact, Bilingualism, and Diglossia  53 has offered a wealth of data regarding the emergence, stability, and the conclusion of diglossia (as examples, for Arabic see Kaye (1994) and Versteegh (1997), for Greek see Frangoudaki (1992) and Silk and Georgakopoulou (2016), and for Swiss German see Keller (1982) and Clyne (1999)). Furthermore, situations of classical diglossia have been identified in many parts of the world and across many historical periods (Hudson 2001). In addition to the cases mentioned by Ferguson, one can think of other situations of classical diglossia such as the ones concerning Latin and the Romance languages (Kabatek 2016; Wright 2010), some South Asian languages, including Tamil (Britto 1991), Bengali (Chakraborti 2003), and Sinhala (Paolillo 1997), and the case of the emergence of Afrikaans (Deumert 2004) as theoretically intriguing cases that warrant the consideration of the concept of diglossia for the understanding of dialect drift and language evolution.

1.1.2  Extended Diglossia/Societal Bilingualism Fishman (1967, 1980) employs the term “diglossia” to describe situations where two linguistic varieties, regardless of their genetic relatedness, coexist in the same speech community but are used in different domains. He specifically describes bilingualism as the individual’s ability to use more than one language, while defining diglossia as an “enduring societal arrangement” (1980: 3) in which each language is assigned a set of functions. As such, he postulates the existence of communities with bilingualism and diglossia at the same time, and communities where only one of the two obtains (Fishman 1967). This type of situation is referred to in the literature as extended diglossia, non‐genetic diglossia or exoglossic diglossia. Fishman referenced the case of Paraguay where Spanish and Guarani are used, even if recent studies are showing that Paraguay is a successful case of societal and educational bilingualism (Gynan 2005). Other situations that have been analyzed under the scope of extended diglossia include the situations of a wide range of minority languages, such as those in the Spanish‐ speaking world (Blas Arroyo 2008) and in the African context (Mkilifi 1978). All the same, serious criticism has been waged against the value of extending the term “diglossia” to cover cases of societal bilingualism based on functional distribution alone (Winford 1985; Ferguson 1991; Hudson 1991, 2002; Sayahi 2014, 2017). Winford (1985: 346) argues that “[s]ince the functions of any two languages controlled by a single speaker are almost always in a partial or total complementary distribution, clearly then, to equate diglossia with bilingualism is not very useful.” While some of the features listed by Ferguson as typical of a diglossic situation can be present in cases of extended diglossia, the reality is that most of them are always present in classical diglossic situations but rarely seen in cases of extended diglossia. Specifically, the first two features that Ferguson described – the functional distribution of the varieties involved and a positive attitude toward the H variety and a negative attitude toward the L variety – are the ones that appear to be more common in both types of situations. Even so, one can argue that in some extended diglossic situations, a positive attitude toward the socially subordinate language can be expected, especially when it serves as an ethnic identity marker. This makes functional distribution the most significant defining feature that both types of situations share. The rest of the features may be present in some contexts of extended diglossia but not in others. For example, in some communities that would fit the description of a case of extended diglossia, the L variety, often the ethnic language of the community, enjoys prestige elsewhere and may even be playing the role of the H variety in other areas (for example, the situation of Spanish in Spanish‐speaking countries vs. that of Spanish in the United States). Equating the functional distribution of any two or more linguistic varieties with them being in a diglossic situation, as evidenced by the use of terms such as triglossia, polyglossia, multilayered diglossia, etc., limits the theoretical import of the concept in understanding mechanisms of language contact and possible contact‐induced change when both varieties

54  Lotfi Sayahi involved are from the same historical language. For the remainder of the chapter, diglossia will be used to refer to the existence of two varieties of the same language, while b ­ ilingualism will be used to refer to the existence of two separate languages within the same community.

1.2  The Question of Genetic Affiliation In his original definition, Ferguson (1959: 336) made it clear that the varieties in the kind of situation he described as diglossic were meant to be “primary dialects” in contrast with a “highly codified […] superposed variety” of the same language. The H variety is standardized and closely monitored by institutions and educated community members, while in the case of the L variety there is a total “absence of normative pressure” (Silva‐Corvalán 1997: 10). The absence of normative pressure in the case of the L variety leaves it open to change that stems from community use, unlike in the case of the H variety which changes according to institutional planning. Another point that is worth highlighting here, and which has implications for the nature of language contact in cases of diglossia and in cases of bilingualism, is that Ferguson (1991) also emphasized that the speakers in a diglossic situation must perceive the varieties involved as part of the same language. This is obviously very different from cases of bilingualism, as we know them. No speaker of Arabic and French in North Africa would argue that they are the same language, but the average speaker of Tunisian Arabic would argue that it is Arabic albeit not the standard variety. Furthermore, speakers of the L variety recognize and contribute to its low status by showing prejudice towards their own native variety in favor of the H variety. This makes influence from other languages less objectionable to the speakers given the lack of purist attitudes towards the L variety, which by definition is seen as an ‘impure’ variety. In the case of Tunisian Arabic, although attitudes have been slowly changing since the transition towards a democratic system following the Revolution of 2011, even teachers still perceive the L variety as an unacceptable system to introduce in education while defending the “authentic” Arabic language. One of the participants cited in Sayahi (forthcoming) states that: “the dialect doesn’t mean much when compared to having competence in other languages ­especially our authentic Arabic language.” This is particularly pertinent to the distinction van Coetsem (1988) establishes between recipient language agentivity and source language agentivity. Van Coetsem (1988: 1) posits that “the notions of loan and borrowing also remain ill‐defined, because the agent of the action is not taken into account” and argues that there is a difference between linguistic transfer in cases when the agent is a speaker of the recipient language and cases when the agent is a speaker of the source language. In the case of diglossia, the L variety is always the native language of the speakers while the H variety is only acquired through formal instruction. This will help predict that any transfer between the L variety and the H variety will, according to van Coetsem’s proposal, take the form of borrowing when the L variety is the recipient language, and imposition when the H variety is the recipient language. It appears, then, that from a methodological standpoint, it would make sense that contact between varieties of the same historical language, which the speakers themselves perceive as one language, be referred to as “contact under diglossia” and contact between clearly distinguishable languages, that are perceived as such by the speakers, should be described as “contact under bilingualism.” Such a distinction is already established when scholars distinguish between language contact and dialect contact, even if diglossic contact is indeed a separate category (Trudgill 1986; Otheguy et al. 2007). The lexical and structural differences that separate the L variety and the H variety are much more significant than differences between most dialects of non‐diglossic languages. The number of divergent features across the English dialect groups are much less of an impediment to mutual intelligibility than the

Contact, Bilingualism, and Diglossia  55 differences between Moroccan Arabic and, for example, Egyptian Arabic which may in fact result in unintelligibility. Yet, the H variety in Morocco, Egypt, and across the Arabic‐ speaking world, is the same standard variety of Arabic.

1.3  The Question of Functional Compartmentalization Both in Ferguson’s (1959) article, and in many subsequent studies, the separation by domains of use of the two varieties in a diglossic situation was believed to be more clear‐cut than is often the case in reality. The fact that both varieties were assumed not to be used within the same domains implied a separation that, if breached, would result in the speakers losing face and the interaction ending in miscommunication. In addition, since both varieties were thought to be mutually exclusive because of a stringent functional compartmentalization, it was expected that there would be no possibility for lexical or structural influence. However, quantitative studies have shown that, indeed, the mutually exclusive distribution of both varieties in a classical diglossic situation is not as straightforward as has been claimed. In fact, Fasold (1984) proposes that the separation in domains of use is never categorical and describes what he refers to as cases of “leaky diglossia.” These include instances where one variety is used in a domain that is traditionally reserved for the other one. For instance, Ferguson mentions that formal domains such as mosques would not see the use of the L variety. However, imams in mosques tend to use Standard Arabic when they read written sermons or recite religious materials, but they also use the local dialect to parse what they say in Standard Arabic for their audience, in many cases composed of illiterate members given widespread illiteracy in the Arabic‐speaking world (Bassiouney 2006; Sayahi 2014). Increasingly, the use of vernacular Arabic in domains that were previously limited to read‐ aloud materials in Standard Arabic has resulted in the use of both the H variety and the L variety in education, media, and even in written form (section 3). The question of functional compartmentalization as the major linchpin of connecting societal bilingualism to diglossia is, in fact, one point that sets apart contact between varieties of the same language and contact between different languages. In many cases of bilingualism, both languages can be acquired formally and are used in education either within the same community or elsewhere. For example, while Spanish may be described as being in a diglossic situation in the United States (Valdés and Geoffrion‐Vinci 1998), it is the second most spoken language in the world and the official language in 21 countries in three different continents. Speakers of Spanish in the United States have growing opportunities to learn and use Spanish with a wealth of materials developed for heritage speakers and opportunities to interact with native speakers in the United States and abroad. In fact, estimates of the number of Spanish speakers in the United States position the country as the second largest Spanish‐speaking country, second only to Mexico (Escobar and Potowski 2015). In contrast, in the case of classical diglossia, the L varieties are not codified or taught to the native population. There are no materials that encourage the use of the Arabic dialects in education anywhere, except perhaps for some textbooks to teach the dialects to foreign students. In fact, Tunisian schoolchildren must put aside their knowledge of Tunisian Arabic as they embark on developing literacy (Sayahi 2015a). In the Maghreb countries in general, education is primarily in Standard Arabic and French, thus marginalizing the L varieties of Arabic in all domains where literacy is required. This is again different from the case of Spanish in the United States, where many Hispanics can use English in informal contexts and Spanish in professional contexts making the situation unlike that of Arabic where nobody uses the H variety at home. The term “diglossia,” even in its extended definition, does not indeed capture the use of English and Spanish in public and private domains in many areas in the United States (Pedraza et al. 1980; Silva‐Corvalán and Enrique Arias 2017).

56  Lotfi Sayahi In sum, what drives the functional compartmentalization in a classical diglossic situation is the attitude and perception of the speakers, the marginalization of L varieties in education because they lack recognition and standardization, and the absence of the H variety in informal domains.

2  Language Contact Phenomena 2.1  Code‐switching and Lexical Insertions While bilingual code‐switching has attracted a wide range of attention from studies analyzing its sociolinguistic significance, discursive functions and syntactic structure making it one of the most investigated topics in linguistics, code‐switching between the H variety and the L variety in a classical diglossic situation has been the subject of only a handful of studies (Eid 1988; Walters 1996a, 1996b; Bassiouney 2006; Boussofara‐Omar 1999, 2003, 2006; Sayahi 2014). In fact, analysis of diglossic code‐switching shows that it presents significant differences from bilingual code‐switching in terms of frequency of occurrence, direction of the switch, and functions. Also important is the question of what type of lexical influence is more common when the linguistic varieties in contact are different languages and when they are varieties of the same language. Poplack and Dion (2012) argue that, in general, inserted lone other‐language items tend to be borrowed while multiword fragments tend to be code‐switched. This section will show that this is also largely the case in situations of diglossic language contact, despite the close typological distance between the varieties involved. One way to compare what happens in each situation is to look at speech samples of speakers in communities where both classical diglossia and bilingualism are present, as is the case in Tunisia.

2.1.1 Code‐switching In this chapter, bilingual code‐switching refers to switching between two clearly distinguishable languages, Tunisian Arabic and French, while diglossic code‐switching refers to switching between the H variety and the L variety of the same language, Tunisian Arabic and Standard Arabic. The differences between the two are visible at the levels of frequency of occurrence, most common types, direction of the switch and main discursive functions. Regarding frequency, in the North African context at least, diglossic code‐switching is overall less frequent than bilingual code‐switching. The reason why it is not as frequent as bilingual code‐switching is that for it to happen, the context of interaction must allow the use of the H variety, a code not used for extemporaneous communication. This limits the contexts for diglossic switching to domains where the use of the H variety is unmarked and where the L variety is increasing its presence. But it does not make diglossic code‐switching itself always a marked choice as postulated by Myers‐Scotton (1986: 411–10), but once the speakers recognize the H variety as an unmarked choice, they feel the need to use it but may switch to the L variety, especially when they are not reading aloud. Contexts where this happens include mosques, where both the H variety and the L variety are used across many parts of the Arab world. Example (1), extracted from a Friday sermon, shows how the imam introduces the quotes from the Quran and explains what they mean in Tunisian Arabic.1 He switches to Standard Arabic to recite the quotes themselves verbatim. It is important to note here that, throughout history, religious institutions, while consecrating the classical language of the religious texts, often have resorted to employing simplified versions of the classical language or the L varieties to reach laypeople, as was the case with Latin (Marín 1984; Wright

Contact, Bilingualism, and Diglossia  57 1982, 1994). Another context for diglossic code‐switching is mass media. Example (2), from a news program on the Tunisian State television, shows how the guest, a medical doctor, switches from Standard Arabic to Tunisian Arabic and even inserts some French items. A third context is the political domain. Example (3), from a session of the Tunisian Assembly, illustrates how a member of the Assembly starts his speech in Standard Arabic and then slips into Tunisian Arabic. (1) wu ha n‐ness ha:ðu:ma illi xðɛ:u ṣ‐ṣi:fa ha:ðiyya wʕidhum rabbi subh̄a:nu wa taʕa:la bi‐l‐falɛ:h̄ wa n‐naʒɛ:h̄ ɛ:ʃ qa:llhum? “ʔittabiʕu: al‐mursali:n ʔittabiʕu: man la: yasʔalukum ʔaʒran wa hum muhtadu:n” wu bayinilhum “wa ma:luni: la: ʔaʕbudu al‐laði: faṭari:?” tawwa insɛ:n rabbi xalqu wu ṣawwru yimʃi: yiʕbid ɣi:ru? (Mohamed Machfar 2009) And the people who obtained this status, God Almighty promised them prosperity and success. What did he say to them? “Follow the messengers, follow those who do not ask you wages and are rightly guided.” And he clarified to them: “And why should I not worship the one who created me?” Now, a person whom God created and shaped goes and worships another? (2) al‐maḍa:r mtɛ:ʕ al‐ʔidma:n ʕala: al‐internet tuðhir ʕala: al‐ ʒasad. famma sɣayra:t yitʃakɛ:u min ʃyɛ:h̄ fi l‐ʕini:n, nuqs fi n‐nðar, famma qillit tarki:z fi d‐dira:sa mtɛ:ʕhum, famma wuʒi:ʕa mtɛ:ʕ ra:s. insɛ:n ki yuqʕid barʃa sur internet wu ʕla l’ordinateur ra:su bɛ:ʃ yu:ʒʕu: (Ben Moallem Hachicha 2018) The damages of internet addiction are noticed in the body. There are little children who complain of dry eyes, bad sight, there is a lack of concentration in their studies, there are headaches. A person, when he spends a lot of time on the internet or on the computer, his head will hurt. (3) kɛ:nit ṣ‐ṣaʃiyyɛ:t tuʕṭa: maʒʒa:nan ʔṣbah̄ at tuba:ʕ bi‐al‐ma:l faqaṭ. yaʕni l‐hadaf min annu nah̄h̄ɛ:u ṣ‐ṣaʃiyyɛ:t hɛ:ða ma wuqiʕiʃ. (Noomane El Euch 2018) Plastic bags were given for free, now they are sold for money only. This means that the goal of removing plastic bags was not met. In cases of bilingual code‐switching, both varieties are used in extemporaneous communication where code‐switching is more likely to happen. Code‐switching between local Arabic varieties and French in daily interactions is very common across North Africa. In fact, in natural conversation, the only code‐switching that one can observe in informal settings is between Tunisian Arabic and French, without involving Standard Arabic. In (4), we have an example where a speaker switches to French at different points while talking about her studies. (4) l‐fac l‐wah̄i:da illi fi:ha l’environnement. θamma wah̄da fi:ha analyse chimique mais muʃ appliquée à l’environnement fi tu:nis, ma naʕriʃ fi Zaghouan je croix. (Female speaker) It is the only college that has environmental science. There is one that has chemical analysis but not applied to the environment in Tunis, in Zaghouan I think. In cases where Standard Arabic is an unmarked code, the switch to Tunisian Arabic carries with it the possibility to switch to French once Tunisian Arabic becomes the base language. Example (5) below is an excerpt from a response by the Minister of the Environment to a question addressed to him in a special session of the Tunisian Assembly. The Minister starts by using Standard Arabic but as he uses the Tunisian Arabic discourse marker maʕnitha, he inserts more French items into a sentence that is clearly in Tunisian Arabic.

58  Lotfi Sayahi (5) miʃ sunʕ faqaṭ miʃ sunʕ wa lɛ:kin h̄ atta z‐zira:ʕa maʕnitha les produits mtɛ:ʕ l‐ biodégradable inaʒʒmu iṣi:ru fi tu:nis une valeur ajoutée kbi:ra hɛ:ða ɛ:ʃ nqu:lu ah̄na. (Mouakhar 2018) Not manufacturing only, not manufacturing but also cultivation. I mean the biodegradable products can be made in Tunisia. It is a great added value, that is what we say. Regarding the direction of the switch, diglossic code‐switching tends to happen more frequently from the H variety, the expected code in the given domain, towards the L variety. This is significant because the speakers often start in the H variety and then move towards the L variety, either because of difficulty communicating in the H variety or because they naturally slip into their native dialect. In the North African context, in the case of bilingual code‐switching, usually the speakers use the L variety as the base language with frequent switches to French given that higher levels of competence are associated with the native language rather than the foreign language (Sayahi 2011). But speakers can switch to French at different switch points and then come back to the L variety, unlike what happens with switching into the L variety from the H variety, where a speaker rarely goes back to the H variety within the same sentence. In general terms, diglossic code‐switching shows less mixing than bilingual code‐switching. Still, scholars have shown, the mixing of the H variety and the L varieties of Arabic is better understood if approached as cases of code‐switching and borrowing as opposed to the emergence of a stable third middle variety (Walters 1996a, 1996b, 2003; Boussofara‐Omar 2006; Sayahi 2014). Regarding the type of the switch, in the case of diglossic code‐switching more instances of intersentential code‐switching can be observed as speakers move from the H variety to using the L variety. In the case of bilingual code‐switching, the switch can take all types of forms depending on the levels of speaker competence in both languages; speakers with higher levels of competence have been shown to switch more often at the intrasentential level than those with less competence (Poplack 1980). Diglossic code‐switching beyond lone items is rarely the case in informal communication where the L variety is the unmarked code, and even these lone items may be better described as borrowings, rather than as instances of code‐switching. Finally, with regard to the possible reasons for speakers to code‐switch, although there is not always a one‐to‐one correspondence between a given instance of code‐switching and a reason that can explain it, bilingual code‐switching has a wide set of functions as described in the literature (Gumperz 1982; Myers‐Scotton 1993, among many others). In contrast, diglossic code‐switching often has four apparent reasons. The first reason is the variable competence in Standard Arabic. As mentioned on p. 55, this is a result of varying levels of education, practice, and comfort with the language. It is understandable that a medical doctor who is trained in a monolingual French system, as is the case of medical schools in Tunisia, finds it hard to talk about her profession in Standard Arabic, a language in which she may have never learned the technical terms. She may switch to her native dialect which in turn allows her to code‐switch to French and insert technical items as needed. The second reason is one that I will call here a parsing function. This happens when a speaker aims to simplify the message and reach people in a much more direct and simplified fashion as in the case of religious discourse. The third reason is quotation, when the speaker quotes directly from written texts, as in the example of the imam in (1) above. Finally, the fourth reason has to do with reflecting ideological stances that favor the Arabo‐Islamic identity by avoiding switching to French in informal contexts and instead inserting items from Standard Arabic. This tends to mark the discourse of more conservative people with a clear religious inclination.

Contact, Bilingualism, and Diglossia  59 It is interesting that in a situation such as the Tunisian case, where both classical diglossia and bilingualism exist, code‐switching happens in parallel manners given that the contact is between Standard Arabic and Tunisian Arabic on the one hand, and Tunisian Arabic and French on the other. That means it is not common to find switching between Standard Arabic and French directly without the meditation of Tunisian Arabic. This has direct implications for the lexicon of Tunisian Arabic as it is the variety that is more open for lexical borrowing from both Standard Arabic and French.

2.1.2  Lexical Insertions As detailed in this volume and elsewhere in the literature on contact linguistics (Winford 2003), lexical change is the most common and easily observable outcome of language contact. There are, however, several differences that can be drawn between processes of lexical borrowing in cases of diglossic contact situations and lexical borrowing in bilingual contact situations. A first distinction has to do with the fact that the typological proximity in cases of classical diglossia and the prestige of the H variety a priori facilitate insertions from the H variety into the L variety. Given that the H variety is the language of religion, literacy, and learned discourse in general, the L variety draws a large part of its learned lexicon from it. The perception by the speakers is that these are two varieties of the same language which facilitates the transfer of technical and specialized vocabulary through a process of vernacularization. Additionally, the changes that a word would require to be adapted into the structure of the L variety may not be as costly as the adaptation of a word from a different language. Inserting lexical items from the H variety, however, rarely deals with concepts that have an equivalent in the L variety. These are usually technical words that have been acquired through education, used by the government and in the administration, or terms adopted in the media that gradually fill a lexical gap in the L variety. Examples in Tunisian Arabic include words related to the administration such as taqa:ʕud ‘retirement,’ muwaððif ‘public servant,’ and ʃahriyya ‘salary.’ This is an argument in favor of the existence of two types of vocabulary in the L variety and the separate language that may eventually develop from it, as is the case for the Romance languages. In Spanish, scholars refer to palabras patrimoniales ‘patrimonial words’ which are words that developed with the language since drifting away from the H variety, and palabras cultas ‘learned words’ which were adapted from Latin when it served as the language of education and knowledge in later stages of the standardization of Spanish (Hualde et al. 2009: 291–2). On the other hand, inserting items from French into Tunisian Arabic can affect more semantic fields and they may even compete with native words, including basic vocabulary. It is true that the line between lexical borrowing and code‐switching can be blurred. But, in many cases, conflict sites show that lone French insertions, although not necessarily phonologically adapted into Tunisian Arabic, in fact show adaptation at the morphosyntactic level. This was proven by Poplack et al. (2015) in their study of lone nouns in Tunisian Arabic that agree with Tunisian Arabic and not French in terms of placement of the demonstrative, plural marking, and the expression of possession. This is also clear in the case of lone French verbs when inserted in Tunisian Arabic as they behave like native verbs despite the significant typological differences between the two languages. One example would be the borrowing of the verb partager ‘to share’ whose use is predominant on Tunisian sites of social media where it is often written in Arabic script and inflected according to Arabic morphological rules. In Tunisia, because bilingual code‐switching is often an unmarked choice in informal domains, loanwords from French may be more successful than those from Standard Arabic, which feel more artificial since they originate in learned discourse. It is noteworthy, however, that in cases of classical diglossia when the L variety starts to be recognized as a legitimate code for formal communication, usually through its use in

60  Lotfi Sayahi official contexts and in written form, it goes through a process of lexical realignment with the H variety. The phenomenon of lexical realignment is in line with Coulmas’ (2002) claims about the role of writing in the evolution of diglossic situations. He argues that “[i]f people choose to become literate and if they choose to speak as they write or to write as they speak, a realignment of codes will be brought about” (Coulmas 2002: 62). By lexical realignment, then, I refer to the heavy borrowing from the H variety into the L variety in cases of classical diglossia where the L variety is being elevated to a national language (Sayahi 2019). Through lexical realignment speakers draw from the H variety lexicon to fill in any lexical gap and avoid resorting to other languages. This process is not carried out in a planned fashion but rather it is a spontaneous reaction from the users themselves. Since code‐switching to a foreign language is usually a highly marked choice in formal contexts, where prepared discourses are read aloud, in Tunisia we have cases of speakers using the Tunisian dialect for official purposes with many insertions from Standard Arabic, even if that means adapting the Standard Arabic form to the morphophonological shape of the L variety (Sayahi 2019). For example, when he addressed the Tunisian people through State television, Prime Minister Youssef Chahed (2018) delivered a prepared speech in Tunisian Arabic without any code‐ switching to French, but with heavy lexical influence from Standard Arabic. The same happened in one of the first official documents published by the Tunisian government in the vernacular where a conscious effort was made to avoid French insertions and replace them with technical words from Standard Arabic (Ministry of Finance 2018). This strategy to realign the L variety with the H variety in formal contexts affects not only new concepts, such as lawh̄a raqmiyya ‘tablet’, but also even established borrowings from French which get replaced by Standard Arabic terms (e.g. bla:ṣa (from French place ‘place’) > maka:n; karta:ba (from French cartable ‘schoolbag’) > mah̄ifḍa; dusiyyɛ:t (from French dossiers ‘files’) > milafɛ:t). In a different scenario, when there is a total disconnect from the H variety, then foreign languages available to the community may become sources of lexical change. This was the case with Maltese where the link with written Arabic was broken abruptly in 1091 and borrowing from Italian and other Romance varieties filled many lexical gaps. The case of Maltese is typical of societal bilingualism where the native language borrows from the prestigious languages with which it is in contact in the absence of access to its historical language. In the other direction, foreign languages that are imposed on new communities may borrow some terms from the local languages, but in cases of diglossic situations the H variety does not borrow lexical items from the L variety. The H variety is closely monitored by official institutions that often operate from outside of one single community or even one single country as in the case of Modern Standard Arabic.

2.2  Structural Influence in Cases of Diglossia Even if the L variety and the H variety share a lot in common at the structural level, the fact that the H variety is not used extemporaneously limits the impact it has on the L variety. The opposite, however, as a result principally of varying levels of competence in the H variety, may result in the L variety of the speaker influencing the rendition of the H variety, in a case of “imposition” as described by van Coetsem.

2.2.1  Phonetic and Phonological Transfer Although Ferguson (1959) postulates that the phonetic inventories of the H variety and the L variety are similar, there are differences that have to do with frequency and diffusion of some segments, in addition to differences at the suprasegmental level. In the case of Standard Arabic and the Arabic dialects, one can mention the case of /q/ as it is realized variably in distinct parts of the Arabic‐speaking world. But, as a result of contact between both varieties,

Contact, Bilingualism, and Diglossia  61 Haeri (1996) found that, in the Egyptian dialect, [q] is being used by educated speakers under the influence of Standard Arabic, while the more common realization is either [ʔ] or [g] depending on the region. Another difference is the glottal stop which has been shown to be preserved in some insertions from the H variety into the L variety (Heath 1989). These cases are not instances of borrowing of a new phone but rather changes in frequency and contexts of occurrence of an already existing sound. In the other direction, Standard Arabic as spoken by native speakers of the different dialects is strongly influenced by the latter. It is relatively easy, in fact, to distinguish speakers of Standard Arabic by region based on influence from their native dialects. This influence concerns segmental and suprasegmental features, including syllable structure and intonation. For example, the realizations of /ʒ/ and /θ/ vary significantly according to the dialect zone, and this carries over into the rendition of Standard Arabic. Looking at bilingual language contact, studies have shown that phonological influence as a result of the nativization of loanwords may lead to what van Coetsem (1988: 98) describes as a “subsidiary phonological borrowing.” This is a reasonably common phenomenon, especially in cases where a sound is missing from a phonemic inventory or is only present as an allophone (Winford 2003; Thomason and Kaufman 1988). In the case of the North African Arabic varieties, we find that the voiced labiodental fricative and the unvoiced bilabial stop have been incorporated along with borrowings from Romance languages, especially in the case of bilingual speakers (Sayahi 2014). A comparable situation is observed in Maltese where some sounds, /p/, /v/, /tʃ/, and /g/ (Borg 1994: 25), which are not present in historical Arabic, are introduced through loanwords from Romance languages.

2.2.2  Morphosyntactic Convergence Since discourse markers are syntactically independent, we find frequent insertions of different discourse markers from the H variety into the L variety in cases of classical diglossia, the same way we find them in bilingual situations. Studies on Spanish in contact with English (Torres 2002; Lipski 2005) have shown that speakers use these markers without necessarily displacing the native ones. In cases of classical diglossia, more discourse markers from the H variety will be inserted when the L variety is used in formal contexts. But in informal contexts, where code‐switching to French is unmarked, French discourse markers can be found frequently. In the case of Tunisian Arabic, the French word donc ‘so’ is particularly frequent and its use is comparable to similar words borrowed in other languages. Regarding morphosyntactic features, the use of Standard Arabic forms such as the complete pronominal system and the corresponding complex verb paradigms, negation structure, and the case system, among many divergent features between the H and the L varieties, is much more marked than the use of lexical insertions. In turn, the H variety tends to be resistant to permanent change as it is taught through textbooks with no native speakers on hand. As with any other cases of second language acquisition, acquiring the H variety and developing advanced competence in it requires extensive training that leads to varying levels of competence. Some cases of influence from the L variety into the H variety include the lack of use of the dual, the use of the vernacular complementizer illi and other function words (see for example the use of miʃ ‘not’ in a Standard Arabic sentence in example (5) above), and even the use of the analytic genitive exponent inexistent in Standard Arabic (Sayahi 2015b). Variation in the H variety, although visible in the speech of its users, reflects both the individual levels of attainment and the effects of the native dialect. These instances of divergence from the rules of the codified language do not make it into textbooks, grammars, or the school system. They are perceived as learner errors and not part of the “real” target language whose mastery is the objective.

62  Lotfi Sayahi

3 Language Maintenance, Language Shift, and Language Genesis Classical diglossia is marked by its stability over generations of speakers and even centuries, as is the case with Arabic. In bilingual situations, language shift can happen over a generation given the appropriate conditions. The direction of the influence between the varieties involved leads to different outcomes that may result in dialect drift and language shift, depending on whether we are dealing with a classical diglossic situation, a situation of bilingualism, or one that involves both scenarios. The evolutionary course of diglossia is divergent from the outcome of contact between two different languages in cases of societal bilingualism. While contact under bilingualism often shows a more common path for the maintenance and nativization of the more prestigious variety, in cases of diglossia the less prestigious variety is the one transmitted naturally and the one that competes with other languages that may be brought into the community. It is also the one that is more likely to develop into a separate language with, in many cases, considerable influence from other languages. By contrast, the standard variety of the diglossic language remains a highly codified system that is only learned through formal instruction and may be more resistant to contact‐ induced change. The case of Maltese is a particularly revealing one in understanding language emergence from within a diglossic situation. Once the diglossic situation ended abruptly as a result of ethno‐religious strife and the expulsion of the Muslim population, the Arabic L variety used on the Maltese islands at the time continued its process of development without the influence of Standard Arabic. This led to a high level of lexical and structural influence from non‐Semitic languages to the degree that Mifsud (1995: 33–4) describes two types of morphologies in Maltese depending on the source language of the vocabulary item: a Semitic Maltese morphology and a non‐Semitic Maltese morphology. This is another major point of divergence between contact among linguistic varieties belonging to the same historical language and contact among varieties that are genetically distant. In cases of classical diglossia, the L variety is the one that spreads to domains of the H variety, eventually displacing it as it gains prestige and recognition. In the case of societal bilingualism, the socially superordinate language, usually the language imposed on native communities, ends up displacing the native variety as many examples around the Spanish‐ speaking world can prove (Sayahi 2018). Eckert (1980), in her description of contact between Gascon and French, reached the conclusion that the functional distribution of these two separate languages was in fact the reason behind the language shift toward French. The stability that is observed in classical diglossic situations – which lasts longer periods of time and often requires major sociopolitical events to change – makes situations of societal bilingualism different. Under bilingualism, one of the languages may be restricted in its domains of use which often leads to a lower value in the linguistic market and the eventual breaking up of natural transmission. The Arabic diglossic situation, on the other hand, has endured for centuries as did the Latin and the Greek situations. This is a fundamental difference between classical diglossia and societal bilingualism with major implications for dialect drift, language genesis, and restructuring in addition to language maintenance and language shift. A final difference is that while contact between two separate languages can be a contact between equally codified and standardized systems, each with its own writing history and literary production, in the case of classical diglossia, the writing of the L variety and its extension to formal domains are pivotal in its evolutionary route. Often coupled with increased nationalistic feelings, writing the local dialect and its use for official purposes will lead to additional recognition and the need for its description, acceptance and use in education (Ferguson 1962). Although this is not the case in any nation of the Arab world today, we do see signs of change in societies where the political system has been reconfigured. In the

Contact, Bilingualism, and Diglossia  63 case of Tunisia, the democratization process has opened the door to more spontaneous communication in Tunisian Arabic, not only in mass media but also in official domains. This in turn has led to discussions about what the future of Tunisian Arabic should be in a process that aims to distinguish the country as an independent democratic nation. As happened with the Romance languages, where the decision to use the vernaculars in institutions of power and knowledge led to their recognition as separate languages, it appears that once the L variety crosses that barrier then its drift from the historical language may accelerate. In turn, once the H variety is relegated to liturgical and literary functions, the reduction in its use, coupled with non‐native competence, may soon render it only marginally significant.

4 Conclusion This chapter argues that contact between varieties that are in a classical diglossic situation is different from contact in situations of bilingualism. Although functional distribution has been used as an argument to conflate both types of contact, a closer look at the processes and outcomes of contact in each type of situation shows that they are fundamentally different. In diglossic situations, the vernacular is much more open to contact‐induced change, as the variety that is acquired as a first language, but at the same time the one with no significant normative pressure. Eventually, as the L variety begins to be used in written form and in formal contexts, it encroaches upon the H variety in contexts that were previously off limits. This signals the beginning of the shift towards the emergence of the L variety as a separate language that may borrow a large number of lexical items as learned vocabulary from the H variety in a process that is described here as one of lexical realignment. In turn, in situations of contact between separate languages, the shift is usually in the opposite direction as speakers abandon their ancestral language to learn the more prestigious variety or the language of the majority. The history of the Romance languages, Modern Greek, and the current situation of Arabic illustrate well the route that the evolution of diglossia takes, showing that a separate language emerges when speakers cease to see their dialect and the standardized variety as the same language. The agentivity of the speakers in writing the L variety and using it in formal contexts, education, and government are clear steps towards the full recognition of the distinctive identity of a linguistic system as a separate language. Studying diglossic situations, past and present, offers a good window into language emergence and the role played by genetic and non‐genetic language contact in this context.

NOTE 1 All Arabic material is reproduced in italics based on the IPA conventions, except for pharyngealized sounds, which are indicated with diacritics placed below the letter, and the voiced palatal approximant, transliterated as y. Standard Arabic material is in bold.

REFERENCES Bassiouney, Reem 2006. Functions of Code‐ switching in Egypt: Evidence from Monologues. Leiden: Brill.

Ben Moallem Hachicha, Maha 2018. Interview. Blas Arroyo, José Luis 2008. Sociolingüística del español: desarrollos y perspectivas en el estudio de

64  Lotfi Sayahi la lengua española en contexto social. Madrid: Cátedra. Borg, Alexander 1994. Observations on some evolutionary parallels and divergences in Cypriot Arabic and Maltese. In Jordi Aguadé, Federico Corriente, and Marina Marugán (eds.), Actas del Congreso Internacional sobre Interferencias Lingüísticas Arabo‐Romances y Paralelos Extra‐Iberos. Zaragoza: Navarro and Navarro, pp. 21–40. Boussofara‐Omar, Naima 1999. Arabic diglossic switching in Tunisia: an application of Myers‐ Scotton’s MLF model (University of Texas at Austin Ph.D. dissertation). Boussofara‐Omar, Naima 2003. Revisiting Arabic diglossic switching in light of the MLF model and its submodels: the 4‐M model and the Abstract Level model. Bilingualism: Language and Cognition 6.1: 33–46 Boussofara‐Omar, Naima 2006. Neither third language nor middle varieties but Arabic diglossic switching. Zeitschrift für Arabische Linguistik 45: 55–80. Britto, Francis 1991. Tamil diglossia: an interpretation. Southwest Journal of Linguistics 10.1: 60–84. Chakraborti, Paromita 2003. Diglossia in Bengali (University of New Mexico Dissertation). Chahed, Youssef. Speech. Clyne, Michael G. 1999. The German Language in a Changing Europe. Cambridge: Cambridge University Press. Deumert, Ana 2004. Language Standardization and Language Change: The Dynamics of Cape Dutch. Philadelphia: Benjamins. Coulmas, Florian 2002. Writing is crucial. International Journal of the Sociology of Language 157: 59–62. Eckert, Penelope 1980. Diglossia. Separate and unequal. Linguistics 18: 1053–64. Eid, Mushira 1988. Principles for switching between Standard and Egyptian Arabic. Al‐Arabiyya 21: 51–80. El Euch, Noomane 2018. Speech. Escobar, Anna Maria and Kim Potowski 2015. El español de los Estados Unidos. Cambridge: Cambridge University Press. Fasold, Ralph W. 1984. The Sociolinguistics of Society. Oxford: Blackwell. Ferguson, Charles A. 1959. Diglossia. Word 15: 325–40. Ferguson, Charles A. 1962. The language factor in national development. Anthropological Linguistics 4. 1: 23–7. Ferguson, Charles A. 1991. Diglossia revisited. Southwest Journal of Linguistics 10: 214–34.

Fishman, Joshua A. 1967. Bilingualism with and without diglossia: diglossia with and without bilingualism. Journal of Social Issues 23.2: 29–38. Fishman, Joshua A. 1980. Bilingualism and biculturism as individual and societal phenomena. Journal of Multilingual and Multicultural Development 1: 3–15. Frangoudaki, Anna 1992. Diglossia and the present language situation in Greece: a sociological approach to the interpretation of diglossia and some hypotheses on today’s linguistic reality. Language in Society 21: 365–81. Gynan, Shaw N. 2005. Official Bilingualism in Paraguay, 1995–2001: an analysis of the impact of language policy on attitudinal change. In Lotfi Sayahi and Maurice Westmoreland (eds.), Selected Proceedings of the 2nd Workshop on Spanish Sociolinguistics. Somerville, MA: Cascadilla, pp. 24–40. Gumperz, John Joseph 1982. Discourse Strategies. Cambridge: Cambridge University Press. Haeri, Nilofaar 1996. The Sociolinguistic Market of Cairo: Gender, Class and Education. London and New York: Kegan Paul, Heath, Jeffery 1989. From Code‐Switching to Borrowing: Foreign and Diglossic Mixing in Moroccan Arabic. London: Kegan Paul International. Høigilt, Jacob, and Gunvor Mejdell (eds.) 2017. The Politics of Written Language in the Arab World: Writing Change. Leiden: Brill. Hudson, Alan 1991. Toward the systematic study of diglossia. Southwest Journal of Linguistics 7: 5–15. Hudson, Alan 2001. Diglossia. In Rajend Mesthrie and R. E. Asher (eds.), Concise Encyclopedia of Sociolinguistics. Amsterdam: Elsevier, pp. 226–31. Hudson, Alan 2002. Outline of a theory of diglossia. International Journal of the Sociology of Language 157.1: 1–48. Hualde, José Ignacio, Antxon Olarrea, Anna María Escobar, and Catherine E. Travis 2009. Introducción a la lingüística hispánica. Cambridge: Cambridge University Press. Kabatek, Johannes 2016. Diglossia. In Adam Ledgeway, and Martin Maiden (eds.), The Oxford Guide to the Romance Languages. Oxford: Oxford University Press. Kaye, Alan S. 1994. Formal vs. informal in Arabic: diglossia, triglossia, tetraglossia, etc., polyglossia multiglossia viewed as a continuum. Zeitschrift für arabische Linguistik 27: 47–66. Keller, Rudolf Ernst 1982. Diglossia in German‐ speaking Switzerland. In William Haas (ed.),

Contact, Bilingualism, and Diglossia  65 Standard Languages: Spoken and Written. Manchester: Manchester University Press, pp. 70–93. Krumbacher, Karl 1902. Das Problem der neugriechischen Schriftsprache. Munich: K. B. Akademie in kommission des G. Franz’schen verlags. Lipski, John 2005. Code‐switching or borrowing? No sé so no puedo decir, You Know. In Lotfi Sayahi and Maurice Westmoreland (eds.), Selected Proceedings of the 2nd Workshop on Spanish Sociolinguistics. Somerville, MA: Cascadilla, pp. 1–15. Machfar, Mohamed. 2009. khutab. Available at www.zitounafm.net/ar/ Marçais, W. 1930. La Diglossie arabe. L’Enseignement public 97: 401–9. Marcos Marín, Francisco 1984. Review of Latín tardío y romance temprano, by Roger Wright. Revista de Filología Española 64: 129–45. Mifsud, Manwel 1995. Loan Verbs in Maltese: A Descriptive and Comparative Study. Leiden: Brill. Ministry of Finance 2018. mizaniyat al‐muwa:ṭin li‐sanat 2018 [Citizen budget 2018]. Mkilifi, Μ. Η. Abdulaziz 1978. Triglossia and Swahili–English bilingualism in Tanzania. In Joshua A. Fishman (ed.), Advances in the Study of Societal Multilingualism. The Hague: de Gruyter Mouton, pp. 129–52. Mouakhar, Riad 2018. Speech. Myers‐Scotton, Carol 1986. Diglossia and codeswitching. In Joshua A. Fishman (ed.), The Fergusonian Impact, vol. 2. Berlin: de Gruyter Mouton, pp. 403–15. Myers‐Scotton, Carol. 1993. Elite closure as a powerful language strategy: the African case. International Journal of the Sociology of Language 103: 149–64. Otheguy, Ricardo, Ana Celia Zentella, and David Livert. 2007. Language and dialect contact in Spanish in New York: towards the formation of a speech community. Language 83: 1–33. Paolillo, John C. 1997. Sinhala diglossia: discrete or continuous variation? Language in Society 26.2: 269–96. Pedraza, Pedro, John Attinasi, and Gerard Hoffman 1980. Rethinking Diglossia. New York: Language Policy Task Force, Centro de Estudios Puertorriqueños, CUNY. Poplack, Shana 1980. ‘Sometimes I’ll start a sentence in Spanish y termino en español’: toward a typology of code‐switching. Linguistics 18.7/8: 581–618. Poplack, Shana and Nathalie Dion 2012. Myths and facts about loanword development. Language Variation and Change 24.3: 279–315.

Poplack, Shana, Lotfi Sayahi, Nahed Mourad, and Nathalie Dion 2015. An exception to the rule? Lone French nouns in Tunisian Arabic. University of Pennsylvania Working Papers in Linguistics 21.2: 177–86. Psichari, Jean 1928. Un pays qui ne veut pas de sa langue. Mercure de France 207: 63–121. Sayahi, Lotfi 2011. Code‐switching and language change in Tunisia. International Journal of the Sociology of Language 211: 113–33. Sayahi, Lotfi 2014. Diglossia and Language Contact: Language Variation and Change in North Africa. Cambridge; Cambridge University Press. Sayahi, Lotfi 2015a. A moving target: literacy development in situations of diglossia and bilingualism. Arab Journal of Applied Linguistics 1.1: 1–15. Sayahi, Lotfi 2015b. Expression of attributive possession in Tunisian Arabic: the role of language contact. In Aaron M. Butts (ed.), Semitic Languages in Contact. Leiden: Brill. 334–47. Sayahi, Lotfi 2017. Diglossia in North Africa. In Mark Aronoff (ed.), Oxford Research Encyclopedia of Linguistics. New York: Oxford University Press. Sayahi, Lotfi 2018. Spanish in contact with other languages and bilingualism across the Spanish‐speaking world. In Kimberly Geeslin (ed.), The Cambridge Handbook of Spanish Linguistics. New York: Cambridge University Press, pp. 459–77. Sayahi, Lotfi 2019. The normalization of Tunisian Arabic use in official discourse: major structural features and lexical realignment. In Enam Al‐Wer and Uri Horesh (eds.), The Handbook of Arabic Sociolinguistics. New York: Routledge, pp. 227–40. Sayahi, Lotfi forthcoming. Language and identity in post‐Revolution Tunisia: between authenticity and commodification. In Reem Bassiouney and Keith Walters (eds.), Routledge Handbook of Arabic and Identity. New York: Routledge. Silk, Michael and Alexandra Georgakopoulou 2016. Standard Languages and Language Standards – Greek, Past and Present. London: Taylor and Francis. Silva‐Corvalán, Carmen 1997 Spanish in Four Continents: Studies in Language Contact and Bilingualism. Washington, DC: Georgetown University Press. Silva‐Corvalán, Carmen and Andrés Enrique Arias 2017. Sociolingüística y pragmática del español. Washington, DC: Georgetown University Press.

66  Lotfi Sayahi Thomason Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Torres, Lourdes 2002. Bilingual discourse markers in Puerto Rican Spanish. Language in Society 31.1: 65–83. Trudgill, Peter 1986. Dialects in Contact. Oxford: Blackwell Valdés, Guadalupe and Michelle Geoffrion‐Vinci 1998. Chicano Spanish: the problem of the “underdeveloped” code in bilingual repertoires. Modern Language Journal 82: 473–501. van Coetsem, Frans 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris Publications. Versteegh, Kees 1997. The Arabic Language. New York: Columbia University Press. Walters, Keith 1996a. Diglossia, linguistic variation, and language change in Arabic. In Mushira Eid (ed.), Perspectives on Arabic Linguistics: Papers from the Annual Symposium on Arabic Linguistics. Amsterdam: John Benjamins, pp. 157–97. Walters, Keith 1996b. Intrasentential codeswitching in diglossic settings and its

implications for linguistic variation and language change. In Jennifer Arnold, Renée Blake, Brad Davidson, Scott Schwenter, and Julie Solomon (eds.), Sociolinguistic Variation: Data, Theory, and Analysis. Selected Papers from NWAV23 at Stanford. Stanford, CA: CSLI, pp. 40116. Winford, Donald 1985. The concept of “diglossia” in Caribbean creole situations. Language in Society 14.3: 345–56. Winford, Donald 2003. An Introduction to Contact Linguistics. Malden, MA: Blackwell Publishers. Wright, Roger 1982. Late Latin and Early Romance in Spain and Carolingian France. Liverpool: Cairns. Wright, Roger 1994. Early Ibero‐Romance: Twenty‐One Studies on Language and Texts from the Iberian Peninsula between the Roman Empire and the Thirteenth Century. Newark, DE: Juan de la Cuesta. Wright, Roger 2010. Bilingualism and diglossia in medieval Iberia (350–1350). In Fernando Cabo Aseguinolaza, Anxo Abuín González, and César Domínguez (eds.), A Comparative History of Literatures in the Iberian Peninsula. Amsterdam: John Benjamins, pp. 333–50.

3 Language Contact and Change through Child First Language Acquisition CARMEL O’SHANNESSY AND LUCINDA DAVIDSON

1 Introduction Processes in child first language acquisition can be a locus of contact‐induced language change, yet they have received little attention in the language contact literature. Age‐group information is important for a full understanding of the mechanisms involved in language contact and change, especially for a close examination of the question of whether there is a break in the transmission of a language from one generation to the next. Research in this area may be informative not only about contact situations, but about human language and cognition more broadly. For children to play a role in contact‐induced language change, a change must have taken place that was led, or significantly influenced, by child speakers, that has remained in, or become, the primary way of speaking of a community of speakers. Evidence for children’s agentivity in a change can only be obtained by comparing adult and child speech production at the same point in time, and finding that the children’s production differs systematically from that of the adults. Views on the roles of child speakers in contact‐ induced language change differ according to the type of language context, and also within similar contexts. Perhaps the most debate is seen in the roles of children in the emergence of creoles, and probably the least is seen in their roles in the emergence of koines. Before turning to the contact literature, some preliminaries about child first language acquisition are needed. In all first language acquisition contexts of typically‐developing children, the children hear or see speech or sign directed to them – the input. There are two main types of theory as to which mechanisms are used to process the input, and other positions that draw on those to differing extents, and the theories are relevant to views of children’s roles in language contact. Nativist theories assume that at least some grammatical structure is innate and available from birth as Universal Grammar (Chomsky 1965, 1981), e.g. tacit knowledge that a language will have the structural categories of noun and verb. As a child receives input, the data are processed according to innate categories, e.g. an agent of an action is linked to a grammatical subject (Pinker 1984). Constructivist theories assume that there is no innate grammatical structure, but that children have the ability to scan the input for recurring patterns and use these to build up structures (Tomasello 2003).

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

68  Carmel O’Shannessy and Lucinda Davidson While on their language acquisition journey, young children at times produce speech that differs from community norms, which are the assumed language learning target for a child (see the discussion in Section 5.2), but usually later conform to the community conventions. An example from English is that many children at some point regularize irregular past tense verb forms, saying go‐ed and bring‐ed instead of went and brought, but later produce went and brought. According to nativist theories, the child’s language environment provides some, but not enough, information to learn all of the rules of a language, so the child also makes use of innate language‐specific knowledge. According to constructivist theories children learn their language by forming hypotheses based on patterns detected in the input, testing the hypotheses against the input, and forming new hypotheses until they match the patterns in the input. In either type of theory, processes that occur when children produce nontarget forms include, for example, over‐regularizations of morphology, as explained above, ­omissions (“I want biscuit”), recombinations (“Can I ’ave a cado?”) and other types of reanalysis (“Nail polish me!”). These processes are produced by monolingual and bi‐ and multilingual children alike. In most cases the novel forms or innovations do not become language change, but are simply part of the acquisition process of each individual child. The acquisition task for a bilingual or multilingual child, that is, a child growing up in a language contact environment, has the additional challenges that the input contains more than one linguistic system, the child learner has to learn to distinguish these, and that at least sometimes elements of only one system are produced at a time. We say “at least sometimes” because in bilingual and multilingual contexts the languages are not always completely separated, e.g. a lexical item from one linguistic system may occur in the grammatical frame of another, and sites of contact may be sites for contact‐induced language change. Also, speakers often switch between elements of two or more languages within an interaction or within a clause, known as code‐switching. Furthermore, the linguistic systems might not be those that are usually thought of as named languages (Otheguy, García, and Reid 2015), but may be systems that combine elements from named languages and/or other linguistic systems. Children receiving bilingual input from birth or soon after are known as simultaneous bilinguals, and the type of acquisition is bilingual first language acquisition (BFLA) (de Houwer 1983, 1990). In BFLA children mix elements from their two languages from an early age, yet they also know that there are two linguistic systems from as early as when they are able to combine two morphemes or words (de Houwer 1990, 2005; Genesee 1989, 2003; Genesee, Nicoladis, and Paradis 1995; Lanza 1992, 1997b; Nicoladis and Genesee 1997). Sometimes children growing up bilingually show a similar type of language combining as in some mixed languages (Bakker, this volume), that is, combining the lexicon of one language with the grammatical morphology of the other (Burling 1959; Lanza 1997a, 1997b, 2004). Crosslinguistic transfer may also occur, and has been seen at the interface of syntax and pragmatics, with the realization of overt pronouns (Müller and Hulk 2001). The nontarget constructions outlined here typically only occur for a limited period of time, and the children later conform to the local community patterns. Nontarget constructions that do not remain in the speech of a community of speakers, and are not then transmitted to the next generation of speakers, are not part of the literature on contact‐induced language change, because no change in a community’s way of speaking has taken place. However they are important to language contact for understanding how children process language in bilingual and multilingual settings. The notion of unchanging community adult norms does not necessarily apply to language contact situations. In both monolingual and multilingual settings there may be language change in progress, and the variability in adult language is likely to be part of the input to children. In language contact settings with change in progress, the variation in adult and child speech may involve more than one named language, and/or newly emerging linguistic systems. These contexts are much less well understood than

Language Contact and Change through Child First Language Acquisition  69 contexts of stable variation (J. Roberts 2002), yet are those where the most evidence of children leading change is considered (Aboh 2015; Bickerton 1984; Mufwene 2001; ­ O’Shannessy 2012, 2013). The major questions at the intersection of language acquisition and language contact are thus: a. Under what sociolinguistic circumstances in a contact situation does the novel speech of children take hold in a community’s way of speaking? b. What kinds of language change are instigated by child speakers in contact situations? In Section  2 we outline theories and empirical case studies of children significantly ­influencing change during the emergence of new languages, specifically creoles and mixed languages. Section 3 discusses the role of children in the emergence of new v ­ arieties of languages, koines, and multiethnolects. In Section  4 we present some examples of child‐influenced change where there has not yet been time to see if the changes will consolidate into new ways of speaking, but it is a possibility. We draw on the previous ­sections to summarize the types of change significantly influenced by child speakers, and  how those speakers’ innovations take hold in a community (Section  5) and offer a conclusion in Section 6.

2 When Children Influence Contact‐induced Change during First Language Acquisition 2.1  New Language Creation – Oral Creoles The most theorizing about the roles of children in contact‐induced language change appears in literature on the emergence of oral creole languages, and ranges from children playing no role or a limited one (Lefebvre 1986, 1998, 2002; Plag 2008, 2009; Siegel 2000, 2008), to their being largely the creators of a creole (Bickerton 1981, 1984). The views that children play little or no role in creole formation posit that creoles are the creations of adults, largely through second language acquisition processes, regardless of whether the lexifier was an actual learning target (Alleyne 1980: 220; DeGraff 2002: 393; Mufwene 2001), and children simply acquire the adult speech patterns as in any other acquisition setting. Within these end points are views that children’s processes of acquisition in contact environments are the same as those in other acquisition environments, but the sociolinguistic setting is distinct (Aboh 2015; DeGraff 2002; Mufwene 2001; O’Shannessy, forthcoming), and/or that children add small, incremental changes to the input they receive, resulting in new or more regular grammatical paradigms (Aboh and Ansaldo 2006; Jourdan 1989; O’Shannessy, forthcoming; Sankoff 1991; Shnukal and Marchese 1983). The greatest role attributed to child speakers in a language contact situation is in Derek Bickerton’s (1981, 1984) theory that creole languages can be created by children nativizing, and in doing so expanding on, the (minimal) input they receive from their surroundings. The term nativization (not to be confused with nativist theories) refers to a cohort of children learning a newly emerging way of speaking as their first or primary language that was not the first language of their parent’s generation. Some dominant theories of creole emergence claim that creole languages evolve from pidgins, perhaps being morphosyntactically expanded over several generations (Arends and Bruyn 1995; S. J. Roberts 2000; Siegel 2008), and finally being nativized by children, by which time they are known as creoles, with native speakers. A pidgin by definition has limited morphosyntactic complexity and vocabulary, and is not a full language (Arends, Muysken, and Smith 1995; Bakker 1995; Holm 1988;

70  Carmel O’Shannessy and Lucinda Davidson Siegel 2008). In contrast, a creole is a full language. There is debate as to whether children nativize a pidgin and through this process expand it into a creole, or nativize an already expanded linguistic system. Bickerton claims that if the only input to a group of very young children is a restricted pidgin, that has been spoken for only one generation, then those children would receive impoverished input, inadequate for full communication and cognitive development. Bickerton claims that a situation such as this would be one where “the human linguistic capacity is stretched to the uttermost” (Bickerton 1984: 3). He invokes a nativist Language Bioprogram Hypothesis to explain how, in order to transform the impoverished pidgin input into a creole, the children must draw on their innate language endowment to fill in the morphosyntactic components missing from the pidgin, since they are not present in the input. In the theory the Language Bioprogram is an innate propensity of every child. If a child receives rich input, e.g. from a fully developed language, properties of the Language Bioprogram will not need to emerge, but if a child receives inadequate input, e.g. from a restricted pidgin, then the Language Bioprogram would emerge and complete the missing configurations. Bickerton posits specific elements that are part of the innate Language Bioprogram, and which automatically come to the fore from the child’s innate cognitive endowment when they are absent from the input. These include features that have been identified as present in Hawai’ian Creole English and common to some Atlantic creoles, e.g. a distinction between specific and nonspecific noun phrase referents, a modal distinction between realis and irrealis, an aspectual distinction between punctual and nonpunctual events (Bickerton 1984: 182). Bickerton analyses differences between Hawai’ian Pidgin English and Hawai’ian Creole English to pinpoint the features that, according to this theory, must have emerged in Hawai’ian Creole English due to children’s additions during nativization of the Pidgin input they received (Bickerton 1984: 15). A criticism of the Language Bioprogram Hypothesis is the premise that a scenario where the only input children received was a restricted pidgin is highly unlikely (DeGraff 2002: 390–1). More likely scenarios, some with empirical evidence, are that children are also spoken to in pre‐existing languages, and their input is multilingual (DeGraff 2002; S. J. Roberts 2000; Shnukal and Marchese 1983; Siegel 2008), and just as complex as in any language learning situation. The claim that Hawai’ian Creole English contains morphosyntactic features that were absent in Hawai’ian Pidgin English and contributing languages, and therefore was not part of the children’s input (Bickerton 1984) has been refuted by S. J. Roberts (2004) and Siegel (2000). They also dispute that Hawai’ian Creole English emerged quickly, and provide evidence that it developed over several generations, nullifying the need for the putative Language Bioprogram Hypothesis to come to the fore. Researchers have argued that while some empirical child language data are consistent with the Language Bioprogram Hypothesis, other child data differs considerably from that proposed by the Bioprogram (Aboh 2015; Mufwene 1999). A strong role of both adult and child learner influence in creole genesis is hypothesized by DeGraff (2002: 391–3) in his theory of an L2A  –  L1A cascade, with Haitian Creole as the example. In this theory adult second language learners learned the lexifier to varying extents, creating a new contact language, and children received the adults’ variable productions as input, along with the ancestral languages. In their processes of first language acquisition they would be the first generation to speak the contact language as a first language. But their production of it would not be identical to the second language productions of the adults, that was part of their input, because of differences in language experience and cognitive processing between adult second language learners and child first language learners. Note that the difference is not in terms of exposure to rich vs. impoverished language input. Along similar lines, several researchers (Aboh 2015; Aboh and Ansaldo 2006; Mufwene 2001), argue that children’s language acquisition in monolingual or multilingual

Language Contact and Change through Child First Language Acquisition  71 environments is essentially the same, but the different types of variable input in different sociolinguistic contexts give rise to slightly altered grammars, and these are then the input to the next generation of child acquirers, and so on. In specific contexts this can result in a creole being formed. Aboh (2015) and Mufwene (2001) claim that every first language acquisition setting is in a sense multilingual, as there is always heterogeneity present in the input to some extent, regardless of whether we think of it in terms of idiolects, varieties, or languages. In this view young child learners always select from competing variants in the input, and their sets of choices forge their idiolect. Aboh and Ansaldo (2006: 44) suggest that innovation through first language acquisition is more likely to occur in a multilingual than a monolingual setting. Empirical data from pidgin‐creole genesis settings show some support for the hypotheses of DeGraff, Mufwene, and Aboh, but little support for the Language Bioprogram Hypothesis (Bickerton 1984). S. J. Roberts (2004: 315) suggests a very different role for children in the development of Hawai’ian Creole English. She suggests three stages of creole development with different roles taken on by each generation of locally born speakers. The generation of adult immigrants were responsible for developing Hawai’ian Pidgin English. The next generation, locally born speakers, spoke it as their second language, in addition to speaking their own first languages. This generation vernacularized the pidgin, that is, adopted it as the main, common language of the community. In this second stage, children from different home language backgrounds talked together in Hawai’ian Pidgin English as a shared language at school, and then used it in their homes, where their ancestral languages were the main languages spoken. In the third stage, the children of that generation nativized the pidgin, which had already been expanded by the adults, and in doing so added regularizations. Here the role of children is split between generations, with the children of immigrants learning the pidgin as a second language, and the next generation nativizing it, that is, learning it as their first language. But each generation of children was spoken to in the pidgin and in other languages, so the input was rich, far from the impoverished scenario Bickerton posits. S. J. Roberts sees the role of children as one of adding regularizations and small, incremental changes to the variety spoken to them, not one of wholesale creation of a linguistic system. There is evidence from other emerging oral language creoles that the role of children is in regularizing patterns and adding incremental changes to the speech spoken to them, which had already been expanded by adult speakers (Jourdan 1989; Shnukal and Marchese 1983). Solomon Islands Pidgin was documented by Jourdan (Jourdan, 1989) at a time when it was the first language of some children, and the main language of others, and most children also spoke another language. The children for whom it was one of their first languages were the first generation to acquire it from birth. Most of the changes in the creole were initiated by urban adults, who alternated between newer and older types of structure. Children then expanded upon and regularized the morphosyntactic variation in adults’ speech. As they did this, structures became more regular and could be used in more abstract contexts (Jourdan 1989: 28). Shifts in the semantics of the word bai in Tok Pisin, the creole that is an official language of Papua New Guinea, were documented by Sankoff (1991). The semantics of the word bai (< Tok Pisin baimbai, < English by and by) were changing from it being a future marker, through to an irrealis marker, and then to a marker of iterative‐habitual and punctual aspect. At the time of documentation adults were speakers of the creole as a second language, and some children had acquired it as their first language. The child first language speakers were leading the change in using bai to mark iterative‐habitual and punctual aspect, that is, they were producing the same constructions as the adults, but much more frequently. Sankoff argues that the input to the children in Tok Pisin overwhelmingly made aspectual rather than tense distinctions, and the children recognized this and brought in “more sweeping changes than their parents were able to do” (Sankoff 1991: 73).

72  Carmel O’Shannessy and Lucinda Davidson Shnukal and Marchese (Shnukal and Marchese 1983) documented the use of Nigerian Pidgin English in a town near Benin, the capital city of Edo State in southern Nigeria. Nigerian Pidgin English was spoken by some children as a shared second language, and by others as one of their first languages. It was spoken in some homes alongside other languages, often when parents had different first languages, but even when their first language was shared. When Nigerian Pidgin English was used in the home as a second language, it was often introduced there by the children, who learned it from their peers, also heard it on television, and read it in magazines and comics. It was then adopted by adults. In other families Nigerian Pidgin English was used consistently by adults and was one of the first languages of the children (Shnukal and Marchese 1983: 19–21). Younger speakers had greater fluency in the language, spoke it at a faster speech rate, and showed more phonological reduction, than older speakers did (Shnukal and Marchese 1983: 23). There were some changes in the production of tone and intonation between children and adults (Shnukal and Marchese 1983: 23). These case studies are linked by the observation that in each instance adult speakers already spoke the contact language as a second language, and it was already quite expanded morphosyntactically, that is, it was not like a restricted pidgin. In each instance the child first language speakers regularized variation, made incremental changes to the adult speech, or produced newer usages more often. What is seen here is that while children played a critical role in the developing language, they did not develop it from scratch, or from lack of input, as the changes they produced were already present in some way in the input they received.

2.2  New Language Creation – Signed Contact Languages Turning to the domain of a signed contact language, Nicaraguan Sign Language emerged in a situation where school‐aged deaf children, with no shared language, were brought together in boarding schools, and created a shared language through interacting with each other using their idiosyncratic home sign systems (Kegl, Senghas, and Coppola 2001; A. Senghas 2003; A. Senghas, Coppola, Newport, and Supalia 1997; A. Senghas, Kita, and Özyürek 2004). The first cohort of school‐aged children created a new common language, and since then subsequent cohorts have regularized variation and increased morphological productivity, through nativizing the input they received from each previous cohort of signers (Kegl et al. 2001; A. Senghas 2003; A. Senghas et al. 1997; A. Senghas et al. 2004; R. J. Senghas, Senghas, and Pyers 2005). The innovations have included hierarchically structured expressions that were not found in the surrounding language environment, and that involve creating discrete elements, and combining them productively (A. Senghas et al. 2004). The emergence of Israeli Sign Language appears to have been motivated by a similar situation. Meir, Israel, Sandler, Padden, and Aronoff (2012) contend that the language developed (80 years ago) in part because of children being brought together and interacting in three deaf schools around the country. Children developed a “gestural communication system that evolved over the years, as it was used by the different cohorts attending these schools” (Meir et al. 2012: 258). A number of sign languages have emerged in small, isolated communities which have a high proportion of congenital deafness. These varieties are termed “village” sign languages (Meir, Sandler, Padden, and Aronoff 2010), and many have seen dramatic change since the advent of formal schooling for deaf children. Whether residential or otherwise, schools were located outside the children’s village, and a standard variety was the language of instruction. In this context, children are exposed to a new language, which they take with them back to the village. This in turn influences the shape of the local sign language. This school‐related contact situation is described for Inuit Uukturausingit, the Inuktitut Sign Language, used by the Inuit in Nunavut, Canada, in contact with American Sign Language (ASL) (Schuit 2012).

Language Contact and Change through Child First Language Acquisition  73 It is also mentioned with respect to Ban Khor Sign Language from northeastern Thailand, in contact with Thai Sign Language (Nonaka 2012), and also Al‐Sayyid Bedouin Sign Language, used in Israel, in contact with Israeli Sign Language (Kisch 2012). Alipur Sign Language, from South India, is another example, in contact with Indian Sign Language (Panda 2012). Panda (2012: 359) explains that some older deaf people in Alipur village have noticed children using signs from the language in the town where the children go to school, as opposed to those local to their village. Increased mobility and access to the internet are also mentioned as factors driving change in this language.

2.3  New Language Creation – Mixed Languages Mixed languages are those that combine the features of (usually) two source languages in systematic ways (see Bakker, this volume). Unlike creoles, they are not borne out of a need to communicate with others whose language(s) one does not speak, but emerge in bilingual contexts, often where there is a desire to signal a new, bilingual‐bicultural identity. They typically combine elements of the source languages largely intact, often in a pattern of retaining the grammar of one language with the lexicon of the other (e.g. Muysken 1994), or, as in the cases reported on here, the verb system of one language with the nominal system of another (Bakker 1994; McConvell and Meakins 2005; O’Shannessy 2005). Until recently, children were not hypothesized to play a role in the emergence of mixed languages, except to faithfully acquire the way of speaking transmitted by adults. One reason for this is the lack of real‐time data from when the new languages were emerging. Another is the observation that since the grammatical structure of both source languages is brought into the new system with little or no alteration, it must have been brought in by speakers who had already completed their acquisition of the grammar of the language, so the agents of change could not have been young children. But empirical data from a very recently emerged mixed language, Light Warlpiri, spoken in a remote Indigenous community in Australia, suggests that young children nativized mixed input directed to them, and that this played a dramatic role in the creation of new morphosyntactic structures in the emerging language (O’Shannessy 2012, 2013). Light Warlpiri emerged in a two‐step process within the last 40 years (O’Shannessy 2012). It is hypothesized that in the first stage, adults were code‐switching between Warlpiri, Kriol (an English‐lexified creole spoken in the north of Australia) and varieties of English. When they spoke to very young children, their code‐switching was in a specific pattern, as part of a baby‐talk register, and the pattern was a string of Warlpiri with pronouns and verbs inserted from Kriol. In the second stage, the young children processed the input as a single system, and added innovations in the verb structure. This resulted in a new language with nominal morphology from Warlpiri, and verbal structure from English and Kriol, with innovations brought in by the children. Lexical items are from all of the sources. A sample sentence is given in (1). Elements from Warlpiri are in italics, and elements from Kriol and/or English are in upright font. (1)

Kala nyarrpara‐rla nyuntu‐ju but where‐LOC 2SG‐TOP ‘But where did you bathe?’

yu‐m bugi? 2SG‐NONFUT bathe (O’Shannessy 2015b: 290)

In (1), the first three words are from Warlpiri, the verb is from Kriol, and the verbal auxiliary yu‐m ‘2sg‐NONFUT’ is the new auxiliary structure. The new structure draws on Warlpiri in terms of semantics (Laughren 2013) and in having a pronominal‐plus‐TMA structure, and it draws on Kriol and English in terms of the phonetic shape of the elements, yu (< Eng/Kriol you/yu), and ‐m ( a‐na ‘1SG‐want’ in English. English‐speaking children produce this contraction at approximately 2;11 to 4;61 years. English‐speaking children later conform to the community target of I wanna. But when Light Warlpiri was emerging, the same reanalysis was made, and other similar reanalyses were made in the verbal auxiliary, involving grammaticalization of a TMA element affixed to a pronoun. In the case of Light Warlpiri the novel constructions diffused through the age cohort of young speakers, who did not then later converge on the adult target. Rather, the innovations remained in the speech of the young speakers of the newly emerging language, and became part of language change (O’Shannessy, forthcoming). Although the innovation in Light Warlpiri is both unusual (involving core verbal structure) and dramatic (creating structure not present in the source languages), O’Shannessy argues that the changes from the point of view of child first language acquirers is only an incremental change to the input the children received. Kriol has preverbal past tense morphemes, bin ‘PAST’ and im‐in ‘3SG-PAST’, but these only occasionally occur in Light Warlpiri, and the new ‐m ‘NONFUTURE’ morpheme is used in past tense contexts. At first glance it could look as though there was no motivation in the input for the child speakers not to use bin ‘PAST’. But a closer examination of the speech styles that likely resemble the input to children at that point in time shows that bin ‘PAST’ was only variably present. The children’s non‐use of bin ‘PAST’, is a regularization of variable input, ‘selection’ in the terms of Aboh (2009), Mufwene (2001), and McConvell (2008a). Speakers of Light Warlpiri have, over time, regularized nominal allomorphic variation in the language (O’Shannessy 2016b), but only some of this has clearly been led by children. Light Warlpiri nominal morphology is taken directly from Warlpiri, and some case‐markers have two (e.g. locative and dative cases) or four (e.g. ergative case) allomorphs. Phonologically reduced forms of some allomorphs occur in both Warlpiri and Light Warlpiri, but in Light Warlpiri the number of allomorphs of some case‐markers has decreased, and for some case‐markers only the phonologically reduced forms are now used. This is in contrast to Warlpiri, where all of the allomorphs still occur. A change in the form of some other suffixes has also occurred, e.g. omitting a final vowel, so that words may end in a consonant (O’Shannessy 2016a, 2016b), where Warlpiri traditionally had only vowel‐ final words (Hale, Laughren, and Simpson 1995). This change on one suffix is led by children (O’Shannessy 2015a). The form of an intensifier suffix, ‐nyayirni ‘very, really’, has a final vowel in traditional Warlpiri, but the final vowel is not always realized, and the form becomes ‐nyayirn. Both forms occur in Light Warlpiri, and children use the ‐nyayirn form significantly more often than adults do (O’Shannessy 2015a: 318). Children also use a reduced form of the diminutive suffix ‐pawu for the traditional form ‐pardu, more often than adults do (O’Shannessy 2015a: 318).

Language Contact and Change through Child First Language Acquisition  75 Regularization of morphological forms in one of the languages spoken by multilinguals can function as a method of differentiating between their languages. The two first languages of these speakers, Light Warlpiri and Warlpiri, share many lexical items and nominal morphology. The regularization of allomorphy in Light Warlpiri makes it more distinct from Warlpiri (O’Shannessy 2015a). Some other points of differentiation are seen in lexical choices. Two synonyms for ‘dog’, maliki and jarntu, have different indexical information attached to them. The word maliki ‘dog’ indexes classic Warlpiri, and traditional stories and songs, and varieties in the southern Warlpiri area. The word jarntu ‘dog’ indexes northern varieties, and not traditional texts. When speaking Warlpiri, children use both maliki and jarntu, but when speaking Light Warlpiri they never use maliki. This is another way of increasing language differentiation (O’Shannessy 2015a). Another mixed language spoken in remote Australia by Indigenous speakers, Gurindji Kriol (McConvell and Meakins 2005; Meakins 2011b), emerged when adults code‐switched in a consistent pattern between Gurindji and Kriol, and the children receiving the code‐ switched input internalized it as a single system (McConvell and Meakins 2005). As the language has unfolded, teenagers have added some innovations, such as combining case‐marking functions from Gurindji with forms from Kriol (Meakins 2011a). For interest we include mention of two putative mixed languages that are believed to have been developed by both adults and school‐aged children, creating and learning them as second languages (de Gruiter 1994b). These are Javindo, which combines Dutch and Javanese (de Gruiter 1994a, 1994b), and Petjo, combining Dutch and Malay (van Rheeden, 1994). In each case, children interacting together when playing largely developed the new ways of speaking. The final example in this section is of continued contact‐induced language change over generations, moving the way of speaking into different language types, including a mixed language type at one point in time. Tiwi is spoken on Bathurst Island and Melville Island in Australia’s Northern Territory. Traditional Tiwi has a complex polysynthetic verb with potentially 18 morphemes. Changes in Tiwi as spoken by people under about age 35 in the 1980s showed less complex verb structure, with fewer morphemes, and some lexical borrowings from English and Kriol (Lee 1987: 148), that is, there has been a move towards more analytical structure. The new system, Modern Tiwi, has been described as a mixed code with verbal structure from Tiwi and nominal structure from English and Kriol (Lee 1987; McConvell 2008b). A style with more changes from Traditional Tiwi is known as New Tiwi. There are two motivations for the development of Modern and New Tiwi that involve child speakers. The first is that young female children lived in a religious mission boarding school from the 1920s to the 1970s, separated from their families, and therefore from adult speakers of Tiwi. The school staff generally did not speak Tiwi, so the children had very reduced access to it, and may have developed the new way of speaking as they grew up, from receiving mixed input and limited input in Tiwi. Second, adults may have used a baby‐talk register to young children that included structures that resemble those of New Tiwi (Lee 1987: 355). More recently, young Tiwi children have been seen to use a speech style that extends the difference from Tiwi (Wilson, Hurst, and Wigglesworth 2018). A case study of two children aged 3;9 and 4;12 years old in a school classroom, has shown increased analytic clause structure and that most of their verbs are from Kriol and English. Some uninflected Tiwi verbs are produced, but there is no use of Tiwi complex verb structure. Both lexical and functional elements are drawn from either Tiwi, English, or Kriol. The data are not compared to data from young adults at the same time point, so it is not clear if the young children are themselves the initiators of the new morphosyntactic constructions. This section has discussed theories about, and empirical evidence of, the role of children in the emergence of new contact languages, specifically creoles and mixed languages. In the next section the roles of children in the emergence of new varieties is discussed.

76  Carmel O’Shannessy and Lucinda Davidson

3  New Dialect Formation The innovations discussed in this chapter do not cause such dramatic overall change as those in the previous section. In the case of koines and multiethnolects, it is usually recognized that the new, conventionalized ways of speaking are mutually intelligible with the pre‐existing varieties, and that a further variety is being added.

3.1  New Dialect Formation – Koines The role that children play in the emergence of koines appears to court considerably less controversy than in some other types of contact‐induced language change. As described by Kerswill (this volume), a koine is a stabilized, composite dialect, formed through the interaction of multiple, mutually intelligible language varieties. In cases of koineization, children are thought to regularize and focus the different speech varieties that they receive as input, and do so in their peer interactions. Two sociolinguistic circumstances are deemed crucial to this process. These are sustained social interaction between speakers of the different language varieties (Siegel 2001: 183), and a shared orientation from speakers towards their new environment (Domingue 1981: 150). Kerswill and Williams’ (2000) research at Milton Keynes, England, describes the role that children of different ages played in the process of koineization that occurred in this new town. Milton Keynes English is thought to have emerged in the speech of second generation inhabitants. Specifically, preadolescent children with strong social connections are identified as leading the focusing process. They are identified as change‐makers owing to the phonology of the new variety being evident in their speech, for example, /θ/ being realized as [f], and /ð/ as [v], intervocalic and final /t/ produced as glottal stops, an absence of post‐ vocalic r, and the pronunciation of (au) as /aʊ/ as opposed to the regionally marked /ɛʊ/ or /ɛɪ/ (Kerswill and Williams 2000: 85–99). While these features were found in the speech of socially integrated eight‐year‐old children, they were not evident in productions from four‐ year‐olds and from older children without strong ties to the social group. The phonologies of these latter groups resembled that of their caregivers, rather than that of their peers. This suggests that a child’s social connectedness was more of a key factor in the emergence of this koine than a speaker’s age or maturational stage. The importance of social integration for koineization can be compared to a situation in which a koine did not develop. In Svalbard (Spitsbergen), a Norwegian Arctic territory, not only was there no stable adult linguistic model for children, there was also little demographic stability (Maehlum 1992). Families rarely stayed in the community for more than 10 years, and within this time they tended to spend extended periods (such as summer holidays) away. The speech of children in Svalbard is described as resembling that of their parents rather than their peers. Amongst themselves, children in Svalbard practiced linguistic accommodation, rather than developing a new, shared variety. This case suggests that koineization requires peer‐group stability. Without this, children do not have the opportunity to establish new group norms. The children of migrants in Høyanger, an industrial town in Norway, also practiced linguistic accommodation with one another (Solheim 2009). A koine emerged here only with the second locally born generation. Unlike Svalbard, there was demographic stability in Høyanger when the second generation was growing up, yet there was also considerable social segregation in the town, between skilled and unskilled workers. In contrast to Milton Keynes, the phonologies of dialects spoken in Høyanger at the time of migration varied starkly from one another. Local dialects existed in the rural Høyanger area which differed considerably from more standard varieties from the east of the country that some migrants brought with them. However, the koine that emerged with the second locally born

Language Contact and Change through Child First Language Acquisition  77 generation in Høyanger reflects the interaction of this diverse input, and is in fact referred to as a “compromise” variety (Solheim 2009: 199). For example, ‘home’, which is [hei:m] in the original local dialects and [jem] in the standard, is realized as [hɛ:m] in the new Høyanger dialect. Similarly, ‘plays’ in local dialects is [spe:lɑ], while it is [spil:ər] in the standard, and [spil:ɑ] in the Høyanger variety (Solheim 2009: 198). The reason behind the delay in koineization in Høyanger therefore appears to be less about the varied nature of the input and more about the social segregation apparent in the community. Koineization occurred in Høyanger only once the community had stabilized and was operating as a cohesive society and culture. What this case of koineization suggests is that a shared culture facilitates the emergence of new varieties, and not simply sustained social contact between speakers of different varieties. It also indicates that social segregation is more of an impediment to the formation of koines than the diversity of varieties spoken in the community. In fact, the situation in Høyanger suggests that orientation to the local community is crucial not only to the formation of koines, but also to their maintenance. Solheim (2009) briefly describes a “post‐koineization” stage in the community, which sees the language of the fourth generation of inhabitants move towards the standard language variety and further away from the local koine. This shift is taken to reflect the increasingly national/global orientation of these children, suggesting that the more outward looking the young population, the further their speech shifts from the local variety. Different again is the case of Dhuwaya, a koine which emerged at Yirrkala, an Aboriginal community in the north of Australia (Amery 1993). Approximately 15 different dialects, of two closely related Australian languages, are spoken in Yirrkala and its surroundings. Each of these dialects is connected to traditional clan membership, and is most readily differentiated by its grammatical morphology. Prior to the establishment of a Methodist mission in 1935, Aboriginal children in the Yirrkala area are thought to have interacted predominantly within their clan (and therefore dialectal) group. Since contact, however, children have been regularly in close contact with children from other groups, through mission‐related institutions such as school and church. In addition to these 15 clan dialects, children in post‐contact Yirrkala also received a particular variety of Dhuwaya in their input, a baby‐talk register referred to as “Baby Dhuwaya.” The koine that emerged amongst child speakers at Yirrkala extends this baby‐talk variety. Dhuwaya sees the regularization of pronominal and demonstrative paradigms, the collapsing of certain grammatical categories, such as the inclusive/ exclusive distinction, and final syllable deletion, particularly after liquids. It also involves the lenition of laterals (/l/ to /y/), which speakers identify as the main characteristic feature of Dhuwaya (Amery 1993: 51). The resulting variety is one that builds on a baby‐talk register, and incorporates many aspects of existing local dialects, without resembling any one variety more than another. Since its development, Dhuwaya has become firmly entrenched in Yirrkala. Yet it was highly stigmatized in the community in the 1980s, one generation after its emergence. Despite the “strong outward opposition” (Amery 1993: 62) to Dhuwaya from many older people (including some speakers of the koine), it became the language of instruction at the school, and is the common language for children and young adults in Yirrkala today. Amery suggests that one reason for Dhuwaya becoming a communilect could be that its use avoids favoring one clan’s dialect – and therefore that clan group – over the others. It is possible, then, that the creation and use of Dhuwaya helped to promote cohesion within the community, despite the negative attitudes from some towards the new language variety. Newly emerging varieties are often subject to stigmatization from speakers who are reluctant to embrace change. But rather than prompting speakers to gravitate towards a more standard variety, negative language attitudes from outsiders may help to elicit, or maintain, a staunch pride in speakers of the new dialect (Beal 1999). The shape of a new variety as it unfolds may reflect the inward or outward orientation of its younger speakers.

78  Carmel O’Shannessy and Lucinda Davidson As these examples illustrate, for a koine to emerge, young members of a community must coordinate with one another in terms of language practices and social orientation, and do so in a stable manner, over time. Speaker attitudes towards the new variety appear to have less influence on the dialect than people’s attitudes towards the community itself.

3.2  New Dialect Formation – Multiethnolects Multiethnolects (cf. Clyne, 2000) refer to new language varieties which develop out of interactions between speakers from various ethnic and linguistic backgrounds. The varieties contributing to the mix are second language varieties of parents and first language varieties of their children. Another key difference from koines is that multiethnolects are viewed as part of speakers’ broader linguistic repertoire. All speakers of multiethnolects are described as having other language varieties at their disposal (Wiese 2009: 784), which they use outside of informal interactions with peers. Research on multiethnolects focuses on young people, predominantly adolescents, in urban environments that have seen recent large‐scale immigration. The centrality of youths in the development and use of multiethnolects is widely accepted. The high degree of variability in a multiethnolect’s makeup is also broadly acknowledged, in that it would be difficult to talk of a clear target variety (Cheshire, Kerswill, Fox and Torgersen 2011). Where disagreement does lie is the extent to which multiethnolects can be considered new language varieties. Some deem them a variable but systematic and productive speech variety (e.g. Wiese 2009), while for others they are an ephemeral “flavor on top” of a stable base language (Nortier and Dorleijn 2013: 25). Across definitions, multiethnolects refer to a recognizable linguistic style (e.g. Freywald, Mayr, Özçelik, and Wiese 2011) that differs from the local standard variety and from the speech of first‐generation immigrants (Quist 2008: 44), which young people from a range of ethnicities use deliberately amongst themselves. Cheshire, Kerswill, Fox and Torgersen et al. (2011) propose two factors which enable the emergence of multiethnolects, one of which is multiethnic friendship groups amongst young people.3 In this way, multiethnolects are similar to koines; peer interaction and social connectedness is crucial to the emergence of both types of language varieties. The desire for sense of group identity, of in‐group solidarity, also appears to play a critical role in the formation of multiethnolects (Dorleijn and Nortier 2008). While koineization can occur when there is a shared orientation to the identity of a broader community, multiethnolects are likely to emerge when adolescents orient to an even more local community. The other key factor identified in the formation of multiethnolects is the presence of children with dominant personalities who are well integrated in the group. It is these individuals who are thought to instigate change and likely have others follow them. This too can be compared to koines, in that socially integrated children are more likely to be the ones driving language change than children who are less well connected with peers. Cheshire and Gardner‐Chloros (2018: 161) also explain that a multiethnolect may emerge when there is no single community model as an acquisition target for the children, and there is a high tolerance of linguistic variation. The children’s speech styles may be fairly varied until they stabilize in adolescence. Detailed descriptions exist of multiethnolects spoken in urban centers around the world that have seen large‐scale immigration. Most of these are in Europe. In Sweden, Rinkebysvenska, ‘Rinkeby Swedish’ is spoken in Stockholm (Kotsinas 1988, 1992, 1998), Rosengårdsvenska, ‘Rosengård Swedish’ in Malmö (Bodén 2004; Ekberg 2011) and Gårdstenish in Gothenburg (Otterup 2011). Københavnsk multietnolekt ‘Copenhagen multiethnolect’ is spoken in a number of areas of Copenhagen, Denmark (Quist 2000, 2008), and a mutliethnolect is also attributed to young people in Oslo, Norway (Opsahl 2009; Svendsen and Røyneland, 2008). In Germany, Kiezdeutsch, ‘Neighborhood/Hood language’ has developed in particular areas of Berlin (Wiese 2009), and Stadtteilsprache ‘District talk’ in Mannheim (Kallmeyer and Keim 2003).

Language Contact and Change through Child First Language Acquisition  79 Urban centers in the Netherlands have seen an increasing use of Straattaal, ‘Street language’ (Appel 1999; Nortier 2001) and of a register labeled ‘Moroccan Flavored Dutch’ (Nortier and Dorleijn 2013), and in mining areas of Belgium a variety called Citétaal, ‘City language’ has emerged (Marzo 2005). Multiethnolects are also described for areas in and around Paris, France (Cheshire and Gardner‐Chloros 2018), and, as already mentioned, Multicultural London English is an emergent variety in London, England (Cheshire et al. 2011). As with all contact language varieties, multiethnolects are products of their local environment. They are directly informed by the particular ethnic and linguistic backgrounds of immigrants, the standard language of the area, and the overarching social structure. Yet despite this context specificity, a number of formal and functional characteristics appear to be common to all multiethnolects reported to date. In terms of structure, one way in which multiethnolects diverge from the local standard varieties is through the reduction or simplification of grammatical categories. In numerous multiethnolects of Dutch and Swedish, the common gender is used in place of neuter (e.g. Cornips 2008; Dorleijn and Nortier 2008; Kotsinas 1988; Marzo 2005; Quist 2008). For example, in Rinkeby Swedish, bord, ‘table’, is combined with the common article, en, whereas in standard Swedish it takes the neuter article, ett bord (Kotsinas 1998: 137). Equally, in new German varieties, ‘my sister’ is mein Schwester, rather than the feminine, meine Schwester (Dirim and Auer 2004: 441). Multicultural London English sees regularization of allomorphy in definite and indefinite articles, with [ə] and [ðə] being used, as well as a glottal stop before word‐initial vowels, instead of standard [ən] and [ði] (Cheshire et al. 2011: 186). Irregular plural forms of adjectives in standard French, such as normaux, are produced as a regular plural, normals (Cheshire, Nortier, and Adger 2015: 10). Less frequent use of articles and prepositions is also common in multiethnolects, e.g. Ich bin [in der] schule, ‘I am [at] school’ (Wiese 2009: 792). The copula is also often not realized, as in Ja ich [bin] aus Wedding, ‘yeah I [am] from Wedding’ (Wiese 2006: 257). These two utterances, both in the Berlin variety, Kiezdeutsch, serve to highlight the variation that is also a common characteristic of multiethnolects. SV word order is another common feature of multiethnolects spoken in Scandinavia and Germany, found in contexts which, in the standard varieties, require inversion of the subject and verb, i.e. when prefaced by an adverbial or in a subordinate clause. But such word order is rare in multiethnolects spoken in the Netherlands, despite Dutch also being a “V2 language” and following the same syntactic rules in this regard (Freywald et  al. 2011). Multicultural Paris French also features non‐standard word order, such as question words following verbs in embedded questions: je sais pas il va où, ‘I don’t know where he’s going’, where où, the final word, is ‘where’, which in standard French cannot be utterance final (Gardner‐Chloros and Secova 2018). Numerous innovations are also documented for multiethnolects. Speakers of ‘Hood language’ in Berlin use the novel particle lassma, from lass uns mal ‘let us (polite)’ (Wiese 2009: 799). In Rosengård Swedish, adolescents have extended uses of sån ‘such’, using it in place of the indefinite articles en/ett (Ekberg 2011). A discourse‐pragmatic innovation is seen in Multicultural London English, in a new quotative expression, ‘this is + speaker’, e.g. This is them “what area are you from, what part?”, this is me “I’m from Hackney” (Cheshire et al. 2011: 172). In this same variety, the function of man is extended to indefinite pronouns, as in It’s her personality man’s looking at (Cheshire, Fox, Kerswill, and Torgersen 2013: 65). Heavy use of verbs such as ‘make’, ‘do’, and ‘be’ are reported in varieties spoken in Germany and Sweden, with light verbs used such as göra sång, lit: ‘make song’, rather than the standard Swedish verb sjunga, ‘sing’ (Kotsinas 1998: 136), and ich mach dich Messer, ‘I’ll knife you’ (lit: I make/ do you knife) rather than using a verb such as stechen ‘stab’ (Wiese 2009: 793). Multiethnolects tend to be largely informed by one or two particular migrant language/s with respect to lexical borrowings, for example, nominals (welke sma?) ‘which girl?’ in Straattaal, spoken in the Netherlands, where sma is ‘girl’ in Sranan (Nortier 2001: 64), verbs

80  Carmel O’Shannessy and Lucinda Davidson (e.g. scassaren, ‘to irritate or tell lies’, in the Belgian multiethnolect, Citétaal, which draws on the Italian verb ‘to destroy’, scassare (Marzo 2005), phrases (e.g. hadi, lan!, Turkish for ‘come on, man!’, in Rinkeby Swedish, (Kotsinas 1998: 136), and discourse markers (e.g. wolla/wallah, derived from Arabic ‘I swear by Allah’, in Oslo (Opsahl 2009) and Copenhagen multiethnolects (Quist 2008: 47). Phonetic features of multiethnolects also differ from the standard language varieties of the area. For example, in the Netherlands, speakers of the multiethnolect realize sibilants in a manner which differs from standard Dutch: a strongly voiced /z/, and pronouncing /s/ as /∫/ when followed by /x/ or /l/ (Nortier 2018). A marked characteristic of Copenhagen multiethnolect for speakers of standard Danish is its lack of stød, the use of glottal restriction to realize a word accent (Quist 2008: 48). The telltale feature of multiethnolect in Malmö, Sweden, is pronouncing the initial sound in checkar, ’to check’ as /t∫/, and dental stops in multiethnolects in France tend to be realized as palatals and affricates (Fagyal and Torgersen, 2018). Prosodic differences from the standard are reported for multiethnolects across Scandinavia (Kotsinas 1998; Quist 2000; Svendsen and Røyneland 2008), and also in Multicultural Paris French (Fagyal and Stewart 2011) and Multicultural London English (Torgersen and Szakay 2011). With respect to the function of multiethnolects, this also appears to be quite similar around Europe. The majority of varieties described are reportedly used as a social resource for self‐ positioning (Quist 2008: 49) within the peer group (Freywald et al. 2011) and/or in opposition to the mainstream culture (Dorleijn and Nortier 2008). Multiethnolects are also used by speakers to pursue and index in‐group solidarity (Doran 2004; Dorleijn and Nortier 2008). While this could also work to exclude non‐members (e.g. Kallmeyer and Keim 2003), this tends to be a less common, or less deliberate, function of the speech style. Other descriptors that speakers of multiethnolects provide is that they are “fun” to speak (e.g. Doran 2004; Marzo 2005), and they help convey a “tough” (e.g. Kallmeyer and Keim, 2003), or non‐ “posh” (Aarsaether, Marzo, Nistov, and Ceuleers, 2015) identity. Multiethnolects appear to be used with friends only. Straattaal, for example, is described as being not suitable for speaking to mothers or teachers (Schoonen and Appel 2005: 104). When a girl of Thai and German descent speaks in “District talk” to a German social worker, she is reprimanded by a peer of Turkish descent: erzähl doch mal gescheit lan ‘tell her smartly, girl’ (Kallmeyer and Keim 2003: 41). This response suggests that using the multiethnolect with someone from outside the peer group is not only inappropriate but it is not “smart” or “clever.” In terms of who speaks a multiethnolect, adolescents in Rotterdam use Straattaal with peers of all other minority groups, but not with people who are ethnically Dutch (Cornips and De Rooij 2013: 139). Københavnsk multietnolekt in Denmark, by contrast, is not restricted to speakers of immigrant backgrounds, with native Danish adolescents reported as using the variety too (Quist 2008: 58). A similar situation to the Copenhagen variety is described for the multiethnolect of Oslo, Norway (Svendsen and Røyneland 2008). Multicultural London English appears to have developed into an ethnically neutral language variety. In a perception test, Londoners could not pick the ethnicity of a speaker with any certainty, and non‐ Londoners tended to label speakers of the multiethnolect simply as Londoners (Cheshire et al. 2013). More evidence of multiethnolects prompting broader changes is the finding that the phonological features of four‐ and five‐year‐olds in London, England, emulate those of Multicultural London English, and diverge from the speech of their adult caregivers (Cheshire et al. 2011: 171). This demonstrates the children’s orientation to the speech of older children. It also indicates that preschool‐aged children can play an active role in promoting language change. Along with Multicultural London English, some other multiethnolects appear to be extending to the broader speech community. For example, in Germany, young4 people’s

Language Contact and Change through Child First Language Acquisition  81 responses to a forced choice perception task suggest that there is a sound change in progress, with the loss of contrast between [ç] and [ ʃ ] in the Berlin metropolitan area, influenced by the multiethnolect Kiezdeutsch (Jannedy and Weirich 2014). This section has outlined the structural patterns and social functions of multiethnolects around contemporary Europe. Predominantly a stylistic resource used by adolescent peer groups, it appears that features of multiethnolects can also spread to the broader community, even to preschool‐aged children.

4  Child‐influenced Changes in Progress As already established in this chapter, children’s novel uses of language do not necessarily lead to language change. Furthermore, innovations that lead to lasting change can only be identified as such with any certainty after the fact, once the variants have already spread. However, in this section we gather together descriptions of children’s language use in various post‐contact contexts, all of which are suggestive of change in progress. The studies describe patterns in children’s language use that differ from that of adults, even though the long‐term effect is not yet known. Some describe innovations in the linguistic form of children’s talk, while others suggest child‐driven change in the function of language. Six young speakers of the mixed language, Gurindji Kriol (see Section 2.3), living in the remote community of Nitjpurru, Australia, are attributed with reorganizing part of the case system of their language (van den Bos, Meakins, and Algy 2017). Aged between 6 and 14 years old, these children have been found to use the nominative case suffix ‐ngku to mark subject and possessor on vowel‐final nominals, and to do so categorically. This innovation mirrors existing allomorphy in the language, in which the suffix ‐tu can mark subject, possessor, and dative on consonant‐final words. Twelve children in neighboring communities, all close relatives of the six at Nitjpurru, and within the same age range, have also been found to use ‐ngku in this novel way. Four of these other children do so in all instances recorded, while the other eight shift between using ‐ngku and ‐yu to mark possession. The authors propose that this variation indicates change in progress. No adult in the communities, including the adult caregivers of these 18 children, use the suffix ‐ngku to mark possession. Mention is also made of a four‐year‐old child, growing up in the same house as some of these older children, who produces ‐ngku to mark subject and possessor. This child’s use hints at L1 acquisition of this feature, which in turn would indicate lasting language change. Muang adolescents in Chiang Mai, northern Thailand, are described as using syncretic speech styles in their interactions with peers. Combining the languages, Kam Muang and Standard Thai, these speakers either produce a “Muang accented” variety of Standard Thai (for example ‘to be asleep’ in Thai, nɔɔn làp, is produced as with a Kam Muang tone, nɔɔn lãp), or insert non‐Muang lexicon (either from Standard Thai or novel hybrid forms, such as jàj kà puŋ for ‘cobweb’, which combines elements from both languages) into utterances with a Kam Muang structure (Howard 2010: 321). Broadly associated with urban Muang youth, this speech style is considered by teens and young adults as “a fun way of speaking with friends” (Howard 2010: 321). Speakers outside of this “youth” age bracket reportedly do not use this speech style. Young children speak predominantly in Kam Muang with peers, while adults decry the syncretic style as muaŋ kəm kəm, ‘inauthentic’ Kam Muang (Howard 2010: 321). This research is based on recordings made in the early 2000s. More recent language data would be required to ascertain whether these “fun” innovations by adolescents, indexing “youth,” have since taken hold within the broader community. Turkish‐German bilingual children aged 4–5 and 8–9 years show transfer of grammatical elements between the two languages (Pfaff 1982). The sociolinguistic profiles of the children correlate with the type and degree of transfer (Pfaff 1982). Turning to phonology, a group of

82  Carmel O’Shannessy and Lucinda Davidson Turkish‐German bilingual children aged 10–12 years use intonation patterns from each language as spoken unilingually in both their Turkish and their German. The bilingual children use the two different patterns to perform specific discourse functions that they do not perform in monolingual speech (Queen 1996, 2001). On Easter Island, Chile, most children of Rapa Nui descent are not productive speakers of Rapa Nui. Young siblings interact mostly in Chilean Spanish, and children are reportedly teased by peers at school if they do not. While many adults still speak Rapa Nui, the shift to Spanish in the community over the last 50 years is “advanced and undeniable” (Makihara 2005a: 729). However, in recent years there has been a push by Rapa Nui to reconstruct their traditional cultural identity, and children are positioned as central players in this move. Makihara (2005a, 2005b) describes a new variety of Rapa Nui Spanish that children speak when interacting with adults. It is purportedly “structurally distinct” (Makihara 2005a: 731) from the Rapa Nui Spanish that adults use, and its key characteristic is that many Spanish lexical items are replaced with Rapa Nui equivalents. Children only use this particular variety of Spanish with adults, and these days adults also use it with children, even those adults who speak Rapa Nui fluently (Makihara 2005b: 732). Such a situation suggests that children play an active role in shaping adults’ linguistic accommodation strategies (Makihara 2005a). Children could also be viewed as encouraging a more inclusive notion of language community amongst adults through their use of this different language variety (Makihara 2005b: 118). Other research explores children’s role in changes to language function, as opposed to language form. In two different Caribbean contexts, Morne‐Carré, a rural village on the island nation of St Lucia, and Dominica, West Indies, adult teachers in the schools try to ­prescribe children’s use of language, encouraging them to speak English. In both sites, English is associated with education, social mobility, and communication with the wider world, whereas the creole, although the language of community and emotional expression, also indexes poverty and a lack of social mobility. Irrespective of their levels of proficiency in the local creole, children in both communities produce it in interactions with peers, and they do so to achieve their own purposes (Garrett 2005, 2011; Paugh 2005a, 2005b). In both contexts, children seem to have extended this adult‐only use of the creole to ideas of authority, autonomy, and the status of being “grown up.” On St Lucia, Kwéyòl is the creole children are reprimanded for speaking (Garrett 2005). However children appear to use Kwéyòl as a resource to assert authority. When within earshot of an adult, children as young as five scold their younger sibling if they use Kwéyòl (Garrett 2007: 246). An interaction between boys aged four to seven years, away from adults, sees them including Kwéyòl in their interactions as a means of positioning themselves within the group, and to project a grown‐up persona (Garrett 2007: 248). Similarly, in Dominica, children as young as three years of age use the local creole, Patwa, with peers for specific functions, e.g. to enact particular adult roles when they are playing. They also use creole to assert authority over one another, through making negative assessments, intensifying their speech, issuing orders, and cursing (Paugh 2005a: 67). The linguistic situation described for Gapun, Papua New Guinea (Kulick 1992), could also potentially be viewed from this perspective. Unlike the Caribbean contexts, the adult‐ directed language choice here is not explicit; children are implicitly socialized into speaking the creole, Tok Pisin, rather than the traditional language of the village, Taiap. However, while the focus of the study is adult‐child interactions, a few mentions of children’s peer talk suggest that these interactions are markedly different from those with adults. That is, children do sometimes use Taiap with peers, whereas they tend to interact with adults in Tok Pisin only. Two examples are provided, of girls, aged six and seven years, who incorporate Taiap into interactions with their respective one‐ and two‐year‐old siblings (Kulick 1992: 218). The child’s use of the vernacular is described as marking utterances as “bossy or funny”

Language Contact and Change through Child First Language Acquisition  83 (Kulick 1992: 218). Such a comment suggests that children are using the language for a slightly different range of functions from adults. These studies describe children in situations of adult‐encouraged language shift, and hint at the subtle changes children make with and to language in their peer interactions. This section has described children in linguistic contexts which appear to be in the midst of change. This change looks to be in large part due to children’s own language choices, and often when in the company of peers. For each of the case studies presented in this section it is not clear if the innovations will remain in the children’s speech beyond childhood. Nevertheless, the studies show that school‐aged children can bring contact‐ induced language changes into their speech as they grow up, as part of the acquisition of their first languages.

5  Typology and Process – Summary 5.1  Summary of Types of Change Instigated by Child Speakers In this section the types of contact‐induced language change that are instigated by child speakers and that take hold in a community’s way of speaking, becoming conventionalized as a new language or variety, are summarized. The main processes are regularization of variation, reanalyses of input patterns to create new structure, and selection of some features over others. The case studies have shown regularization of morphological paradigms (Jourdan 1989; O’Shannessy 2013; Shnukal and Marchese 1983), and where the speakers are bilingual, regularization in one of their languages more than the other, creating greater differentiation ­between them (O’Shannessy 2015a). In several cases children have made incremental morphosyntactic changes (Jourdan 1989; O’Shannessy, forthcoming; S. J. Roberts 2004; Shnukal and Marchese 1983), and in some cases children have selected features in the input from changes in progress and pushed the change further (O’Shannessy 2015a; Sankoff 1991). Children have increased morphology through grammaticalization processes (O’Shannessy, 2013; A. Senghas et  al. 2004), and reanalyzed structure by finding patterns in the input, deducing a new meaning for an element, and then regularizing it (O’Shannessy 2013). They have combined the forms from one language with the function of another (Meakins 2011b; O’Shannessy 2016b). In one case they have used lexical items from one language that increased the amount of analytic structure in their productions (Lee 1987, 1988). In some instances, when code‐switched input was presented to children in the input in a frequent and very consistent pattern, they internalized it as a single system, although this is rare (McConvell and Meakins 2005; O’Shannessy 2012). Children have regularized word order across clause types (Kotsinas 1998; Quist 2008). They have introduced contractions of multiword constructions to create new word structure (Wiese 2009), and made lexical choices that increase differentiation between their languages (O’Shannessy 2015a). In phonology children have made phonological reductions, and changed tone and intonation slightly (Shnukal and Marchese 1983), or created new functions for existing intonation patterns (Queen 1996, 2001). Children have increased the use of phonotactic patterns of one language in the other (O’Shannessy 2016b). In koines they have regularized phonological variation from input from several dialects (Amery 1993; Kerswill and Williams 2000), and in multiethnolects they have changed the quality of vowels and consonants (Cheshire et  al. 2011; Wiese 2009). Many of these are similar to processes that we see in child first language acquisition or bilingual first language acquisition – regularization, reanalysis, language mixing, crosslinguistic transfer, and other nontarget constructions. The data lead us to conclude, then, that

84  Carmel O’Shannessy and Lucinda Davidson children’s language processing in different types of input context is largely similar (Aboh 2015; DeGraff 2002; O’Shannessy, forthcoming), and that what changes is the type of input and the sociolinguistic setting, which favors or disfavors the innovations taking hold in a community’s way of speaking.

5.2 How Children’s Novel Speech Productions Conventionalize in a Community’s Way of Speaking In each type of contact‐induced language change outlined in this chapter where there is empirical evidence of children’s speech production, except for Section 4, novel productions by children (ranging from three‐year olds to adolescents) have spread through their peer cohort and remained in their speech as they grew up, creating a new variety or a new language. The question then is why this happens in these environments, and either not at all, or much less often, in other environments. Most of the changes have taken place in clearly bounded communities of speakers (O’Shannessy 2013; Ross 2013). Time spent interacting with peers, and a desire to form a peer‐group identity is present in many of the case studies, and may apply to both the adult speakers instigating the first step of a two‐step process as well as the children subsequently adding changes. In many of the communities, children may model themselves and their speech on peers, rather than on adults’ speech styles. As McConvell explains for the Australian cases, … young children are also subject to linguistic input from older children, and this peer influence can eclipse that of the parental and grandparental generation. Children and teenagers may deliberately choose not to emulate parents or the old language. Instead they select or build a language variety of their own from among the models available. If there is no counter‐weight from the old language, then this peer‐group talk can form the basis of the language of the rising generation. (McConvell 2008b: 240)

Once a new way of speaking among peers becomes popular, to speak another way would be to diverge from one’s peers (cf. Giles, Coupland, and Coupland 1991), which members of tightly‐knit friendship groups are typically reluctant to do. In some of the cases described here the children spend a lot of time playing with each other in multi‐age groups from a relatively young age, so have strong reasons to emulate other young speakers over older speakers. In addition, the stability of peer interactions over a considerable period of time has been seen to be a factor favoring change. As a new way of speaking is conventionalized among its young speakers, it can come to signal their identity as a group. When a way of speaking performs an identity function it is an important part of its speakers’ linguistic repertoire. The communities often have few prescriptive norms in forms of schooling, literacy ­avenues such as newspapers etc., and adults often have few literacy requirements for their participation in the workforce (cf. Siegel 1985). Children may also have a high degree of social independence from a young age and not be subject to prescription in terms of speech styles and constructions (Bavin and Shopen 1991; Hamilton 1981; McConvell 1991; O’Shannessy 2012). In the case of Light Warlpiri, there is also a sociolinguistic culture of being creative with language use (O’Shannessy 2013). Where there is empirical data on the input to young children at the time they were acquiring the language, it suggests that there is in most cases a two‐stage process, with the adult input already containing many of the features that became part of the new, nativized way of speaking. None of the empirical data suggests an abrupt break between the properties of the input and the properties of the children’s productions. In some cases, those of

Language Contact and Change through Child First Language Acquisition  85 Light Warlpiri (O’Shannessy 2012) and Tiwi (Lee 1987), the children may have been spoken to in a distinct style as part of a baby‐talk register. The children then produced the same distinctive features and retained them in their way of speaking as they grew up. Regardless of a baby‐talk register, there is evidence in most instances that children were spoken to in forms that fairly closely resembled the way of speaking that conventionalized in the community, whether the input was from an adult generation or from peers.

6 Conclusion In summary, this chapter has reviewed empirical data from settings where children have played an influential role in contact‐induced language change. They mostly show a two‐ stage process, where a significant amount of reanalysis had been undertaken by adult speakers and was part of the input to the children. The children then acted on that input, changing it as they did so. They most commonly regularized variation, made incremental reanalyses on the adult productions, and selected some features over others to reproduce more frequently. The processes seen in these instances of child‐influenced contact‐induced language change are either the same as, or similar to, those seen in other contexts of child first language acquisition or child bilingual first language acquisition. This suggests that children’s processing in each kind of situation is similar, and the different types of input influence the way in which children act on the input they receive, e.g. if there is variation within a morphological paradigm the children may make it more regular. It also suggests that the factors favoring or disfavoring children’s innovations becoming part of a community’s new way of speaking lie in the sociolinguistic environment. Existing change in progress, with variation, in a delineated community, where children interact together frequently and have a reason to show a distinct identity, favors child‐influenced contact‐ induced language change.

NOTES 1 2;11 means two years, 11 months; and 4;6 means 4 years, 6 months. 2 3;9 means three years, nine months; 4;1 means four years, one month. 3 But see Evaldsson (2005) and Evaldsson and Cekaite (2010) for multiethnic groups of 7–9 year olds in a school context who orient to the standard language and use deviance from it as a resource to ostracize peers. 4 No precise ages were provided.

REFERENCES Aarsaether, Finn, Stefania Marzo, Ingvild Nistov and Evy Ceuleers 2015. Indexing locality: contemporary urban vernaculars in Belgium and Norway. In Jacomine Nortier and Bente Ailin Svendsen (eds.), Language, Youth and Identity in the 21st Century Cambridge: Cambridge University Press, pp. 249–70. Aboh, Enoch O. 2009. Competition and selection. That’s all! In Enoch O. Aboh and Norval Smith

(eds.), Complex Processes in New Languages. Amsterdam: Benjamins, pp. 317–44. Aboh, Enoch O. 2015. The Emergence of Hybrid Grammars. Language Contact and Change. Cambridge: Cambridge University Press. Aboh, Enoch O. and Ansaldo, Umberto. 2006. The role of typology in language creation. In Umberto Ansaldo, S Matthews, and Lisa Lim

86  Carmel O’Shannessy and Lucinda Davidson (eds.), Deconstructing Creole. Berlin: Benjamins, pp. 39–66. Alleyne, Mervyn C. 1980. Comparative Afro‐ American: An Historical‐comparative Study of English‐based Afro‐American Dialects of the New World. Ann Arbor, MI: Karoma. Amery, Rob. 1993. An Australian Koine – Dhuwaya, a variety of Yolnu Matha spoken at Yirrkala in North‐East Arnhem Land. International Journal of the Sociology of Language 99: 45–64. Appel, René. 1999. Straattaal: De mengtaal van jongeren in Amsterdam. Toegepaste taalwetenschap in artikelen 62.1: 39–55. Arends, Jacques and Bruyn, Adrienne. 1995). Gradualist and developmental hypotheses. In Jacques Arends, Pieter Muysken, and Norval Smith (eds.), Pidgins and Creoles: An Introduction. Amsterdam: Benjamins. Arends, Jacques, Muysken, Pieter, and Smith, Norval (eds.) 1995. Pidgins and Creoles: An Introduction. Amsterdam: Benjamins. Bakker, Peter. 1994. Michif, the Cree‐French mixed language of the Metis buffalo hunters in Canada. In Peter Bakker and Maarten Mous (eds.) Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam: IFOTT, pp. 13–33. Bakker, Peter. 1995. Pidgins. In Jacques Arends, Pieter Muysken, and Norval Smith (eds.), Pidgins and Creoles: An Introduction. Amsterdam: Benjamins, pp. 25–40. Bavin, Edith L. and Tim Shopen 1991. Warlpiri in the 80s: an overview of research into language variation and child language. In Suzanne Romaine (ed.) Language in Australia, 104–17. Beal, Joan C. 1999. “Geordie nation”: attitudes to dialect and regional identity in the north‐east of England. Lore and Language 17.1–2: 33–48. Bickerton, Derek 1981. Roots of language. Ann Arbor, MI: Karoma. Bickerton, Derek 1984. The Language Bioprogram Hypothesis. The Behavioral and Brain Sciences 7: 173–221. Bodén, Petra 2004. A New Variety of Swedish? Paper presented at the Speech Science and Technology. Burling, Robbins 1959. Language development of a Garo and English‐speaking child. Word 15: 45–68. Cheshire, Jenny, Susan Fox, Paul Kerswill and Eivind Torgersen 2013. Language contact and language change in the multicultural metropolis. Revue Française de Linguistique Appliquée 18.2: 63–76.

Cheshire, Jenny and Penelope Gardner‐Chloros 2018. Introduction: multicultural youth vernaculars in Paris and urban France. Journal of French Language Studies 28.2: 161–4. Cheshire, Jenny, Paul Kerswill, Sue Fox, and Eivind Torgersen 2011. Contact, the feature pool and the speech community: the emergence of Multicultural London English. Journal of Sociolinguistics 15.2: 151–96. Cheshire, Jenny, J. M. Nortier, and David Adger 2015. Emerging multiethnolects in Europe. Queen Mary’s Occasional Papers Advancing Linguistics 33. Chomsky, Noam 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Chomsky, Noam 1981. Lectures on Government and Binding. Dordrecht: Foris Publications. Clyne, Michael 2000. Lingua franca and ethnolects in Europe and beyond. Sociolinguistica 14.1: 83–9. Cornips, Leonie 2008. Losing grammatical gender in Dutch: the result of bilingual acquisition and/or an act of identity? International Journal of Bilingualism 12(1–2): 105–24. Cornips, Leonie and Vincent A. De Rooij 2013. Selfing and othering through categories of race, place, and language among minority youths in Rotterdam, the Netherlands. Multilingualism and Language Diversity in Urban Areas: Acquisition, Identities, Space, Education 1: 129. de Gruiter, Miel 1994a. Javindo, a contact langauge in pre‐war Semarang. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam: IFOTT, pp. 151–9. de Gruiter, Miel 1994b. Javindo, a contact language in pre‐war Semarang. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam, pp. 151–9. de Houwer, Annick 1983. Some aspects of the simultaneous acquisition of Dutch and English by a three‐year‐old child. Nottingham Linguistic Circular 12.2: 106–29. de Houwer, Annick 1990. The Acquisition of Two Languages from Birth: A Case Study. Cambridge: Cambridge University Press. de Houwer, Annick 2005. Early bilingual acquisition: focus on morphosyntax and the separate development hypothesis. In Judith F. Kroll and Annette B. M. de Groot (eds.), Handbook of Bilingualism: Psycholinguistics Approaches. Oxford: Oxford University Press, pp. 30–48.

Language Contact and Change through Child First Language Acquisition  87 DeGraff, Michel 2002. Relexification: a reevaluation. Anthropological Linguistics 44.4: 321–414. Dirim, İnci and Peter Auer 2004. Türkisch sprechen nicht nur die Türken: über die Unschärfebeziehung zwischen Sprache und Ethnie in Deutschland. Berlin: de Gruyter Mouton. Domingue, N. 1981. Internal change in a transplanted language. Studies in the Linguistic Sciences 4.2: 151–9. Doran, Meredith. 2004. Negotiating between Bourge and Racaille: Verlan as youth identity practice in suburban Paris. Negotiation of Identities in Multilingual Contexts, 93–124. Dorleijn, Margreet and Jacomine Nortier 2008. The hand and the glove: code and style as bilingual options among young people of Turkish and Moroccan descent in the Netherlands. In Vally Lytra and J. Normann Jørgensen (eds.), Multilingualism and Identities Across Contexts: Cross‐disciplinary Perspectives on Turkish‐speaking Youth in Europe. Copenhagen: University of Copenhagen, pp. 109–50. Ekberg, Lena 2011. Extended uses of sån (“such”) among adolescents in multilingual Malmö, Sweden 1. In Roger Källström and Inger Lindberg (eds.), Young Urban Swedish: Variation and Change in Multilingual Settings. Gothenburg, Sweden: University of Gothenburg, pp. 49–66. Evaldsson, Ann‐Carita 2005. “Staging insults and mobilizing categorizations in a multiethnic peer group, Discourse and Society 16.6: 763–86. Evaldsson, Ann‐Carita and Asta Cekaite 2010. “‘Schwedis’ he can’t even say Swedish” – subverting and reproducing institutionalized norms for language use in multilingual peer groups. Pragmatics 29.4: 587–605. Fagyal, Zsuzsanna and Christopher Stewart 2011. Prosodic style‐shifting in preadolescent peer‐ group interactions in a working‐class suburb of Paris. Friederike Kern and Margret Selting (eds.), Ethnic Styles of Speaking in European Metropolitan Areas Amsterdam: Benjamins, pp. 75–99. Fagyal, Zsuzsanna and Eivind Torgersen 2018. Prosodic rhythm, cultural background, and interaction in adolescent urban vernaculars in Paris: case studies and comparisons. Journal of French Language Studies 28.2: 165–79. Freywald, Ulrike, Katharina Mayr, Tiner Özçelik, and Heike Wiese 2011. Kiezdeutsch as a multiethnolect. Ethnic Styles of Speaking in European Metropolitan Areas, 45–73.

Gardner‐Chloros, Penelope and Maria Secova 2018. Grammatical change in Paris French: in situ question words in embedded contexts. Journal of French Language Studies 28.2: 181–207. Garrett, Paul B. 2005. What a language is good for: language socialization, language shift, and the persistence of code‐specific genres in St. Lucia. Language in Society 34.3: 327–61. Garrett, Paul B. 2007. Language socialization and the (re) production of bilingual subjectivities. In Monica Heller (ed.), Bilingualism: A Social Approach. Basingstoke: Palgrave Macmillan, pp. 233–56. Garrett, Paul B. 2011. Language socialization and language shift. In Alessandro Duranti, Elinor Ochs, and Bambi Schieffelin (eds.), The Handbook of Language Socialization. Oxford: Blackwell. Genesee, Fred. 1989. Early bilingual development: one language or two? Journal of Child Language 16: 161–79. Genesee, Fred 2003. Rethinking bilingual acquisition. Bilingualism: Beyond Basic principles. Festschrift in Honour of Hugo Baetens Beardsmore, pp. 204–28. Genesee, Fred, E. Nicoladis, and J. Paradis 1995. Language differentiation in early bilingual development. Journal of Child Language 22: 611–31. Giles, Howard, Nikolas Coupland, and Justine Coupland 1991. Accommodation theory: communication, context and consequence. In Howard Giles, Justine Coupland, and Nikolas Coupland (eds.), Contexts of Accommodation: Developments in Applied Sociolinguistics. Cambridge: Cambridge University Press, pp. 1–68. Hale, Kenneth, Mary Laughren, and Jane Simpson 1995. Warlpiri. In Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld, and Theo Vennemann (eds.), An International Handbook of Contemporary Research, vol. 2. Berlin: de Gruyter, pp. 1430–49. Hamilton, Annette. 1981. Nature and Nurture: Aboriginal Child‐rearing in North‐central Arnhem Land. Canberra: Australian Institute of Aboriginal Studies. Holm, John. 1988. Pidgins and Creoles, vol. 1. Cambridge: Cambridge University Press. Howard, Kathryn M. 2010. Social relationships and shifting languages in northern Thailand. Journal of Sociolinguistics 14.3: 313–40. Jannedy, Stefanie and Melanie Weirich 2014. Sound change in an urban setting: category

88  Carmel O’Shannessy and Lucinda Davidson instability of the palatal fricative in Berlin. Laboratory Phonology 5.1: 91–122. Jourdan, Christine 1989. Nativization and anglicization in Solomon Islands Pijin. World Englishes 8.1: 25–35. Kallmeyer, Werner and Inken Keim 2003. Linguistic variation and the construction of social identity in a German‐Turkish setting. In Jannis K. Androutsopoulos and Alexandra Georgakopoulou (eds.), Discourse Constructions of Youth Identities, vol. 110. Amsterdam: Benjamins, p. 29. Kegl, Judy, Ann Senghas, and Marie Coppola 2001. Creation through contact: Sign language emergence and sign language change in Nicaragua. In Michel DeGraff (ed.), Language Creation and Language Change: Creolization, Diachrony and Development. Cambridge, MA: MIT Press, pp. 179–237. Kerswill, Paul and Ann Williams 2000. Creating a new town koine: children and language change in Milton Keynes. Language in Society 29: 65–115. Kisch, Shifra 2012. Demarcating generations of signers in the dynamic sociolinguistic landscape of a shared sign language: The case of the Al‐Sayyid Bedouin. In Ulrike Zeshan and Connie de Vos (eds.), Sign Languages in Village Communities: Anthropological and Linguistic Insights. Berlin: de Gruyter Mouton, pp. 87–126. Kotsinas, Ulla‐Britt 1988. Immigrant children’s Swedish – a new variety? Journal of Multilingual and Multicultural Development 9.1–2, 129–40. Kotsinas, Ulla‐Britt. 1992. Immigrant adolescents’ Swedish in multicultural areas. In C. Palmgren, K. Lövgren, and G. Bolin (eds.), Ethnicity in Youth Culture. Stockholm: Stockholms Universitet, pp. 43–62. Kotsinas, Ulla‐Britt. 1998. Language contact in Rinkeby, an immigrant suburb Jugendsprache. langue des jeunes, youth language. Frankfurt‐ am‐Main: Peter Lang Europäischer Verlad der Wissenschaften, pp. 125–48. Kulick, Don 1992. Language Shift and Cultural reproduction: Socialization, Self and Syncretism in a Papua New Guinean Village. Cambridge: Cambridge University Press. Lanza, Elizabeth. 1992. Can bilingual two‐year‐ olds code‐switch? Journal of Child Language 19: 633–58. Lanza, Elizabeth 1997a. Language contact in bilingual two‐year‐olds and code‐switching: language encounters of a different kind? International Journal of Bilingualism 2.1: 135–62.

Lanza, Elizabeth 1997b. Language Mixing in Infant Bilingualism: A Sociolinguistic Perspective. Oxford: Oxford University Press. Lanza, Elizabeth 2004. Language Mixing in Infant Bilngualism: A Sociolinguistic Perspective, 2nd edition. Oxford: Oxford University Press. Laughren, Mary 2013. A revised analysis of Warlpiri verb inflections plus auxiliary combinations: their makeover in “Light” Warlpiri. In John Henderson, Marie‐Eve Ritz, and Celeste Rodríguez Louro (eds.), Proceedings of the 2012 Conference of the Australian Linguistic Society. Perth, WA: Australian Linguistic Society, pp. 1–14. Lee, Jennifer 1987. Tiwi Today: A Study of Language Change in a Contact Situation. Canberra: Pacific Linguistics. Lee, Jennifer 1988. Tiwi: a language struggling to survive. Aboriginal language use in the Northern Territory: 5 reports, vol. Series B vol 13. Darwin, Australia: Summer Institute of Linguistics, Australian Aborigines and Islanders Branch, pp. 75–96. Lefebvre, Claire. 1986. Relexification in Creole Genesis Revisited Substrata versus Universals in Creole Genesis. Amsterdam: Benjamins, pp. 279–300. Lefebvre, Claire 1998. Creole Genesis and the Acquisition of Grammar: The Case of Haitian Creole. Cambridge: Cambridge University Press. Lefebvre, Claire 2002. Relabelling: a major process in language contact. Journal of Language Contact, Thema 2: 91–111. Maehlum, Brit. 1992. Dialect socialization in Longyearbyen, Svalbard (Spitsbergen): a fruitful chaos. In Ernest H. Jahr (ed.), Language Contact. Theoretical and Empirical Studies. Berlin: de Gruyter. Makihara, Miki. 2005a. Being Rapa Nui, speaking Spanish children’s voices on Easter Island. Anthropological Theory 5.2: 117–34. Makihara, Miki 2005b. Rapa Nui ways of speaking Spanish: language shift and socialization on Easter Island. Language in Society 34.5: 727–62. Marzo, Stefania. 2005. Between Two Languages: The Linguistic Repertoire of Italian Immigrants in Flanders. In James Cohen, Kara T. McAlister, Kellie Rolstad, and Jeff MacSwan (eds.), Proceedings of the 4th International Symposium on Bilingualism. Somerville, MA: Cascadilla Press, pp. 154–59. McConvell, Patrick 1991. Understanding language shift: a step towards language maintenance. In Suzanne Romaine (ed.),

Language Contact and Change through Child First Language Acquisition  89 Language in Australia. Cambridge: Cambridge University Press, pp. 143–55. McConvell, Patrick. 2008a. Language mixing and language shift in Indigenous Australia. In Jane Simpson and Gillian Wigglesworth (eds.), Children’s Language and Multilingualism: Indigenous Language use at Home and School. London: Continuum. McConvell, Patrick 2008b. Mixed languages as outcomes of code‐switching: Recent examples from Australia and their implications. Journal of Language Contact 2: 187–212. McConvell, Patrick and Felicity Meakins 2005. Gurindji Kriol: a mixed language emerges from code‐switching. Australian Journal of Linguistics 25.1: 9–30. Meakins, Felicity 2011a. Borrowing contextual inflection: evidence from northern Australia. Morphology 21.1: 57–87. Meakins, Felicity 2011b. Case‐marking in Contact: The Development and Function of Case Morphology in Gurindji Kriol, vol. 39. Amsterdam: Benjamins. Meir, Irit, Assaf Israel, Wendy Sandler, Carol A. Padden, and Mark Aronoff 2012. The influence of community on language structure: evidence from two young sign languages. Linguistic Variation 12.2: 247–91. Meir, Irit, Wendy Sandler, Carol A. Padden, and Mark Aronoff 2010. Emerging sign languages. In Marc Marschark, Patricia Elizabeth Spencer, and Peter Nathan (eds.), Oxford Handbook of Deaf Studies, Language and Education, vol. 1. Oxford: Oxford University Press, pp. 267–80. Mufwene, Salikoko 1999. On the Language Bioprogram Hypothesis: hints from Tazie. In Michel DeGraff (ed.), Language Creation and Language Change: Creolization, Diachrony, and Development. Cambridge, MA: MIT Press. Mufwene, Salikoko 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Müller, Natascha and Aafke Hulk 2001. Crosslinguistic influence in bilingual acquisition: Italian and French as recipient languages. Bilingualism: Language and Cognition 4.1: 1–21. Muysken, Pieter 1994. Media Lengua. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam: IFOTT, pp. 201–5. Nicoladis, Elena and Fred Genesee 1997. The role of parental input and language dominance in bilingual children’s code‐mixing. Proceedings of the Annual Boston University Conference on Language Development 21.2: 422–32.

Nonaka, Angela M. 2012. Language ecological change in Ban Khor, Thailand: an ethnographic case study of village sign language endangerment. Sign Languages in Village Communities: Anthropological and Linguistic Insights 4: 277. Nortier, Jacomine 2001. Murks en Straattaal [Murkish and Street Language]. Amsterdam: Prometheus. Nortier, Jacomine 2018. Language and identity practices among multilingual Western European youths. Language and Linguistics Compass 12.5): e12278. Nortier, Jacomine and Margreet Dorleijn 2013. Multi‐ethnolects: Kebabnorsk, Perkerdansk, Verlan, Kanakensprache, Straattaal, etc. In Peter Bakker and Yaron Matras (eds.), Contact Languages: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 237–70. Opsahl, Toril. 2009. Wolla I swear this is typical for the conversational style of adolescents in multiethnic areas in Oslo. Nordic Journal of Linguistics 32.2: 221–44. O’Shannessy, Carmel 2005. Light Warlpiri: a new language. Australian Journal of Linguistics 25.1: 31–57. doi:10.1080/07268600500110472 O’Shannessy, Carmel 2012. The role of code‐ switched input to children in the origin of a new mixed language. Linguistics 50.2: 328–53. O’Shannessy, Carmel 2013. The role of multiple sources in the formation of an innovative auxiliary category in Light Warlpiri, a new Australian mixed language. Language 89.2, 328–53. O’Shannessy, Carmel 2015a. Multilingual children increase language differentiation by indexing communities of practice. First Language 35(4–5): 305–26. O’Shannessy, Carmel 2015b. Typological and social factors influencing a new mixed language, Light Warlpiri. Code‐switching Between Structural and Sociolinguistic Perspectives, 289–304. O’Shannessy, Carmel 2016a. Distributions of case allomorphy by multilingual children. Linguistic Variation 16.1: 68–102. O’Shannessy, Carmel 2016b. Entrenchment of Light Warlpiri morphology. Loss and Renewal: Australian Languages Since Colonisation 13: 217. O’Shannessy, Carmel forthcoming. How ordinary child language acquisition processes can lead to the unusual outcome of a mixed language. International Journal of Bilingualism. Otheguy, Ricardo, Ofelia García, and Wallis Reid 2015. Clarifying translanguaging and deconstructing named languages: a

90  Carmel O’Shannessy and Lucinda Davidson perspective from linguistics. Applied Linguistics Review 6.3: 281–307. Otterup, Tore 2011. “It’s simply a gift” – multilingualism as an individual and societal resource. In Roger Källström and Inger Lindberg (eds.), Young Urban Swedish: Variation and Change in Multilingual Settings. Gothenburg, Sweden: University of Gothenburg, pp. 173–94. Panda, Sibaji 2012. Alipur Sign Language: a sociolinguistic and cultural profile. Sign Languages in Village Communities: Anthropological and Linguistic Insights. Berlin: de Gruyter Mouton. Paugh, Amy L. 2005a. Acting Adult: Language Socialization, Shift, and Ideologies in Dominica, West Indies. In James Cohen, Kara T. McAlister, Kellie Rolstad, and Jeff MacSwan (eds.), Proceedings of the 4th International Symposium on Bilingualism. Somerville, MA: Cascadilla Press, pp. 180–20, Paugh, Amy L. 2005b. Multilingual play: Children’s code‐switching, role play, and agency in Dominica, West Indies. Language in Society 34.1: 63–86. Pfaff, Carol W. 1982. Constraints on language mixing: intrasentential code switching and borrowing in Spanish/English. In Jon Amastae and Lucia Elias‐Olivares (eds.), Spanish in the United States: Sociolinguistic Aspects. Cambridge: Cambridge University Press, pp. 264–97. Pinker, Stephen 1984. Language Learnability and Language Development. Cambridge, MA: Harvard University Press. Plag, Ingo 2008. Creoles as interlanguages: syntactic structures. Journal of Pidgin and Creole Languages 23.2: 307–28. Plag, Ingo 2009. Creoles as interlanguages: word‐formation. Journal of Pidgin and Creole Languages 24.2: 339–62. Queen, Robin 1996. Intonation in Contact: A Study of Turkish‐German Bilingual Intonation Patterns. Austin: University of Texas at Austin. Queen, Robin 2001. Bilingual intonation patterns: evidence of language change from Turkish‐ German bilingual children. Language in Society 30.1: 55–80. Quist, Pia. 2000. Ny københavnsk "multietnolekt." Om sprogbrug blandt unge i sprogligt og kulturelt hetergene miljøer. Danske Talesprog: 143–212. Quist, Pia 2008. Sociolinguistic approaches to multiethnolect: language variety and stylistic practice. International Journal of Bilingualism 12(1–2): 43–61.

Roberts, Julie 2002. Child language variation. In J. K. Chambers, Peter Trudgill, and Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. Oxford: Blackwell, pp. 333–48. Roberts, Sarah J. 2000. Nativization and the genesis of Hawaiian Creole. In John McWhorter (ed.), Language Change and Language Contact in Pidgins and Creoles. Amsterdam: Benjamins, pp. 257–300. Roberts, Sarah J. 2004. The emergence of Hawai’i Creole English in the early 20th century: the sociohistorical context of creole genesis. (Stanford University Ph.D. dissertation). Ross, Malcolm. 2013. Diagnosing contact processes from their outcomes: the importance of life stages. Journal of Language Contact 6.1: 5–47. Sankoff, Gillian. 1991. Using the future to explain the past. In Francis Byrne and Thom Huebner (eds.), Development and Structures of Creole Languages: Essays in honor of Derek Bickerton, pp. 61–74. Schoonen, Rob and René Appel. 2005. Street language: a multilingual youth register in the Netherlands. Journal of Multilingual and Multicultural Development 26.2: 85–117. Schuit, Joke 2012. Signing in the Arctic: external influences on Inuit Sign Language. In Ulrike Zeshan and Connie de Vos (eds.), Sign Languages in Village Communities: Anthropological and Linguistic insights. Berlin: de Gruyter Mouton, pp. 181–208. Senghas, Ann 2003. Intergenerational influence and ontogenetic development in the emergence of spatial grammar in Nicaraguan Sign Language. Cognitive Development 18.4: 511–31. Senghas, Ann, Marie Coppola, Elissa L. Newport, and Ted Supalia 1997. Argument structure in Nicaraguan Sign Language: the emergence of grammatical devices. In Elizabeth Hughes, Mary Hughes, and Annabel Greenhill (eds.), Proceedings of the 21st Annual Boston University Conference on Language Development. Somerville, MA: Cascadilla, pp. 550–61. Senghas, Ann, Sotaro Kita, and Aslı Özyürek 2004. Children creating core properties of language: evidence from an emerging sign language in Nicaragua. Science 309.5731: 56–56c. Senghas, R. J., A. Senghas, and J. E. Pyers 2005. The emergence of Nicaraguan Sign Language: questions of development, acquisition, and evolution. In Sue Taylor Parker, Jonas Langer, and Constance Milbrath (eds.), Biology and

Language Contact and Change through Child First Language Acquisition  91 Knowledge Revisited: From Neurogenesis to Psychogenesis. New York: Psychology Press, pp. 287–306. Shnukal, Anna and Lynell Marchese 1983. Creolization of Nigerian Pidgin English: Aaprogress report. English World‐Wide 4: 17–26. Siegel, Jeff 1985. Koines and koineization. Language in Society 14.3: 357–78. Siegel, Jeff 2000. Substrate influence in Hawai’i Creole English. Language in Society 29: 197–236. Siegel, Jeff. 2001. Koine formation and creole genesis. In Norval Smith and Tonjes Veenstra (eds.), Creolization and Contact (pp. 175–198). Amsterdam: Benjamins. Siegel, Jeff 2008. The Emergence of Pidgin and Creole Languages. Oxford: Oxford University Press. Solheim, Randi 2009. Dialect development in a melting pot: the formation of a new culture and a new dialect in the industrial town of Høyanger. Nordic Journal of Linguistics 32.2: 191–206. Svendsen, Bente Ailin, and Unn Røyneland 2008. Multiethnolectal facts and functions in Oslo, Norway. International Journal of Bilingualism 12(1–2): 63–83. Tomasello, Michael 2003. Constructing a Language: A Usage‐based Theory of Language Acquisition. Cambridge, MA: Harvard University Press.

Torgersen, Eivind and Anita Szakay 2011. A study of rhythm in London: Is syllable‐timing a feature of Multicultural London English? University of Pennsylvania Working Papers in Linguistics 17.2, 19. Van den Bos, Jackie, Felicity Meakins, and Cassandra Algy 2017. Searching for “Agent Zero”: the origins of a relative case system. Language Ecology 1.1: 4–24. Van Rheeden, Hadewych 1994. Petjo: the mixed language of the Indos in Batavia. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam, pp. 223–37. Wiese, Heike 2006. “Ich mach dich Messer”: Grammatische Produktivität in Kiez‐Sprache. Linguistische Berichte 207: 245–73. Wiese, Heike 2009. Grammatical innovation in multiethnic urban Europe: new linguistic practices among adolescents. Lingua 119.5, 782–806. Wilson, Aidan, Peter Hurst, and Gillian Wigglesworth 2018. Code‐switching or code‐mixing? Tiwi children’s use of language resources in a multilingual environment. In Gillian Wigglesworth, Jane Simpson, and Jill Vaughan (eds.), Language Practices of Indigenous Children and Youth: the Transition from Home to School. London: Palgrave Macmillan, pp. 119–46.

4 Contact and Grammaticalization1 BERND HEINE AND TANIA KUTEVA 1  Introduction: Contact‐induced Grammaticalization Language contact may have a wide range of implications for the languages involved, and it may affect virtually any component of language structure (Thomason and Kaufman 1988). It manifests itself in the transfer of linguistic material from one language to another, typically involving the following kinds of transfer: (1) Kinds of linguistic transfer: a. form, that is, sounds or combinations of sounds, b. meanings (including grammatical meanings) or combinations of meanings, c. form–meaning units or combinations of form–meaning units, d. syntactic relations, that is, the order of meaningful elements, e. any combination of (a) through (d). Our interest in this chapter is with (1b) and, to a lesser extent, with (1c), for a comprehensive model of linguistic transfer involving 1(a) through 1(d), readers are referred to Kuteva (2017). Following Weinreich (1953: 30–1); see also Heine and Kuteva (2003, 2005, 2006), the terms model language (ML) and replica language (RL) are used for the languages acting as the source (or donor) and the target (or recipient) of transfer, respectively. Weinreich’s term replication stands for kinds of transfer that do not involve phonetic substance of any kind, that is, for (1b) and (1d) – for what traditionally is referred to with terms such as structural borrowing or (grammatical) calquing. Thus, by grammatical replication we mean a process whereby speakers create a new grammatical meaning or structure in language R on the model of language M by using the linguistic resources available in R. The term borrowing is reserved for transfers involving phonetic material, either on its own (1a) or combined with meaning (1c).2 Furthermore, we will distinguish between replication restricted to the lexicon (= lexical replication) and replication that concerns grammatical meanings or structures (= grammatical replication). As has been shown in Heine and Kuteva (2003, 2005, 2006), grammatical replication is essentially in accordance with principles of grammaticalization; however, there are a few cases that are not, and the term restructuring has been proposed for the latter. Besides, some cases of what has been traditionally referred to as polysemy copying may also involve grammatical meanings; such cases can, too, be subsumed under grammatical replication. The model of contact‐induced transfer we will be using here can summarily be represented as in Figure 4.1.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

94  Bernd Heine and Tania Kuteva Contact-induced linguistic transfer

Replication

Grammatical replication

Contact-induced grammaticalization

Ordinary CIG

Replica CIG

Borrowing

Lexical replication

Polysemy copying

Restructuring

Rearrangement

Loss

Figure 4.1  Main types of contact‐induced linguistic transfer (CIG = contact‐induced grammaticalization).

A theoretical position underlying the majority of works on grammaticalization is that grammaticalization is a language‐internal process, and that grammaticalization contrasts with language contact, which is viewed as providing an alternative explanation for language change processes. Notice that the very introduction of the notion of grammaticalization was an attempt to complement the taxonomy of the three major types of language change identified by the Neogrammarians and de Saussure, namely sound change, analogy, and borrowing. Thus when Meillet (1912) first proposed the term grammaticalization (which – unlike analogy – creates new categories for which no previous patterns exist) he set a standard for treating grammaticalization as something other than contact‐induced phenomena (such as borrowing). A perhaps extreme version of the view that language‐internal change and language contact are mutually exclusive can be seen in the assertion that the former is portrayed as “natural,” whereas contact‐induced change (Ross 2003), or language‐external change, is “non‐natural” (for an explicit articulation of this standpoint, see Trudgill 1983: 102). In this spirit, Lass (1997: 199, 209) argues that whenever there is a possibility of dual or multiple origin for a given feature in a language, the most “parsimonious” explanation for a linguistic innovation is one in terms of internal development (or “endogeny,” Filppula 2003) rather than of language contact, that is, of external development: endogenous changes, he argues, occur in any case “whereas borrowing is never necessary.” In the model that we propose, language contact and grammaticalization are not mutually exclusive; rather, they may work in conspiracy with each other. Accordingly, both language‐ internal change and contact‐induced change are natural. Our model thus is an attempt to capture the inner dynamics of a complex process involving no fewer than the following factors: (2) a. universals of human conceptualization and grammaticalization, b. language contact, and c. the situation of discourse. The goal of grammaticalization theory is to describe grammaticalization, that is, the way grammatical forms arise and develop through space and time, and to explain why they are structured the way they are. Contact‐induced grammaticalization is a grammaticalization

Contact and Grammaticalization  95 process that is due to the influence of one language on another; instances of it have been ­discussed in a wide range of studies (see especially Campbell 1987; Stolz 1991; Haase 1992; Heine 1994; Harris and Campbell 1995; Nau 1995; Bisang 1996; Kuteva 1998, 2000; Dahl 2000; Heine and Kuteva 2003, 2005, 2006; Stolz and Stolz 2001; Aikhenvald 2002; Kuteva et al. 2019), even if a number of these studies do not use a framework of grammaticalization theory. There is a widespread assumption among linguists that grammatical structure, or syntax, cannot be “borrowed,” that is, transferred from one language to another. This assumption is reflected in a recent survey article by Sankoff (2002), who concludes that “whether or not ‘grammar’ or ‘syntax’ can be borrowed at all is still very much in question …many students of language contact are convinced that grammatical or syntactic borrowing is impossible or close to it” (Sankoff 2002; see also Silva‐Corvalán 2007). We consider this no longer to be an issue, considering that there is by now abundant evidence to demonstrate that both grammar – syntax included – can be “borrowed” or, as we will say here, replicated (see e.g. Ramisch 1989; Ross 1996, 2001, 2003; Johanson 1992, 2002; Aikhenvald 2002; Heine and Kuteva 2003, 2005, 2006), and the present chapter will provide further evidence in support of this observation.

2  Process versus Outcome Our main concern in this chapter is with grammatical replication as an outcome, for which there is some cross‐linguistic evidence, and we will have little to say about the process leading to this outcome since it is still in the main poorly understood. The following remarks are meant to provide at least some general understanding of the nature of this process, which has both a sociolinguistic and a linguistic component. We assume that at the beginning of the process as a sociolinguistic phenomenon there typically is spontaneous replication in bilingual interaction, where an individual speaker – consciously or unconsciously – propagates novel features in the replica language that have been influenced by some other language (or dialect). Spontaneous replication, described with references to notions such as “speaker innovation” (Milroy and Milroy 1985: 15), is highly idiosyncractic and the vast majority of instances of it will have no effect on the language concerned, being judged as what are commonly referred to as “speech errors.” But some instances may catch on: being taken up by other speakers and used regularly, they may become part of the speech habits of a group of speakers (early adopters), and they may spread to other groups of speakers – in exceptional cases even to the entire speech community. Still, this process does not necessarily lead to linguistic change: such innovations may remain restricted to some specific period of time, being abandoned either by the very speakers who introduced them or by the next generation of speakers. It is only if an innovation acquires some stability across time that grammatical replication has taken place. All the data that are available suggest that grammatical replication as a linguistic process is grounded in discourse pragmatics and semantics rather than in syntax. It may concern a single lexical or grammatical item (item extension), a cluster of several items (pattern extension), or a natural class of items (category extension). Contact‐induced innovations tend to be confined to an increase in frequency of use and an extension of existing pieces of discourse structure to new contexts, some of which are associated with novel meanings, and they also tend to be confined to specific semantic and/or lexical features. Some form of consistency is attained when, as a result of language contact, clusters of discourse pieces turn into new use patterns, or when existing minor use patterns turn into major ones, as sketched in Table  4.1. It is only a minor fraction of new use patterns that will eventually develop further into new distinct grammatical structures of grammatical replication, such as functional categories or new forms of syntax.

96  Bernd Heine and Tania Kuteva Table 4.1  From minor to major use pattern in grammaticalization (Heine and Kuteva 2005, 2006). Stage

Frequency

Context

Meaning

0 I

Low frequency Increase in frequency High frequency

Restricted Extension to new contexts Context generalization

Weakly grammaticalized An additional, more grammatical meaning may emerge in the new contexts Generalization of the new grammatical meaning

II

With these remarks we wish to draw readers’ attention to the fact that discussion in this chapter takes care only of a highly limited range of phenomena that have to be taken into account when studying contact‐induced grammaticalization in particular and grammatical replication in general.

3 Ordinary versus Replica Contact‐induced Grammaticalization A major type of grammatical replication is contact‐induced grammaticalization. A distinction between two kinds of contact‐induced grammaticalization is proposed by Heine and Kuteva (2005: ch. 3). Speakers may create a new use pattern or category that is equivalent to a corresponding category in the model language either by drawing on universal principles of grammaticalization (= ordinary contact‐induced grammaticalization), or else by replicating the process that they observe in the model language (= replica contact‐induced grammaticalization). The following example illustrates the significance of ordinary contact‐induced grammaticalization. Eastern Oceanic languages of northern and central Vanuatu commonly distinguish a durative aspect indicating that an act is in progress. Apparently in an attempt to find an equivalent for such a category in Bislama, the English‐based pidgin of Vanuatu, speakers used an expression commonly recruited cross‐linguistically to develop progressive and durative aspect markers (Kuteva et al. 2019). They chose a use pattern involving their verb stap ‘stay, be present, exist’3 to develop a durative aspect marker, which appears in the same syntactic slot as the durative markers in the Oceanic model languages4 (Keesing 1988, 1991: 328); cf. the following sentences, where (3a) illustrates the replica and (3b) the model structure:5 (3) a. Bislama (English‐based pidgin; Keesing 1991: 328) em i‐ stap pik‐ im he he‐ DUR dig‐ TRS ‘He’s in the process of digging yams.’ b. Vetmbao (Malekula, Oceanic; Keesing 1991: 328) naji ng‐ u‐ xoel dram. he he‐ DUR‐ dig yam ‘He’s in the process of digging yams.’

yam. yam

That this is an instance of ordinary contact‐induced grammaticalization is suggested by the fact that the Oceanic durative markers did not provide any model for how to create a durative marker in Bislama. The pidgin speakers therefore drew on a universal conceptual

Contact and Grammaticalization  97 strategy (see Bybee, Perkins, and Pagliuca 1994) by grammaticalizing a lexical verb to an aspect marker that is functionally and syntactically equivalent to that of the Oceanic model languages. The significance of the distinction between ordinary and replica contact‐induced grammaticalization can be illustrated with example (4), which involves the latter. In the course of half a millennium of contact with Italian, speakers of Molisean (Molise Slavic, a language derived from Croatian and spoken on the eastern coast of central‐southern Italy) created an indefinite article by grammaticalizing their numeral for ‘one’.6 In doing so, they appear to have replicated a situation that they observed in their model language, Italian (Breu 2003a). That the articles in the two languages are largely equivalent can be shown with the following example, where both languages have an optional zero article (Ø) in certain generic uses: (4) Molisean versus Italian (Breu 2003a: 42) Molise Slavic Ona je na študentesa./Ø profesoresa. Italian Lei è una studentessa./Ø professoressa. ‘She is a student/a professor.’ But Molisean speakers did not acquire a definite article, that is, they did not similarly replicate the Italian definite article. Breu attributes this to the fact that Italian provided a model for the indefinite but not for the definite article: whereas the Italian numeral un‐ ‘one’ and the indefinite article have similar forms, and hence allowed Molisean speakers to establish a conceptual link between the two, the Italian demonstratives show no formal similarity with the definite article. Accordingly, Molisean speakers acquired an indefinite article via replica grammaticalization, but creating a definite article would have required ordinary grammaticalization, which these speakers did not make use of, hence they did not develop a definite article. Another example of replica contact‐induced grammaticalization comes from Balkan Turkish (Heine and Kuteva 2006; Kuteva and Heine 2012; Kuteva 2017). It concerns the development of a wh‐interrogative word into a relative clause marker. Note that wh‐relativization appears to be an areal typological feature specific to Europe; if attested outside Europe, this is primarily in non‐European languages that have had some history of contact with European languages. Notice also that the historical development of the wh‐interrogative into a relativizer involves a grammaticalization path with three distinct stages: (a) interrogative stage, (b) complementizer stage, and (c) relativizer stage (Heine and Kuteva 2006: ch. 6). In some language varieties, some wh‐words have undergone (b) and (c), (e.g. Romance ­languages), in other varieties, however, we can say that – with respect to some wh‐interrogatives – only (b) has been reached. This is the case with the Silesian dialect of Polish, for instance, where the wh‐form for animates, fto ‘who’ can only be used as an interrogative and as a complementizer, but not as a relativizer (Kocur 2005). On the other hand, there are ­language varieties which represent a much more advanced stage of the wh‐relativization process. Balkan Turkish, for instance, is a high‐contact variety of Turkish, spoken on the Balkan Peninsula, where the other languages spoken (Bulgarian, Greek, Albanian, Romanian, Macedonian) belong to the best‐studied Sprachbund in the world, the Balkan Sprachbund. Let us recall that Standard Turkish consistently employs the so‐called gap relativization strategy, with no overt case‐marked reference to the head noun within the relative clause (Comrie and Kuteva 2005). What translates English relative clauses in Standard Turkish are inherited, attributive gerundial constructions preceding the head noun: (5) Turkish (Comrie and Kuteva 2005: 496) [kitab‐ ɩ al‐ book‐ ACC buy‐ ‘the student who bought the book’

an] PRT

öğrenci student

98  Bernd Heine and Tania Kuteva (6) Turkish (Comrie and Kuteva 2005: 496) [öğrenci‐ nin al‐ student‐ GEN buy‐ ‘the book which the student bought’

dɩğ‐ NMZ‐

ɩ] 3SG

kitap book

Unlike Standard Turkish, in Balkan Turkish we find a different relative clause construction. First, it employs a wh‐form ne ‘what’, which originally could only be used as an interrogative word, and has only recently taken on the function of a relativizer. Second, unlike the inherited preposed gerundial constructions used to mark relativization in Standard Turkish, the wh‐form, ne ‘what’, when used as a relative clause marker in Balkan Turkish, is postposed, just like wh‐words functioning as relative clause pronouns in the other Balkan Sprachbund languages (and other European languages) as illustrated below: (7) Balkan Turkish (Matras 1998: 94) şu araba kimindir bura that car who:GEN:COP here ‘Whose is that car that is parked here?’

ne REL

(8) Balkan Turkish (Matras 1998: 94) eski konuşma ne onlar old speech REL 3PL:NOM ‘It is an old language that they speak.’

duruyor? stand:CONT:3SG

konuşurlar speak:HAB:3PL

As has been argued in the typological literature already, wh‐relativization is an areal characteristic of European languages, therefore it is perfectly justifiable to assume that the exceptional situation of Balkan Turkish using a wh‐word as a relativizer is the result of the language contact that this dialect of Turkish has had with the Balkan Sprachbund languages over centuries. More precisely, as Heine and Kuteva (2006) and Kuteva and Heine (2012) argue, in this particular variety, language contact must have triggered a grammatical innovation, and this innovation  –  once set in motion  –  has followed a grammaticalization path observed in the Balkan Sprachbund (as well as in other European) languages. The observations presented in Matras (1998: 94–6) indicate that Balkan Turkish has indeed gone almost – but not fully – the whole grammaticalization path established in the model European languages for wh‐relativization. One can say that once the contact with the Balkan Sprachbund languages triggered the grammaticalization of the interrogative form ne ‘what’ in Balkan Turkish, this grammaticalizing structure clearly underwent (b) (complementizer stage), as indicated by uses going back to Ottoman Turkish (with parallels in modern Standard Turkish): (9) Ottoman Turkish (Matras 1998: 96) ekim siz what:CONJ 2PL:NOM ‘We shall do whatever you wish.’

dilersenüz wish.COND:2PL

edevüz do:OPT:1PL

Moreover, it appears to have entered (c) (the relativizer stage) as well, all the more so that in some contexts it has become an obligatory marker of relativization, with no alternative relative construction marker in the Balkan Turkish dialect. However, as Matras (1998: 94–5) points out, the Balkan Turkish “relative clause is not yet fully integrated into [a] pre‐planned, coherent complex construction.” In other words, in the case of wh‐relativization in Balkan Turkish we are dealing with a marker of an incipient relativizer stage, that is, a situation where a major part – but not the whole – of a grammaticalization process was replicated, over time, from ML (European Languages spoken on the Balkan Peninsula) into RL

Contact and Grammaticalization  99 (Balkan Turkish), the result being that a new relativization strategy, wh‐relativization, was added to the existing gerundial, the “gap” relativization strategy in the morphosyntax of this high‐contact language variety.

4 Contact‐induced Grammaticalization vs Polysemy Copying Breu (2003b) reports the following case of grammatical replication in the contact situation between Italian, the dominant language, and Molisean, which has been in contact with Italian for roughly 500 years: the Italian verb portare is polysemous, meaning both ‘carry’ and ‘drive a car’, while the pre‐contact Slavic verb nosit only meant ‘carry’. Speakers of Molisean replicated the Italian polysemy by extending the meaning of nosit to include both ‘carry’ and ‘drive a car’. This process is called polysemy copying by Heine and Kuteva (2005, section 3.2). Polysemy copying (usually discussed under the rubric of calquing or loan translation) is fairly common in lexical replication, as in the present example, but appears to be rare in grammatical replication, where a more complex process tends to be involved, namely contact‐induced grammaticalization (see Heine and Kuteva 2005: 100–3 and Heine 2012 for detailed discussion). The following example may illustrate the latter process. Slavic languages are renowned for their lack of articles, that is, of markers whose primary function it is to express definite or indefinite reference. The situation is, however, slightly more complex. There is a gradually decreasing degree of grammaticalization of definite and indefinite articles, respectively, as we move from Western European to Central European to Eastern European language varieties (i.e. languages, dialects, and regional varieties). In other words, with reference to the situation among the Slavic languages, there is hardly any discernible grammaticalization in the easternmost languages (Standard Russian, Belarusian, Ukrainian). The more one moves west and south, the higher the degree of grammaticalization of articles becomes. Given this geographical linguistic distribution, Heine and Kuteva (2006) argue that language contact was instrumental in the grammaticalization of articles in the majority of the cases discussed, especially for the following reasons. First, this is overwhelmingly the conclusion reached by experts who have dealt in some detail with the languages concerned (e.g. Breu 1994, 2003a, 2004; Lötzsch 1996). Second, there is sociolinguistic evidence to the effect that it is exactly those languages that are known to have had a long history of contact with article languages such as German, Italian, or Greek that exhibit at least minor use patterns of articles. With reference to whether it is possible to reduce grammatical replication to a question of polysemy copying, this case is of interest for the following reason. From a grammaticalization perspective, it is the following stages that mark the gradual pragmatic and semantic evolution of many indefinite articles (Heine 1997: 70–82): 1. An item serves as a nominal modifier denoting the numerical value ‘one’ (numeral). 2. The item introduces a new participant presumed to be unknown to the hearer and this participant is then taken up as definite in subsequent discourse (presentative marker). 3. The item presents a participant known to the speaker but presumed to be unknown to the hearer, irrespective of whether or not the participant is expected to come up as a major discourse participant (specific indefinite marker). 4. The item presents a participant whose referential identity neither the hearer nor the speaker knows (nonspecific indefinite marker). 5. The item can be expected to occur in all contexts and on all types of nouns except for a few contexts involving, for instance, definiteness marking, proper nouns, predicative clauses, etc. (generalized indefinite article).

100  Bernd Heine and Tania Kuteva Grammaticalization theory would predict that if a language has reached a given stage then it has also passed through all preceding stages, and this evolutionary scale can be used synchronically as an implicational scale. Both German and Italian have definite articles of stage 4, but neither has stage 5. Slavic languages on the other hand are widely believed to lack articles. That this belief is in need of revision has been demonstrated most of all by Breu (2003a) in his seminal analysis of Slavic “micro‐languages.” We are not able to do justice to the fine‐grained analysis presented by Breu (2003a); readers are referred to this work for many more details. The following discussion is confined to one of the micro‐languages spoken at the western periphery of the Slavic‐speaking territory, namely Upper Sorbian,7 which has been in contact with the dominant language German for nearly one millennium. In the examples below, sentences from Upper Sorbian (US) are given, followed by German (G) and English translations; there are no glosses in Breu’s publication (the markers in question are printed in bold, Ø stands for lack of article). Like German, Upper Sorbian has a fully grammaticalized stage 2, 3, and 4 indefinite article that can be traced back to the numeral jen‐ ‘one’.8 As the following examples show, their use is obligatory (where Ø = no article and *Ø = the article may not be omitted). Example (a) illustrates the presentative use (stage 2), characteristic of openings in tales, (b) the specific indefinite stage 3, and (c) the nonspecific indefinite stage 4. With abstract and generic referents as well, Upper Sorbian shows roughly the same degree of grammaticalization as the German indefinite article does, cf. (d). (10) Upper Sorbian (Breu 2003a: 37–52) a. Stage 2 US To běše jemo jena stara žona. G Es war einmal eine alte Frau. ‘Once upon a time there was an old woman.’ b. Stage 3 US Najmóle jo jen tołsty muž nutř šišoł. G Plötzlich kam ein dicker Mann herein. ‘Suddenly a fat man came in.’ c. Stage 4 US Dyš tybe jen pólcaj słóši, tón tybe zašperwe. G Wenn dich ein Polizist hört, wird er dich einsperren. ‘If a policeman hears you, he’ll arrest you.’

*Ø *Ø *Ø *Ø *Ø *Ø

Breu (2003a) describes a parallel case from Molisean of Italy, which also can be assumed to have lacked article‐like grammatical forms prior to language contact with the article language Italian. Like the Upper Sorbian numeral for ‘one’, Molisean ‘one’ disposes of a paradigm of morphophonological distinctions, one difference being that the Molisean forms have long and short forms as well, e.g. je’na versus na (nominative masculine singular). Like Upper Sorbian, Molisean shows roughly the same degree of grammaticalizaton of the numeral, having developed a stage 4 indefinite article of the same kind as the model language Italian. The reader is referred to Breu (2003a) for examples; suffice it to illustrate the more advanced stages. In the examples below, sentences from Molisean (M) are given, followed by Italian (I) and English translations (once again there are no glosses; the markers in question are printed in bold, Ø stands for lack of article). The (a) example shows the use with a stage 3 article with an abstract noun, while (b) shows a generic use of stage 4, where use versus non‐use of the indefinite article appears to be lexically determined, see also (4) repeated here as (11b). Note that in both examples, the replica and the model languages agree to the extent that both can be used with and without an article.

Contact and Grammaticalization  101 (11) Molisean (Breu 2003a: 42) a. M Jo, sa jima na strah! I Ahi, ho avuto una paura! ‘Boy, was I scared!’ b. M I

or Ø or Ø

Ona je na študentesa./Ø profesoresa. Lei è una studentessa./Ø professoressa. ‘She is a student/a professor.’

As these examples show, Molisean speakers, like Upper Sorbian speakers, have carried their numeral through all stages of grammaticalization, developing a stage 4 indefinite article largely equivalent to the Italian model. But the situation is different in other Slavic languages. Table 4.2 shows in particular two things. First, it is exactly those languages which have had the most intense contact with languages having full‐fledged (stage 4) indefinite articles that also have created corresponding articles. At one end there are Upper Sorbian and Molisean; at the other end there are Ukrainian and Belorusian, both languages with the least amount of contact with article languages. Second, Table 4.2 also shows that the contact‐induced grammaticalization proceeds in one direction from one stage to the next, where a new stage is built on the stage immediately preceding it. Synchronically, this fact can be described in the form of an implicational scale of the following kind: if a given article has stage X then it also has all preceding stages. Note that we are restricted here to nonstandard, colloquial, varieties of the languages concerned. As Breu (2003a) has shown for Upper Sorbian, an entirely different picture would arise if Standard Upper Sorbian were chosen. This example confirms what has been observed in other cases where we have comparative data on grammatical replication, e.g. on the possessive perfect (Heine and Kuteva 2006: ch. 4) or on auxiliaries in European languages (Heine and Miyashita 2008). Rather than replicating a grammatical category in toto, speakers start out with the replication of the initial stages of grammaticalization, and it requires a situation of long and intense contact for the replica category to attain the same degree of grammaticalization as the corresponding category of the model language. This constraint on contact‐induced grammatical replication suggests that, at least in cases such as the ones mentioned, there is really no polysemy copying; rather, what language contact triggers is a gradual process from less to increasingly more grammatical structure – a process that may end up in a fully equivalent replica category, as we observed in Upper Sorbian and Molisean. In most of the cases that have been Table 4.2  Degree of grammaticalization from numeral ‘one’ to indefinite article in selected Slavic languages. (sources: Breu 2003a; Heine and Kuteva 2006, ch. 3).

Stage

Function

Upper Sorbian

Molise Slavic

Mace‐ donian

Czech, Bulgarian

1 2 3 4

Numeral ‘one’ Presentative Specific indefinite Nonspecific indefinite

+ + + +

+ + + +

+ + +

+ + (+)

(sources: Breu 2003a; Heine and Kuteva 2006, ch. 3)

Serbian, Croatian, Polish, Ukrainian, Russian Belorusian + (+)

+

102  Bernd Heine and Tania Kuteva reported on grammatical replication, however, the process does not run through its full course (see also the discussion of wh‐relativization in Balkan Turkish above). Rather, the replica categories tend to be less grammaticalized than the corresponding model categories – a generalization that provides one of the diagnostics of contact‐induced grammaticalization (Heine and Nomachi 2013). There exist, however, examples where polysemy copying involving grammatical meanings seems to be the most plausible explanation behind particular instances of grammatical replication. It is these instances – rare as they are – that we will now turn to. Let us recall that in contact‐induced grammaticalization, what is replicated is – usually – part(s) of a grammaticalization process with some kind of a lexical (or less grammatical) linguistic structure at its beginning point and one or several grammatical (or more grammatical) structures along the grammaticalization path that a particular grammaticalization process follows. Unlike contact‐induced grammaticalization, in the case of polysemy copying, both the lexical (or less grammatical) structure and the grammatical (or more grammatical) structure(s) get replicated simultaneously, wholesale, into the replica language. Let us illustrate this by means of an example from the contact situation described in Keesing (1991), which involves Oceanic languages, i.e. the model, and the English‐based pidgin of Vanuatu, Bislama, i.e. the replica language. A number of northern and central Vanuatu languages (the model) have a grammatical structure which constitutes a reduplicated form of the verb ‘go’ as a marker indicating the passage of time in a narrative. In Epi (one of the model languages), for instance, the reduplicated form bababa of the verb of motion ba ‘go’ expresses this meaning. Likewise, in Bislama (the replica language), when the verb go (< Eng. go) is reduplicated as go‐go‐go it is used as a discourse marker which – again – stands for the passage of time in narratives (see Heine and Kuteva 2005: 100). Thus this is a case where the polysemy of the verb go as a lexical structure (in its spatial meaning) and as a grammatical structure (when reduplicated) has been copied from the model languages into the replica language. Notice, however, that unlike contact‐induced grammaticalization, which is characterized by intermediate stages of development, the above example does not appear to involve intermediate stages. Examples like the above indicate that we can regard polysemy copying – when one of the meanings involved is grammatical  –  as a branch of grammatical replication distinct from contact‐induced grammaticalization (see also Kuteva 2017).

5 Propelling versus Accelerating Forces in Language Contact We distinguish two kinds of forces moving grammaticalizing structures along grammaticalization paths: propelling forces and accelerating forces. Propelling forces, which are causally responsible for grammatical change to happen, include universal mechanisms of change such as metaphor and metonymy, which are usually realized through inference (Bybee et al. 1994: 281–93) and context‐induced re‐interpretation (Heine, Claudi, and Hünnemeyer 1991). It has been recognized by students of grammaticalization that such universal mechanisms “drive” grammaticalizing material along grammaticalization paths. What usually goes unrecognized in the mainstream literature on grammaticalization, however, is that language contact – rather than being the “non‐natural” alternative to the “natural” universal cognitive factors in language change – can also be a propelling force. It, too, can drive a lexical or a grammatical morpheme along a grammaticalization path. Which of the two forces, or to which extent each of them, is at work in a given case of contact‐induced language change is hard to determine on the basis of the little information that exists on attested cases of language contact processes.

Contact and Grammaticalization  103 The following might be an instance of a propelling force. A grammaticalization process that can be said to be characteristic of “Standard Average European” languages (Haspelmath 2001) is the one leading from question words to markers of clause subordination, and from interrogative constructions to complement and adverbial clause constructions, cf. English Who came? versus I don’t know who came; for more details of this process, see Heine and Kuteva (2006: ch. 6; 2007b: ch. 5). This process appears to be largely a European and an Indo‐ European phenomenon; if attested elsewhere, this is likely to be in languages that have had some history of contact with European languages. Documented examples of a process whereby a non‐Indo‐European language grammaticalized one or more of its question words to clause subordinators on the model of a European language include the following (see Heine and Kuteva 2006: ch. 6): (12) a. Basque on the model of Spanish, Gascon, and French (Hurch 1989: 21; Trask 1998: 320), b. Balkan Turkish on the model of Balkanic languages (Matras 1998), c. the North Arawak language Tariana of northwestern Brazil on the model of Portuguese (Aikhenvald 2002: 183), d. the Aztecan language Pipil of El Salvador on the model of Spanish (Campbell 1987: 259–60). From what we know about the nature of the replica languages concerned it is unlikely that a process from interrogation to subordination marker would have happened without language contact. Hence, there is reason to maintain that in these cases, language contact was a sine qua non for grammatical change to take place – in other words, this appears to be an example where language contact must have acted as trigger and a propelling force for grammaticalization. It is more difficult to identify clear cases where language contact acted as an accelerating force in grammatical change, that is, where it was causally responsible for speeding up a change that would have happened anyway: to be sure, the literature on language contact offers a number of examples that are intuitively suggestive of acceleration, but there are hardly any documented cases that would allow us to establish beyond reasonable doubt that language contact was the only, or at least the decisive, factor in contact‐induced grammaticalization, or any other kind of grammatical replication. Problems that one is confronted with here concern, on the one hand, the question of how degrees of speed in grammatical change can be measured; on the other hand, they also relate to the fact that contact‐induced grammatical change is a complex process involving a range of different factors, and it is in most cases unclear which of the factors contributes exactly what to a given process. Still, there are examples that can be taken to illustrate the effects of acceleration. One characteristic of many Indo‐European languages, not commonly found in other language families, can be seen in the presence of suffixal inflections (degree markers) on adjectives to express degrees of comparison, that is, to form comparative and superlative categories (cf. English small, small‐er, small‐est) or, alternatively, by suppletive forms (cf. good, better, best), both of which tend to be referred to as synthetic forms or constructions. These forms have turned out to be vulnerable to loss and replacement, and in a number of Indo‐European languages, especially the Romance languages, they have been replaced with free forms, that is, with analytic constructions (cf. English more, most beautiful). This process can be described as one of loss of synthetic constructions, but it is also one of grammaticalization of an analytic construction involving a degree adverb to a comparative construction which replaces the synthetic construction. Replacement has affected many European languages and in most cases there is no clear evidence that language contact played any role. But there are also cases

104  Bernd Heine and Tania Kuteva where contact apparently was an accelerating factor. Such cases include in particular Germanic languages in contact with Romance languages, where the former have lost their synthetic comparatives except for a few relics: (13) a. Analytic forms were still rare in Northern English in the fifteenth century, while in Southern English, which was more strongly exposed to French influence, these forms had become current at least a century earlier (Danchev 1989: 169). b. In German, the synthetic comparative construction is still well established, e.g. schön ‘beautiful’, schön‐er (‘beautiful‐COMP’) ‘more beautiful’. But in the dialect of Luxembourg, which has a history of centuries of intense contact with French, instances of an analytic comparative can be found, e.g. mehr schön (‘more beautiful’), presumably on the model of French plus beau (‘more beautiful’) (Alanne 1972; Danchev 1989: 170). c. The Dutch varieties spoken in Flanders, including Algemeen Nederlands, have been massively influenced by French, and this influence has also had the effect that the analytic pattern with the particle meer ‘more’ is used increasingly on the model of French plus, at the expense of the inherited synthetic construction using the comparative suffix ‐er (Taeldeman 1978: 58). d. Molisean has a 500‐years history of intense contact with the language of the host country, Italian, and one of the results of contact has been that the speakers of Molisean have given up the conventional Slavic synthetic construction by replicating the analytic Italian construction, using veče ‘more’ on the model of Italian più ‘more’ as degree marker. Thus, while Standard Croatian uses the synthetic form lyepši ‘more beautiful’, Molisean has veče lip (‘more beautiful’) instead (Breu 1996: 26; see also Breu 1999). More examples of this kind can be found in Heine and Kuteva (2006: ch. 2). What they seem to suggest is that the grammaticalization of analytic comparative constructions at the expense of inherited synthetic constructions, while representing an ongoing process in a number of European languages, can be accelerated by language contact.

6  Grammaticalization Areas In the course of language contact, a process of grammatical replication may affect three or even more languages, thereby giving rise to a grammaticalization area. A grammaticalization area exists when there is a group of geographically contiguous languages that have undergone the same grammaticalization process as a result of language contact. The following example may give an illustration of it. Passive markers in European languages commonly use “be” or “become” as auxiliaries, with the main verb being encoded as a perfect participle form or some equivalent of it. But, as Ramat (1998) shows, Rhaeto‐ Romance, Italian, and the Bavarian dialect of German share a periphrastic passive construction based on the grammaticalization of the lexical verb “come” to a passive auxiliary, cf. (14): (14) The Alpine “come”‐passive (Ramat 1998: 227–8) a. Ladin (Rhaeto‐Romance) Cô vain fabricheda la scuola nouva. [here comes built the school new]9 ‘Here the new school is being constructed.’

Contact and Grammaticalization  105 b. Italian Qui viene costruita la scuola nuova. [here comes built the school new] ‘Here the new school is being constructed.’ c. Bavarian (German) Då kummt de nei(e) Schul gebaut. [here comes the new school built] ‘Here the new school is being constructed.’ That this is an example of a grammaticalization area is suggested by the following observations: the constructions in the three languages concerned are the result of the same general process of grammaticalization from a construction [“come” + perfect participle] to a passive construction, where a lexical verb for “come” gave rise to a passive auxiliary. And this process must have involved language contact, for the following reasons. First, “come”‐passives do not appear to be cross‐linguistically common10 and it is therefore statistically unlikely that these three neighboring languages should have undergone such a rare process independently of one another. Second, common genetic inheritance can be ruled out as a contributing factor: the “come”‐passive cannot be traced back to earlier stages of Romance or Germanic languages. And third, the languages concerned can be shown to share a history of contact; Ramat (1998: 227–8) notes that there have been centuries of contact between southern Germany and northern Italy and that it was the Romance type that “has influenced the geographically contiguous Bavarian passive.” Accordingly, areal diffusion of a grammaticalization process among these geographically adjacent communities offers the most reasonable explanation. A number of such areas are discussed in Heine and Kuteva (2005, section 5.2). As pointed out there, grammaticalization areas may arise in two different ways: either there is replication leading from the model language (M) to the replica language (R1), which again serves as a model for another replica language (R2), as sketched in (15a) below, or the process leads straight from the model language to two different replica languages (R1, R2), cf. (15b). The synchronic outcome is in both cases the same, namely a grammaticalization area consisting of the three languages M, R1, and R2. (15) Patterns of transfer in grammaticalization areas a. M > R1 > R2 b. M > R1 > R2

7  Change in Typological Profile Contact‐induced grammaticalization can vary considerably in the way it contributes to language change. On the one hand, it may affect only a limited spectrum of grammatical structures of the replica language; on the other hand, its effects can be more dramatic in that it can be instrumental in changing the typological profile of a language.11 Examples of such contact‐induced changes in typological profile are discussed in Heine and Kuteva (2005: ch. 5); the following example may illustrate the kind of changes that are involved in this process. Over three hundred years of language contact in Sri Lanka between the Dravidian language Tamil on the one hand and Malay and Portuguese on the other had the effect that the latter two, nowadays referred to as Modern Sri Lanka Malay and Modern Sri Lanka Portuguese, respectively, developed into creoles with basic SVO

106  Bernd Heine and Tania Kuteva word order. More recently, the two have replicated grammatical structures of Tamil to the extent that they have turned into verb‐final languages, developed roughly the same case system, evidentials, quotative markers, semantics of verb markers, etc. – just like the model language Tamil (Bakker 2000). Grammaticalization played an important role in the Tamilization process. The following example shows how the change in typological profile from a creole to a Tamil‐like structure was achieved. Neither of the two creoles had a grammaticalized case system. In an attempt to replicate the case system of Tamil, speakers of the erstwhile creoles used a universal process of grammaticalization in developing adpositional, adverbial, and other independent forms into case suffixes, using the lexical resources that were available in the creoles. Thus, Portuguese pera/para ‘for’ was grammaticalized to an accusative/dative case suffix ‐pə, Portuguese junto ‘together’ to a locative case suffix ‐(u)nto, or Portuguese sua, seu ‘his, her, its’ to a genitive suffix ‐su(wə) in Modern Sri Lanka Portuguese, thereby matching the corresponding Tamil case markers (Bakker 2000: 32).

8  Constraints on Contact‐induced Grammatical Change In a number of works on language contact, most recently in Thomason (2007), it has been argued that there are no absolute linguistic constraints on contact‐induced change. In fact, there is now a wealth of data on language change processes in situations of language contact which all point in the same direction, namely that linguistic change in contact situations is fairly unconstrained. Work on grammatical replication, some of which is discussed in this chapter, shows however that this claim may be in need of revision. The observations made in previous work by the present authors (Heine and Kuteva 2003, 2005, 2006) suggest that grammatical change in language contact is shaped by universal principles of grammaticalization (see especially Heine et al. 1991; Bybee et al. 1994; Hopper and Traugott 2003), and that these principles impose a number of constraints on what is, and what is not, a possible contact‐induced grammatical change. Constraints are chiefly of the following kinds. First, they concern the choices that speakers make when looking in the replica language for translational equivalents of use patterns and categories that they find in the model language. For example, in a number of languages, a new future tense category has been created on the model of some other language (Heine and Kuteva 2005, Section 3.3). In creating this category, speakers have drawn typically on a universally available strategy by grammaticalizing e.g. a verb for ‘go (to)’, ‘come (to)’, or ‘want’, but not a verb for, say, ‘hit’ or ‘cut’. And in order to replicate an indefinite article of the model language, they most likely will select their numeral for ‘one’, and a demonstrative attribute to replicate a definite article, rather than grammaticalizing ‘one’ to a definite and a demonstrative to an indefinite article. Second, constraints relate to directionality in contact‐induced grammatical change. To use the examples just mentioned, it has happened time and again that a language has developed a future tense12 via the grammaticalization of ‘go (to)’ or ‘want’, etc., but it is highly unlikely that language contact will have the effect that a future tense marker develops into a verb for ‘go (to)’ or ‘want’, an indefinite article into a numeral for ‘one’, or a definite article into a demonstrative – in other words, grammatical change in language contact situations is essentially unidirectional. And third, the development of contact‐induced grammaticalization proceeds along a step‐by‐step sequence, as we showed in our example of indefinite articles in Slavic languages (section 4). Accordingly, it is unlikely that speakers will replicate a highly grammaticalized use pattern of a model language unless they have earlier replicated less grammaticalized use patterns.

Contact and Grammaticalization  107 To be sure, these are constraints that are not restricted to language contact; rather, they concern grammatical change in general. Grammaticalization is a ubiquitous process and is essentially unidirectional, irrespective of whether it takes place language‐internally or in situations of language contact.

9  Sociolinguistic versus Linguistic Factors Language contact is based typically on the social interaction among speakers (or writers) using more than one language (or dialect). Accordingly, sociolinguistic parameters provide an important basis for describing and understanding contact‐induced linguistic change, as has been aptly pointed out by many writers since Thomason and Kaufman (1988). One may wonder, however, whether much is gained if the study of language contact is reduced to sociolinguistic methodology. Studies on language contact differ greatly on whether they use a sociolinguistic or a linguistic framework, or a combination of both. All these frameworks have provided valuable insights, and there is no intrinsic reason to assume that any of them is superior to any other. Which approach is most appropriate depends on which aspect of language contact one decides to study and on the kinds of problems one wants to solve. Grammatical replication is a linguistic process and, accordingly, has been approached primarily by means of linguistic methodology (Heine and Kuteva 2005, 2006). While being ultimately the result of social processes, an analysis of it is less dependent on sociolinguistic variables than many other manifestations of language contact. This can be shown with the following example. A paradigmatic sociolinguistic situation of language contact is one where one of the languages involved is pragmatically dominant (Matras 1998: 285) or functions as a dominant code (Johanson 199, 2002) and serves primarily as L2 for L1 speakers of the dominated code. As has been pointed out by the authors concerned, this distinction is of help for understanding some aspects of both bilingual behavior and contact‐induced language change; cf. e.g. Johanson’s (1992, 2000: 165–6; 2002: 3) distinction between adoption and imposition. With reference to grammatical replication, however, this distinction is of secondary importance: Replication processes, as we described them in this chapter, may proceed in much the same way from L2 to L1 and from L1 to L2, or from dominant to dominated and from dominated to dominant code. In the former case, speakers use their L2 as a model for replication in their L1, while in the latter case they replicate features of their L1 in their (use of) L2. There are in fact a number of cases showing that grammatical replication may proceed in both directions (see Heine and Kuteva 2005, Section 6.3; 2006, Section 7.2). One case concerns Austronesian languages such as Tigak that appear to have provided a model for replication in the lingua franca Tok Pisin, but Tok Pisin also served as a model for Tigak (Jenkins 2002). Another example concerns the Basque language, which has replicated a wide range of grammatical structures from the dominant Romance languages Spanish and French, as well as from Gascon (see Hurch 1989; Haase 1992, 1997); at the same time, however, Basque also acted as a model language for grammatical replication by Spanish speakers in the Basque country (Urrutia Cárdenas 1995). Further examples are not hard to come by: one may mention language contact on the Channel island Guernsey, where English both served as a model language for the local Norman variety Guernésiais, and as a replica language, developing a number of new grammatical structures on the model of Guernésiais (Ramisch 1989; Jones 2002); situations of language contact between Turkic and Iranian languages, where both served at the same time as model and as replica languages (Soper 1987); or the contact setting in northwestern Amazonia, where the North Arawak language Tariana acted as a replica language for a wide range of grammatical structures from the East Tucanoan

108  Bernd Heine and Tania Kuteva language Tucano, but also provided the model for replication in the Tariana L2 variety of Portuguese (Aikhenvald 2002). The study of a larger range of contact situations suggests indeed that L2 > L1 replication is about as common as L1 > L2 replication (see Heine and Kuteva 2005 for examples), and that the processes are roughly of the same kind. It would therefore seem that in this kind of linguistic change, sociolinguistic variables only play a minor role and can largely be ignored in the study of grammatical replication – even if ultimately contact‐induced change is triggered by sociolinguistic factors.

10 Conclusions All the information that is available on language contact suggests that grammatical replication in general, and contact‐induced grammaticalization in particular, are far more common than has previously been assumed. This is partly due to the fact that grammaticalization is a ubiquitous, universally attested process that occurs in the same way language‐internally and language‐externally, and that in many cases it turns out to be hard, if not impossible, to determine whether a given grammaticalization process was contact‐induced or not (for a set of diagnostics, see Heine and Kuteva 2007a; Heine and Nomachi 2013). It is also due to the fact that, compared to other forms of contact‐induced change, grammatical replication has so far received little scholarly attention. Especially the earlier stages of the process, where new use patterns evolve in language contact, are still largely a terra incognita.

NOTES 1 T. Kuteva would like to thank the Humboldt Foundation, UCL as well as SOAS, University of London, and the National Social Sciences Fund of the Republic of China 15BYY107 for their support. 2 There are many alternative terminologies; for example, Thomason and Kaufman (1988) or Thomason (2001: 93) use borrowing for both kinds of transfer. 3 Bislama stap is historically derived from English stop. 4 Our interpretation of this case rests on Keesing (1988, 1991). Note, however, that this grammaticalization is not confined to Bislama, it can also be observed in other English‐based pidgins such as Tok Pisin and Solomons Pijin, and it remains unclear whether the process looked at here took place independently of these other cases. 5 The following abbreviations are used in glosses: dur = durative; trs = transitivizer; COMP = comparative. 6 The numeral has a range of different morphophonological variants depending on case, gender, and number; the nominative masculine singular form, for example, is je’na or, in its short form, na (Breu 2003a: 34). 7 All data analyzed by Breu (2003a) are from a rural, spoken variety of Upper Sorbian, which differs considerably from Standard Upper Sorbian, especially with reference to the phenomena looked at here. 8 The Upper Sorbian form for jen‐ ‘one’ is made up of a complex morphophonological paradigm on the basis of distinctions of six cases, two numbers (singular, plural), and three genders (masculine, feminine, and neuter) (see Breu 2003a: 36). 9 Ramat (1998) does not provide glosses for these examples; the use of parentheses indicates that the glosses are ours. 10 The only other case we are aware of is found in Maltese. 11 A “change in typological profile” obtains when a language as a result of grammatical replication experiences a number of structural changes to the effect that that language is structurally clearly different from what it was prior to language contact (Heine and Kuteva 2006, section 7.2).

Contact and Grammaticalization  109 12 Note that we are using here and elsewhere a shorthand convention. If we say that “a language has developed,” then it goes without saying that languages cannot do such things, rather it is the speakers of this language that do them.

REFERENCES Aikhenvald, Alexandra Y. 2002. Language Contact in Amazonia. New York: Oxford University Press. Alanne, Eero 1972. Zur Rolle der syntaktischen Interferenz der verwandten und unverwandten Sprachen. Neuphilologische Mitteilungen 73: 568–74. Bakker, Peter 2000. Convergence intertwining: an alternative way towards the genesis of mixed languages. In Dicky G. Gilbers, John A. Nerbonne, and Jos Schaeken (eds.), Languages in Contact (Studies in Slavic and General Linguistics, 28). Amsterdam: Rodopi, pp. 29–35. Bisang, Walter 1996. Areal typology and grammaticalization: processes of grammaticalization based on nouns and verbs in East and Mainland South East Asian languages. Studies in Language 20.3: 519–97. Breu, Walter 1994. Der Faktor Sprachkontakt in einer dynamischen Typologie des Slavischen. In Hans Robert Mehlig (ed.), Slavistische Linguistik 1993. Munich: Otto Sagner, pp. 41–64. Breu, Walter 1996. Überlegungen zu einer Klassifizierung des grammatischen Wandels im Sprachkontakt (am Beispiel slavischer Kontaktfälle). Sprachtypologie und Universalienforschung 49.1: 21–38. Breu, Walter 1999. Die Komparation im Moliseslavischen. In René Métrich, Albert Hudlett, and Heinz‐Helmut Lüger (eds.), Des racines et des ailes: théories, modèles, expériences en linguistique et en didactique. Nancy: ANCA, pp. 37–63. Breu, Walter 2003a. Der indefinite Artikel in slavischen Mikrosprachen: Grammatikalisierung im totalen Sprachkontakt. In Holger Kuße (ed.), Slavistische Linguistik 2001. Munich: Sagner, pp. 27–68. Breu, Walter 2003b. Bilingualism and linguistic interference in the Slavic‐Romance contact area of Molise (Southern Italy). In Regine Eckardt, Klaus von Heusinger, and Christoph Schwarze (eds.), Words in Time: Diachronic Semantics from Different Points of View. Berlin: de Gruyter Mouton, pp. 351–73.

Breu, Walter 2004. Der definite Artikel in der obersorbischen Umgangssprache. In Marion Krause and Christian Sappok (eds.), Slavistische Linguistik 2002. Munich: Sagner, pp. 9–57. Bybee, Joan L., Revere D. Perkins, and William Pagliuca 1994. The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago: University of Chicago Press. Campbell, Lyle 1987. Syntactic change in Pipil. International Journal of American Linguistics 53.3: 253–80. Comrie, Bernard and Tania Kuteva 2005. Relativization on subjects. In Martin Haspelmath, Matthew Dryer, David Gill, and Bernard Comrie (eds.), The World Atlas of Linguistic Structures (WALS). Oxford: Oxford University Press, pp. 494–7. Dahl, Östen 2000. The grammar of future time reference in European languages. In Östen Dahl (ed.), Tense and Aspect in the Languages of Europe (Empirical Approaches to Language Typology, Eurotyp 20–6). Berlin: De Gruyter Mouton, pp. 309–28. Danchev, Andrei 1989. Language change typology and adjectival comparison in contact situations. Folia Linguistica Historica 9.2: 161–74. Filppula, Marku 2003. The quest for the most “parsimonious” explanations: endogeny vs. contact revisited. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 161–73. Haase, Martin 1992. Sprachkontakt und Sprachwandel im Baskenland: die Einflüsse des Gaskognischen und Französischen auf das Baskische. Hamburg: Buske. Haase, Martin 1997. Gascon et basque: bilinguisme et substrat. Sprachtypologie und Universalienforschung 50.3: 189–228. Harris, Alice C. and Lyle Campbell 1995. Historical Syntax in Cross‐Linguistic Perspective. Cambridge: Cambridge University Press. Haspelmath, Martin 2001. The European linguistic area: Standard Average European. In Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An

110  Bernd Heine and Tania Kuteva International Handbook, vol. 2 (Handbücher zur Sprach‐ und Kommunikationswissenschaft 20.2). New York: Walter de Gruyter, pp. 1492–510. Heine, Bernd 1994. Areal influence on grammaticalization. In Martin Pütz (ed.), Language Contact and Language Conflict. Amsterdam: Benjamins, pp. 55–68. Heine, Bernd 1997. Possession: Sources, Forces, and Grammaticalization. Cambridge: Cambridge University Press. Heine, Bernd 2012. On polysemy copying and grammaticalization in language contact. In Claudine Chamoreau and Isabelle Léglise (eds.), Cross‐linguistic Tendencies in Contact‐ Induced Change: A Typological Approach Based on Morphosyntactic Studies. Berlin: de Gruyter Mouton, pp. 125–66. Heine, Bernd, Ulrike Claudi, and Friederike Hünnemeyer 1991. Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press. Heine, Bernd and Tania Kuteva 2003. Contact‐ induced grammaticalization. Studies in Language 27: 529–72. Heine, Bernd and Tania Kuteva 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva 2006. The Changing Languages of Europe. Oxford: Oxford University Press. Heine, Bernd and Tania Kuteva 2007a. Identifying instances of contact‐induced grammatical replication. Paper presented at the symposium “Language Contact and the Dynamics of Language: Theory and Implications,” Max Planck Institute for Evolutionary Anthropology, Leipzig, 10–13 May. Heine, Bernd and Tania Kuteva 2007b. The Genesis of Grammar: A Reconstruction. Oxford: Oxford University Press. Heine, Bernd and Hiroyuki Miyashita 2008. Accounting for a functional category: German drohen “to threaten.” Language Sciences 30: 53–101. Heine, Bernd and Motoki Nomachi 2013. Contact‐induced replication: some diagnostics. In Martine Robbeets and Hubert Cuyckens (eds.), Shared Grammaticalization. (Studies in Language Companion Series 132). Amsterdam: Benjamins, pp. 67–100. Hopper, Paul J. and Elizabeth C. Traugott 2003. Grammaticalization, 2nd edition. Cambridge: Cambridge University Press. Hurch, Bernhard 1989. Hispanisierung im Baskischen. In Norbert Boretzky, Werner

Enninger, and Thomas Stolz (eds.), Beiträge zum 5. Essener Kolloquium über “Grammatikalisierung: Natürlichkeit und Systemökonomie,” vol. 1 (Essener Beiträge zur Sprachwandelforschung 7). Bochum: Brockmeyer, pp. 11–35. Jenkins, Rebecca Sue 2002. Language Contact and Composite Structures in New Ireland, Papua New Guinea. Ann Arbor: UMI Dissertation Services. Johanson Lars 1992. Strukturelle Faktoren in türkischen Sprachkontakten. In Sitzungsberichte der Wissenschaftlichen Gesellschaft an der Johann‐Wolfgang‐Goethe‐Universität Frankfurt am Main. Stuttgart: Steiner, pp. 169–299. Johanson, Lars 2000. Linguistic convergence in the Volga area In Dicky G. Gilbers, John Nerbonne and Jos Schaeken (eds.), Languages in Contact. (Studies in Slavic and General Linguistics, 28.) Amsterdam, Atlanta, GA: Rodopi, pp. 165–78. Johanson, Lars 2002. Contact‐induced change in a code‐copying framework. In Mari C. Jones and Edith Esch (eds.), Language Change: The Interplay of Internal, External and Extra‐ Linguistic Factors. Berlin: de Gruyter Mouton, pp. 285–313. Jones, Mari C. 2002. Mette a haout dauve la grippe des Angllais: convergence on the Island of Guernsey. In Mari C. Jones and Edith Esch (eds.), Language Change: The Interplay of Internal, External and Extra‐linguistic Factors (Contributions to the Sociology of Language 86). Berlin: de Gruyter Mouton, pp. 143–68. Keesing, Roger M. 1988. Melanesian Pidgin and the Oceanic Substrate. Stanford: Stanford University Press. Keesing, Roger M. 1991. Substrates, calquing and grammaticalization in Melanesian Pidgin. In Elizabeth C. Traugott and Bernd Heine (eds.), Approaches to Grammaticalization, vol. 1. Amsterdam: Benjamins, pp. 315–42. Kocur, Justine 2005. Entwicklung der WH‐ Relativpronomen im Englischen und Polnischen: ein Vergleich (University of Düsseldorf MA Thesis). Kuteva, Tania 1998. Large linguistic areas in grammaticalization: auxiliation in Europe. Language Sciences 20.3: 289–311. Kuteva, Tania 2000. Areal grammaticalization: the case of the Bantu‐Nilotic borderland. Folia Linguistica 34.3–4: 267–83. Kuteva, Tania 2017. Contact and borrowing. In Adam Ledgeway and Ian Roberts (eds.). Cambridge Handbook of Historical Syntax. Cambridge: Cambridge University Press.

Contact and Grammaticalization  111 Kuteva, Tania and Bernd Heine 2012. An integrative model of grammaticalization. In Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds), Grammatical Replication and Borrowability in Language Contact. (Trends in Linguistics, Studies and Monographs, 242.) Berlin: de Gruyter Mouton, pp. 159–98. Kuteva, Tania, Bernd Heine, Bo Hong, Haiping Long, Heiko Narrog, and Seongha Rhee 2019. World Lexicon of Grammaticalization, 2nd revised and extensively updated edition. Cambridge: Cambridge University Press. Lass, Roger 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Lötzsch, Renate 1996. Interferenzbedingte grammatische Konvergenzen und Divergenzen zwischen Sorbisch und Jiddisch. Sprachtypologie und Universalienforschung 49.1: 50–9. Matras, Yaron 1998. Utterance modifiers and universals of grammatical borrowing. Linguistics 36.1: 281–331. Meillet, Antoine 1912. L’évolution des formes grammaticales. Scientia (Rivista di Scienzia) 12.26 (6): 384–400. Milroy, Lesley and James Milroy 1985. Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339–84. Nau, Nicole 1995. Möglichkeiten und Mechanismen kontaktbewegten Sprachwandels unterbesonderer Berücksichtigung des Finnischen (Edition Linguistik 08). Munich: LINCOM Europa. Ramat, Paolo 1998. Typological comparison and linguistic areas: some introductory remarks. Language Sciences 20.3: 227–40. Ramisch, Heinrich 1989. The Variation of English in Guernsey/Channel Islands (Bamberger Beiträge zur Englischen Sprachwissenschaft 24). Frankfurt‐am‐Main: Peter Lang. Ross, Malcolm D. 1996. Contact‐induced change and the comparative method: cases from Papua New Guinea. In Mark Durie and Malcolm D. Ross (eds.), The Comparative Method Reviewed: Regularity and Irregularity in Language Change. New York: Oxford University Press, pp. 180–217. Ross, Malcolm D. 2001. Contact‐induced change in Oceanic languages in North‐West Melanesia. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press, pp. 134–66.

Ross, Malcolm D. 2003. Diagnosing prehistoric language contact. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 174–98. Sankoff, Gillian 2002. Linguistic outcomes of language contact. In J. K. Chambers, Peter Trudgill, and Natalie Schilling‐Estes (eds.), Handbook of Language Variation and Change. Oxford: Blackwell, pp. 338–68. Silva‐Corvalán, Carmen 2007. The limits of convergence in language contact. Paper presented at the symposium “Language Contact and the Dynamics of Language: Theory and Implications,” Max Planck Institute for Evolutionary Anthropology, Leipzig, 10–13 May. Soper, John David 1987. Loan Syntax in Turkic and Iranian: The Verb Systems of Tajik, Usbek, and Qashqay. Ann Arbor: UMI Dissertation Services. Stolz, Christel and Thomas Stolz 2001. Mesoamerica as a linguistic area. In Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, vol. 2 (Handbücher zur Sprach‐ und Kommunikationswissenschaft 20.2). New York: Walter de Gruyter, pp. 1542–53. Stolz, Thomas 1991. Sprachbund im Baltikum? Estnisch und Lettisch im Zentrum einer sprachlichen Konvergenzlandschaft (Bochum‐ Essener Beiträge zur Sprachwandelforschung 13). Bochum: N. Brockmeyer. Taeldeman, Johan 1978. Französisch‐flämische Sprachinterferenz in Flandern. In Per Sture Ureland (ed.), Sprachkontakte im Nordseegebiet: Akten des 1. Symposions über Sprachkontakt in Europa, Mannheim 1977 (Linguistische Arbeiten 66). Tübingen: Max Niemeyer, pp. 43–66. Thomason, Sarah G. 2001. Language Contact: An Introduction. Edinburgh/Washington, DC: Edinburgh University Press/Georgetown University Press. Thomason, Sarah G. 2007. Social and linguistic factors as predictors of contact‐induced change. Paper presented at the symposium “Language Contact and the Dynamics of Language: Theory and Implications,” Max Planck Institute for Evolutionary Anthropology, Leipzig, 10–13 May. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press.

112  Bernd Heine and Tania Kuteva Trask, Robert L. 1998. The typological position of Basque: then and now. Language Sciences 20.3: 313–24. Trudgill, Peter 1983. On Dialect: Social and Geographical Perspectives. Oxford: Blackwell. Urrutia Cárdenas, Hernán 1995. Morphosyntactic features in the Spanish of the Basque Country.

In Carmen Silva‐Corvalán (ed.), Spanish in Four Continents. Washington, DC: Georgetown University Press, pp. 243–59. Weinreich, Uriel 1953. Languages in Contact: Findings and Problems. The Hague: de Gruyter Mouton.

5 Contact and Language Convergence ANTHONY P. GRANT

1  Kinds of Convergence Language convergence, which is an important facet of Contact‐Induced Linguistic Change (hereafter CILC), is an ambiguous term which can be interpreted in at least two if not three senses. In the first and more frequent sense, a language system becomes in part more like another (usually more prestigious or powerful) language system as a result of contact‐ induced change. This occurs because people who speak different languages but who come into contact frequently and for a range of different purposes may take over features of various kinds from one another’s language over a period of time. In the rare second sense, a language system is gradually absorbed (in part rather than in whole) into another language system, to the extent that a new language evolves which contains essential features of both. The first process is attested globally, the second is less common but has been attested in a number of cases in various parts of the world. (A third sense is found in work by Raymond Hickey, for instance Hickey (2011), which draws upon examples from Irish and Irish English. This states that convergence consists of the coming together of internal and external factors in the development of a linguistic feature which is attested in the language in which it arises and also in a language which is in a position to influence the first language. This resembles the first sense, except that here the source language already exemplifies the construction in question). An introductory section presents the two major senses of language convergence, and the commoner form of LC will be explored in Section 2, with examples from languages which have undergone interesting instances of convergence in various parts of their structures (Bloomfield’s 1933 concept of intimate borrowing is relevant here). Anything can be transferred from one linguistic system to another in a case of convergence, although a lot of attention in the literature has been paid to phenomena other than unidirectional lexical borrowing, which is often part of a case of convergence.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

114  Anthony P. Grant As Matthews (1956) shows, convergence can result from bidirectional influence: Finnic Livonian absorbed a large number of loans from Baltic, and then exerted a strong influence upon Latvian. This was the language by which Livonian (its speakers having lost political power and prestige) was then strongly influenced, absorbing not only lexicon (including Latvian loans from Low and High German and Russian) but also syntactic constructions and numerous items of derivational morphology. The earlier stages of this contact affect levels of the language which are most basic. This case is an example of bidirectional influence dependent upon the relative balance of power in that part of the Baltic littoral from one century to another. Similar remarks apply to the relations between Quechua and Aymara in the Andes (Adelaar 2012); the direction of influence seems to have shifted over the centuries, and although all Quechua varieties had a solid core of Aymara elements, some, such as Cuzco Quechua, acquired even more in later centuries, with unusual borrowing patterns and some replicated phonemes. The second kind of convergence, discussed in Section 3, can arguably be illustrated by Para‐Romani varieties from Europe and the Near East, but also with some languages which border on the territory of “stable mixed languages,” such as the Amazigh Valley Berber‐ Songhay mixed languages of West Africa. Furthermore, it can be illustrated with cases of languages whose major components are discrete but nonetheless closely related, such as Stedsk (lit. ‘city‐ish’, i.e. urban dialects of Frisian in the Netherlands, van Bree 1994), or Tojol‐ ab’al of Mexico. Distinctions and similarities will be teased out, and in the fourth section, a conclusion, attempts will be made to arrive at general patterns of what happens, what rarely happens, and what never happens in instances of extreme language convergence. After some discussion of views on convergence, instances of linguistic convergence at several levels are presented from a range of languages, together with a discussion of language creation through convergence, and a discussion of the effects of metatypy (Ross 1996). This is followed by a brief account of linguistic areas, and a conclusion.

1.1  Some Views on Linguistic Convergence Linguistic convergence in at least the first sense – systemic approximation of one language to another, usually more prestigious language – has been discussed by several investigators. The seminal article by Haugen (1950) talked about “loan translations” as “calquing” (French calque ‘footprint’), and the issue was also recognized in Weinreich (1953). Matras (2010), working from a functionalist standpoint, infused with the linguistic ideas of Karl Bühler concerning the nature of the speaker’s relationship with the linguistic system, draws in his work on the concept of pivot grammar (as found in the work on first language acquisition work by Braine 1963). Lucas (2015) draws in his discussion of contact‐induced linguistic change upon the work of Frans van Coetsem and his concepts of the two transfer types, borrowing and imposition, and their interaction with the donor and recipient languages: substrate influence can be seen as imposition whereas influence from other languages is more indicative of borrowing. In such a model of convergence the question arises of whether the source language or the recipient language has agentivity. As Grant (2002) has pointed out, any kind of contact‐induced linguistic change or CILC in any language can be expressed as being transfer of fabric (this denotes the borrowing of morphemes, including lexemes) or transfer of pattern (the borrowing of ways of using morphemes), and the two kinds of transfer are certainly not mutually exclusive within the same construction. They may, and often do, co‐occur. Furthermore, both kinds of transfer can result in words or constructions which then take on linguistic lives of their own, undergoing changes quite independent of the source material or source languages. Once a word or a construction has been taken over from one linguistic system into another, its subsequent adventures and changes are part of the history of that new linguistic system, and these developments may not be replicated or paralleled or pre‐empted in the language from which the

Contact and Language Convergence  115 term was originally taken. The history of the English verb STOP, which ultimately derives from a Greek word referring to the substance ‘tow’ (used to stuff holes in ships), is an eminent example of this (OED Online, s.v. STOP1). Indeed we may or should ask of any potential instance of CILC the following questions: Does this instance involve transfer of fabric? Does this instance involve transfer of pattern? If the answer to either or both of these is “Yes”, then we are dealing with a case of CILC. Though not focusing on a single language, Aikhenvald (2002) is an important study of aspects of CILC, including morphosyntactic features, among some languages in Amazonia, in which the author presents copious data from her fieldwork on a number of languages, including Tariana and Baniwa, in addition to her theory of CILC and her valuable concept of “system‐changing” and “system‐preserving” features. This is an interesting set of studies of convergence in an area where lexical borrowing is nonetheless generally dispreferred. Another work which exemplifies morphological and syntactic convergence across unrelated languages is Soper (1987), comparing features of Iranian Tajik and Turkic Uzbek and Qashqai; the former has largely remodeled its verbal tense‐aspectual system by emulating the structures of the latter languages but using inherited Iranian morphs. Sasse (1985) and Haase (1992) are superb studies of structural and other kinds of convergence for those who read German.

2  Some Strata of Linguistic Convergence Pitting convergence against borrowing as opposite sides of a coin is a less fruitful viewpoint than regarding them as opposite ends of a continuum with the mere transfer of fabric at one end and the mere transfer of patterns at the other (Heath 1984 and Grant 2003 provide these labels; see also the work of Nau 1995). The reason why the continuum approach is effective is because so many instances of pattern transfer involve the use of an item or items of transferred fabric, be it a lexical item used lexically or structurally, or a borrowed structural item. Heath (1984) introduced the term pattern transfer, which led Grant (2002, 2003) to develop the complementary (not opposing) term transfer of fabric to refer to the replication of morphs (including lexicon) from one language to another. The impact of one linguistic system upon another includes the impact of pattern transfer added to the impact of fabric transfer. Some distinctions between fabric and pattern are given in Figure 5.1, and examples of several of these occur in the chapter. Examples of Fabric Sounds (phonetic content) Some intonational elements Morphemes Lexicon Derivational cycles Material used to construct discourse Most adstrate effects Paradigmatic items (mostly)

System-affecting features (some) Borrowing (mostly) System-preserving changes

Examples of Patterns Sound patterns (segmental and canonical phonology) Suprasegmental phonology Morphemic arrangements (syntax) Semantic fields Meanings of derivative forms (often)Pragmatics Many substrate effects Syntagmatic items (mostly) Weinreich 1953 (mostly) Ross 1996 (mostly) System-affecting features Imposition System-affecting changes (to various degrees) (mostly; often one morph replaces another)

Figure 5.1  Some binary or multiple divisions in the classification of instances of language contact.

116  Anthony P. Grant Contact-induced linguistic transfer ↓ ↓ Borrowing Replication ↓ ↓ Grammatical replication Lexical replication ↓ ↓ ↓ ↓ Contact-induced Restructuring Loss Rearrangement grammaticalization

Figure 5.2  Main types of contact‐induced linguistic transfer (after Heine and Kuteva 2008: 59).

A picture of the kinds of influences which can occur in CILC can be seen in Figure 5.2, taken from Heine and Kuteva (2008, 59), see also Heine and Kuteva, this volume. All of these can be argued to be part of the set of convergence phenomena. This framework is extensive, flexible, and logically rooted but it is not complete – for example, it does not cover instances of phonetic or phonological borrowing except by default (as they occur in instances of restructuring, or as part of the consequences of lexical borrowing). But within its terms of reference it is insightful and brings a degree of welcome order and logical hierarchy to the field. Typological change, and furthermore any syntactic change, is not always the result of CILC, though the wide‐ranging paper by Campbell, Bubenik, and Saxon (1988) presents some robust examples of typological syntactic change through CILC, occurring in several members of families such as Iranian or Ethiosemitic, or in individual languages such as Cappadocian Greek.

2.1  Phonetic and Phonological Convergence Phonetic convergence in the truest sense (that is, contrasting with phonological convergence) can be found in Creek/Muskogee of Oklahoma. Originally the sound written was pronounced as a voiceless bilabial fricative [ɸ], but under the influence of English this has become [f], a labiodental fricative (Haas 1940: 149) – a fact I confirmed in a brief period of fieldwork with the late Margaret Mauldin in 1995. In another case where contact has been intense, Slater (1999) points out that the effect of Sinitic on some languages of the Qinghai Gansu Sprachbund has been so pervasive, as Mangghuer, a Mongolic language, only employs the syllables (onset‐rhyme combinations) which are also available in the local form of Mandarin. Especially interesting evidence of this comes from the Bantu Nguni languages of South Africa, most emphatically Xhosa but also Zulu and Swati, which have absorbed hundreds of words from the Khoe language !Ora or Korana. Many of these have come into Xhosa with an admittedly somewhat reduced but nonetheless still impressive range of the pulmonic ingressive and velaric ingressive consonants which they had in their source language. (However, Korana permits word‐final consonants, which Xhosa does not.) These sounds at least partly entered the Nguni languages through their employment in the hlonipha speech avoidance style in which wives were supposed not to utter the syllables which comprised their husbands’ names, and were to substitute other syllables (Lanham 1964; Ownby 1985; Herbert 1990). For instance Korana /awa ‘to gather together’ is taken into Xhosa as i‐/áwa, written as ‘church; Sunday’. The semantics of this form and its secondary meaning of ‘Sunday’ resemble those of Zulu i‐sonto ‘church; Sunday’, which is also a loan but this time from Afrikaans sondag ‘Sunday’. Xhosa phonology did not absorb syllable‐final consonants of the kinds !Ora possessed, and it retains a basic CV syllable structure. Nor did it take over all the consonants or even all the vowels which Korana has. Nor yet can all words in Nguni which contain velaric ingressive consonants be traced back to Khoisan languages.

Contact and Language Convergence  117 Table 5.1  Some cardinal numerals in Malay, Acehnese, and Tsat. Malay

Acehnese

Tsat

English

satu dua empat lima tujuh sembilan

sa duwa pɯǝt limʌŋ tujoh sikurɯǝŋ

sa33 thua11 pa:Ɂ24 ma33 su55 thua11 pa:nɁ42

‘one’ ‘two’ ‘four’ ‘five’ ‘seven’ ‘nine’

Sometimes the effects of one language system upon another are subtle and not easy to detect. We may think, for instance, of the use of a retroflex postalveolar sibilant in Quapaw, formerly spoken in Oklahoma (Rankin 1987). This instance of phonetic convergence is part of the evidence of Quapaw being influenced by Muskogean and other languages of the Lower Mississippi Sprachbund, a grouping which is discussed in Kaufman (2019) but to which Quapaw does not itself belong. Within Siouan, Quapaw belongs to the Dhegiha subgroup of Mississippi Siouan, together with Kansa, Osage, and Omaha‐Ponca. All these languages have postalveolar sibilants, but only that of Quapaw is retroflex, just as has been recorded for Muskogean Choctaw and Chickasaw, the isolate Tunica, and some other languages in the linguistic area. In another case, Hull (2000) points out that Tetum Dili, the lingua franca of East Timor, has changed its earlier *w to /b/ because it is spoken in an area where the substratum language was Mambae, which lacks /w/. Sometimes the effects are more dramatic: Hollyman (1986) points out that the Polynesian language Fagauvea has more than doubled its stock of consonant phonemes as a result of borrowing heavily from its neighbor, the Oceanic but non‐Polynesian language Iaai. With respect to suprasegmental phenomena, we note the case of Hainan Cham or Tsat, an Austronesian language related to Malay and Acehnese, developing full‐fledged tonality through contact with other languages (Thurgood, Thurgood, and Li 2014). Some cardinal numerals are given as examples (the numerals after the words indicate tone values, see Table 5.1). Tsat has gone even further than Chinese insofar as all Hainan Cham syllables must bear a tone.

2.2  Cases of Morphosyntactic Convergence A number of cases of morphological convergence have been noted in the literature; there are even more observed cases of syntactic convergence. Sometimes the ability to claim that a particular case study instantiates morphological convergence is limited by other considerations. For instance, it is tempting but probably unjustified to attribute the spread in Middle and Modern English of ‐es (Old English ‐as) as a noun plural marker to the influence of Norman French ‐(e)s, which also expressed plurality is nouns (and which, like the Old English suffix, is inherited from Proto‐Indo‐European). Nonetheless Middle English used the affix to pluralize nouns (such as feminine ones, and many masculine and neuter ones) which could not take it in Old English. Modern English has largely shed the grammatical gender which Middle English had on its nouns, and nearly all Present‐Day English nouns which pluralize overtly take ‐(e )s. Cases of morphological convergence nonetheless exist. Dawkins (1916) described a number of Cappadocian Greek dialects, some of which, such as Fertek, remodeled the synthetic Greek

118  Anthony P. Grant Table 5.2  Some case forms in Turkish, Cappadocian Greek, and Standard Modern Greek (Sasse: 1992: 66, some forms slightly adjusted).

sg nom/acc gen pl nom/acc gen

Turkish

Ulağaç

Standard Modern Greek

adam adam‐ın adam‐lar adam‐lar‐ın

átropos atropos‐yu atropos‐ya atropos‐ya‐yu

anθropos anθropu anθropi anθropon

oblique noun pluralization on agglutinative Turkish lines (Matras 2010 illustrates this). The tendency was taken even further in the last stages of Ulağaç Cappadocian Greek: Sasse (1992, 66) gives an example from data he collected, which show that the last speakers of this variety had converged their Greek structure so that it was now isomorphic with its Turkish equivalent: The plural of átropos, formerly something like atropi, is now atropos‐ya, etc. We have now arrived at an exact equivalent of the Turkish declension pattern (Table 5.2). The partial convergence of one language towards another through wholesale inflectional or derivational morphological borrowing may also occur. Seifart has conducted primary research on the Colombian Arawakan language Resígaro, which has absorbed many inflectional morphemes from the unrelated (Witotoan) Bora language, including a whole set of over three dozen noun‐class markers (itself a morphological innovation in the language), but which has borrowed very little lexicon from Bora or for that matter from Spanish (see Seifart 2015). An example of syntactic convergence or transfer of pattern, taken from Taylor (1977, 5), will demonstrate what a change in element ordering or in the sequence of morphemes looks like within a construction. As far as we can tell from our rather sparse structural, textual, and syntactic records of what is the now extinct Dominican variety of Island Carib (which is actually not a Cariban language at all, but an Arawakan one, and its modern form is generally referred to as Garifuna) it used the element order Demonstrative plus Noun (Rat 1898). The Central American correlate of this language, which is used in Belize, Honduras, and Guatemala and which was formerly also spoken in eastern coastal Nicaragua, uses Noun plus Demonstrative. A pair of parallel examples follows. (1) Thukúra hįáru (Dominican Island Carib) this woman ‘this woman’ (2) Hįáru tugúra (Belizean Garifuna) woman this ‘this woman’ Dominican Garifuna preserves the element order which was recorded for seventeenth‐ century Island Carib. The change in ordering in Belizean Garifuna must have come as the result of intense pressure from (and generations of increasing bilingualism in, and domination by) Dominican Creole French (which has the same element order as Dominica Garfuna had. This creole probably acquired it under the influence of those African languages, such as Kwa languages, where demonstrative adjectives are postposed to their nouns). (3) fãm sa‐a (Dominican Creole French) woman this‐DEF ‘this woman’

Contact and Language Convergence  119 The actual forms of the demonstrative adjectives which were used in the creole were not transferred from there into that variety of Island Carib, which had retained demonstratives of its own, but their element order was. What makes this all the more strange is that it was the Dominican form of Island Carib, which was exposed to (and which was ultimately replaced by) Dominican Creole French and which preserved the original element order, while Central American Island Carib or Belizean, Honduran and Guatemalan Garifuna lost this order. That language has for the most part had little contact with Creole French since the 1790s, calques the Creole French order. An example of morphosyntactic convergence which also illustrates the effects of borrowing and the ways in which borrowed elements can develop new roles in the structure of their recipient languages comes from Uto‐Aztecan. San Pedro Jícora Nawa of Durango, Mexico, hereafter SPJ, has been extremely influenced by Spanish over the centuries, to the extent that the typically Mesoamerican NP‐NP possessive construction which is customarily found in other Nawa varieties has largely been replaced by one which incorporates the use of Spanish de with a Nawa article. It has gone further down this path than some other Nawa varieties. In Michoacán Nahuatl, Spanish de is optionally used in NP‐NP constructions, though not when the possessor is a proper name: (4a–b) i‐tsotsomahli (de) mo‐tah‐tsin / 3SG‐clothes of your‐father‐HONORIFIC / ‘your father’s clothes’ /

mo‐tah‐tsin your father‐HON ‘your father’s clothes’

i‐tsotsomahli 3SG‐clothes

(5) i‐lahketili de Rikarda 3SG‐loom of Ricarda ‘Ricarda’s loom’ (Sischo 1979: 341). Both Mexicanero and Hidalgo varieties of Nawa, here illustrated from questionnaires from the Archivo de Lenguas Indígenas de México (responses to questions 512–21), preserve affixal personal possessive markers. Now Hidalgo Nawa sometimes uses Spanish de to reinforce or double‐mark the construction: (6) N i‐tlapachih‐ka de n i‐kal o‐ Ø‐ The 3SG‐roof‐ABS   of the 3SG‐house  pret‐ Ø‐ The 3SG‐cover‐Nominalization    of the 3SG‐house ‘The roof of the house fell in.’ (Lastra 1980: 119)

wets fall

But generally it does not: (7) N i‐kawayo in xwan o‐ Ø‐ The 3SG‐horse the John   PRET‐3SG‐ ‘John’s horse died’.

mik died  (Lastra 1980: 118)

Possessive constructions involving two full NPs (rather than NP de NP constructions using borrowed Spanish de where this has the sense of ‘made of’, and where one or both nouns may but need not be of Spanish origin) are rather infrequent in the texts in Preuss ed. Ziehm (1967–76). Such examples as I have found use the traditional Mesoamerican construction. Modern Mexicanero uses de more consistently for this kind of construction(and for others,

120  Anthony P. Grant including sometimes the marking of relative clauses in a way not licensed by Spanish). Surface phonemic transcriptions are followed by morphophonemic analyses: (8) i‐techo de in kal uwets(i) i‐techo de in kal u‐wets(i) 3SG‐roof of the house fell in (9) Umih u‐Ø‐miki PRET‐3SG‐died

ni‐kabayo de in xwan. ni‐kabayo de in xwan 3SG‐horse of the John died (Canger 2001: 151)

Canger’s example 512 is an exception to this principle and is a rare example of the construction of NP‐NP through recursion: (10a, b)  Mexicanero: i‐tenko in komál tapan‐ki Hidalgo: i‐tenko in komal‐e, λapan‐ki. (‐e marks absolutive) (Lastra 1980: 118) Both: 3SG‐rim the griddle it.broke‐adjectivizer. ‘The rim of the griddle is broken’. (Canger 2001: 151; these instantiate example 512) Multiple or recursive possession is illustrated in Canger (2001: 151, example 516), and Lastra (1980). Two variants are offered, with glosses (morphophonemic forms are same as surface forms here): (11a, b) In techo de in kal de no‐ta‐tsi u‐Ø‐wetsi The roof of the house of 1SG‐father‐DIM PRET‐3SG‐fall In techo de nikal de notatsi u‐Ø‐wetsi The roof of 1SG‐house of 1SG‐father‐DIM PRET‐3SG‐fall. ‘The roof of my father’s house fell in.’ Note also the use in Mexicanero of a Spanish borrowing for ‘roof’. But the material for Hidalgo Nawa offers one form for this construction: (12) N i‐tlapachih‐ka de n  i‐kal   no‐tata o‐Ø‐ wets of the   3SG‐house  1SG‐father PRET‐Ø‐ fall The 3SG‐roof‐ABS ‘The roof of my father’s house fell in.’ (Lastra 1980: 118; in both cases example 515) Convergence can lead to divergence when the language which absorbs an element begins to use it in ways not found in the source language. Chamorro borrowed the preposition para ‘for ’ from Spanish, but has extended its use into a new role as a future marker: (13) Para hu hanao asta gi i gipot. FUT I go up.to at the party ‘I’m going to go to the party.’ (Topping, Ogo, and Dungca 1975: 163). But free grammatical/functional elements can also be borrowed with their initial purpose intact. An example is don, borrowed as a completive marker from Western Caribbean Creole

Contact and Language Convergence  121 English into the heavily anglicized Pearl Lagoon Basin Miskitu language of eastern coastal Nicaragua: (14) Yang don  tak‐ri  inska ba klin mun‐i I CMPLT emerge‐past fish the clean cause‐verbal‐noun. ‘I’ve finished cleaning the fish.’ (Mark Jamieson, personal communication.) Sometimes convergence actually consists of compromise: speakers of two languages share an available but not dominant exponence of an important linguistic feature, such as word order whereas their dominant word orders differ. Consequently the second favorite wins out in the language which emerges (this is the case for instance in some syntactic features in Berbice Dutch Creole, where OV order was the second favorite in both Eastern Ijo and Dutch and came to dominate in Berbice Dutch Creole: Kouwenberg 1992). An excellent treatment of a case of syntactic convergence with ample illustrations is by Nadkarni (1975).

2.3  Semantic and Pragmatic Convergence Convergence of semantic properties is not uncommon; the structures of a language’s semantic field may be rearranged as the result of contact with speakers of another language. One such example is that of the Middle and Modern English kinship system. The distinctions in Old English between ‘father’s brother’ and ‘mother’s brother’ on the one hand, and between ‘mother’s sister’ and ‘father’s sister’ on the other, which were lexicalized in Old English, were lost when English took over uncle, aunt from French, where such a distinction between types of uncles and aunts was not made. English semantics thus converged slightly towards that of French. Nonetheless, although the use in English of compound terms for relations by marriage echoes those of French, the use of the affix –‘in‐law’ is derived neither from French by calquing nor by borrowing the idiom from Norse, the source of the word ‘law’; the construction is an internal development within English. Here, as elsewhere, borrowed items can develop a life of their own. Another example of subtlety in the transfer of a feature is taken from a form of South Siberian Turkic, languages which have in the past couple of centuries been strongly influenced by Russian which is now replacing them. Gregory Anderson noted that in his fieldwork he found speakers of the Khakas language who had been born in the 1970s. In their speech they used the genitive (rather than the traditional Turkic ablative) with their verb ‘to fear’, on the model of Russian bojat’sja (Anderson 2005: 180–1). Other features of verb argumentation can be transferred from one language to another: an example is the form for ‘to play’, in German spielen but in standard Yiddish shpiln zikh, the verb being made reflexive on the lines of coterritorial Polish bawić się (Lockwood 1965: 244). An interesting case of semantic convergence without extensive borrowing is that of the last speakers of Natchez in Mississippi, members of a tribe who were defeated militarily by the French in 1729 and then dispersed widely (some into slavery in the West Indies). Despite the fact that all the last few speakers of Natchez, the last of whom died in the 1950s, were multilingual (in Natchez, Creek, Cherokee, and latterly English), lexical items from other languages account for less than 1% of the recorded vocabulary of about 4,000 items. (Kimball 2003 describes Natchez structure.) There are a number of calques, however, which as instances of loan‐translation exemplify semantic convergence of a less prestigious language with a more prestigious one: Natchez ɁayaM with a word‐final voiceless nasal means both ‘bedbug’ and ‘ax’ on the model of the more prestigious Cherokee galuysdi. (Natchez data are from Mary Haas’ fieldnotes, now housed in the American Philosophical Society, Philadelphia; Cherokee form is from Feeling and Pulte 1975.)

122  Anthony P. Grant An instance where English has influenced a language through extensive calquing is that of Kashaya, or Southwestern Pomo, in north‐central California. Kashaya contains hundreds of loans from Spanish, a couple of dozen more which ultimately come from Russian (via Eskimoan or Athabaskan languages spoken by Alaskan workers who were involved in the Russian fur trade in that area), but not many from English. Convergence here has been through loan‐translation from English into Kashaya; Oswalt lists a number of instances of one such case, involving the calquing of Kashaya mo:‐ as a means of replicating uses of English ‘run’ in phrases and idioms where such usage would once have been alien to Kashaya: Oswalt (1985). Instances of speaker of a language taking over pragmatic techniques from a more prestigious language system are legion, especially if the speech act in question relates to a new medium, such as the openings and valedictions in letter‐writing, or the structuring of a sermon. An instance from spoken languages will instantiate pragmatic convergence. Terms for the polite expression ‘please!’ used to support a request can involve convergence as well as borrowing in European languages: Albanian ju lutem, Estonian palun, Latvian lū dzu, Polish proszę all mean ‘I ask (you)’ (often originally ‘I pray’ or ‘I implore’), cf. German bitte, Serbian molim, Greek parakaló. On the other hand, there are calques on French in English please/if you please (where please is itself a loanword from French), Dutch als het U belieft (and Afrikaans asseblief) and Lëtzebuergesch wann Iech gelifft.

2.4  The Issue of Metatypy Mixing of elements and the transfer of patterns of many kinds but with little fabric transfer, occurs in a number of cases, and this has come to be known as metatypy (Ross 1996). A well‐ documented instance is Takia, a North New Guinea Oceanic language (a member of the Bel group) of the Austronesian family. Takia semantics, clausal syntactic and morphological typology and idiomatic structure have been heavily influenced and reshaped by such processes. As a result they resemble those of its erstwhile non‐Austronesian neighbor on Karkar Island off the North New Guinea coast, namely Waskia. This is a language in which many speakers of Takia were once bilingual but from which it has borrowed practically no morphemes of any kind. However, Takia you, the normal word for ‘water’, derives from Waskia or Bargam, another non‐Austronesian language. The phrasal, clausal and sentential syntax of Takia has been changed over the centuries to resemble that of the formerly more prestigious Waskia. Many phrases, idioms, and even sentences occurring in Takia are directly translatable into Waskia simply by substituting one set of morphemes for another without changing the order in which they occur. Takia is not a mixed language or a creole (Ross 1996), and it is clear that its closest relatives are the other members of the Bel group of North New Guinea languages such as Gedaged, a language used in local missions, and one from which Takia has also borrowed a considerable amount of lexicon. Some examples of Takia being metatypized towards Waskia syntactic structures (taken from Ross 1996: 191) are given below. (15a–d) Takia: Waskia: Takia: Waskia:

Waskia tamol an Waskia kadi mu Waskia man DETERMINER ‘the Waskia man’ tamol tubun uraru en kadi bi‐biga itelala pamu man (PLURAL‐) big two this ‘these two big men’

Contact and Language Convergence  123 Takia: Waskia:

Kai sa‐n ab Kai ko kawam Kai CLASSIFIER‐his house POSTPOSITION ‘Kai’s house’

Takia: Waskia:

yai tamol an ida ane kadi mu ili I man DETERMINER with‐him ‘the man and I’

Further research by Malcolm Ross on language contact involving Takia and related and neighboring languages has made several important points clear. First, the Bel languages as a group are certainly not immune to borrowing lexicon from other Austronesian languages or from non‐Austronesian languages and to being structurally influenced by these languages – Dami, for instance, has borrowed lexicon much more heavily from non‐Austronesian languages. Second, Waskia has borrowed more lexicon from Oceanic languages (including maybe from Takia and certainly earlier stages of Oceanic languages) than the Bel languages have taken in lexicon from Waskia. Third, Takia has borrowed from other languages (from German, from Takia’s close relative Gedaged, and latterly from Tok Pisin) as much or more lexicon than it has taken from Waskia. Ross also notes that Takia also took from Bargam or Mugil, another non‐Austronesian language of the north coast of New Guinea, more lexicon than it has borrowed from Waskia, but Bargam has not shaped Takia’s semantics or morphosyntactic typology as much as Waskia has done. It is likely that the greatest number of lexical resemblances between Waskia and Takia are the numbers of Tok Pisin loans which they share.

2.5 A Second Kind of Convergence: Two Become One. Convergence as Language Creation Some languages appear to have developed from the continuous convergence of two or more languages, in such a way that not even the inflectional morphological system can be traced to a single ancestral language. Generally we know far less about the paths of development of these languages than we wish, with the result that the ways in which they emerged will remain uncertain and the province of guesswork. Some of them such as Stedsk/Town Frisian (van Bree 1994) combine elements from languages which were already very closely related. Many languages which have been listed together with pidgins or creoles, for example, may be better classified as being languages of other kinds. For example, Europe has hosted a number of Para‐Romani varieties. These combine quantities of Romani lexicon (including some function words) with a structurally regularized version of the morphosyntax (and generally also the segmental and canonical phonology) of the local language. Their lexica often also contain words of unknown origin, some of them probably artificially created, but some others traceable to non‐Romani slangs. Some of the interest which these languages have garnered results from them often being the only attested Romani linguistic material from a particular area. It must be noted that Romani varieties which have preserved the original Indic‐derived morphology are themselves etymologically diverse in their lexica. Even though over 75% and almost all the productive inflectional morphology of these Romani varieties derives from Indic, they all contain a large number of loans from Iranian (mostly Farsi), Armenian, Greek, and South Slavic, and often also from other languages with which speakers of Romani were formerly coterritorial, such as Romanian and German. Smart and Crofton (1873) and latterly Matras (2010; Angloromani), Krinková (2015; Iberian Romani), and Carling, Lindell, and Ambrazaitis (2014; Scandoromani) are good guides to some

124  Anthony P. Grant Para‐Romani varieties. Bakker and Cortiade 1991 is a general account of Para‐Romani varieties, including details of some languages which are not listed here. There is a small group of Berber‐Songhay mixed languages (so‐called Northern Songhay languages) which are spoken in the Azawagh Valley in Mali and Niger (with an outlier, Korandje, being spoken in the village of Tabelbala in western Algeria). The basic morphology and a few hundred words of the high‐frequency lexicon of these languages derive from Songhay languages, while some less frequent structural features and the bulk of the everyday lexicon derives from Berber (Amazigh) languages, not merely from the coterritorial Tuareg varieties, with further lexical strata from Fulfulde, French, and especially Arabic. These are what we may call Core‐Periphery mixed languages, with a core of morphemes (lexical and grammatical) deriving from one source and a larger and more peripheral body of material from another. What distinguishes these from other languages which have borrowed heavily (such as English) is the thinness of the core. At most 300 morphs (lexical or structural) of Sonmghay origin have been recorded in Tadaksahak. Benítez‐Torres and Grant (2017) provide details and a bibliography. Tojol‐ab’al or Tojolabal (‘correct‐language’) of Chiapas, Mexico, formerly often known as Chaneabal, is such a case. It is unmistakeably a Mayan language, but dispute has simmered as to which of the five major branches of Mayan it belongs (the assertion that it is Tzeltalan in Dienhart 1989 is disputed by Campbell 1990). I do not claim to have a definitive answer to Danny Law’s question (extensively explored in Law 2017, see also Law, this volume) as to whether Tojol‐ab’al is a Tzeltalan or Chujean Mayan language, or to unravel how it came about. However, we observe from Law’s analysis that most of the personal possessive prefixes on nouns are Tzeltalan. Some of the basic vocabulary is also distinctly Tzeltalan rather than pan‐Mayan or Chujean (though much is exclusive to Tojol‐ab’al, much else is typically Chujean ‐ indeed Tojol‐ab’al is closer to Chuj with respect to basic vocabulary – and some is pan‐Mayan or largely so, and of course the language contains numerous loans from Spanish). Law (2017) points out on the basis of a 1200‐lexical item sample that overall Tojol‐ab’al lexicon shows more similarities with Tzeltal than with Chuj, however. Law (2017) shows that among the elements which Tojol‐ab’al shared either with Tzeltalan or with Chujean (Q’anjob’alan) but not with both, some forms are clearly Tzeltalan because they are innovations rather than retentions there. These include the marker for the agentive noun, the potential completive marker, the negative existential form, the 1pl inclusive personal suffix, and the corresponding 1pl exclusive suffix, the phonologically identical incompletive intransitive and transitive markers, the antipassive suffix, the noun plural marker (and the form used in pluralizing adjectives), the plural imperative marker and the irrealis suffix. On the other hand, some are by the same measure clearly Chuj or Chujean‐Q’anjob’alan because they have been innovated in those languages too and are also found in Tojol‐ab’al: the general preposition, the future potential affix, the progressive auxiliary suffix, the nominal plural for humans, the 2pl suffix, the interrogative particle, and the forms of the directionals. The problems involved in assigning Tojol‐ab’al to a particular branch of Mayan are not born of a dearth of data on it or on its potential closest relatives, all of which are richly documented. One problem is the assignment of relative weights to particular features as being diagnostic f membership in one or another branch of Mayan. There are plural person markers from Tzeltalan and from Chujean. The antipassive particle is Tzeltalan in form, mediopassive is Chujean, as is the progressive marker. Unless we have seriously misunderstood the development of Mayan languages, either the exclusively Tzeltalan or the exclusively Chujean elements (probably the former set) must have been borrowed into Tojol‐ab’al, and this indicates that any kind of features can be transferred between typologically similar languages which share a common ancestor even if they are in no way mutually intelligible.

Contact and Language Convergence  125

3  A Note on Linguistic Areas Ever since the pioneering work of Jernej Kopitar on structural similarities among some of the languages in the Balkans (Kopitar 1829), linguists have been interested in bundles of structural and/or lexical phenomena which cross genealogical boundaries within a particular geographical or cultural area. These features are held to be shared among some sets of languages which belong to more than one family (some of which may be isolates), yet which do not characterize all the languages in all the genealogical groups which exhibit these features. Some members of at least one family whose members participate in this may lack the features in question. Such collections of shared features are known as linguistic areas or Sprachbünde or more recently as convergence areas (Hickey 2017). Linguistic areas, which are built up out of shared phonetic, phonological, morphosyntactic, semantic, and pragmatic features, often but not always accompanied by lexical borrowing, are dynamic rather than static. They can change their features and arguably their membership over time – and, like the Clear Lake linguistic area in north‐central California (Sherzer and Bauman 1972), they can come to an end when their component languages become dormant. Indeed some linguists (notably Campbell 2017) have observed that it can be difficult to define them satisfactorily. Matras, McMahon, and Vincent (2006) and especially Hickey (ed. 2017) are valuable introductions to this subject and contain numerous case studies of linguistic areas from around the world, while Masica (1976) gives a good account of problems inherent in defining a linguistic area (in this case the Indian subcontinent). Emeneau (1956) is a classic study of India as a linguistic area, while Campbell, Smith‐ Stark, and Kaufman (1986) is a detailed account of the Mesoamerican Sprachbund. Friedman (2017) is the most recent account of the Balkan linguistic area. Romani Tosk Albanian Greek Bulgarian Macedonian Torlak Serbian Romanian Aromunian Megleno-Romanian East Rumelian Tk.

mangav dua θeló iskam sakam oču vreau voi voi isterim

te džav të shkoj na paó da otida da odam da idem să merg s(i) njergu să/si merg gideyim

Figure 5.3  “I want to go” in languages of the Balkan Sprachbund (Friedman 2017). Note that Slavic or Turkic languages outside this area use the infinitive rather than subjunctive constructions in this phrase: Russian xoču idti, Turkish gitmek isterim.

Figure  5.3 gives a taste of one feature from the Balkan Sprachbund. Even then, there are caveats: Geg Albanian does not employ this construction, while East Rumelian Turkish does not use a subordinating conjunction here. As for Romani, the construction depicted here for a Balkan Romani lect has been recorded in Romani dialects as far as Wales and is common to most of them. Areality comes with provisos.

4 Conclusions Linguistic convergence can be seen as both a process – the transmission of features from one language to another through copying or replication (for which “borrowing” is a well‐ established but rather unsatisfactory term) and the result of the cumulation of such

126  Anthony P. Grant transfers. Unsurprisingly, it is easier for a feature to pass into one language from another if the two are typologically similar (whether or not they are genealogically related to one another) than it is if they are very different typologically (even if they can be shown to descend from the same proto‐language and thus be genealogically interrelated). Anyone who thinks that the typological patterning of structures within a language family is always uniform should bear the following fact in mind. No fewer than seven of the 17 actually attested orders which combine various exponents of Subject‐Verb‐Object, Genitive‐Noun, Noun‐Adjective and Preposition/Postposition out of the 24 theoretically possible ones discussed in Greenberg (1963), which were reviewed by Campbell, Bubenik, and Saxon (1988), can be found among members of the Indo‐European language family. Meanwhile even a casual comparison of the typological profiles of English and Hindi, at phonological, morphological, syntactic and semantic levels, will show how much two languages which ultimately derive from the same proto‐language can diverge in the space of a few millennia. The bigger the language family is, the greater is the potential for typological diversity. (This potential need not be realized, of course.) Convergence is a term which means different things to different linguists; even shorn of its connections with lexical borrowing and realigned to considerations of pattern transfer between language systems, it shows that the limits on such transfer are few, if any. As some of the cases in this chapter have shown, it is also clear from a study of issues in linguistic convergence that features or elements which have been transmitted into another language through partial convergence, once assimilated, may go on to play roles in their recipient language which are unknown in the source language. It is not easy to remove borrowing or indeed grammaticalization from the picture of convergence as all these intertwine (see Heine and Kuteva, this volume). And as the case of Tojol‐ab’al demonstrates, determining what is inherited and what is borrowed is rarely as easy as one may hope.

REFERENCES Adelaar, Willem 2012. Modeling convergence: towards a re‐examination of the history of Quechuan‐Aymaran interaction. Lingua 122: 461–69. Aikhenvald, Alexandra 2002. Language Contact in Amazonia. Oxford: Oxford University Press. Anderson, Gregory D. 2005. Language Contact in South Central Siberia. Berlin: Harrassowitz. Bakker, P. and M. Cortiade 1991. In the Margin of Romani. Gypsy Languages in Contact. Amsterdam: Publikaties van het Instituut voor Algemene Taalwetenschap. Benítez‐Torres, Carlos M. and Anthony Grant 2017. On the origins of some Northern Songhay mixed languages. Journal of Pidgin and Creole Languages 32: 263–303. Bloomfield, Leonard 1933. Language. New York: Henry Holt. Braine, Martin 1963. The ontogeny of English phrase structure: the first phase. Language 39: 1–13. Campbell, Lyle 2017. Why is it so hard to define a linguistic area? In Raymond Hickey (ed.),

Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 19–39. Campbell, Lyle, Vít Bubenik, and Leslie Saxon 1988. Word order universals: refinements and clarifications. Canadian Journal of Linguistics/ Revue canadienne de linguistique 33: 209–230. Campbell, Lyle, Thomas Smith‐Stark, and Terrence Kaufman 1986. Mesoamerica as a linguistic area. Language 62: 530–70. Campbell, Lyle 1990. Review of Diehart 1989. Language 66: 865–6. Canger, Una 2001. Mexicanero de la Sierra Madre occidental. Mexico City: El Colegio de México. Carling, Gerd, Lenny Lindell, and Gilbert Ambrazaitis 2014. Scandoromani: Remnants of a Mixed Language. Leiden: Brill. Dawkins, Richard M. 1916. Modern Greek in Asia Minor. Oxford: Oxford University Press. Dienhart, John M. 1989. The Mayan Languages: A Comparative Vocabulary. Odense: Odense University Press. Emeneau, Murray 1956. India as a linguistic area. Language 32: 1–16.

Contact and Language Convergence  127 Feeling, Durbin and William Pulte 1975. Cherokee‐English Dictionary. Tahlequah: Cherokee Tribe of Oklahoma. Friedman, Viktor A. 2017. Languages of the Balkans. Oxford Online Encyclopedia of Linguistics. Oxford: Oxford University Press. Grant, Anthony P. 2002. Fabric, pattern, shift and diffusion: what changes in Oregon Penutian languages can tell historical linguistics. In Laura Buszard‐Welcher (ed.), Proceedings of the Meeting of the Hokan‐Penutian Workshop June 17–18, 2000, University of California at Berkeley. Berkeley: Survey of California and Other Indian Languages 2002, pp. 33–56. Grant, Anthony P. 2003. Review of King (2000). Word 54: 251–56. Greenberg Joseph H. 1963. Some universals of grammar with particular reference to the order of meaningful elements. In Joseph H. Greenberg (ed.), Universals of Language 2. New York: Academic Press, pp. 73–113. Haas, Mary R. 1940. Ablaut and its function in Muskogee. Language 16: 141–50. Haase, Martin 1992. Sprachkontakt und Sprachwandel im Baskenland: die Einflüsse des Gaskognischen und Französischen auf das Baskische. Hamburg: Buske. Haugen, Einar 1950. On the analysis of lexical borrowing. Language 26: 240–61. Heath, Jeffrey 1984. Language contact and language change. Annual Review of Anthropology 13: 367–84. Herbert, Robert K. 1990. The sociohistory of clicks in Southern Bantu. Anthropological Linguistics 34: 295–315. Heine, Bernd and Tania Kuteva 2008. Constraints on contact‐induced linguistic change. Journal of Language Contact THEMA 2, 57–90. Hickey, Raymond 2017. Areas, areal features and areality. In Raymond Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 1–15. Hickey, Raymond (ed.) 2011. Blackwell Handbook of Language Contact. Oxford: Blackwell. Hickey, Raymond (ed.) 2017. Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press. Hollyman, Kenneth J. 1986. De muna fagauvea: dictionnaire fagauvea‐français. Auckland: Te Reo Monographs. Hull, Geoffrey 2000. The historical phonology of Tetum. Studies in the Languages and Cultures of East Timor 3: 158–212. Kaufman, David 2019. The Lower Mississippi as a linguistic area. In Anthony P. Grant (ed.), The

Oxford Handbook of Language Contact. Oxford: Oxford University Press. Kimball, Geoffrey 2003. Natchez. In Heather Hardy and Janine Scancarelli (eds.), Native languages of the Southeastern United States. Lincoln: University of Nebraska Press, pp. 385–453. Kopitar, Jernej 1829. Albanische, walachische und bulgarische Sprache. Jahrbücher der Literatur (Vienna) 46: 59–106. Kouwenberg, Silvia 1992. From OV to VO: linguistic negotiation in the development of Berbice Dutch Creole. Lingua 88: 163–299. Krinková, Zuzana 2015. From Iberian Romani to Iberian Para‐Romani Varieties. Prague: Karolinum. Lanham, Leonard W. 1964. The proliferation and extension of Bantu phonemic systems influenced by Bushman and Hottentot. In Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: de Gruyter Mouton, pp. 382–91. Lastra, Yolanda 1980. El náhuatl de Acaxochitlán, Hidalgo. Mexico City: El Colegio de México. Law, Danny 2017. Language mixing and genetic similarity: the case of Tojol‐ab’al. Diachronica 34.1: 40–78. Lockwood, W. B. 1965. An Informal History of the German Language. London: Heffer. Lucas, Christopher 2015 Contact‐induced language change. The Routledge Handbook of Historical Linguistics, edited by Claire Bowern and Bethwyn Evans. London: Routledge, pp. 519–36. Masica, Colin 1976. Defining a Linguistic Area: South Asia. Chicago: University of Chicago Press. Matras, Yaron 2010. Romani in Britain: The Afterlife of a Language. Edinburgh: Edinburgh University Press. Matras, Yaron, April M. S. McMahon, and Vincent, Nigel (eds.) 2006. Linguistic Areas. Convergence in Historical and Typological Perspective. New York: Palgrave Macmillan. Matthews, William K. 1956. The Latvian element in modern Livonian. In Margarethe Woltner and Herbert Bräuer (eds.), Festschrift für Max Vasmer zum 65. Geburtstag. Wiesbaden: Otto Harrassowitz, pp. 307–18. Nadkarni, Mangesh V. 1975. Bilingualism and syntactic change in Konkani. Language 51: 672–83. Nau, Nicole 1995. Möglichkeiten und Mechanismen kontaktbewegten Sprachwandels. Unter besonderer Berücksichtigung des Finnischen. Munich: Lincom Europa.

128  Anthony P. Grant Oswalt, Robert 1985. The infiltration of English into Indian. International Journal of American Linguistics 51: 527–9. Ownby, Carolan Postma 1985. Early Nguni history: the linguistic evidence and its correlation with archaeology and oral tradition (University of California, Los Angeles Ph.D. thesis). Preuss, Konrad T. 1967–76. Nahua‐Texte aus San Pedro Jicora in Durango, edited by Elsa Ziehm. Berlin: Gebrüder Mann. Rankin, Robert A. 1987. Quapaw: areal and genetic affiliations. In William F. Shipley (ed.). In Honor of Mary Haas. Berlin: de Gruyter Mouton, pp. 629–50. Rat, Joseph Numa 1898 The Carib Language as Now Spoken in Dominica, West Indies Journal of the Anthropological Institute of Great Britain and Ireland, 27: 293–315. Ross, Malcolm 1996. Contact‐induced change and the Comparative Method. In Mark Durie and Malcolm Ross (eds.), The Comparative Method Reviewed. Oxford: Oxford University Press, pp. 180–217. Sasse, Hans‐Jürgen 1985. Sprachkontakt und Sprachwandel: die Gräzisierung der albanischen Mundarten Griechenlands. Papiere zur Linguistik 32: 37–95. Sasse, Hans‐Jürgen 1992. Language decay and contact‐induced change: similarities and differences. In Matthias Brenzinger (ed.). Language Death: Factual and Theoretical Explorations with Special Reference to East Africa. Berlin: De Gruyter, pp. 59–89. Seifart, Frank 2015. Tracing social history from linguistic and ethnographic data: the prehistory of Resígaro contact with Bora. Mundo Amazónico 6: 97–110. Sherzer, Joel and Richard Bauman 1972. Areal studies and culture history: language as a

key to the study of historical culture contact. Southwestern Journal of Anthropology 28: 131–52. Sischo, William 1979. Michoacán Nahuatl. In Ronald A. Langacker (ed.), Uto‐Aztecan Grammatical Sketches volume 2: Modern Aztec Grammatical Sketches. Dallas: Summer Institute of Linguistics, pp. 307–80. Slater, Keith W. 1999. Language mixing in China’s Qinghai‐Gansu Sprachbund. Paper presented at the Aarhus Mixed Languages Workshop, May 1999. Smart, Bath C. and Henry T. Crofton. 1873. The Dialect of the English Gipsies. London: Asher. Soper, John David 1987. Loan syntax in Turkic and Iranian (University of California at Los Angeles Ph.D. dissertation). Taylor, Douglas R. 1977. Languages of the West Indies. Baltimore: Johns Hopkins University Press. Thurgood, Graham, Ela Thurgood, and Fengxiang Li. 2014. A Grammatical Sketch and Phonology of Hainan Cham. Berlin: de Gruyter Mouton. Topping, Donald N., Pedro M. Ogo, and Bernadita C. Dungca. 1975. Chamorro‐English Dictionary. Honolulu: University Press of Hawai’i. van Bree, Cor. 1994. The development of so‐ called Town Frisian. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: 14 Studies in Language Intertwining. Amsterdam: IFOTT, pp. 69–82. van Coetsem, Frans 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter. Weinreich, Uriel. 1953. Languages in Contact: Findings and Problems. The Hague: de Gruyter Mouton.

6 Contact and Linguistic Typology OLIVER BOND, HELEN SIMS‐WILLIAMS, AND MATTHEW BAERMAN 1 Introduction As an enterprise, linguistic typology is perhaps best understood in terms of the shared ­ideology of its practitioners. For instance, Nichols (2007) proposes that despite having the trappings of a sub‐discipline, what actually connects those who identify as typologists is not a unified approach to the investigation of language, but a set of traits and experiences that inform their work. She argues that while typologists come in different flavors, they are likely to have conducted fieldwork of some sort, work with grammars (and possibly archives), and be empirically oriented. Typologists are interested in capturing diversity across languages, and explaining this variation.1 Given that researchers working on language contact ­phenomena frequently share a similar approach, it is natural therefore that their research should often have a typological flavor. While typologists seek to understand why our current linguistic landscape looks the way it does, work on language contact is concerned with a very specific part of this endeavor, namely, what the interplay between languages in contact can reveal about the structure of linguistic systems and the patterns of usage that lead to their convergence. The relationship between contact linguistics and typology has evolved as the two fields have matured. By the mid‐twentieth century, typology was largely concerned with what is possible in language, that is, discovering what the universals of language might be. This spawned a large body of work attempting to establish probabilistic statements about the relationships between variables observed across languages (e.g. word order typology, including Greenberg 1963; Hawkins 1983; Tomlin 1985; Dryer 1992). In part, typology’s agenda was to classify languages into types based on the presence or absence of a particular property (so‐called “holistic” typology). Yet its overarching aim was to reveal which ­patterns across languages could be attributable to genetic relationships and which are cross‐ linguistically observable, and potentially universal. Much important work on language contact being conducted at the same time focused on the structural parameters of contact – explaining the grammatical prerequisites for borrowing, code‐switching, and the diffusion of structural features (Weinreich 1953; Moravcsik 1978; Poplack 1980; Thomason and Kaufman 1988). Work in defining areas of linguistic convergence and the evidence for them has been at the core of research in both fields.2 For instance, this is seen in the influential work of Dryer (1989) and Nichols (1992) in explaining and accounting for the influence of areality (see contributions in Hickey ed., 2017) in the distribution of typological variables. Taking areal effects

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

130  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman seriously in the discipline led to such works as the World Atlas of Language Structures and WALS Online (Haspelmath, Dryer, Gil, and Comrie 2005; Dryer and Haspelmath 2013) which maps different language characteristics in a diverse sample of languages (wals.info) and the Atlas of Pidgin and Creole Language Structures (APiCS) and APiCS Online (https://apics‐ online.info/) which charts variables in languages known to have formed through intense contact (Michaelis, Maurer, Haspelmath, and Huber 2013a, 2013b). Contemporary typology does not simply ask “what is possible?” or “what dependencies exist,” but examines “what is where why?”, with reference to historical and other factors affecting the distribution of language properties (Bickel 2007). Language contact is thus a major explanatory device of typology and understanding the ways in which languages with different typological profiles are affected by contact‐induced change is a shared goal. The extent to which a language shows contact effects that influence its typological profile depends on a large number of factors, such as the geographic and social isolation of a speech community, migration and trading patterns, the types of societies that people live in (i.e. the types of cultural practices that facilitate or prohibit contact) and the extent to which multilingualism prevails. While these parameters may well have been ignored in much mainstream linguistic work of the previous century in favor of an oversimplified but more tractable approach to language (monolingualism over multilingualism, “competence” over “performance” and genetic inheritance over alternative explanations), language contact as an explanation for linguistic configurations has a natural place within contemporary typology, alongside language‐internal change and “cognitive pressures” on the shape of language (Muysken 2016: 538–9). Change though language contact and hybridity processes are increasingly relevant in building theoretical models of language structure (e.g. Aboh 2015). In studying contact‐induced change (Winford 2005; Kossmann 2015) typologists can begin to understand which features of language are readily acquired or arise through contact. Similarly, it allows us to investigate whether contact‐induced loss of a linguistic property follows the same pathways as regular internal change (Sims‐Williams, forthcoming). Rather than provide an overview the history of the relationship between typology and language contact, we focus here on recent developments in research on language contact in relation to contemporary thought in linguistic typology. While the typological approach to linguistic research is amenable to any sub‐field of linguistics, we focus here on morphological typology (Section 2) and examine some of the ways in which our understanding of the diversity of morphological systems can help provide support for contact‐based explanations of linguistic variables (Sections 3–6). Conclusions arising from these considerations are presented in Section 7.

2  Contact and Morphological Typology While typology in general is concerned with identifying sets of variables (i.e. parameters by which languages vary) and developing probabilistic theories explaining their distribution (Bickel 2007: 242), morphological typology relates these goals to the shape, properties, and distribution of morphological systems. In its perhaps most familiar form, dating back to the early nineteenth century, morphological typology is portrayed as classifying languages into types according to two main parameters: (i) whether a language is largely isolating or synthetic in its expression of grammatical categories; and (ii) whether those categories are expressed agglutinatively or fusionaly. While these are useful broad labels for describing general tendencies in the morphological makeup of a language, contemporary morphological typology is considerably more sophisticated, and like the rest of typology, is interested in revealing fine‐grained variation and empirically motivated explanations, specifically for the organizational principles underlying inflectional systems and patterns of word‐formation.

Contact and Linguistic Typology  131 One aspect of this endeavor is to discover the means of morphological expression available in languages and to understand the principles underlying its organization by examining different morphological systems and the extent to which they deviate from a baseline. This is normally taken to be agglutinative, concatenative morphology, since this represents the most transparent type of organization. In recent work by morphologists, this baseline has been further extended to encompass the principles that provide a reference point for comparing the organization of morphological paradigms. This baseline is known as “canonical inflection” (Corbett 2005, 2007). When morphological systems of different types or with different kinds of subsystems come to be in contact, there are several ways in which they may be affected. In the simplest cases, borrowed words simply join the lexicon of the recipient language, but of particular interest is the way in which they are integrated into its morphological system (Section 3). When morphological material from a donor language is borrowed, a different set of concerns prevail (Section 4). Contact can also result in emergent and adapted morphological systems (Section 5) or in the loss of inflection distinctions (Section 6).

3 Integration of Borrowed Stems into Morphological Systems When words are borrowed from a donor language they typically undergo some form of adaption in order to be integrated into the recipient language. When that language has inflectional classes and/or gender, the borrowed form must be assigned to one of each of these classes. By way of example, consider contact between Khoisan languages – a cover‐term used to refer to three distinct language families of southern and eastern Africa: Kx’a, Khoe‐Kwadi, and Tuu (Güldemann and Fehn 2014) – and a cluster of Bantu languages (“southwest Bantu”) spoken in the Kavango‐Zambezi frontier area, in the borderlands between Nambia, Angola, Zambia, and Botswana (Gunnick, Sands, Pakendorf, and Bostoen 2015). Khoisan loans into these Bantu languages are of particular interest because of their behavior with respect to noun class assignment. While most (non‐Bantu) nominal loans into southwest Bantu languages become part of the default noun class (9/10), or classes where there is a phonological motivation for the class membership, Khwe (Khoe‐Kwadi) and Ju (Kx’a) loans are frequently integrated into various other classes through the addition of a (phonologically unpredictable) noun class prefix, as with Mbukushu di‐nǀánu from Khwe nǀánú ‘button’ and Fwe mu‐ǀáwa from Ju ǁáú (Kalahari Currant, sp. Rhus tenuinervis) (Gunnick et al. 2015: 207–11, 224, 227). This is noteworthy, because it is uncommon in southwest Bantu languages for non‐Khoisan loans to be morphologically integrated into classes that are ­neither the default nor phonologically motivated. Gunnick et al. (2015) propose that Khwe and Ju borrowings were integrated into southwest Bantu morphosyntax through “paralexification” (Mous 2001). Paralexification is a contact‐ induced phenomenon where a borrowed word form is said to be added to an existing synonymous lexical entry. Through this process it acquires the lexical/grammatical characteristics of an item that already exists in the recipient language. Mous (2001) proposes that paralexification differs from regular lexical borrowing in that the borrowed form must share not only the meaning, but also other semantic (e.g. metaphorical extensions), morphological (e.g. inflectional class) and syntactic (e.g. argument structure) properties of the existing lexical item (although this is subject to possible change over time). That synonyms in the recipient (or matrix language in the case of code‐switching) act as a model for integration of borrowed items is intuitive (particularly in second language acquisition). However it is not the only way, since borrowings into Bantu languages are typically added to default noun classes (Mous 2003: 215), usually the class without a noun class prefix in the singular.

132  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman Genuine paralexification results in a non‐canonical morphological phenomenon, ­ verabundance (cf. Thornton 2011). Overbanudance occurs when more than one inflected o word form exists in a single paradigm cell of a lexeme for a given combination of inflectional features. In the case of southwest Bantu languages, the relevant inflectional category is number, the expression of which is determined by the noun class assigned to the borrowed item. While the two forms have identical morphological characteristics, the overabundant forms differ in terms of their stem. Overabundance may turn out to be a common property of languages in contact. It can arise when a borrowed stem or affix competes directly with a pre‐existing form, but also when two or more morphological systems fuse through language mixing (Meakins and Wilmoth 2018).

4 Augmentation and Substitution of a Morphological System Through Contact Although rare when compared with lexical borrowing, there are many documented instances of morphological borrowing through contact (see Seifart 2013, 2017 for a large cross‐linguistic sample, and Gardani et  al. 2014 for a useful overview), which can be ­distinguished according to a range of parameters that together provide a framework for typologizing contact‐induced change.

4.1  Means of Augmentation and Substitution Focusing here on word structure, we can distinguish the borrowing of morphological exponents (i.e. the borrowing of morphological matter), known as MAT‐borrowing, from the borrowing of patterns of exponence (i.e. the borrowing of patterns of word structure), also known as PAT‐borrowing (Matras and Sakel 2007a, 2007b, 2007c; Sakel 2007, see also the discussion in Grant, this volume). Typically, the two occur together. When inflectional elements (such as affixes) are borrowed in combination with other elements, such as the stem that hosts them, matter and pattern are borrowed simultaneously. Various names for this phenomenon have emerged in recent literature on morphological borrowing. Kossmann (2015) calls the process of borrowing stems together with any dependent morphological material concomitant borrowing, while Evans (2016) calls it hosted borrowing. The borrowing of a morphological exponent together with a host from the source language has also been referred to as trojan horse borrowing (Meakins 2011). In such cases a morphological exponent is originally borrowed as part of a morphologically complex loanword, and later becomes used on non‐native stems through a process of language‐internal analogical extension. indirect borrowing (Seifart 2015) of this kind involves two distinct hypothetical stages: in stage 1, a number of morphologically complex loans are borrowed into the lexicon of the recipient language. In stage 2, an analysis of the internal structure of the loans occurs, such that affixes come to be used productively on native stems. Morphology is said to be indirectly borrowed because it first enters the language as part of a larger structure. Seifart (2015) distinguishes this from direct borrowing (cf. Kossmann’s (2015) isolated borrowing), whereby speakers identify an affix in a donor language, and use it with native stems in the recipient language without any intermediate stage. It is likely that (nearly) all examples of phoneme borrowing is indirect, but borrowing of morphological material may, in theory, be either. The evidence base for morphological borrowing is contingent on the process involved. For Seifart (2012: 473) morphological formatives are only demonstrably borrowed into a language when they are used on at least some native stems, i.e. when their use is not restricted to equally borrowed stems. But establishing whether a borrowing is direct or indirect is a

Contact and Linguistic Typology  133 more complex task. Seifart (2015) provides criteria to assess the extent to which individual instances of affix borrowing are direct or indirect based on their distribution across complex loanwords and hybrid formations in corpora. He concludes that this distinction can be viewed as a scale, rather than a dichotomy. The scale considers “the extent to which speakers of the recipient language rely (i) on their knowledge of the donor language (direct borrowing) and (ii) on complex loanwords within their native language (indirect borrowing).”

4.2  Borrowing of Morphological Subsystems The morphological system and the typological profile of a language are most dramatically affected when complete grammatical subsystems are borrowed at once. A particularly vivid example of this is described by Seifart (2012, 2015) for Resígaro, an Arawakan language of Brazil, which has borrowed a small number of morphosyntactic subsystems from Bora, a neighboring language isolate. A substantial amount of inflectional and derivational material that cannot be traced to the Arawakan family has been borrowed into Resígaro, including 20 noun class and gender markers, six number markers, and eight bound grammatical roots that are used to form numerals, demonstratives, and other pro‐forms. Examples of borrowed class and gender markers are provided in Table 6.1. Seifart (2012: 479, 487) observes that borrowed noun class markers are used more frequently, and more productively on Arawakan noun stems, than on Bora stems (which only comprise around 5% of Resígaro vocabulary), with many borrowed noun class markers occurring only on native stems. Table 6.1  Borrowed noun class and gender markers (Seifart 2012: 483–4).

Gloss

Bora

Proto‐Bora‐ Muinane

Resígaro

Resígaro example

cm.tree cm.hole cm.river

‐ʔɛ ‐ʔɛhɯ ‐ʔi

*‐ʔɛ *‐ʔɛhɯ *‐ʔi

‐ʔɛ ‐ʔɛhɯ ‐ʔi

cm.oblong cm.pillar cm.3dim cm.bush cm.root

‐ʔo ‐ʔahkɯ ‐ba ‐bahɯ ‐baʲhkɛ

(*‐ʔo) *ʔahkɯ *‐ba *‐bahɯ *‐baikkɛ

‐ʔo ‐aakɯ ‐ba ‐bahɯ ‐bákɛ́

cm.branch

‐ɡʷáhka

*‐gákka

‐gáka

cm.plank cm.house cm.disc cm.tube cm.stick cm.pointed cm.day cm.leg cm.river cm.round feminine sg

‐ɡʷa ‐ha ‐hɨ ‐hɯ ‐i ‐ko ‐ko:hɨ ‐kɯba mo:a ‐ɯ ‐(pi)dʒɛ

*‐gai *‐ha *‐hɨ *‐hɯ *‐i *‐ko *‐ko:hɨ *‐kɯba *moáai (*‐ɯ) *‐ge

‐gɯ ~ ‐ga ‐ha ‐hi ‐hɯ ‐i ‐ko ‐koohi ‐kɯʔba ‐moóa ‐ɯ ‐(pi)dʒɛ

koóna‐ʔɛ ‘almond tree’ tɛ́‐ʔɛhɯ ‘this hole’ tɛ́‐ʔi hiβíi ‘water‐star’ (proper name) níítsi‐ʔo ‘cigarette’ tɛ́‐aakɯ ‘this pillar’ síʔí‐bá ‘another (box, log, etc.)’ tɛ́‐bahɯ ‘bush’ pipíigí‐ʔɛ‐bákɛ́ ‘root of pijuayo palm tree’ koóna‐ʔɛ‐gáka ‘almond tree branch’ βáʔa‐gɯ́ ‘machete’ dʒídʒáá‐há‐koba ‘big house’ daʔm̥ óótsi‐hí ‘hat’ kaanɛ́‐hɯ́ ‘doing this (story)’ goʔkoótsi‐í ‘mortar’ tɛɛ́‐ko ‘this (stick)’ tsá‐kóóhi ‘one day’ hí‐kɯ́ʔba ‘this leg’ tɛ́‐moóa ‘this river’ sí‐ɯ ‘another (pebble, etc.)’ phai‐pídʒɛ́ ‘old woman’

134  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman This type of borrowing is particularly interesting from a typological perspective because the borrowed morphology has agreement‐like properties. Substantive modifiers, used anaphorically, must agree with the noun class of the noun they modify, as illustrated in (1), where the inflected property words bear the same class marker as the entity they modify. (1) mɛtsoíigí‐ɯ ʤíʤaá‐ɯ́‐koba ripe‐CM.ROUND big‐CM.ROUND‐AUG ‘Into what was a ripe, big (fruit) she entered.’ (Seifart 2012: 485; sacha_res 062)

ʤoβí‐giko be‐ALL

tsó‐kaʔpʰaaβɯ́ 3SG.F‐enter

Resígaro has also borrowed a set of animate dual suffixes, which also mark gender. Dual is not attested as a number value in any other Arawakan language (Seifart 2012: 487) and this therefore stands as a highly likely instance of the borrowing of a grammatical feature value. The suffixes are used on nouns to indicate inherent number and gender properties of a stem in both languages, as in (2). (2) a. Bora okáhi‐mɯ́pɨ tapir‐DL.F ‘two female tapirs’ (Seifart 2012: 487)

b. Resígaro an̥óógi‐ mɯ́pí tapir‐DL.F ‘two female tapirs’

They are also used on verbs to indicate the number of the predicate subject, as shown in (3). (3) a. Bora pɛ:‐mɯ́pɨ go‐DL.F ‘They went.’ (Seifart 2012: 488)

b. Resígaro nɛ‐ʔpí‐mɯ́pí 3PL‐GO‐ DL.F ‘They went.’

There are two further striking properties that characterize the results of contact between Bora and Resígaro: (i)  most of the borrowed morphological material form tightly integrated subsystems that express quantification (nominal classification, number marking and quantification proper); (ii)  despite the fact that sustained contact must have been a precursor to large‐scale ­borrowing, the other subsystems of grammar have hardly been influenced at all. This raises questions about the borrowability of different types of inflection. A subject to which we now turn.

4.3  Limits of Morphological Borrowability The limits on borrowability and congruence have been the focus of much cross‐linguistic work on language contact (Weinreich 1953; Moravcsik 1978; Thomason and Kaufman 1988; Myers‐Scotton 1993; Sebba 1998; Matras 2007). It is widely acknowledged that lexical borrowing is more pervasive than morphological borrowing, for instance, and this is thought to be related to (i) the structural autonomy of the borrowed item, and (ii) its semantic autonomy.

Contact and Linguistic Typology  135 Gardani (2008) observes that contextual inflection (e.g. agreement and structural case morphology), is more difficult to borrow than inherent inflection, and “the transfer of entire inflectional paradigms has long been regarded as the last challenge to morphological borrowability” (Gardani, Arkadiev, and Amiridze 2014: 11). But there are examples that demonstrate that it is possible, as seen with borrowing into Resígaro (Section  4.2). Wholesale borrowing of two verb paradigms as well as the prefixal paradigm of reflexive pronouns into Marrku (or Marrgu, a recently extinct language, most probably an Iwaidjan, non‐Pama‐ Nyungan language) is discussed in Evans (2016), while Pakendorf (2009, this volume) explores the borrowing of the assertive‐presumptive paradigm from Sakha (Turkic) into Even (Northern Tungusic), both spoken in Siberia. When contextual inflection is borrowed, it tends not be at the level of a single affix, but rather a whole subsystem, or at least a substantial part thereof. To account for this observation, Seifart (2012: 475) proposes the following principle: Principle of Morphosyntactic Subsystem Integrity (PMSI) Borrowing of paradigmatically and syntagmatically related grammatical morphemes is easier than borrowing of the same number of isolated grammatical morphemes.

While framed as a principle of organization, he suggests that the PMSI is probably motivated by a more general tenet that avoids restructuring tightly integrated morphosyntactic subsystems, such that borrowing complete, or largely complete, paradigms may ultimately be less disruptive than piecemeal morphological borrowing.

5 Influence of Morphological Profile on Emergent vs Existing Systems The augmentation or substitution of the morphology of a language through contact assumes that recipient and donor varieties can be distinguished. In the case of mixed languages, which emerge through a process of fusion of existing systems, the languages are perhaps best considered in terms of their role in code‐switching. Here we consider the effect of the morphology of complex predicates on the emergence of two mixed languages of Australia, Gurindji Kriol, and Light Warlpiri, before considering this in relation to parallels in contact‐ induced change observed in Murrinhpatha (Mansfield 2016).

5.1  The Emergence of Gurindji Kriol and Light Warlpiri Gurindji Kriol and Light Warlpiri are mixed languages (O’Shannessy and Davidson, this volume) that have arisen in the last 35–40 years as a result of the conventionalization of code‐ switching between an Australian Aboriginal language in the Ngumpin‐Yapa group of Pama‐ Nyungan (Gurindji, Warlpiri) and a contact variety of English (Kriol/Aboriginal English) (McConvell and Meakins 2005; Meakins 2011; O’Shannessy 2005, 2012). Today, they are the main everyday languages of the Gurindji and Warlpiri people, spoken alongside their source languages, in the inland north of Australia. While the two mixed varieties show striking similarities (for instance, their nominal syntax is predominantly Gurindji or Warlpiri rather than Kriol in origin), they differ in terms of the way that the verbal inventory has been constructed: one third of all verbs in Gurindji Kriol have a Gurindji origin, while just seven Light Warlpiri verbs have a Warlpiri source. Meakins and O’Shannessy (2012) propose that typological differences in the verbal structures of the two traditional languages is responsible for this outcome.3

136  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman

5.2  Complex Predicates in Gurindji Gurindji has a system of coverbs that combine with one of 31 “inflecting verbs” to form complex predicates. The inflecting verb is obligatory; it can form a predicate without a coverb, and is the host for inflectional morphology encoding Tense‐Aspect‐Mood (TAM) information. Coverbs are an open class with hundreds of members. They do not inflect, other than by taking the continuative suffix ‐karra. They rarely form predicates without an inflecting verb, but can do so in imperatives and reduced subordinate clauses (Meakins and O’Shannessy 2012: 228–9). While coverbs and inflecting verbs are generally adjacent, as in (4), they do not form a phonological word and they can also be separated by other constituents, as in (5). jarrwaj pu‐nya karu (4) marluka‐ngku   ngu=Ø=Ø old.man‐ERG CAT=3SG.NOM=3SG.ACC spear pierce‐PST child[abs] ‘The old man speared the child.’ (Meakins and O’Shannessy 2012: 228; ES: FHM110: Picture elicitation) (5) jarrwaj‐karra ngu=rna=rla pu‐nga‐ni nyila=ma spear‐CONT CAT=1SG.NOM=3OBL pierce‐IMPF‐PST that=TOP ‘There I was prodding at it.’ (Meakins and O’Shannessy 2012: 228; VD: FM09_a127: Description) This suggests that these two components are relatively independent. Most Gurindji coverbs have a “loose nexus” with the inflecting verb (cf. Bowern 2014), since all verbs can be separated from the inflecting verb, and nearly all coverbs can occur before or after the inflecting verb (see Meakins and O’Shannessy 2012: 229–30 for brief discussion of exceptions).

5.3  Complex Predicates in Warlpiri In contrast to Gurindji, Warlpiri is characterized by a large inventory of inflecting verbs. There are around 120 inflecting verbs in Warlpiri which form a “tight nexus” with their coverb (where present). Evidence for this can be seen in syntax and phonology: coverbs must precede the inflecting verb and can occur together with inflecting verbs in clause initial position, which can only contain a single constituent (Nash 1982: 175–82). Together they form a phonological phrase. Since they are not independent words, coverbs in Warlpiri appear as the first element in a bipartite stem, as demonstrated in (6) for the coverb rdilyki ‘break’ and the inflecting verb pantu ‘pierce’. (6) ngula‐ngu patulu‐ngu‐ju rdilyki‐pantu‐rnu that‐ERG bottle‐ERG‐TOP break‐pierce‐PST ‘That bottle pierced the tyre, damaging it.’ (Meakins and O’Shannessy 2012: 235; ERGstroyWA29)

taya‐ju tyre‐TOP

Some coverbs are highly restricted in their distribution: they occur immediately before a single inflecting verb, and cannot be separated from it by any other element, nor occur with any other inflecting verb. Others are less restricted in terms of their ability to be non‐adjacent, and can combine with different inflecting verbs (to varying degrees) to produce different semantic effects. In each case the coverb precedes the inflecting verb, hence the label “preverb” given to these verbs in descriptions of Warlpiri. For instance, the coverb jaa‐ ‘open’ can combine with a few different inflecting verbs, but this is nonetheless restricted.4

Contact and Linguistic Typology  137

5.4 Emergence of New Verbal Systems Through Contact with Kriol Kriol is the name given to a group of mutually intelligible contact‐varieties spoken as a first language by up to 30,000 Aboriginal people in northern Australia (Meakins 2014). It contributes much of the verbal structure of both Gurindji Kriol and Light Warlpiri. Complex predicates in Kriol consist of an auxiliary bearing TAM information and a lexical verb, which older generations of Gurindji speakers treat as structurally congruent equivalents to Gurindji inflecting verbs and coverbs respectively when code‐switching (Meakins 2010: 26–9; Meakins and O’Shannessy 2012: 230–2). In the mixed language Gurindji Kriol, Gurindji origin coverbs are treated like Kriol lexical verbs. They can occur with the full range of Kriol TAM auxiliaries and clitics (Meakins and O’Shannessy 2012: 227), as shown in (7) where the Gurindji coverb ngalyakap ‘lick’ (marked with the Gurindji continuative suffix –karra) occurs in a complex predicate with the Kriol past tense auxiliary bin. (7)

dat kamel‐tu i bin ngalyakap‐karra the camel‐ERG 3SG PST lick‐CONT ‘The camel licked the kangaroo.’ (Meakins and O’Shannessy 2012: 227)

kengkaru kangaroo

Meakins and O’Shannessy (2012) propose that the structural independence of coverbs in  Gurindji provided a congruent structural point for code‐switching in bilinguals and  the  consequent integration of coverbs as lexical verbs in complex predicates in Gurindji Kriol. In contrast, Warlpiri coverbs are not found in complex predicates in Light Warlpiri. This appears to be because Warlpiri and English/Kriol verbal components are not interchangeable in code‐switching (O’Shannessy 2006: 67–8). Warlpiri coverbs are not attested when the code‐switching matrix language is Kriol/ English, presumably because English lexical verbs do not have a bipartite stem, and they therefore cannot accommodate a coverb within their morphological/phrasal structure. When the matrix language in code‐switching is Warlpiri, English/Kriol verbs can be integrated as a coverb with one of two inflecting verbs. The inchoative verb ‐jarrimi is used with intransitive verbs (8), while causative ‐mani is used to integrate transitive verbs (9). (8) yama‐ngka ka slip‐jarri‐mi shade‐LOC IMPF sleep‐INCHO‐NPST ‘He is sleeping in the shade.’ (intransitive) (Meakins and O’Shannessy 2012: 237; A10:ergstoryWA10.3) (9) marda‐rni ka‐npa wati kuja ka puj‐i‐ma‐ni toyota hold‐NPST IMPF‐2SG.NOM man REL IMPF push‐EP‐CAUS‐NPST car ‘Have you got a man who is pushing a car?’ (transitive) (Meakins and O’Shannessy 2012: 237; A40:ergcardWA40A37) In these cases, the English/Kriol independent verb stem is treated as a bound coverb. But this pattern is not observed in Light Warlpiri, where English origin verbs can be used as lexical verbs without a Warlpiri origin component. Meakins and O’Shannessy (2012) suggest the lack of Warlpiri origin coverbs in Light Warlpiri stems from the fact that Kriol/English verbs and Warlpiri coverbs are not ­considered equivalent word classes by speakers as a result of differences in their structural

138  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman autonomy. This demonstrates that differences in the morphological structure of word classes may have a considerable effect on the emergence of new structures in a mixed language. While Warlpiri coverbs are not attested in Light Warlpiri, seven verbs have their source in a Warlpiri inflecting verb: panti‐ ‘pierce’, kati‐ ‘press’, punta‐ ‘steal/take from’, paka ‘hit’, kiji ‘throw kick’, winji‐ ‘pour, spill’ and pali ‘die’ (Meakins and O’Shannessy 2012: 233). Owing to a change in Warlpiri phonotactics under English influence, inflected forms of the verb have been reinterpreted as verb stems (e.g. panti‐rni ‘pierce‐npst’ > patirn) which are then available to bear the highly productive Kriol transitive marker –im which can attach to stems of any origin, as in (10).5 (10) ah jinta‐ju watiya‐ng i=m ah one‐TOP thorn‐ERG 3SG=NFUT ‘Ah! A thorn pierced (his) foot!’ (Meakins and O’Shannessy 2012: 233; C03_6_3)

pantirn‐im pierce‐TR

wirliya foot

In terms of structure, these verb forms are integrated on a par with Kriol origin verbs.

5.5  Coverbs in Murrinhpatha Mansfield (2016) examines similar verbal structures in Murrinhpatha, a Daly language of northern Australia, in which the expansion of the light‐verb construction used to accommodate borrowings has expanded its range dramatically within just a few decades as a result of contact with English and Kriol. The verbal morphology of Murrinhpatha (as recorded for speakers born between 1900 and 1960) is largely agglutinative in nature, but with a core of highly irregular stem modifications (Walsh 1976; Blythe 2009; Nordlinger 2010; Street 1987). Although this structure goes back historically to a coverb and accompanying inflecting verb (Blythe 2009: 139; Nordlinger 2010: 338) similar to those described for Gurindji and Warlpiri, the parts are no longer separable and for many agglutinative verbs the relationship between the two elements has become semantically opaque. Examples of agglutinative stems are exemplified in (11) and (12). Each agglutinative verb consists of an initial classifier stem fused with subject and TAM morphology, and a lexical stem, as well as various other inflectional exponents (Nordlinger 2010). (11) Ku were dirran‐ngi‐ngarlbarl ANIM dog 3SG.watch‐1SG.O‐bark ‘The dog is barking at me.’ (Mansfield 2016: 398; Fieldnotes 2014‐09) (12) Pan‐pe‐lip 3SG.slash‐head‐hit ‘She hit him on the head.’ (Mansfield 2016: 398; 2013‐01_pict.elicit) Eleven of the 38 classifier stems can also be used as independent “simple verbs,” e.g. pirrim “it stands” (Mansfield 2016: 405). A third type of predicate – phrasal verb constructions – are formed from a coverb and one of four “simple” verbs stems: the so‐called light verbs “do,” “sit,” “be,” and “go.” As in other languages of the region, the coverb carries the main lexical content expressed by the predicate. In Murrinhpatha, the light verb is inflected for tense and argument inflections and also

Contact and Linguistic Typology  139 Table 6.2  Verb structure types in older and younger speaker samples (Mansfield 2016: 420).

Older Younger

Agglutinative

Phrasal

Simple

Total

213 69% 181 56%

1 (1 inherited coverb) 0.3% 65 (16 inherited coverb/49 English) 20%

94 31% 77 24%

303 323

contributes the telicity of the predicate. Examples are given in (13) and (14) with an inherited Murrinpatha coverb and a borrowed English one. (13) Wilili wurran walk 3SG.go ‘He’s walking.’ (Mansfield 2016: 419) (14) Meikit mam ngamimarda=thu make.it 3S.do other.side=hither ‘He made it all the way across.’ (Mansfield 2016: 409; 2011‐07‐25_LP‐GD) Using small corpora of speech for older and younger Murrinpatha speakers, Mansfield shows that there are substantially more examples of phrasal verbs in the speech of younger speakers than older speakers of Murrinhpatha, with English coverbs predominating over inherited ones (Table 6.2). Mansfield (2016: 398) proposes that the phrasal verb construction observed in the speech of younger Murrinhpatha speakers has been “rapidly reinvigorated from a marginal structure to a major one” as a result of contact with English and Kriol. The use of light‐verb constructions for lexical borrowings is motivated by the requirement to be congruent, and acts as a strategy to integrate items that do not meet the morphological profile of synthetic verbs in the recipient language. Mansfield suggests that recent lexical borrowing has triggered a change in the formation of verbal structures, stimulating a shift from synthetic expression of verbs to phrasal forms, with a more transparent bipartite structure. While the use of light verbs for lexical borrowing is a common cross‐linguistic strategy (e.g. Wohlgemuth 2009), the Murrinpatha pattern is notable in terms of the extent to which speakers are embracing this alternative means of expression for inherited content too, and thus changing the typological profile of the verbal system in the process.

6  Loss of Inflection Through Contact One way that contact can reshape a morphological system is to drive it out of existence, something that inflectional morphology in particular is susceptible to. It has been proposed that contextual inflection is lost more quickly and borrowed less readily than inherent inflection (Gardani 2008; Roberts and Bresnan 2008), but little is currently known about the ways in which this proceeds in different morphological systems and whether contact‐ induced loss of inflection proceeds in the same way as regular language‐internal change. We illustrate here what contact‐induced loss can reveal about the organization of morphological systems through the loss of gender in Cappadocian Greek.

140  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman

6.1  The Loss of Gender in Cappadocian Greek Cappadocia in central Asia Minor was home to Greek speakers from the Hellenistic period until the 1923 population exchange between Greece and Turkey which moved the region’s remaining Greek Orthodox inhabitants to Greece.6 This was always a heavily multilingual region (Janse 2002) but was subject to particularly strong influence from Turkish from the eleventh century, when invasions of Seljuk Turks cut Cappadocian Greek speakers off from the Byzantine empire in the West. This meant that the Turks had a much greater linguistic and cultural influence here than in other Greek‐speaking communities of Asia Minor, corroborated by the large‐scale conversion of Greek Orthodox Christians to Islam from at least 1437 (Dawkins 1916: 1; Thomason and Kaufman 1988: 215–16). The sociolinguistic landscape during this period was characterized by widespread bilingualism, with a bias towards Turkish as the dominant language. This extended contact had profound effects on the local varieties of Greek, documented by Dawkins (1916), who conducted extensive fieldwork in the region between 1909 and 1911, and who also described the sociolinguistic situation in detail. One particularly striking feature of Cappadocian Greek dialects, from the point of view of the standard language, is the lack of agreement in the noun system. While the standard variety has agreement in gender (masculine/feminine/neuter), number (singular/plural) and case (nominative/accusative/ genitive) marked on all nominal elements, Cappadocian has generalized what is historically a neuter nominative‐accusative form for all agreement targets, including pronouns, without any corresponding change to the form of agreement controllers. This is usually attributed to contact with Turkish, which has neither gender nor case agreement (though it marks case on nouns and pronouns), and exhibits number agreement only in pronouns. The only distinction that remains marked in Cappadocian agreement is that of singular versus plural number, supporting the claim that contextual inflection is lost more quickly and borrowed less readily than inherent inflection (see Gardani 2008, 2012), and Johanson (2002: 109) specifically in relation to Turkish contact).

6.2  Turkish‐Greek Convergence in Cappadocia The Turkish influence on Cappadocian Greek is clearly discernible: so much so that Cappadocian (sometimes grouped together with other Asia Minor Greek varieties) features as a textbook example of “heavy borrowing” (or perhaps more appropriately “code‐ mixing”: Janse 2002) due to extended contact in several general treatments of language contact (Thomason and Kaufman 1988; Matras 2009; Drinka 2010; Winford 2005, 2010). For a general overview of Turkish features in Cappadocian, see Dawkins (1916), Thomason and Kaufman (1988: 215–22), Janse (2009). Here it suffices to say that Turkish influence is visible at all linguistic levels. Often‐cited examples include extensive lexical borrowing, various types of assimilation to the Turkish phonological system including the (limited) transfer of Turkish vowel harmony, the development of differential object marking on the Turkish model, calquing of Turkish agglutinative noun structures and verbal periphrases, the development of Turkish‐style possessive suffixes out of possessive pronouns, and shifts towards SOV structures, which is the basic word order in Turkish, but marked in Standard Modern Greek. These examples include a number of instances of loss by assimilation: in phonology, for instance, Cappadocian has lost dental fricatives, and in morphology it has lost the distinction between perfect and past perfective, both of which are absent in Turkish. The extent of this influence is vividly expressed by Kontosopoulos in his survey of Greek dialects: “whoever hears… the Cappadocian dialect doesn’t know whether he is dealing with Turkish in a Greek mouth, or Greek in a Turkish mouth.”7 Such convergence is likely to

Contact and Linguistic Typology  141 have been led by bilinguals, speaking a Greek heavily influenced by the grammatical ­structures of their dominant language, Turkish (Winford 2005: 402–9; cf. Thomason and Kaufman 1988: 218).

6.3  The Loss of Gender The loss of gender in Cappadocian Greek has been treated as a fairly unambiguous example of change due to contact by Dawkins (1916: 115) and most scholars who have followed him (e.g. Janse 2009), for obvious reasons. Cappadocian is the only variety of Greek to have lost gender completely: outside Asia Minor, Greek is comparatively conservative in preserving the Indo‐European gender system (Matasović 2004: 74), although it has undergone a number of innovations by comparison with the ancient language (Morpurgo Davies 1968). It is also the variety which has both been in closest contact with Turkish, a genderless language, and cut off from the rest of the Greek‐speaking world. As we have seen, this contact had a deep impact on other areas of grammar; it therefore seems highly unlikely, at least on the face of it, that the loss of gender in Cappadocian could have been unrelated to Turkish influence. Turkish has no gender system of any kind, beyond an animacy distinction in interrogative pronouns (cf. English who/what?). It lacks even semantic gender in third person pronouns, the sole relic of Indo‐European gender in a number of languages (e.g. English he/she/it). Standard Modern Greek, on the other hand, makes a three‐way gender distinction between masculine, feminine, and neuter, with gender agreement marked on determiners, pronouns, and adjectives (used both attributively and predicatively), and some numerals. Agreement is strictly syntactic (Holton et al. 2012: 317, 603–7). Gender is assigned primarily on the basis of inflection class, but the three genders have a semantic core which is reflected by the usual strategy for assigning gender to loanwords: e.g. mánatzer ‘manager’ (masculine or feminine, depending on the referent), but asansér ‘lift, elevator’ (neuter, from French ascenseur, masculine) (Holton et al. 2012: 323–5). By contrast, in Cappadocian Greek, historically neuter forms have been generalized in all agreement targets, without any accompanying change to the form of agreement controllers8 (15) (Dawkins (1916: 346), cited by Karatsareas 2011: 148). (15) Ʒíleps‐e ke t állo ʃinífsa. Píj‐e envied‐3SG and the.N other.N sister‐in‐law(F) went‐3SG so dóma. to.the roof ‘And the other sister‐in‐law was jealous. She also went to the roof.’

ki and

etó it.N

This can be explained as the result of grammatical pattern borrowing led by Greek‐Turkish bilinguals. Further evidence for the role of contact comes from studies on the acquisition of Greek gender by bilinguals and L2 speakers, which show both that they struggle with gender agreement and that they tend to extend the neuter when in doubt (e.g. Seaman 1972 for Greek in contact with English; Georgalidou et al. 2011 for Turkish, and further references and discussion in Karatsareas 2011: 168–70).

6.4  Diachronic Pathways for the Loss of Gender Recent research by Karatsareas (2009, 2011, 2013, 2016) has re‐examined a number of alleged Turkisms in Cappadocian Greek in the light of dialectological evidence. He rightly points out that for assessing the degree of Turkish influence, the appropriate counterpoint for Cappadocian is not Standard Modern Greek but the other Asia Minor Greek dialects, with which it shares a more recent common heritage. With one exception, these diverge

142  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman considerably from other Greek varieties in their gender systems. He uses this comparative evidence to reconstruct a pathway for the loss of gender in Cappadocian, arguing convincingly that it was gradual and structured rather than abrupt (Karatsareas 2009, 2011, 2014). In his account, gender loss in Cappadocian is the end point of two diachronic processes, shared to a certain degree by all Asia Minor Greek dialects apart from Silliot. The first of these is the development of semantic agreement (Karatsareas 2014), moving gradually along two trajectories: Corbett’s (1979) agreement hierarchy (16), and Sasse’s (1993) individuation hierarchy (17) (spreading from right to left in both cases). Semantic agreement is attested in Pontic, formerly spoken on the southern shores of the Black Sea, and Rumeic, the dialect of Greek communities in Mariupol (Ukraine) descending ultimately from Pontic settlers. Rumeic has gone a step further than Pontic, in that semantic agreement is almost entirely complete, while Pontic retains syntactic agreement in the definite article when it immediately precedes human‐denoting nouns. (16) attributive – predicate – relative pronoun – personal pronoun (17) proper names – humans – animals – tangible objects – abstracts – mass nouns This meant that inflection class was no longer a reliable predictor of agreement. This cleared the way for the second diachronic process, namely the generalization of neuter agreement, to make its way from contexts like (18) to contexts like (19) (Karatsareas 2009: 222–3). (18) to kalon the.N good.N ‘the good chicken’

to the.N

kosara chicken(F)

(19) i kalesa the.F good.F ‘the good woman’

i the.F

jineka woman(f)

In Cappadocian this was universal, leading to the loss of the gender system, but neuter agreement with masculine and feminine animate nouns is also found to an extent in all the Asia Minor Greek dialects apart from Silliot. This is clearest in Pharasiot (the dialect of Faraşa and surrounding villages to the east of Cappadocia), which preserves tripartite syntactic agreement in the definite article immediately preceding a noun, and has some residual uses of the feminine personal pronoun, but otherwise has neuter agreement like Cappadocian. In Pontic and Rumeic, neuter agreement can be triggered by human‐denoting nouns, particularly feminine plurals. In Pontic, female animals pattern with inanimates in triggering neuter agreement, while male animals trigger masculine agreement like male humans. Both in the spread of semantic agreement and neuter agreement, the dialectal evidence shows a  tendency for the feminine to be affected before the masculine, and the plural before the  singular. The definite article holds onto distinctions for the longest, when it comes ­immediately before the noun.

6.5  Internal Versus External Explanations Karatsareas (2009, 2011, 2013, 2016) calls for a shift of focus away from language contact and towards language‐internal factors which contributed to the resemblances between Cappadocian Greek and Turkish. He argues persuasively that the loss of gender in Cappadocian was not purely contact‐induced, as has been suggested, but was made possible by the previous development of semantic agreement, which was already underway in the common ancestor of Pontic and Cappadocian before Turkish contact in Cappadocia

Contact and Linguistic Typology  143 intensified (Karatsareas 2014: 94–5). Nonetheless he acknowledges that this dialectological approach “does not preclude language contact as a major contributing factor in the subsequent development of specific innovations” (2013: 210).9 And indeed, it is hard to deny Turkish contact a large role in the story. It is the two Asia Minor Greek dialects which have been subject to the most intense Turkish influence, Pharasiot and Cappadocian, where gender loss has gone furthest. While internal factors may have predisposed Cappadocian to lose gender, subsequent pattern borrowing from Turkish can explain not only the loss of gender, but also the loss of case agreement. Furthermore Cappadocian Greek is only one of a number of languages in contact with Turkic to lose adjective agreement, along with Persian, Mongolian, and Western Armenian (Johanson 2002: 109). This case study illustrates how the loss of an inflectional system can reveal aspects of its inner structure that otherwise may remain invisible: its gradual dismantling follows covert fault lines and hierarchies set down by the grammatical system. Compare Jingulu (Meakins and Pensalfini 2016) where language obsolescence is leading to the structured breakdown of an originally four‐gender system, with masculine agreement optionally triggered by nouns of all genders, and neuter agreement optionally triggered by either neuter or vegetable nouns. The breakdown of the gender system can be attributed either to simplification as the result of language obsolescence, or specifically of contact with genderless languages, but either way it is channeled along lines set down by language‐internal categories and hierarchies. This agrees with the suggestion of Janse et al. (2011: 238) that change to gender systems may reveal aspects of synchronic grammar, because changes originate in decisions made by speakers, which reflect how they understand the linguistic system they are working with. Even if the end result of loss can be viewed as a construction originating externally, the way that it is achieved may elucidate both language‐internal structure and universal hierarchies.

7 Conclusion Language contact is an important explanatory tool for understanding the distribution of typological variables, and must be taken into consideration as a possible influence when constructing probabilistic theories accounting for cross‐linguistic diversity. We have examined four different case studies in which language contact can be used to account for the morphological profile of languages in four different regions of the world: southern Africa, Brazil, northern Australia and Turkey. In each case contact explanations help us to understand the distribution of variables of interest to typologists. Together, a broad understanding of typological characteristics of recipient and donor languages, and a cross‐linguistic view on the circumstances that facilitate contact provide insights into the role morphology can play in contact explanations of the distribution of variables. First, we sought to demonstrate that lexical borrowings can be integrated into the morphological system of a donor language in various ways. Frequently borrowed forms are assigned to a default class, or a semantically or phonologically motivated class. Sometimes, class membership is based on a (near) synonymous model, resulting in overabundance in the paradigms of existing forms on the recipient language, rather than distinct lexical entries. The second case study looked at the augmentation of a morphological system through borrowing into Resígaro from Bora (Seifart 2012, 2015). Resígaro differs from other Arawakan languages in that it has borrowed a large amount of morphological material from the neighboring isolate Bora. Large‐scale inflectional and derivational borrowing of this kind appears to be possible providing that grammatical subsytems are borrowed intact. How the morphological typology of a language will be effected in contact situations is dependent on the properties of the morphosyntactic systems of the varieties in contact, as well as those social factors that facilitate contact, as shown in our third case study of complex

144  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman predicates in languages of northern Australia. Gurindji Kriol and Light Warlpiri are both mixed languages that draw on resources from Kriol and Aboriginal English on the one hand, and the Ngumpin‐Yapa languages, Gurindji and Warlpiri on the other (Meakins and O’Shannessy 2012). Each of these languages integrates verbal elements of Ngumpin‐Yapa origin in different ways, based on the extent to which complex predicates consist of morphologically bound or independent verbal stems. Similar issues were shown to be important in Murrinhpatha, which appears to be exhibiting change in its morphological profile, shifting towards the analytic expression of complex predicates (Mansfield 2016). Our final case study strove to show that the typological properties of a language may also be changed through contact that leads to loss, rather than augmentation or reorganization. Studying contact‐induced change leading to the loss of an inflectional system can reveal aspects of its inner structure that otherwise may remain invisible. We argued that the loss of gender in Cappadocian Greek may be part of a more general loss of adjective agreement through contact between Turkic languages and unrelated languages of Eurasia.

Abbreviations The following abbreviations are used in glossing: 1 = first person, 2 = second person, 3 = third person, 3DIM = three‐dimensional shape, ABS = absolutive, ACC = accusative, ALL = allative case, ANIM = animate, AUG = augmentative, CAT = catalyst, CAUS = causative, CM = noun class marker, CONT = continuative, DL = dual, EP = epenthetic syllable, ERG = ergative, F = feminine, IMPF = imperfective, INCHO = inchoative, LOC = locative, N = neuter, NOM = nominative, NFUT = non‐future, NPST = non‐past, O = object, OBL = oblique, PL = plural, PST = past, REL = relativizer, S = subject, SG = singular, TOP = topic, TR = transitive. Information listed after the source of examples indicates the session name, annotation number, activity type, etc., as provided in the original source.

NOTES 1 For an overview of the history of typology see Shibatani and Bynon (1995). 2 See Koptjevskaja‐Tamm (2010) for a discussion of the role of areality in linguistic typology and language contact research. 3 See McConvell (2010: 776–7) for discussion of diffusion of complex verbs in Non‐Pama‐Nyungan languages to East Ngumpin languages (including Gurindji). 4 Two further class of elements referred to as “adverbial coverbs” and “dative adjunct coverbs” (Laughren 2010; Nash 1982) can be separated from and postposed to the inflecting verb, but since they appear to have different distributional and semantic properties to coverbs proper, they will not be considered further here. 5 For discussion of intransitive pali ‘die’ which does not bear the transitive suffix, see Meakins and O’Shannessy (2012: 239). 6 In the twenty‐first century, small communities of Cappadocian speakers survive in Central and Northern Greece (Janse 2009; Vassalou et  al., forthcoming). Until their discovery, the dialect was thought to have been extinct since the 1960s (e.g. Kontοsopoulos 1994 [2008]: 7). 7 “Ὅποιος ἀκούει – ἢ μᾶλλον διαβάζει, γιατὶ σήμερα δεν μιλιοῦνται πιὰ τὰ ἰδιώματα αὐτά, ἀφοῦ ὅλοι σχεδὸν οἱ φορεῖς τους, πρόσφυγες τοῦ 1922, ἔχουν πεθάνει – τὴν καππαδοκικὴ διάλεκτο, δὲν ξέρει ἂν ἔχει νὰ κάνη μὲ τούρικια σὲ ἑλληνικὸ στόμα ἢ μὲ ἑλληνικὰ σὲ στόμα τούρκικο” (Kontosopoulos 1994 [2008]: 7). 8 Relics of gender are retained in three Cappadocian varieties, with biases towards retention in the feminine gender, in human nouns, and in the definite article. For details see Karatsareas (2011: 148–50).

Contact and Linguistic Typology  145 9 Similar sentiments were expressed by Dawkins: “the Turkish influence found already existing a loss of grammatical gender or at least a tendency to lose grammatical gender, and carried this further to its own conditions of total absence of any distinctions of gender” (1916: 116).

REFERENCES Aboh, Enoch Oladé 2015. The Emergence of Hybrid Grammars. Language Contact and Change. Cambridge: Cambridge University Press. Blythe, Joe 2009. Doing referring in Murriny Patha conversation (University of Sydney Ph.D. thesis). Bickel, Balthasar 2007. Typology in the twenty‐ first century: major current developments. Linguistic Typology 11: 239–51. Bostoen, Koen and Bonny Sands 2012. Clicks in south‐western Bantu languages: contact‐ induced vs. language‐internal lexical change. In Matthias Brenzinger and Anne‐Maria Fehn (eds.), Proceedings of the 6th World Congress of African Linguistics Cologne 2009, 129–40. Cologne: Rüdiger Köppe. Bowern, Claire 2014. Complex predicates in Australian languages. In Harold Koch and Rachel Nordlinger (eds.), The Languages and Linguistics of Australia: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 263–94. Corbett, Greville G. 1979. The agreement hierarchy. Journal of Linguistics 15.2: 203–24. Corbett, Greville G. 2005. The canonical approach in typology. In Zygmunt Frajzyngier, Adam Hodges, and David S. Rood (eds.), Linguistic Diversity and Language Theories. Amsterdam: Benjamins, pp. 25–49. Corbett, Greville G. 2007. Canonical typology, suppletion and possible words. Language 83: 18–42. Dawkins, R. M. 1916. Modern Greek in Asia Minor: A Study of the Dialects of Siĺli, Cappadocia and Pharasa, with Grammar, Texts, Translations and Glossary. Cambridge: Cambridge University Press. Drinka, Bridget. 2010. Language contact. In Silvia Luraghi and Vit Bubenik (eds.), The Continuum Companion to Historical Linguistics. London: Continuum, 325–45. Dryer, Matthew S. 1992. The Greenbergian word order correlations. Language 68: 81–138. Dryer, Matthew S. and Martin Haspelmath (eds). 2013. The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://wals.info.

Evans, Nicholas 2016. As intimate as it gets? Paradigm borrowing in Marrku and its implications for the emergence of mixed languages. In Felicity Meakins and Carmel O’Shannessy (eds), Loss and Renewal: Australian Languages Since Contact. Berlin: de Gruyter Mouton, pp. 29–56. Gardani, Francesco 2008. Borrowing of Inflectional Morphemes in Language Contact. Frankfurt‐am‐ Main: Peter Lang. Gardani, Francesco 2012. Plural Across Inflection and Derivation, Fusion and Agglutination. The Origins of Bound Morphology. Leiden: Brill. Gardani, Francesco, Peter Arkadiev, and Nino Amiridze 2014. Borrowed morphology: an overview. In Francesco Gardani, Peter Arkadiev, and Nino Amiridze (eds.), Borrowed Morphology. Berlin: de Gruyter Mouton, pp. 1–24. Georgalidou, Marianthi, Vassilios Spyropoulos, and Hasan Kaili 2011. Spoken varieties of Greek in the bilingual Muslim community of Rhodes. In Proceedings of the 7th International Conference on Greek Linguistics. University of York. Greenberg, Joseph H. 1963. Some universals of grammar with particular reference to the order of meaningful elements. In Joseph H. Greenberg (ed.), Universals of Human Language. Cambridge, MA: MIT Press, pp. 73–133. Güldemann, Tom and Anne‐Maria Fehn (eds.) 2014. Beyond “Khoisan.” Historical relations in the Kalahari Basin. Amsterdam: Benjamins. Haspelmath, Martin, Matthew S. Dryer, David Gil, and Bernard Comrie (eds.) 2005. The World Atlas of Language Structures. Oxford: Oxford University Press. Hawkins, John A. 1983. Word Order Universals. New York: Academic Press. Hickey, Raymond (ed.) 2017. The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press. Holton, David, Peter A. Mackridge, Irene Philippaki‐Warburton, and Vassilios Spyropoulos 2012. Greek: A Comprehensive Grammar of the Modern Language, 2nd edition. London: Routledge.

146  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman Igartua, Iván 2006. Genus alternans in Indo‐ European. Indogermanische Forschungen 111: 56–70. Janse, Mark 2002. Aspects of bilingualism in the history of the Greek language. In James Noel Adams, Mark Janse, and Simon Swain (eds.), Bilingualism in Ancient Society: Language Contact and the Written Text. Oxford: Oxford University Press, pp. 332–90. Janse, Mark 2009. Greek‐Turkish language contact in Asia Minor. Études Helleniques‐ Hellenic Studies, 17(1), 37–54. Janse, Mark, Brian D. Joseph, and Gunter De Vogelaer 2011. Changing gender systems: a multidisciplinary approach. In Gunter De Vogelaer and Mark Janse (eds) Folia Linguistica 45.2: 237–44. Johanson, Lars 2002. Structural Factors in Turkic Language Contacts. Richmond: Curzon. Karatsareas, Petros 2009. The loss of grammatical gender in Cappadocian Greek. Transactions of the Philological Society 107.2: 196–230. Karatsareas, Petros 2011. A study of Cappadocian Greek nominal morphology from a diachronic and dialectological perspective (unpublished Ph.D. dissertation, University of Cambridge). Karatsareas, Petros 2013. Understanding diachronic change in Cappadocian Greek: the dialectological perspective. Journal of Historical Linguistics 3.2: 192–229. Karatsareas, Petros 2014. On the diachrony of gender in Asia Minor Greek: the development of semantic agreement in Pontic. Language Sciences 43: 77–101. Karatsareas, Petros 2016. Convergence in word structure: revisiting agglutinative noun inflection in Cappadocian Greek. Diachronica 33: 31–66. Kontosopoulos, Nikolaos G. 1994 [2008]. ∆ιάλεκτοι και ιδιώματα της Νέας Ελληνικής [Dialects and idioms of Modern Greek] (Revised edition). Athens: Grigoris [5th edition 2008]. Koptjevskaja‐Tamm, Maria 2010. Linguistic typology and language contact. In Jae Jung Song (ed.), The Oxford Handbook of Linguistic Typology. Oxford: Oxford University Press. Kossmann, Maarten 2015. Contact‐induced change. In Matthew Baerman (ed.), The Oxford Handbook of Inflection. Oxford: Oxford University Press, pp. 251–71. Laughren, Mary 2010. Warlpiri verbs of change and causation: the thematic core. In Mengistu Amberber, Brett Baker, and Mark Harvey (eds.), Complex Predicates: Cross‐linguistic

Perspectives on Event Structure. Cambridge: Cambridge University Press, pp. 167–236. Mansfield, John 2016. Borrowed verbs and the expansion of light verb phrases in Murrinhpatha. In Felicity Meakins and Carmel O’Shannessy (eds.), Loss and Renewal: Australian Languages Since Colonisation. Berlin: de Gruyter Mouton, pp. 397–424. Matasović, Ranko 2004. Gender in Indo‐European. Heidelberg: Universitätsverlag Winter. Matras, Yaron 2007. The borrowability of grammatical categories. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross‐Linguistic Perspective. Berlin: de Gruyter Mouton, pp. 31–74. Matras, Yaron. 2009. Language Contact. Cambridge: Cambridge University Press. Matras, Yaron and Jeanette Sakel (eds). 2007a. Grammatical Borrowing in Cross‐linguistic Perspective. Berlin: de Gruyter Mouton. Matras, Yaron and Jeanette Sakel 2007b. Introduction. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross‐ Linguistic Perspective. Berlin: de Gruyter Mouton, pp. 1–13. Matras, Yaron and Jeanette Sakel 2007c. Investigating the mechanisms of pattern replication in language convergence. Studies in Language 31.4: 829–65. McConvell, Patrick 2010. Contact and indigenous languages in Australia. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐Blackwell, pp. 770–94. McConvell, Patrick and Felicity Meakins 2005. Gurindji Kriol: a mixed language emerges from code‐switching. Australian Journal of Linguistics 25.1: 9–30. Meakins, Felicity 2010. The development of asymmetrical serial verb constructions in an Australian mixed language. Linguistic Typology 14.1: 1–38. Meakins, Felicity 2011. Borrowing contextual inflection: evidence from northern Australia. Morphology 21.1: 57–87. Meakins, Felicity 2014. Language contact varieties. In Harold Koch and Rachel Nordlinger (eds.), The Languages and Linguistics of Australia: A Comprehensive Guide. Berlin, Germany: de Gruyter Mouton, pp. 365–416. Meakins, Felicity and Carmel O’Shannessy. 2012. Typological constraints on verb integration in two Australian mixed languages. Journal of Language Contact 5.2: 216–46. Meakins, Felicity and Rob Pensalfini 2016. Gender bender: superclassing in Jingulu gender marking. In Felicity Meakins and

Contact and Linguistic Typology  147 Carmel O’Shannessy (eds), Loss and Renewal: Australian Languages Since Colonisation. Berlin: de Gruyter Mouton, pp. 425–50. Meakins, Felicity and Sacha Wilmoth forthcoming. Complex cell‐mates: Morphological overabundance resulting from language contact. In Peter Arkadiev and Francesco Gardani (eds.), The Complexity of Morphology. Oxford: Oxford University Press. Michaelis, Susanne Maria, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) 2013a. The Atlas of Pidgin and Creole Language Structures. Oxford: Oxford University Press. Moravcsik, Edith 1978. Universals of language contact. In Joseph H. Greenberg (ed.), Universals of Human Language. Stanford: Stanford University Press, pp. 94–122. Morpurgo Davies, Anna 1968. Gender and the development of the Greek declensions. Transactions of the Philological Society 67, 12–36. Mous, Maarten 2001. Paralexification in language intertwining. In Norval Smith and Tonjes Veenstra (eds.), Creolization and Contact. Amsterdam: Benjamins, pp. 113–23. Mous, Maarten 2003. The Making of a Mixed Language: The Case of Ma’a/Mbugu. Amsterdam:Benjamins. Muysken, Pieter 2016. Language contact: Trojan horse or new potential for cross‐fertilization? Linguistic Typology 20.3: 537–45. doi:10.1515/ lingty‐2016–0025 Myers‐Scotton, Carol 1993. Social Motivations for Code‐switching: Evidence from Africa. Oxford: Oxford University Press. Nash, David 1982. Warlpiri verb roots and preverbs. In Stephen Swartz (ed.), Papers in Warlpiri Grammar: In Memory of Lothar Jagst. Darwin: SIL, 165–216. Nichols, Johanna 1992. Linguistic Diversity in Space and Time. Chicago: University of Chicago Press. Nichols, Johanna 2007. What, if anything, is typology? Linguistic Typology 11, 231–8. Nordlinger, Rachel 2010. Verbal morphology in Murrinh‐Patha: evidence for templates. Morphology 20.2: 321–41. O’Shannessy, Carmel 2005. Light Warlpiri: A New Language. Australian Journal of Linguistics 25.1): 31–57. O’Shannessy, Carmel 2006. Language Contact and Children’s Bilingual Language Acquisition: Learning a Mixed Language and Warlpiri in Northern Australia (University of Sydney Ph.D. dissertation). O’Shannessy, Carmel 2012. The role of code‐ switched input to children in the origin of a

new mixed language. Linguistics 50.2): 305–40. Pakendorf, Brigitte 2009. Intensive contact and the copying of paradigms: An Even dialect in contact with Sakha (Yakut). Journal of Language Contact 2, 85–110. Poplack, Shana 1980. Sometimes I’ll start a sentence in Spanish y termino en español: toward a typology of code‐switching. Linguistics 18: 581–618. Roberts, Sarah J. and Joan Bresnan 2008. Retained inflectional morphology in pidgins: A typological study. Linguistic Typology 12: 269–302. Sakel, Jeanette 2007. Types of loan: Matter and pattern. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross‐linguistic Perspective. Berlin: de Gruyter Mouton, pp. 15–30. Sasse, Hans‐Jürgen 1993. Syntactic categories and subcategories. In Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld, Theo Vennemann (eds.), Syntax: Ein internationales Handbuch zeitgenössischer Forschung [An International Handbook of Contemporary Research], vol. 1. Berlin: de Gruyter, pp. 646–86. Seaman, P. David 1972. Modern Greek and American English in Contact. The Hague: de Gruyter Mouton. Sebba, Mark 1998. A congruence approach to the syntax of codeswitching. The International Journal of Bilingualism 2.1: 1–19. Seifart, Frank 2012. The Principle of Morphosyntactic Subsystem Integrity in language contact: evidence from morphological borrowing in Resígaro (Arawakan). Diachronica 29.4, 471–504. Seifart, Frank 2013. AfBo: A world‐wide survey of affix borrowing. Leipzig: Max Planck Institute for Evolutionary Anthropology. Avalable at http://afbo.info. Seifart, Frank 2015. Direct and indirect affix borrowing. Language 91.3: 511–31. doi:10.1353/ lan.2015.0044 Seifart, Frank 2017. Patterns of affix borrowing in a sample of 100 languages. Journal of Historical Linguistics 7.3: 389–431. doi:10.1075/ jhl.16002.sei Shibatani, Masayoshi and Theodora Bynon (eds.) 1995. Approaches to Language Typology. Oxford: Oxford University Press. Sims‐Williams, Helen and Matthew Baerman forthcoming. A typological perspective on the loss of inflection. In Svenja Kranich and Tine Breban (eds.), Lost in Change: Causes and

148  Oliver Bond, Helen Sims‐Williams, and Matthew Baerman Processes in the Loss of Grammatical Constructions and Categories. Amsterdam: Benjamins. Street, Chester 1987. An Introduction to the Language and Culture of the Murrinh‐Patha. Darwin: Summer Institute of Linguistics. Thomason, Sarah G, and Terrence Kaufman 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley, CA: University of California Press. Thornton, Anna M. 2011. Overabundance (multiple forms realizing the same cell): a non‐ canonical phenomenon in Italian verb morphology. In Martin Maiden, John Charles Smith, Maria Goldbach, and Marc‐Olivier Hinzelin (eds.), Morphological Autonomy: Perspectives from Romance Inflectional Morphology. Oxford: Oxford University Press, pp. 358–81. Tomlin, Russell 1985. Basic Word Order: Functional Principles. London: Croom Helm. Vassalou, Nikoleta, Dimitrias Papazachariou, and Mark Janse forthcoming. The Vowel

System of Mišótika Cappadocian. In Thanasis Georgakopoulos and Theodossia‐Soula Pavlidou (eds.), Proceedings of the 12th International Conference on Greek Linguistics. Newcastle: Cambridge Scholars Publishing. Walsh, Michael 1976. The Murinypata language of north‐west Australia (Australian National University Ph.D. thesis). Weinreich, Uriel 1953. Languages in Contact: Findings and Problems. The Hague: de Gruyter Mouton. Winford, Donald 2010. Contact and borrowing. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐Blackwell, pp. 170–87. Winford, Donald 2005. Contact‐induced changes: classification and processes. Diachronica 22.2: 373–427. doi:10.1075/ dia.22.2.05win Wohlgemuth, Jan 2009. A Typology of Verbal Borrowings. Berlin: de Gruyter Mouton.

7 Contact and Language Shift RAYMOND HICKEY 1 Introduction Language documentation, available for at least the past two millennia, is testimony to a plethora of contact situations varying at least according to initial conditions, the social factors operative during contact, the number and typology of the languages involved, and the attested outcome of contact. Such situations can be classified on the basis of shared features and in any taxonomy language shift will be prominent (Hickey 2020). A shift scenario, at its most basic, is defined by contact in which a community initially speaks language A and at a later point in time has moved to language B. Close in type are situations in which two or more languages become typical for a community, i.e. bilingualism / multilingualism, or cases in which a language ceases to be spoken – language death – either by all its speakers dying out1 or by a community switching entirely to another language.2 For the latter scenario just consider the early Indo‐European migrations. Movements of subgroups of this family into new geographical locations usually meant that the pre‐Indo‐European populations were “absorbed,” i.e. that they shifted in language (and culture) to the branch of Indo‐ European they were exposed to (Drinka, this volume). This shift may have been partial or complete, for instance, on the Iberian peninsula it was partial with Basque remaining but in the British Isles it was complete, all pre‐Indo‐European languages having disappeared before written records became available (Eska, this volume). Cases of language shift may have lasted into history, making the “absorption” more visible, as was the case with Etruscan in Italy or Gaulish in area of present‐day France. Whether the Indo‐European branches still show traces of this early contact and shift is much disputed (Vennemann 2010). But going on shift scenarios today and assuming that the same principles of contact applied then as now, one can assume that an effect of earlier groups on later groups took place if the size of the shifting population was sufficient for the features of its shift variety to influence the language they were shifting to. This is not always the case, however, so a note of caution should be struck here. Moving forward to recent history one can see in the Anglophone world that language shift did not always leave traces of the original language(s). The considerable shift of native Americans to English has not affected general forms of English in either the USA or Canada. What may occur is that the shift variety establishes itself as a form in its own right, as a focused variety with a stable speech community, cf. South African Indian English (Mesthrie 1992), but even then there is usually a further approximation to supraregional forms of English which reduce the specific profile of the shift variety, cf. Australian Aboriginal English and Maori English in New Zealand.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

150  Raymond Hickey

1.1  Language Shift Varieties as a Typological Class? Research in recent years into varieties of English have posed the question of whether there are certain types which in principle form a typological class, notably creoles (McWhorter 1998, 2008). The “exceptionalist” view of creoles (McWhorter 2008: 289, 2018; Bakker, Daval‐ Markussen, Parkvall, and Plag 2011; Parkvall 2008) revolves around the fact that creoles are new languages created from input pidgins, themselves simplified forms of a lexifier language combined with one or more indigenous substrates. It is the scenario of language creation, typically arising on slave plantations, which has led authors to make claims such as the following: “creoles being less grammatically complex … reflect … more closely than older languages the basic core of the human language capacity” (McWhorter 2008: 289). While the scenario of segregation and isolation, which obviously obtained on slave‐based plantations, formed a social situation not replicated elsewhere, the language shift environment of non‐prescriptive language acquisition, usually as a second language in adulthood, is arguably close to that in which creoles arose (see the detailed discussion in Hickey 1997a). In the words of Mesthrie (1996, 95) “language shift varieties may well provide ‘a missing link in the chain of possible contact varieties’…, especially between creole and non‐creole.” In the context of creoles, the central question has been whether there are certain key features which are characteristics of all, or nearly all, creoles and which can be traced to their essential nature as new languages. Underlying this question is the assumption that new languages would embody universals of human language, core features which, with the increasing age of languages, opacify or disappear. Quite often the features are not new but represent certain values for parameters in languages, such as SVO (subject‐verb‐object) for the parameter “word order” in normal declarative sentences.3 The other word orders attested in the world’s languages today would thus be seen as resulting from other forces, such as topicalization or the post‐position of qualifiers, operating over time and leading to orders other than SVO. There have been proponents and opponents of the view of creoles as special instances of new languages. Among the former are Derek Bickerton (Bickerton 1975, 1984) and John McWhorter (McWhorter 1998, 2008, 2012) and among the latter one finds Salikoko Mufwene (Mufwene 2001) and Michel DeGraff (DeGraff 2003, 2005). In an attempt to relativize the prototypical view of creoles recent research has examined examples from outside the two geographical regions from which most attestations derive, the Atlantic and the Pacific arenas, see Buchstaller, Holmberg, and Almoaily (2014). For the present study, data is presented from the language shift from Irish to English over the past few centuries. A re‐consideration of the question of whether shift varieties represent a typological class comparable to that assumed by its proponents for creoles is offered in the conclusion. The data for the current study is largely taken from the present, i.e. long after the shift took place for the majority of Ireland’s population. However, diachronic material is available to attain some time depth for the issue of language shift. Nonetheless, certain factors must be left unaddressed because of a lack of suitable data for an assessment. For instance, the classic study of language shift from Hungarian to German in Oberwart in the Burgenland, a border region of Hungary and Austria (Gal 1979), considers the differential behavior of men and women in the shift situation. But no such considerations are possible for Irish and Irish English given the dearth of relevant data for this question.

1.2  Motivation for Shift The motivation for language shift and the circumstances under which it takes place will vary from case to case but there is sufficient common ground for generalizations to be made about

Contact and Language Shift  151 language shift and for the analysis of a single instance to be of broader value to the study of language shift as a whole. For the following discussion the language shift which took place in Ireland, roughly ­between the early seventeenth century and the late nineteenth century, will be considered. This is a shift from the original language of the vast majority in Ireland, Irish, to English, a language which was imported to Ireland in the late twelfth century and which is now (early twenty‐first century) the language of over 99% of the Irish population, in both the north and south of the country,4 English has not always been the dominant language in Ireland. Initially, this was Anglo‐ Norman and then Irish so that in the fourteenth and fifteenth centuries, English had receded to the east coast and was only found in any strength in the towns. But after significant

Areas with largest Irish-speaking populations

Brought by seventeenth-century planters

Ulster Scots

Mid-Ulster English

Donegal Derry

Belfast Sligo Dundalk Westport Midlands Connemara

Dublin

Galway

East coast dialect area Limerick Kerry

Tralee

Waterford Cork

Southwest and west

Forth and Bargy archaic dialect, died out early nineteenth century

Map 7.1  Language and dialect divisions in Ireland.

Area of original settlement in late twelfth century by Anglo-Normans

152  Raymond Hickey victories for the English at the beginning of the seventeenth century, after the settlement of the north with Lowland Scots and people from northern and northwestern England and after the Cromwellian campaigns and settlements of the 1640s and early 1650s, the fortunes of the Irish language and culture declined and were to wane steadily in the centuries between that time and the present. Currently, Irish is spoken natively by not more than 50,000 people (probably by fewer) in three main pockets on the western seaboard. The Irish spoken by these individuals is very strongly influenced by English (see Section 8). But in the main, this chapter will deal with the rise of vernacular varieties of Irish English during the shift period of the past few centuries. The information presented here is intended to highlight key aspects of language shift. For reasons of space nothing like a comprehensive treatment can be offered. A much more detailed analysis of contact phenomena can be found in the central section of Hickey (2007, ch. 4).

1.3  The Nature of the Shift The shift in Ireland must have involved considerable bilingualism over several centuries. The native language for the majority of the population was initially Irish and recourse to this was always there. English would have been used in contact with English speakers (administrators, bailiffs, or those few urbanites who only spoke English). There was also considerable interaction between the planters (English/Scottish settlers) and the native Irish, certainly in the countryside where this group of English speakers had settled. Indeed there may be grounds for assuming that a proportion of the planters by the mid‐seventeenth century would have had at least a rudimentary knowledge of Irish. They would have been a source of bilingualism for the native Irish population, at the interface between themselves and those planters without any Irish. However, this source of bilingual interaction should not be overestimated. There would seem to be little evidence for the view that key features of Irish English arose through the interaction with bilingual people of English origin. The planters in the south of Ireland5 numbered a few thousand at the most while there were several ­million Irish speakers, probably between 7 and 8 million before the onset of the Great Famine (1845–8). Furthermore, the view that the planters were cared for by Irish nurses and had contact with the children of the native Irish is supported by authors like Bliss (1976: 557). The ultimate effect of this would have been to render the language of the planters more like that of the native Irish so that no specific variety of planter English arose. The language shift did not progress evenly across the centuries. Major external events, chiefly famine and emigration, accelerated the pace. During such setbacks, Irish quickly lost ground which it was not to recover. Famine struck throughout the eighteenth century, especially in the 1720s and in 1740–1, and emigration from Ulster was considerable during this century, though this largely involved settlers of Scottish origin who moved to North America. The most significant blow to the Irish language was the Great Famine of the late 1840s which hit the poorer rural areas of Ireland hardest. The twin factors of death and emigration reduced the number of Irish speakers by anything up to 2 million in the course of less than a decade. The famine also brought home to the remaining Irish speakers the necessity to switch to English to survive in an increasingly English‐speaking society and to prepare for possible emigration.

2  What Can Be Traced to Contact? It goes without saying that there is no proof in contact linguistics. If a structure in one language is suspected of having arisen through contact with another, then a case can be made for contact when there is a good structural match between both languages. Take as an example the phrases at the beginning of the following sentences6 which have an exact equivalent in Irish:

Contact and Language Shift  153 (1) a. More is the pity, I suppose. (TRS‐D, S42, M) Is mór an trua, is dóigh liom. [is big the pity, is suppose with‐me] b. Outside of that, I don’t know. (TRS‐D, W42‐2, F) Taobh amuigh de sin, níl a fhios agam. [side out of that, not‐is know at‐me] c. There’s a share of jobs alright. (TRS‐D, S7, M) Tá roinnt jabanna ann, ceart go leor. [is share jobs‐GEN in‐it right enough] However, the case for contact as a source, at least as the sole source, is considerably weakened if the structure in question is attested in older forms of the language which has come to show it. There are features in Irish English of this type, that is they could have a source either in older forms of English taken to Ireland or in Irish through contact. An example of this is provided by doubly marked comparatives. In Irish, comparatives are formed by pre‐placing the qualifier níos ‘more’ and inflecting the adjective as well. For instance, déanach ‘late’, which consists of the stem déan‐ and the stem‐extending suffix ‐ach, changes to déanaí in the ­comparative although the comparative particle níos is used as well: (2) Beimid ag teacht níos déanaí. ‘We will be coming later.’ [will‐be‐we at coming more later] This double marking may have been transferred in the language shift situation. But such marking is also typical of earlier forms of English (Barber 1997 [1976]: 200–1) and may well have been present in input forms of English in Ireland. It is still well attested today as in the following examples: (3) a. He’s working more harder with the new job. (WER, F50+) b. We got there more later than we thought. (DER, M60+) In such cases it is impossible to decide what the source is, indeed it is probably more sensible to postulate a double source, and to interpret the structure as a case of convergence.

3  The Search for Categorial Equivalence An essential distinction in any consideration of contact‐induced change is the difference between the presence of a category in a certain language and the exponence of this category. For instance, the category “future” exists in the verb systems of both English and Irish but the exponence is different, i.e. via an auxiliary will/shall in the first language, but via a verb‐stem suffix in the second language. This type of distinction is useful when comparing Irish English with Irish, for instance when comparing habitual aspect in both languages, as can be seen from Table 7.1.

3.1  Category and Exponence When shifting to another language, temporarily or permanently, adults expect the same grammatical distinctions in the target which they know from their native language. To this end they search for equivalents in the target language to categories they are already familiar with. This process is an unconscious one and persists even with speakers who have considerable target language proficiency. If the categories of the outset language are semantically

154  Raymond Hickey Table 7.1  Category and exponence in Irish and Irish English. Category

Exponence in Irish English

Exponence in Irish

Habitual

1.  do(es) be + V‐ing They do be fighting a lot. 2.  bees (northern) The lads bees out a lot. 3.  verbal ‐s (first person) I gets tired of waiting for things to change.

bíonn + nonfinite verb form Bíonn siad ag troid go minic. [is‐HAB they at fighting often]

motivated then the search to find an equivalent in the target is all the more obvious. Here is an illustration. In Irish there is a distinction between the second person singular and plural pronoun but not in standard English. In the genesis of Irish English, speakers would seem to have felt the need for this nonexistent distinction in English and three solutions to this quandary arose: (4) a. the use of available material, yielding you # ye (ye available from early English input) b. the analogical formation of a plural: you # youse < you + {S} (not attested before early to mid‐nineteenth century) c. a combination of both (a) and (b) as in you # yez < ye + {S} (not found before mid‐nineteenth century) In all these cases the search for an equivalent category of second person plural was solved in English by the manipulation of material already in this language. At no stage does the Irish sibh [∫ivʲ] ‘you‐PL’ seem to have been used, in contrast, for instance, to the use of West African unu ‘you‐PL’ found in Caribbean English (Hickey 2003). Apart from restructuring elements in the target, speakers can transfer elements from their native language. This transfer of grammatical categories is favored, if the following conditions apply: 1. 2. 3. 4.

The target language has a formal means of expressing this category. There is little variation in the expression of this category. The expression of this category is not homophonous with another one. The category marker in the outset language can be identified  –  is structurally ­transparent – and can be easily extracted from source contexts.

Before looking at a case where transfer did actually take place, one where it did not is presented as it can be seen that the complete lack of equivalence precluded any transfer to English. Irish has a special form of the verb, known as the “autonomous”, [AUT] in the glosses in (5), a finite verb which is not bound to a particular person, i.e. which is agentless: (5) a. Táthar ann a cheapann go bhfuil an ceart aige. [is‐AUT in‐it that think that is the right at‐him] ‘There are people who think he is right.’ b. Bristear an dlí go minic. [break‐AUT the law often] ‘The law is often broken.’

Contact and Language Shift  155 c. Cailleadh anuraidh í. [lost‐AUT she last‐year] ‘She died last year.’ d. Rugadh mac di. [born‐AUT son to‐her] ‘She bore a son.’ Neither in present‐day contact English nor in the textual record for Irish English is a direct transfer of the autonomous form of Irish attested. Agentless finite verb forms are unknown in English and, furthermore, the means of expressing agentivity in Irish, via a compound form of preposition + pronoun  –  see example (5d)  –  is not available in English either. Instead the internal means found in English, the passive, generic sentences with there, are and were used.

3.2  Attested Cases of Shift Where transfer is attested it is worth considering just how this may have taken place. In a language shift situation, transfer must first occur on an individual level, perhaps with ­several individuals at the same time. But for it to become established, it must be accepted by the community as a whole. If such transfer is to be successful, then it must adhere to the principle of economy: it must embody only as much change in the target as is necessary for other speakers in the community to recognize what native structure it is intended to reflect. To illustrate how this process of transfer is imagined to have occurred in the historical Irish context, consider the example of the immediate perfective formed by using the prepositional phrase tar éis ‘after’. (6) Tá siad tar éis an obair a dhéanamh. [is they after the work comp do] ‘They are after doing the work’, i.e. ‘They have completed the work.’ The pivotal elements in this construction are listed in (7); the complementizer a is of no semantic significance: (7) a. adverbial phrase tar éis ‘after’ b. nonfinite verb form déanamh ‘doing’ c. direct object obair ‘work’ It would appear that the Irish constructed an equivalent to the output structure using English syntactic means. Item (a) was translated literally as ‘after’, (b) was rendered by the nonfinite V‐ing form yielding sentences like They’re after doing the work. With a translation for tar éis and a corresponding nonfinite form the task of reaching a categorial equivalent would appear to have been completed. Importantly, the Irish word order “object + verb” was not carried over into English (*They’re after the work doing). In putting the case against transfer Harris (1991: 205) argues that the order of nonfinite verb form and object is different in Irish and English and hence that transfer is unlikely to have been the source. However, the aim in the contact situation was to arrive at a construction which was functionally equivalent to that in the outset language. A word order such as that in John is after the house selling would not only unnecessarily flout the sequence of verb and object in English (unnecessary as it would not convey additional information) but also give rise to possible confusion with the resultative perfective which in Irish English is realized by means of a past participle following its object.

156  Raymond Hickey In the transfer of structure during language shift, it would seem both necessary and sufficient to achieve correlates to the key elements in the source structure. Another instance of this principle can be seen with the resultative perfective of Irish English. (8) Tá an obair déanta acu. ‘They have finished the work.’ [is the work done at‐them] IrEng: ‘They have the work done.’ Essential to the semantics of the Irish construction is the order “object + past participle.” Consequently, it is this order which is realized in the Irish English equivalent. The prepositional pronoun acu ‘at‐them’ (or any other similar form) plays no role in the formation of the resultative perfective in Irish, but is the means to express the semantic subject of the sentence. As this is incidental to the perfective aspect expressed in the sentence, it was neglected in Irish English. The immediate perfective with after does not appear to have had any model in archaic or regional English (Filppula 1999: 99–107). With the resultative perfective, on the other hand, there was previously a formal equivalent, i.e. the word order “object + past participle.” However, even if there were instances of this word order in the input varieties of English in Ireland this does not mean that these are responsible for its continuing existence in Irish English. This word order could just as well have disappeared from Irish English as it had in forms of mainland English (van der Wurff and Foster 1997). However, the retention in Irish English and the use of this word order to express a resultative perfective can in large part be accounted for by the wish of Irish learners of English to reach an equivalent to the category of resultative perfective which they had in their native language. Another issue to consider is whether the structures which were transferred still apply in the same sense in which they were used in previous centuries. It would be too simplistic to assume that the structures which historically derive from Irish by transfer have precisely the same meaning in present‐day Irish English. For instance, the immediate perfective with after has continued to develop shades of meaning not necessarily found in the Irish original as Kallen (1989) has shown in his study.

4  The Prosody of Transfer The case for contact should be considered across all linguistic levels. In particular it is beneficial to consider phonological factors when examining syntactic transfer (Hickey 1990: 219). If one looks at structures which could be traced to transfer from Irish, then one finds in many cases that there is a correspondence between the prosodic structures of both languages. To be precise, structures which appear to derive from transfer show the same number of feet and the stress falls on the same major syntactic category in each language (Hickey 1990: 222). A simple example can illustrate this. Here the Irish equivalent is given, although that is not of course the immediate source of this actual sentence as the speaker was an English‐speaking monolingual. (9) A … don’t like the new team at all at all. (WER, M55+) [ˌ ˈ ˌ ˈ] Ní thaitníonn an fhoireann nua le hA … ar chor ar bith. [ˌ ˈ ˌ ˈ] [not like the team new with A … on turn on anything]

Contact and Language Shift  157 The repetition of at all at all creates a sentence‐final negator which consists of two stressed feet with the prosodic structure WSWS (weak‐strong weak‐strong) as does the Irish structure ar chor ar bith. This feature is well established in Irish English and can already be found in the early nineteenth century. Consider now the stressed reflexives of Irish which are suspected by many authors (including Filppula 1999: 77–88) of being the source of the Irish English use of an unbound reflexive. (10) ˌAn ˈbhfuil [INTERROG is ‘Is he himself in today?’ IrEng: ‘ˌIs ˌhimˈself ˈin ˌtoˈday?’

ˌsé ˈfhéin he self

ˌisˈtigh in

ˌinˈniu? today]

The strong and weak syllables of each foot are indicated in the Irish sentence and its Irish English equivalent in (10). From this it can be seen that the Irish reflexive is monosyllabic and, together with the personal pronoun, forms a WS foot: ˌsé ˈfhéin [he self]. In Irish English the equivalent to this consists of a reflexive pronoun on its own: ˌhimˈself, hence the term “unbound reflexive” as no personal pronoun is present. If both the personal and reflexive pronoun were used in English, one would have a mismatch in prosodic structure: WS in Irish and SWS (ˈhe ˌhimˈself) in Irish English. One can thus postulate that the WS pattern of ˈhimˈself was interpreted by speakers during language shift as the prosodic equivalent of both the personal pronoun and reflexive pronoun of Irish ˌsé ˈfhéin and thus used as an equivalent of this.7 Another example of prosodic match can be seen with the immediate perfective of Irish English (discussed on p. 155) which corresponds, in the number of stressed syllables, to its Irish equivalent. (11) a.

b.

She’s after breaking the glass. [ˈ ˈ  ˌ ˈ ˌ ˌ ˈ] Tá sí tréis an ghloine a bhriseadh. [ˈ ˌ  ˈ ˌ ˈ ˌ]  ˌ ˌ ˈ He’s after his dinner. [ˈ ˈ ˌ ˌ ˈ ˌ] Tá sé tréis a dhinnéir. [ˈ ˌ ˈ ˌ ˈ ˌ]

3 stressed syllables (feet) 3 stressed syllables (feet) 2 stressed syllables (feet) 2 stressed syllables (feet)

This consists in both languages of three or two feet depending on whether the verb is understood or explicitly mentioned (it is the number of stressed syllables which determines the number of feet). In both languages a stressed syllable introduces the structure and others occur for the same syntactic categories throughout the sentence. A similar prosodic correspondence can be recognized in a further structure, labeled “subordinating and” (Klemola and Filppula 1992), in both Irish and Irish English. (12) a. He went out ‘He went out b. Chuaigh sé amach [went he out

ˈand ˈit ˈraining. although it was raining.’ ˈagus ˈé ag cur ˈbáistí. and it at putting rain‐GEN]

Again there is a correlation between stressed syllable and major syntactic category, although the total number of syllables in the Irish structure is greater (due to the number of weak syllables). The equivalence intonationally is reached by having the same number of feet, i.e. stressed syllables, irrespective of the distance between them in terms of

158  Raymond Hickey intervening unstressed syllables. And again, it is a stressed syllable which introduces the ­subordinate clause. Another case, where prosodic equivalence can be assumed to have motivated a nonstandard feature, concerns comparative clauses. These are normally introduced in Irish by two equally stressed words ˈná ˈmar ‘than like’ as in the following example: (13) Tá sé i bhfad níos fearr anois [is it further more better now ‘It’s now much better than it was.’

ˈná ˈmar not like

a bhí. COMP was]

Several speakers from Irish‐speaking regions, or those which were so in the recent past, show the use of than what to introduce comparative clauses: (14) a. b. c. d.

It’s far better than what it used to be. (TRS‐D, W42‐1, F) To go to a dance that time was far better than what it is now. (TRS‐D, W42‐1, F) Life is much easier than what it was. (TRS‐D, W42‐1, F) They could tell you more about this country than what we could. (TRS‐D, S7, M)

It is true that Irish mar does not mean ‘what’, but what can introduce clauses in other instances and so it was probably regarded as suitable to combine with that in cases like those in (14). From the standpoint of prosody, ˈthan ˈwhat provided a combination of two equally stressed words which match the similar pair in equivalent Irish clauses. The use of than what for comparatives was already established in the nineteenth century and is attested in many emigrant letters such as those written from Australia back to Ireland, e.g. the following in a letter from a Clare person written in 1854: I have more of my old Neighbours here along with me than what I thought (Fitzpatrick 1994: 69). It is also significant that the prosodically similar structure like what is attested in the east of Ireland where Irish was replaced by English earliest, e.g. There were no hand machines like what you have today (speaker from Lusk, Co. Dublin).

5  Coincidental Parallels Despite the typological differences between Irish and English there are nonetheless a number of unexpected parallels which should not be misinterpreted as the result of contact. Some cases are easy, such as the homophony between Irish sí /∫i:/ ‘she’ and English she (the result of the vowel shift of /e:/ to /i:/ in early modern English). A similar homophony exists for Irish bí ‘be’ and English be, though again the pronunciation of the latter with /i:/ is due to the raising of English long vowels. Other instances involve parallel categories, e.g. the continuous forms of verbs in both languages: Tá mé ag caint léi [is me at talk‐NONFINITE with‐her] ‘I am talking to her.’ Indeed the parallels among verbal distinctions may have been a trigger historically for the development of nonstandard distinctions in Irish English, i.e. speakers during the language shift who found equivalents to most of the verbal categories from Irish expected to find equivalents to all of these. An example of this is habitual aspect, which is realized in Irish by the choice of a different verb form (bíonn [habitual] versus tá [non‐habitual]): (15) Bíonn sé ag caint léi. ‘He talks to her repeatedly.’ [is‐HABITUAL he at talking with‐her] IrEng: ‘He does be talking to her.’

Contact and Language Shift  159 Another coincidental parallel between the two languages involves word order, despite the differences in clause alignment which both languages show. In both Irish and English prepositions may occur at the end of a clause. A prepositional pronoun is the most likely form in Irish because it incorporates a pronoun which is missing in English. (16) An buachaill a raibh mé ag caint leis. [the boy that was I at talking with‐him] ‘The boy I was talking to.’ Further parallels may be due to contact which predates the coming of English to Ireland. For example, the use of possessive pronouns in instances of inalienable possession is common to both English and Irish: (17) Ghortaigh sé a ghlúin. [injured he his knee] ‘He injured his knee.’ This may well be a feature of Insular Celtic which was adopted into English, especially given  that other Germanic languages do not necessarily use possessive pronouns in such contexts, cf. German Er hat sich am Knie verletzt, lit. ‘He has himself at‐the knee injured’ (Hickey 2012a, 2012b).

6  What Does Not Get Transferred? If the expectation of outset language categories in the target language is a guiding principle in language shift, then it is not surprising to find that grammatical distinctions which are only found in the target language tend to be neglected by speakers undergoing the shift. The reason for this neglect is that speakers tend not to be aware of grammatical distinctions which are not present in their native language, at least this is true in situations of unguided adult learning of a second language. What is termed here “neglect of distinctions” is closely related to the phenomenon of underdifferentiation which is known from second language teaching. This is the situation in which second language learners do not engage in categorial distinctions which are present in the target language, for instance when German speakers do not distinguish between when and if clauses in English (both take wenn in German). This neglect can be illustrated by the use of and as a clause co‐ordinator with a qualifying or concessive meaning in Irish English (see remarks on subordinating and p. 157): (18) Chuaigh sé amach agus é [went he out and IrEng: ‘He went out and it raining.’ ‘He went out although it was raining.’

ag it

cur at putting

báistí. rain‐GEN]

To account for the neglect of distinctions in more detail, one must introduce a distinction between features which carry semantic value and those which are of a more formal character. Word order is an example of the latter type: Irish is a consistently post‐specifying language with VSO (verb‐subject‐object) as the canonical word order along with Noun + Adjective, Noun + Genitive for nominal modifiers. There is virtually no trace of post‐specification in Irish English, either historically or in present‐day contact varieties of English in Ireland. The use of the specifically Irish word order would, per se, have had no informational value for Irish speakers of English in the language shift situation.

160  Raymond Hickey Table 7.2  Convergence scenarios in the history of Irish English. 1a Source 1 (English) independent of Source 2 (Irish) English development Irish What are you reading? Céard atá tú a léamh? [what that‐is you COMP reading] Outcome in Irish English: continuous verb phrases maintained 1b Source 1 (English) independent of Source 2 (Irish) older English input Irish Are ye ready? An bhfuil sibh réidh? [INTERROG are you‐PL ready] Outcome in Irish English: Distinct second person plural pronoun maintained 2 Source 1 (English) provides form and Source 2 (Irish) semantics English input Irish (periphrastic/emphatic) Bíonn sé amuigh ar an bhfarraige. He does live in the west. [is‐HABITUAL he out on the sea] Outcome in Irish English: Habitual is established, He does be out on the sea. 3 Failed convergence: Source 1 (dialectal English) shares feature with Source 2 (Irish) English input Irish They were a‐singing. Bhí siad ag canadh. [were they at singing] Outcome in Irish English: A‐prefixing does not establish itself

Another example, from a different level of language, would be the distinction between palatal and nonpalatal consonants in Irish phonology. This difference in the articulation of consonants lies at the core of the sound structure of Irish. It has no equivalent in English and the grammatical categories in the nominal and verbal areas which it is used to indicate are realized quite differently in English (by word order, use of prepositions, suffixal inflections, etc.). An awareness of the semantic versus formal distinction helps to account for other cases of non‐transfer from Irish. For instance, phonemes which do not exist in English, such as /x/ and /ɣ/, have not been transferred to English, although there are words in Irish English, such as taoiseach ‘prime minister’, pronounced [ˈt ̪i:∫ǝk], with a final [‐k] and not [‐x], which could have been carriers of the [x] sound into English during the language shift period. Although the /k/ versus /x/ distinction is semantically relevant in Irish, it would not be so in English and hence transfer would not have helped realize any semantic distinctions in the target language.8 A further conclusion from these considerations is that the source of a sound like /x/ in Ireland can only be retention from earlier varieties of English. This explains its occurrence in Ulster Scots and in some forms of Mid‐Ulster English, but also its absence elsewhere, although it is present in nearly all dialects of Irish (Hickey 2011, 2014).

7  Interpreting Vernacular Features There has been much discussion of the role of English input and transfer from Irish in the genesis of Irish English, most of which has centered around suggested sources for vernacular features. Some developments in English ran parallel to the structure of Irish and so appeared in Irish English not so much by transfer, which implies a mismatch between outset and target language, but simply by equivalence. One such development of the later modern period (Beal 2004: 77–85) is the be + V‐ing (progressive) construction as in What are you reading? This would have represented an appropriate equivalent to Irish Ceárd atá tú a léamh? [what that‐is

Contact and Language Shift  161 you COMP reading] or Ceárd atá á léamh agat? [what that‐is at‐its reading at‐you]. Another development is the rise of group verbs (phrasal verbs – transitive and intransitive, prepositional verbs and phrasal‐prepositional verbs, Denison 1998: 221). These types of verb occur widely in Irish, e.g. Ná bí ag cur isteach orthu [not be at put in on‐them] ‘Don’t be disturbing them.’ Indeed calques on the English phrasal and prepositional verbs are a major source of loans from English into Irish today (see p.163). The cases just cited represent instances of convergence, i.e. developments in two ­languages which result in their becoming increasingly similar structurally. Convergence can be understood in another sense which is relevant to the genesis of specific features of Irish English. This is where both English input and transfer from Irish have contributed to the rise of a feature. This applies to the do(es) be habitual where English input provided periphrastic do and Irish the semantics of the structure and its co‐occurrence with the expanded (‐ing) form. Mention should also be made of features which exist in Irish and in nonstandard varieties of English in England, but not, curiously, in Irish English. The best example of this is a‐prefixing (see (3) in Table 7.2). This is recorded for southwestern British English, e.g. I be a‐singing (Elworthy 1877: 52–3, West Somerset). Such structures look deceptively Irish: the sentence could be translated directly as Bím ag canadh [is‐HABITUAL‐I at singing]. However, a‐prefixing does not occur in modern Irish English and is not attested in the textual record of the past few centuries to any significant extent. Montgomery (2000) is rightly skeptical of a possible Celtic origin of this feature, contra Dietrich (1981) and Majewicz (1984) who view the transfer interpretation favorably.

8  The Influence of English on Contemporary Irish The discussion thus far has concerned transfer from Irish to English during the formative period of Irish English. In present‐day Ireland the Irish language has no influence on English but the reverse is very much the case. There are virtually no monoglots of Irish left, except perhaps for very few rural speakers of traditional dialect in the three remaining pockets of historically continuous Irish in the northwest, west, and southwest of Ireland (see Map 7.1 p.151). The remainder of the native speakers are good bilinguals with a command of English almost indistinguishable from English speakers in Ireland. Given that the social position of Irish is precarious, despite official support from the government and many additional institutions, the pressure of English on the language is very considerable. After all one has on the one hand a rural language spoken by less than 50,000 and on the other the dominant world language. This situation means that English exercises a significant influence on the structure of Irish, something which set in during the nineteenth century (Stenson 1993). For native speakers the influence is not so much felt in phonology or in morphology, given the considerable differences between the two languages on these levels. Furthermore, the lexis of Irish has many loans from English which go back to the late Middle Ages (Hickey 1997b) and which have been adapted to Irish. The lexical influence of English is obvious in code‐switching (Stenson 1991), i.e. the direct use of English words in Irish sentences, and in obvious calques: (19) a. Níl muid an‐happy faoi. [not‐is we very‐happy under‐it] ‘We are not very happy about it.’ b. Croí‐bhriste a bhí sí mar gheall ar an toradh. [heart‐broken COMP was she on account of the result] ‘She was heart‐broken over the result.’

162  Raymond Hickey Pragmatic markers, such as well, just, now, really, are commonly inserted into Irish sentences, either at the beginning or end or at a clause break: (20) a. Tá sé níos diocra, just, ná mar a cheap mé. (CCE‐W, M65+) [is it more difficult just than what COMP thought I] ‘It’s more difficult, just, than I thought.’ b. Well, tá mé ag súil leis an earrach now. (CCE‐S, M60+) [well is I at looking with the spring now] ‘Well, I am looking forward to spring now.’

8.1  Transfer and Typological Fit In syntax, the influence of English is strongest, despite the typological differences between the two languages.9 There are certain structural parallels between Irish and English, which facilitate the transfer of English patterns into Irish, points of typological fit between the two languages. English phrasal verbs and verbs with prepositional complements are particularly common in Irish (Stenson 1997; Veselinović 2006). Often they are translated (21a) or they are integrated into Irish by having the productive verb‐forming ending ‐áil attached (21c): (21) a. Bhí sí déanta suas mar cailleach. (CCE‐W, F55+) [was she done up as a witch] ‘She was done up as a witch.’ b. Sheasfadh sé amach i d’intinn. (CCE‐W, F55+) [stand‐COND it out in your‐mind] ‘It would stand out in your mind.’ c. Ná bí ag rusháil back amáireach. (CCE‐W, F55+) [not be at rushing back tomorrow] ‘Don’t be rushing back tomorrow.’ Typical word order in Irish has changed in some cases under the influence of English. Previously, it was normal to find adverbials in phrase‐ and sentence‐final position (indicated in parentheses below). Under the influence of English, adverbs (underlined below) are drawn closer to the elements they modify. This can be a verb (22a) or a predicative adjective (22b): (22) a. Ní fhaca mé riamh rud mar sin (riamh). [not saw I ever thing like that (ever)] ‘I never saw anything like that.’ b. Tá a seanathair fós beo (fós). [is her grand‐father still alive (still)] ‘Her grandfather is still alive.’ This pattern also applies to the order of verb objects. Direct objects previously occurred after prepositional objects in final position, but it is increasingly common to find the order typical of English, namely direct object + prepositional object: (23) Chonaic mé í thíos ar an trá [í]. [saw I her down on the strand (her)] ‘I saw her down on the strand.’ A further instance is the position of interrogative elements which can occur word‐finally in English and are found more and more often in this position in Irish:

Contact and Language Shift  163 (24) a. Chun ceann nua a dhéanamh, nó céard? [in‐order‐to one new COMP do or what] ‘In order to do a new one, or what?’ b. Clúdódh sé céard? [cover‐COND it what] ‘It would cover what?’ The readiness of Irish to adopt syntactic patterns of English is also seen in direct translations of English idioms. These are usually translated word for word, something which is possible in quite a number of cases: (25) a. Thóg sé tamall fada, ceart go leor. [took it time long right enough] ‘It took a long time sure enough.’ b. Bhí orm súil a choinneáil ar an am. [was on‐me eye COMP keep on the time] ‘I had to keep an eye on the time.’ c. Caithfidh tú d’intinn a dhéanamh suas. [must you your‐mind COMP make up] ‘You have to make up your mind.’ Sentence‐initial absolute constructions are also increasingly common, introducing a type of sentence which is more typical of English than of traditional Irish: (26) a. Ag fanacht san iarthar, dúirt an t‐aire stáit inné go … ‘Staying in the west, the minister of state said yesterday that …’ b. Le bheith fírinneach, níl ach droch‐sheans ann. ‘To be truthful, there is only a slight chance of it.’ The examples just discussed show how permeable the syntax of Irish is, despite the obvious typological differences between Irish and English (VSO word order, post‐modification). Such examples have occurred between the two languages through contact, not through shift (they are found with speakers who continue to use their native Irish). This situation of contact with the super‐dominant language English has meant that Irish has been influenced in many subtle, infiltrating ways as well. Speakers establish lexical equivalences between Irish and English and this can lead to the English range and application of words spreading into Irish. A good example is the verb faigh ‘get’. The Irish word corresponds to the English word in its meaning of ‘acquire’ (27a). But it is increasingly being used in the inchoative sense of ‘get’ which is ousting the Irish verb éirigh ‘rise’ which is traditionally used in this sense: (27) a. Fuair sí bronntanas óna máthair. [got she present from‐her mother] ‘She got a present from her mother.’ b. Tá sé ag fáíl níos fuaire anois. (modern) Tá sé ag éirí níos fuaire anois. (traditional) [is it at getting/rising more colder now] ‘It is getting colder now.’ The situation of the Irish language vis‐à‐vis English today shows that, under pressure, a ­language can be infiltrated syntactically by a co‐existing dominant language. The ease with

164  Raymond Hickey which this infiltration can take place has to do with the typological fit of the two languages at certain points in their grammar. Thus the similarity of Irish and English phrasal verbs obviously facilitated the calquing of examples from the latter language in the former one. In aggregate, these external influences can lead fairly rapidly to typological change and thus illustrates a scenario in which language convergence can take place. This convergence is especially obvious in the Irish of non‐native speakers where the agency in the contact scenario of both languages is reversed. An important issue here is the quantitative relationship of native to non‐native speakers. Given that the latter are much more numerous than the former, and indeed are motivated in using the language, the future of the language may not just be some typological convergence of Irish English but the wholesale replacement of first language Irish by second‐language Irish (Walsh 2020), resulting from a contact situation in which the speed of change is greatly accelerated given that non‐native speakers do not necessarily heed such central grammatical features of the language as grammatical gender or the palatal # non‐palatal distinction in its phonology.

9 Conclusion The data from Irish and English considered in this chapter shows how syntactic material can be transferred from one language to another. For the historical language shift scenario in Ireland it furthermore shows how unguided second language acquisition means that a search for categories in the target language which speakers know from the outset language is a dominant feature of the shift process. Where the equivalents reached by individuals or small groups are accepted by the community they can establish themselves as focused features of a later shift variety. This in itself is a process which involves the loss of many earlier features found closer to the time of the language shift. For instance, emigrant letters from the nineteenth century exhibit vernacular Irish English of the time which allowed the absence of an indefinite article as in instances like before I had Ø letter from him; she got to be Ø very stout girle; thair is Ø chance in this countery (Hickey 2019). This is a transfer feature from Irish, which only has a definite article, but not an indefinite one. However, this feature has been lost from Irish English and has not been attested at any time in the twentieth century. The consideration of loss and survival of transfer elements provokes the more general question of what types of feature survive the language shift process. Is it possible to predict what features are retained in the shift variety once the dust settles? Looking at the transfer features from Irish to English considered in this chapter, the answer would seem to have to do with (i) semantic motivation for the continuing existence of a feature, e.g. the immediate or resultative perfectives of Irish English and/or with (ii) the formal marking of central verbal categories such as habitual aspect, cf. the do+be habitual of vernacular Irish English. A further consequence of the survival of such features is that a fairly uniform geographical distribution of transfer phenomena can be found, promoting areality (Hickey 2012c, 2017a, 2017b) across the island of Ireland despite internal divisions such as that between the north and south. The question was raised at the beginning of this chapter whether language shift varieties represent a typological class of their own. To answer this question positively, there must be sufficient features which are unique to shift varieties, i.e. “additive features,” and which appear irrespective of their occurrence in either the substrate or superstrate inputs which engender a shift variety. From the consideration of shift varieties it would seem that the most robust “additive” features are aspectual categories, above all habitual and perfective, which

Contact and Language Shift  165 surface in shift varieties even if they are not present in input sources. Given the absence of much overt grammatical expression in shift varieties, the explicit marking of aspectual categories is all the more remarkable. Word order is, in principle, a relevant feature in this context, but given that English has SVO, the use of SVO in Irish English is of no diagnostic value. What is remarkable, however, is that the VSO word order of the input language Irish has apparently never been transferred to any variety of Irish English despite the appearance of other substrate features in the shift varieties. The remaining signal features of shift varieties, such as the general lack of inflectional morphology can be sufficiently accounted for as due to the elimination of redundancy, a key process during initial stages in the genesis of shift varieties in situations of unguided adult second language acquisition. Low target language competence levels among speakers partaking in the shift can be responsible for other features dictated by pragmatic and discourse factors in early shift‐variety usage, e.g. topicalization via fronting, a common feature of Irish English (Hickey 2007: 267–9). The question of typological class may ultimately be one of linguistic labeling. Any insight to be gained from such labeling must be evident from the primary data of the shift, i.e. the resultant shift variety. Here one can see that syntactic features, not present in the target language, emerge and are motivated by the search for categorial equivalence by the speakers shifting to the target language. This insight is of value as it shows how speakers in a non‐prescriptive environment can manipulate the target language to reach equivalents to features known from the outset language in the shift scenario.

NOTES 1 Cf. the Northwest Caucasian language Ubykh whose last speaker is known to have died in Turkey in the late twentieth century. 2 Examples for the latter are known from history; at present instances are constantly increasing due to the all‐pervasive influence of major European languages such as English or Spanish on minority languages throughout the world. 3 Just what the default values for certain parameters are has been a contested matter, see Kihm (2008). 4 This figure excludes the many foreigners who are now living in Ireland, immigrants to the country, above all from the 1990s and early 2000s. In practical terms, one can say that the 99% referred to consists of those people who are Irish and whose parents were Irish as well. 5 The situation in the north of Ireland was quite different because there Scottish and northern English settlers had come in considerable numbers during the seventeenth century. 6 The sample sentences provided in this chapter stem from various data collections of the author, both for Irish and for English. These are the following: CCE = A Collection of Contact English, DER = Dublin English Recordings, WER = Waterford English Recordings. In addition there are a few other abbreviations: M = male, F = female. Before a number, W = West, S = South. TRS‐D stands for Tape Recorded Survey of Hiberno‐English Speech – Digital. This collection is based on recordings made by colleagues in the Department of English, Queen’s University, Belfast. 7 Later, a distinct semanticization of this usage arose whereby the unbound reflexive came to refer to someone who is in charge, the head of a group or of the house, etc. 8 These remarks refer to language shift. Of course, in a borrowing situation, a sound can enter a language with word(s) which show it, e.g. nasal vowels in German which are contained in French loans. Furthermore, new phonotactic combinations may enter with loanwords, e.g. [∫ ] before an obstruent or nasal in words of Yiddish origin in English, cf. schmooze, schmuck. 9 This has been registered for some time by Irish scholars, at least since the mid‐twentieth century, cf. Ó Cuív (1951: 54–5).

166  Raymond Hickey

REFERENCES Bakker, Peter, Aymeric Daval‐Markussen, Mikael Parkvall, and Ingo Plag 2011. Creoles are typologically distinct from non‐creoles. Journal of Pidgin and Creole Languages 26.1: 5–42. Barber, Charles 1997 [1976]. Early Modern English. Edinburgh: Edinburgh University Press. Beal, Joan C. 2004. English in Modern Times 1700–1945. London: Arnold. Bickerton, Derek 1975. Dynamics of a Creole System. Cambridge: Cambridge University Press. Bickerton, Derek. 1984. The Language Bioprogram Hypothesis. The Behavioral and Brain Sciences 7: 173–88. Bliss, Alan J. 1976. The English language in early modern Ireland. In Theodore W. Moody, Francis X. Martin, and Francis J. Byrne (eds.), A New History of Ireland, vol. 3: Early Modern Ireland (1534–1691). Oxford: Clarendon Press, pp. 546–60. Buchstaller, Isabelle, Anders Holmberg, and Mohammed Almoaily (eds.) 2014. Pidgins and Creoles beyond Africa‐Europe Encounters. Amsterdam: John Benjamins. DeGraff, Michel 2003. Against creole exceptionalism, Language 79.2: 391–410. DeGraff, Michel 2005. Linguists’ most dangerous myth: the fallacy of creole exceptionalism, Language in Society 34.4: 533–91. Denison, David 1998. Syntax. In Suzanne Romaine (ed.), The Cambridge History of the English Language, vol. 4: 1776–1997. Cambridge: Cambridge University Press, pp. 92–326. Dietrich, Julia C. 1981. The Gaelic roots of a‐prefixing in Appalachian English. American Speech 56: 314. Elworthy, Frederic T. 1877. The grammar of the dialect of West Somerset. Transactions of the Philological Society 79: 143–257. Filppula, Markku 1999. The Grammar of Irish English: Language in Hibernian Style. London: Routledge. Fitzpatrick, David 1994. Oceans of Consolation: Personal Accounts of Irish Migration to Australia. Cork: Cork University Press. Gal, Susan 1979. Language Shift: Social determinants of linguistic change in bilingual Austria. San Francisco: Academic Press. Harris, John 1991. Conservatism versus substratal transfer in Irish English. In Peter Trudgill and J. K. Chambers (eds.), Dialects of English: Studies in Grammatical Variation. London: Longman, pp. 191–212.

Hickey, Raymond 1990. Suprasegmental transfer: on prosodic traces of Irish in Irish English. In Jacek Fisiak (ed.), Further Insights into Contrastive Linguistics. Amsterdam: John Benjamins, pp. 219–29. Hickey, Raymond 1997a. Arguments for creolisation in Irish English. In Raymond Hickey and Stanisław Puppel (eds.), Language History and Linguistic Modelling, A Festschrift for Jacek Fisiak on his 60th Birthday. Berlin: de Gruyter Mouton, pp. 969–1038. Hickey, Raymond 1997b. Assessing the relative status of languages in medieval Ireland. In Jacek Fisiak (ed.), Studies in Middle English. Berlin: de Gruyter Mouton, pp. 181–205. Hickey, Raymond 2003. Rectifying a standard deficiency: pronominal distinctions in varieties of English. In Irma Taavitsainen and Andreas H. Jucker (eds.), Diachronic Perspectives on Address Term Systems. Amsterdam: John Benjamins, pp. 345–74. Hickey, Raymond 2007. Irish English: History and Present‐Day Forms. Cambridge: Cambridge University Press. Hickey, Raymond 2011. The Dialects of Irish, Study of a Changing Landscape. Berlin: de Gruyter Mouton. Hickey, Raymond 2012a. Early English and the Celtic hypothesis. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 497–507. Hickey, Raymond 2012b. Assessing the role of contact in the history of English. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 485–96. Hickey, Raymond 2012c. English in Ireland. In Raymond Hickey (ed.), Areal Features of the Anglophone World. Berlin: de Gruyter Mouton, pp. 79–107. Hickey, Raymond 2014. The Sound Structure of Modern Irish. Berlin: de Gruyter Mouton. Hickey, Raymond 2017a. Areas, areal features and areality. In Raymond Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 1–15. Hickey, Raymond 2017b. Britain and Ireland. In Raymond Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 270–303.

Contact and Language Shift  167 Hickey, Raymond 2019. Mining vernacular correspondence for linguistic insights. In Raymond Hickey (ed.), Keeping in Touch. Emigrant Letters across the English‐speaking World. Amsterdam: Benjamins, pp. 1–24. Hickey, Raymond 2020. Shift varieties as a typological class? A consideration of South African Indian English. In Raymond Hickey (ed.), English in Multilingual South Africa. The Linguistics of Contact and Change. Cambridge: Cambridge University Press, pp. 265–87. Kallen, Jeffrey L. 1989. Tense and aspect categories in Irish English. English World‐Wide 10: 1–39. Kihm, Alan 2008. Creoles, markedness and default settings: an appraisal. In Silvia Kouwenberg and John Kingler (eds.), The Handbook of Pidgin and Creole Studies. Malden, MA: Wiley, pp. 411–39. Klemola, Juhani and Markku Filppula 1992. Subordinating uses of “and” in the history of English. In Matti Rissanen, Ossi Ihalainen, Terttu Nevalainen, and Irma Taavitsainen (eds.) History of Englishes: New Methods and Interpretations in Historical Linguistics. Berlin: de Gruyter Mouton, pp. 310–18. Majewicz, Elżbieta 1984. Celtic influences upon English and English influences upon Celtic languages. Studia Anglica Posnaniensia 27: 45–50. McWhorter, John 1998. Identifying the creole prototype: vindicating a typological class, Language 74.4: 788–818. McWhorter, John 2008. Deconstructing creole. Review article of Umberto, Ansaldo, Stephen Matthews and Lisa Lim Deconstructing Creole. Amsterdam: John Benjamins. Journal of Pidgin and Creole Languages 23.2: 289–306. McWhorter, John 2012. Case closed? Testing the feature pool hypothesis, Journal of Pidgin and Creole Languages 27.1: 171–82. McWhorter, John 2018. The Creole Debate. Cambridge: Cambridge University Press. Mesthrie, Rajend 1992. English in Language Shift: The History, Structure and Sociolinguistics of South African Indian English. Cambridge: Cambridge University Press. Mesthrie, Rajend 1996. Language contact, transmission, shift: Indian South African English. In Vivian de Klerk (ed.) Focus on South Africa. Amsterdam: John Benjamins, pp. 79–98. Montgomery, Michael 2000. The Celtic element in American English. In Hildegard L. C. Tristram

(ed.), The Celtic Englishes II. Heidelberg: Carl Winter, pp. 231–64. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Ó Cuív, Brian 1951. Irish Dialects and Irish‐ Speaking Districts. Dublin: Institute for Advanced Studies. Parkvall, Mikael 2008. The simplicity of creoles in a cross‐linguistic perspective. In Matti Miestamo, Kaius Sinnemäki, and Fred Karlsson (eds.), Language Complexity: Typology, Contact, Change. Amsterdam: Benjamins, pp. 265–85. Stenson, Nancy 1991. Code‐switching vs. borrowing in Modern Irish. In P. Sture Ureland and George Broderick (eds.), Language Contact in the British Isles: Proceedings of the Eighth International Symposium on Language Contact in Europe. Tübingen: Niemeyer, pp. 559–79. Stenson, Nancy 1993. English influence on Irish: the last 100 years. Journal of Celtic Linguistics 2: 107–28. Stenson, Nancy 1997. Language contact and the development of Irish directional phrase idioms. In Anders Ahlqvist and Vera Čapková (eds.), Dán do oide [A poem for a mentor]: Essays in Memory of Conn R. Ó Cléirigh. Dublin: Linguistics Institute of Ireland, pp. 559–78. Vennemann, Theo 2010. Contact and prehistory: the Indo‐European Northwest. In Raymond Hickey (ed.), Handbook of Language Contact. Malden, MA: Wiley‐Blackwell, 380–405. Veselinović, Elvira 2006. How to put up with cur suas le rud and the bidirectionality of contact. In Hildegard L. C. Tristram (ed.), The Celtic Englishes IV. Potsdam: Potsdam University Press, pp. 173–90. van der Wurff, Wim and Tony Foster 1997. Object‐verb word order in 16th century English: a study of its frequency and status. In Raymond Hickey and Stanisław Puppel (eds.), Language History and Linguistic Modelling: A Festschrift for Jacek Fisiak on his 60th Birthday. Berlin: de Gruyter Mouton, pp. 439–53. Walsh, John 2020. The Irish language and contemporary Irish identity. In Raymond Hickey and Carolina P. Amador‐Moreno (eds.), Irish Identities. Sociolinguistic Perspectives. Berlin: de Gruyter Mouton, pp. 21–44.

8 Contact and Lexical Borrowing PHILIP DURKIN 1  What is Lexical Borrowing? Lexical borrowing denotes the process by which lexical items from one language (the donor language) are replicated in another language (the recipient language). Sometimes it is used more restrictively to refer specifically to instances where the agentivity is taken to be on the part of the recipient language. (See Section 10 on contrasts sometimes drawn with imposition.) The most salient type of lexical borrowing is a loanword, e.g. English sushi from Japanese sushi, where the Japanese lexeme is duplicated in English (with some accommodation to the sound system of English). The suitability of the terms “borrowing” and “loanword” is often challenged on the very reasonable grounds that nothing is removed from the donor language, either temporarily or permanently; rather, a new lexical item is created in the recipient language by duplicating what is found in the donor language. However, in spite of the transparent unsuitability of the underlying metaphor, “borrowing” and “loanword” (and a whole constellation of other terms featuring “loan”) remain usual in both popular and (most) technical use, and seem likely to remain so. The agentivity in lexical borrowing is, of course, only in a rather abstract sense on the part of “a language.” It is individual speakers who adopt a word from another language in their speech or writing; often (although not always) large numbers of individual speakers will be doing this separately from one another. Almost always, this initial act of borrowing explains only how the word has got into the usage of a (usually very small) minority of speakers of the language, or, to put this another way, how this newly borrowed word has entered the language’s variation pool of alternative ways of expressing the meaning in question (because there will always be alternatives, even if only in the form of longer phrasal explanations of the relevant meaning). In order to enter general usage in the language, the word must then become adopted by a wider range of speakers, by a process of internal borrowing. These two processes, of initial interlinguistic transfer and subsequent intralinguistic spread, are closely interconnected but distinct and may show different motivation as well as different mechanisms. Nonetheless, this distinction is frequently ignored in the literature, which typically operates on assumptions about the agentivity of a language, in the abstract. The implications of this, particularly for the imprecision of many of the apparently categorical distinctions between different types of borrowing frequently made in the literature, will be a recurrent theme in this chapter.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

170  Philip Durkin

2  Borrowing of Features Other Than Lexemes This chapter will focus on lexical borrowing, but many other elements and features may also be borrowed. Some are typically conveyed through or as a result of lexical borrowing, such as (in phonology) the creation of new phonemic contrasts as a result of borrowing, or (in morphology) the development of new productive affixes as a result of borrowing. These phenomena are typically referred to as borrowing, although it is often truer to say that they are results of lexical borrowing: in the two examples given, new phonemic contrasts or new productive affixes may arise in the recipient language duplicating ones found in the donor language, but the vector for this duplication is the borrowing of lexical items. Borrowing of larger structures, such as syntactic ones, are also often linked to lexical borrowing (e.g. if borrowing of verbs results in changes to patterns of verbal complementation).1 Affixes are often described as being “borrowed” from one language to another. This is true in the sense that an affix that originates in one language can be replicated in another language, becoming productive in new formations in that language. The process involved is however very different from that involved in a loanword entering a language, and once the process has been considered in more detail it may seem more accurate to describe it as a possible consequence of lexical borrowing, rather than as a type of borrowing in its own right. To take a familiar example, English has acquired many productive affixes that originated in French, Latin, or Greek (in many cases, two or even all three of these can reasonably be identified as co‐donors). Thus, ‐ment is common as the ending of words borrowed from French and/or Latin (such as ornament, judgment, or testament) from early Middle English (i.e. by the beginning of the fourteenth century); already in the fourteenth and fifteenth centuries it is found forming a small number of new nouns in combination with words not of French or Latin origin, such as eggment or hangment; by the sixteenth century such formations become more common, such as acknowledgment, amazement, atonement, or wonderment. Such formations, where there is no suspicion that a French or Latin etymon may have existed but escaped the lexicographical record, are the clearest evidence that –ment has become productive in English. We may thus say that ‐ment has been “borrowed” into English, but what we mean by this is that the borrowing of large numbers of words ending in ‐ment has led by analogy to ‐ment becoming a word‐forming element in English. In fact, for such analogy to operate, at least one of two further things is required: there must also either be extensive borrowing of the parent word stems and/or other words from the same family, so that borrowing of, for instance, commence as well as commencement makes clear the function of ‐ment as an affix (and more specifically its frequent function in forming nouns on verbs), or there must be a command of the donor language among a sufficiently large and influential portion of the population that is sufficiently deep to involve awareness of the morphological composition of borrowed words; in practice, in this particular historical contact situation both of these are likely to have applied, but the former was probably much the most influential factor. Theoretically, it is perhaps possible that knowledge of the morphological structures of the donor language could in isolation lead to the creation of a new affix in the recipient language, without the need for any borrowings containing that affix: in English, it is conceivable that such formations could be found among coinages of new technical vocabulary from elements of Latin or Greek origin in the sciences, but even here the usual pattern is that new word‐forming elements arise on the pattern of complete words borrowed from Latin or Greek. In the case of some affixes arising as a result of borrowing there is a restriction to forming new words only on stems originating in the language(s) from which the affix has originated. For example, ‐al becomes productive in English in forming adjectives and nouns by analogy with its function in words borrowed from French and Latin, but its productivity is almost

Contact and Lexical Borrowing  171 entirely restricted to words of French or Latin origin (or, sometimes, Greek, following a pattern found in Latin); thus the adjectives occasional, professional, or circumstantial, and the nouns acquittal or refusal are thus just a few among very many examples of words formed in ‐al within English (so far as the evidence in English, French, and Latin allows one to assess) on the model of words borrowed into English already having this suffix, but examples of formations in ‐al on words of native (i.e. Germanic) or other origins are vanishingly rare (bestowal and betrothal may be two such, and bridal and burial apparently show remodeling of existing words using this suffix). In the classificatory system for borrowings developed by Haugen (1950a, 1950b, 1953) and still frequently referenced and summarized, such formations are classified as a type of derivational blend, but if such words are included in a classification of borrowings then it must be acknowledged that this is being taken in a very broad sense that includes intralinguistic creations that show the effects of borrowings (a point made e.g. by Winford 2003: 44).

3  Semantic Borrowing There are several different major types of semantic borrowing commonly identified, as well as some minor types (which in some instances may be better regarded as aspects of the effects of lexical borrowing). All types of semantic borrowing are relatively little studied in comparison with loanwords, largely because they are often very difficult to detect with any degree of confidence. Usually easiest to identify are loan translations or calques, where the structure of a word in the donor language is replicated in the recipient language. Thus French prêt‐à‐porter is a calque on English ready‐to‐wear (or just possibly on ready‐for‐wear), reproducing its structure and semantics item‐for‐item (by contrast, English prêt‐à‐porter shows a subsequent loanword from French). Similarly, Old English ælmihtiġ (modern English almighty) probably shows a calque on Latin omnipotēns, its components all and mighty reproducing the components omni‐ ‘all’ and potēns ‘mighty, powerful’ of the Latin word. Where the resemblance is too close to be attributed to chance, and we can rule out influence from a third language, we can usually identify a calque with reasonable confidence. Identifying the direction of influence can sometimes be more of a challenge, but near‐certainty is often provided by such criteria as: documentary evidence that the model existed much earlier in the donor language than the recipient; knowledge of the direction of cultural influence; or even historical knowledge about the circumstances of coining of the model (or a mixture of these). Typically, the evidence will all be, in the strictest sense, non‐linguistic, which is perhaps one reason why this area of study remains somewhat on the fringes of historical linguistics. Loan translations are sometimes distinguished terminologically from loan creations, a somewhat confusing term for a situation where a word is created in a language specifically in order to provide a counterpart for a word in another language, but the elements from which it is created do not correspond at all semantically to those which make up the word it renders, for instance (in a famous example from Weinreich 1953: 51) Yiddish mitkind ‘sibling’ (literally ‘a with‐child’) the creation of which was motivated by the need for a term equivalent to English sibling, German Geschwister, etc. In practice, many formations sit somewhere on a cline between these two types: rather approximate loan translations are very common, and may even be more common than completely exact ones. Semantic loans, where a word in one language replicates a meaning shown by a word with which it shares another meaning or meanings or in another language, are widespread but particularly difficult to identify with certainty. Very often, where two words share one or more core meanings in two languages, newly innovated meanings in one of the languages will be replicated in the other, by what is essentially a process of proportional analogy. For

172  Philip Durkin instance, Old English þrōwung ‘suffering’ acquired the additional meaning ‘(Christ’s) passion’ almost certainly on the model of the range of meanings shown by Latin passiō ‘suffering, (Christ’s) passion’; we can have reasonable certainty in this instance, given that we know that the new meaning was acquired with conversion to Christianity, and that Latin was the main vector for this. In a not uncommon pattern, semantic influence can be identified from a cognate that can be ruled out (more or less confidently) as direct etymon on formal grounds. Thus English popular is a borrowing from Latin populāris, but a number of its senses show clear influence from the related French populaire. Once we move beyond specific specialized meanings acquired in a known historical and cultural context, certainty is often much more elusive. We may be able to detect with reasonable confidence that close intercultural relationships play a part in words in two languages sharing a range of meanings, but the precise details of influences can be much harder to identify. Frequently, influence can run counter to the direction of an earlier borrowing; for example, German Kultur was borrowed from Latin cultū ra in the seventeenth century, but in the nineteenth century (as shown primarily by the work of social and cultural historians) its semantic development exercised significant influence on the development of French culture (a medieval classicizing replacement of Old French couture, regularly developed from Latin cultū ra) and English culture (a medieval borrowing from French and Latin), and similarly on the semantics of related words across languages of western Europe. Similar patterns may be found in cases where there is no etymological relationship between the words concerned: for example, German Geschichte and French histoire show many shared meanings reflecting many instances of cross‐linguistic influence, while in English these meanings are now distributed between history (the exact counterpart of French histoire) and its variant story.

4  Loan Blends A separate category of loan blend is frequently identified (especially following the influential classification in Haugen 1953) for cases where adaptation occurs at the point of borrowing. Thus, English neurotize shows a borrowing of French neurotiser with substitution of the suffix ‐ize for its French equivalent ‐iser. We can be sure that this happened in the case of this rare technical word, because we know (on the basis of non‐linguistic evidence) that it was coined in French by the scientist Vanlair in 1882, and we also know that the origin of the ‐t‐ in its stem, faithfully reproduced in the English borrowing, is unexplained, hence we can rule out a calque in English. In most cases, it is far harder to rule out alternative explanations. Additionally, even in this instance we may prefer to classify the substitution of ‐ize for ‐iser as simply a part of the naturalization of a loanword, similar to the (usually unremarked) replacement of the French infinitive inflection ‐er by zero in English.2

5  Motivations for Borrowing Most elementary discussions of lexical borrowing identify the two principal motivations for borrowing as need and prestige: languages borrow words either because they are needed for describing newly encountered things, processes, actions, etc., or because at the time when the loan occurs the donor language has particular prestige in the minds of speakers of the recipient language. Several things are worth noting here. First, a need for a new word can be (and often is) met in ways other than by borrowing: in assessing motivation, the harder (and often essentially unanswerable) question is typically why in a particular instance lexical borrowing has been preferred over various alternatives. Thus, English tomato (borrowed

Contact and Lexical Borrowing  173 ultimately from Nahuatl tomatl via Spanish tomate) provides a classic example of borrowing of a word for a newly encountered thing, but a short survey of other names for the tomato found in English and other European languages in the period of its introduction to Europe reveals a variety of alternative naming strategies, as exemplified by Italian pomodoro, literally ‘apple of gold’, or the now obsolete French pomme d’amour and the English calque on this love apple, or again German Paradeiser (now almost exclusively Austrian usage), earlier Paradiesapfel, literally ‘paradise apple’.3 Prestige does often play a part in determining whether borrowing is preferred over alternatives, but typically “prestige” is being used here as a cover term for a set of interrelated factors, including: cultural influence of speakers of the donor language; predominance of speakers of the donor language in a particular field to which the borrowing belongs semantically; the general state of social, cultural, trading, etc. relations between speakers of the two languages; how widespread knowledge of the donor language is among speakers of the recipient language, or influential groups among those speakers (i.e., if one person adopts a word from this language, how likely is it that others will understand this as a result of their own knowledge of the donor language). For example, English during the Middle English period showed massive borrowing of lexis from French (usually specifically Anglo‐Norman) and Latin. These languages were indeed prestigious in comparison with English, but we gain a much more informed understanding both of the extent and the nature of the borrowing that occurred when we consider that in this period: almost all literate English speakers also had expert knowledge of Anglo‐Norman and Latin; for several centuries, Latin appears to have been taught in grammar schools through the medium of Anglo‐Norman, which was learned at a much earlier age by those destined for a grammar‐school education;4 for centuries after the Norman Conquest, Latin was the formal language of legal written record as well as of the church, while Anglo‐Norman was also preferred over English in many written and some formal spoken functions (e.g. in legal pleading).5 We can perhaps collapse all of this under the catch‐all of “prestige,” but only if we are using this as a shorthand for the linguistic, social, and cultural relations between two speaker communities, and the degree and nature of bilingualism or multilingualism in those communities.

6  Immediate and Remote Etymons An important distinction often ignored is between the immediate and remote etymons of a loanword. Thus in the example of tomato looked at in Section 5, the immediate etymon of English tomato is Spanish tomate and its remote etymon is Nahuatl tomatl. In terms of the historical contact situation, the English loanword resulted from contact with Spanish, not Nahuatl, and hence the word is a Spanish loanword in English, although many speakers may well have been aware that it ultimately had a more “exotic” origin. (In this instance word form is one of several criteria – linguistic as well as non‐linguistic – that put the transmission via Spanish really beyond doubt. When borrowed in the seventeenth century the English word was tomate; the modern form tomato probably results from associative influence of potato, an “exotic” word for a food originating from the New World.) The case of English tomato is fairly clear‐cut: as well as the formal difference between the English word and Nahuatl tomatl, we know that the word was well embedded in English long before any extensive direct contacts between speakers of English and speakers of Nahuatl. Thus the contexts of use of menina ‘young girl, young lady‐in‐waiting’ in English point strongly to borrowing both directly from Portuguese and via Spanish. On such clear‐ cut instances of multiple input, see Section 7. In many other instances, the balance of probabilities is less clear, and we simply cannot rule out the possibility that a remote etymon may also partly have played a role as a direct etymon.

174  Philip Durkin

7  Multiple Inputs and Polygenesis Many loanwords show input from more than one language. In some instances this is clearly proved by considerations of word form, and the only difficulty is whether we should regard the distinct forms as belonging to the same word, or as showing synonyms of ultimately the same origin. Thus South African English babalaas (adjective) ‘hung‐over’, (noun) ‘hang‐over’, is a borrowing from Afrikaans babalaas, which is in turn from Zulu ‐bhabhalazi, the stem of ibhabhalazi (which is itself from an Afrikaans word that ultimately originates from European Dutch). The South African English word also occurs in the form babalazi, showing input directly from Zulu ‐bhabhalazi. (Here the alternative explanation would be that the form babalazi was also transmitted via Afrikaans, but supporting evidence for this is lacking.) Such multiple inputs are particularly common in communities where three or more languages are in frequent use. In fact, a significant proportion of the words in high‐frequency use in contemporary published written English result from borrowing from both French and Latin in the multilingual situation during the Middle English period outlined in Section 5. Some particularly clear examples include accept, application, capital, element, general (adjective), history, natural, nature, present (verb and adjective), question, use (noun and verb).6 Identification of multiple inputs has often been eschewed in traditional etymological lexicography, sometimes being seen as a last resort where a single etymon cannot be satisfactorily settled upon. However, if the focus is on individual language users as outlined in Section 1, such multiple inputs seem an entirely natural result of contact situations in which individual speakers are familiar with what is, from an etymological perspective, ultimately the same word, with equivalent form and equivalent meanings, in more than one language. It is also likely that many words that show input from only one donor language nonetheless result from an accretion of multiple different instances of importation from the donor language by different speakers, just as language‐internal derivative formations on common patterns are likewise likely to result from the convergence of multiple instances of lexical innovation by different individuals in comparable communicational situations, rather than radiating from a single instance of innovation by one individual. An indicator for this among loanwords is the frequent accretion of multiple different word forms and multiple different meanings from a donor language over a short period of time: convergence of multiple instances of borrowing seems much more plausible than a single instance of borrowing ­followed by multiple instances of formal or semantic influence from the donor.7

8  Borrowing vs Code‐switching Almost all lexical borrowing assumes a degree of bilingualism: unless at least some speakers of the donor language and recipient language can understand one another’s language to at least a limited extent, lexical borrowing is hardly likely (simply pointing and naming could perhaps support very limited borrowing of names for things or actions that can easily be indicated by pointing). However, the term code‐switching is normally reserved for situations where speakers who are both completely fluent in two languages switch between these languages during a conversation. It is debated to what extent communication of this sort gives rise to lexical borrowing. In particular, there remains considerable controversy around whether single‐word insertions of foreign‐language words by bilingual speakers are always loanwords (nonce or one‐off loanwords, if they are not adopted elsewhere), or whether they are sometimes not loanwords at all but single‐word code‐switches (probably now a minority view).8 It is uncertain whether this distinction has as much force as it may appear to have when presented in abstract terms: if such single‐word insertions are always loanwords, it is still worth noting that such loans are more likely to occur among bilingual speakers

Contact and Lexical Borrowing  175 addressing other bilinguals who can confidently be relied upon to know the donor‐language word just as well as the speaker does (and hence less strong motivation is required for a loanword to be introduced); if, on the other hand, one classifies such instances as single‐word code‐switches, should the use of a particular lexical item become habitual among the bilingual community (because it fulfills a particular communicational need) it is hard to see why this use should not subsequently spread to monolingual speakers, and this subsequent intralinguistic spread is likely to show very similar processes to those that apply for any loanword.

9 What Does it Mean to Say that a Language has Borrowed a Word? A key question, but one surprisingly seldom addressed head‐on except as regards the debate concerning loanwords versus single‐word code‐switches, is how we determine that a word has been borrowed into another language. Words that enter the basic, everyday, or high‐­ frequency vocabularies of the recipient language are largely unproblematic in this regard. So are words that show some degree of morphological adaptation to the recipient language (such as suffix substitution, e.g. neurotize discussed on p. 172), even if they occur with low frequency. Most relatively rare words pose more of a problem, especially if they are used in a technical field that has some particular association with the donor language; e.g. if English‐ speaking wine connoisseurs use French terms such as cuvée, en primeur, or terroir, or (if they are keen on German wine) German terms such as Auslese, Eiswein, or Spätlese, does this mean that these words and phrases have been borrowed into English? In deciding this, what (if any) weighting do we give to such questions as degree of phonological or morphological adaptation (there will always be some), or orthographic features such as retention of accents in French or initial capitals in German, or use of italics? Is the speaker monolingual or bilingual? But here we may encounter some circularity: a lover of German wine may have an expert command of an extensive specialist vocabulary of German wine terms, but extremely limited active or passive command of German beyond this. Is this an example of very limited and specialized bilingualism? Or, particularly if a broadly similar command of such terminology is shared by other enthusiasts who are essentially monolingual English speakers, does this mean that all of these words are loanwords, albeit restricted to a very small and specialized group of speakers and confined to a very specific lexical field? Lexicographers have to make pragmatic decisions in such cases, and are normally very aware of how pragmatic such decisions are, in any particular instance involving weighing against one another large numbers of factors that may seem to point in contrary directions. Crucially, the balance of these factors for a particular lexeme can change over time: there may be markers of greater assimilation, such as writing cuvee rather than cuvée, or using terroir in extended metaphorical application to the impact of region of production on products such as cheese or sausages rather than wine; very typically, there is gradual increase in frequency of use among an ever wider group of speakers. If we only conclude that we have a borrowing when we are at some point along this cline, then this has significant implications for how we define borrowing. Most significantly, a narrow focus only on the first speakers of the recipient language to use a particular lexeme may mean that we are neglecting the key developments that lead us to conclude that a borrowing has actually occurred. A separate but related question is whether any systematic distinctions can be drawn between words that show greater or lesser degree of naturalization or adaptation to the recipient language. Dictionaries struggle with whether to flag “unnaturalized” or “semi‐naturalized” words as a distinct category; increasingly, there is a tendency to abandon any such attempt, since different indicators so often conflict, vary in the usage of different individuals or speaker communities, and change over often short timescales.9 In German lexicography, and more

176  Philip Durkin broadly in (traditional) German philological scholarship, a distinction has long been drawn between the categories of Lehnwort and Fremdwort, where the defining characteristic of Lehnwörter are that they show accommodation to the phonology and morphology of the borrowing language and may give rise to new derivative formations, whereas Fremdwörter remain unassimilated phonologically.10 In practice, these categories strain against the variable degree of assimilation with regard to these different criteria shown by individual words: some foreign‐ language words conform entirely to the phonological norms of the borrowing language, hence phonological assimilation is at most marginal; some languages – such as English – have very limited inflectional morphology, hence many loans (for instance uncountable nouns) have no opportunity to show assimilation to the inflectional morphology of the borrowing language; semantics and word class both have a huge impact on whether a word gives rise to derivative formations in the recipient language. Haugen terms such assimilation “substitution” (e.g. of a native phoneme for one in the donor language, or of native inflectional morphology for that of the donor language) and distinguishes this from the importation of foreign elements unchanged, noting that “every loan [is] part importation and part substitution” (1953: 388). Additionally, words that show little or no assimilation to the recipient language may nonetheless be very widely familiar and in common use: for instance, English appendix (from Latin appendix) shows little phonological assimilation (because little is required), retains for many speakers the “alien” plural appendices with modification of the word stem, and lacks derivatives in common use (appendicitis being better regarded as a related loan from the international Latin terminology of medicine, while in terms of wider affiliations append and appendage are only distant relatives resulting from earlier borrowing), and yet it is widely familiar both denoting a part of the body and denoting an adjunct, addition, or appendage (especially to a text).

10  Language Shift, and Imposition Most of the discussion in this chapter has assumed a situation of language maintenance, i.e. linguistic transfer between communities who continue speaking the two languages concerned. It is an interesting question whether different considerations apply in situations where speakers of the donor language are in the process of switching to the recipient language: this is often not regarded as borrowing at all, and it may be unhelpful to think in terms of donor and recipient. A distinction is often made between borrowing and imposition, the essential difference being one of agency: borrowing is by speakers of the recipient language, while imposition denotes the process of speakers of one language bringing words (or other features) into another language in a situation of language switch. It is often assumed that in a situation of language switch lexical borrowing is likely to be very light, while importation of other features (for instance prosodic or syntactic ones) may be considerable (Hickey, this volume). However, imposition is also sometimes adduced as an explanation for cases in which very basic and/or closed‐class vocabulary enters a language in a situation where there is no obvious explanation from the perspective of “need” and the donor language has low prestige (indeed, is in the process of being abandoned by its speakers). For example, Townend (2002) makes a persuasive case for imposition during the period of language switch, explaining in this way the many everyday and basic level words that entered Middle English from Scandinavian varieties during the period when Scandinavian‐speaking settlements were shifting to English (including, probably, the third person plural pronouns they and them and possessive adjective their), whereas earlier borrowings into Old English reflect need more closely in featuring words for new legal concepts, military or naval terminology, etc. A challenge for such an approach is how to account for the subsequent spread of such items within the recipient language, to populations who have never switched from the language in which the words originated. In other words, is it enough simply to have an explanation of

Contact and Lexical Borrowing  177 how an item has entered the variation pool of a language, or do we also need to account for how it has gained frequency and spread in competition with alternative means of realizing the same meaning? It may be that such interlinguistic spread is simply relatively frictionless in comparison with movement between languages, or it may be that explanations should be sought in sociolinguistic situations or population movements in the borrowing language, or alternatively that functional considerations may explain spread. (For example, albeit still controversially, the spread of third person plural they in English is often explained as an instance of therapeutic change, restoring distinctions between pronoun forms after phonological change had resulted in some varieties in homophony of the inherited pronoun forms for third person singular masculine, third person singular feminine, and third person plural.11)

11  Borrowed Words in Basic Meanings Most examples discussed in this chapter have been drawn from English, a language with a particularly richly documented (and examined) history. Identification of lexical borrowings at a time depth beyond rich written records for the language in question is much more problematic, as considerations of word form must be relied upon to identify loanwords, and even these can be unclear particularly where the genetic affiliations of a language (such as membership of a wider language family, etc.) are uncertain or unknown. An approach that has been much used in this area involves identifying a list of basic meanings, and the words which normally realize each of these meanings in the language in question, these words being assumed to be likely to be particularly resistant to borrowing. The lists devised by Morris Swadesh have been much used. These were put together largely on the basis of Swadesh’s intuitions and practical experience.12 More recently an ambitious attempt has been made to produce a new list of this type, the Leipzig‐Jakarta List of Basic Vocabulary, on an empirical basis, deriving a list of those meanings most resistant to borrowing from an extensive cross‐linguistic survey of the etymological origins of the words realizing basic meanings across a wide range of languages.13 The empirical method employed to generate the Leipzig‐Jakarta list has had the very valuable consequence of also shedding considerable light on the degree of borrowing that is typical in basic meanings cross‐linguistically, creating an approximate scale against which individual languages can be assessed. Thus, from this perspective English emerges rather far up in the list of “high borrowers,” while Mandarin Chinese stands out as by far the lowest borrower among the languages surveyed.14

12  Borrowing Hierarchies and Borrowability The extent of lexical borrowing shown in a particular contact situation, and its nature – as for instance how much basic vocabulary is borrowed, or whether borrowing is largely ­confined to nouns or extends to adjectives, verbs, or to closed‐class grammatical items – is often placed on a scale, ranging from light to heavy, often correlated with whether the sociolinguistic relationship between the recipient language and the donor language is taken to be substratal, adstratal, or superstratal.15 The likelihood of borrowing of particular word classes is often presented separately as a borrowing hierarchy. Winford (2003) ­presents the following hierarchy based on Muysken (1981), proceeding from most to least likely to show borrowing:16 nouns > adjectives > verbs > prepositions > co‐ordinating conjunctions > quantifiers > determiners > free pronouns > clitic pronouns > subordinating conjunctions

178  Philip Durkin Such scales and hierarchies are certainly useful in gaining an impression of what is typical in various borrowing situations. They may, of course, be subject to review in the light of further empirical work. Thus, a finding that emerged from the extensive cross‐linguistic work undertaken in preparing the Leipzig‐Jakarta List of Basic Vocabulary (see Section 11) was that adjectives and adverbs appear “almost as hard to borrow as verbs” (Tadmor et al. 2010: 231). Additionally, when non‐linguistic factors such as influence of the speakers of the donor language in specific cultural or technological spheres are taken into account (compare Section 5), the record of lexical borrowing from one language to another may often correspond only extremely approximately to what a borrowing scale or general measures of borrowability might predict.

NOTES 1 Compare Winford (2003: 79–89). For a useful overview of the effects of borrowing on phonology, syntax, and morphology in the history of English, with copious examples, see Miller (2012). 2 For discussion of numerous problematic cases see Durkin (2009) 137–9. 3 See Durkin (2009) 145–7. 4 See Ingham (2012). 5 See overview in Durkin (2014). 6 See extensive discussion in Durkin (2014) 236–53. 7 See further Durkin (2009: 68–73). 8 In favor of regarding these as single‐word code‐switches see e.g. Myers‐Scotton (2002) or Thomason (2003); against this view, see e.g. Poplack, Sankoff, and Miller (1988), Poplack and Meechan (1998), Poplack (2004), Gardner‐Chloros (2010), Poplack and Dion (2012). 9 On such dilemmas and changing approaches in the history of the Oxford English Dictionary’s editorial practices see especially Ogilvie (2013). 10 For contextualization in German linguistic and cultural history see especially Chambers and Wilkie (1970: 70–1). 11 See summary in Durkin (2014: 175–9). For a recent attempt to explain they, their, them as endogenous forms not involving borrowing see Cole (2018). In the view of the present writer borrowing from Scandinavian remains the most convincing explanation, but Cole usefully highlights how receptive the existing English pronoun system was to this new form, hence this much‐cited example in the literature on lexical borrowing must always be treated carefully. 12 See especially McMahon and McMahon (2005), Wichmann and Grant (2012). 13 See Haspelmath (2009), Tadmor (2009), Tadmor et al. (2010). 14 Tadmor (2009). 15 See especially Thomason and Kaufman (1988), Thomason (2001), Thomason (2003). 16 Reproduced in Winford (2003: 51) with discussion and references to earlier such hierarchies devised by other scholars.

REFERENCES Chambers, W. Walker and John R. Wilkie 1970. A Short History of the German Language. London: Methuen. Cole, Marcelle 2018. A native origin for present‐ day English they, their, them. Diachronica 35 165–209.

Durkin, Philip 2009. The Oxford Guide to Etymology. Oxford: Oxford University Press. Durkin, Philip 2014. Borrowed Words: A History of Loanwords in English. Oxford: Oxford University Press.

Contact and Lexical Borrowing  179 Gardner‐Chloros, Penelope 2010. Contact and code‐switching. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐Blackwell, pp. 188–207. Haspelmath, Martin 2009. Lexical borrowing: concepts and issues. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages : A Comparative Handbook. Berlin: de Gruyter Mouton, pp. 35–54. Haspelmath, Martin and Uri Tadmor (eds.) 2009. Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter Mouton. Haugen, Einar 1950a. Problems of bilingualism, Lingua 2: 271–90. Haugen, Einar 1950b. The analysis of linguistic borrowing. Language 26: 210–31. Haugen, Einar 1953. The Norwegian Language in America: A Study in Bilingual Behavior. Philadelphia: University of Pennsylvania Press Ingham, Richard 2012. The Transmission of Anglo‐ Norman: Language History and Language Acquisition. Amsterdam: Benjamins. McMahon, April and Robert McMahon 2005. Language Classification by Numbers. Oxford: Oxford University Press. Miller, D. Gary 2012. External Influences on English: From its Beginnings to the Renaissance. Oxford: Oxford University Press. Musyken, Pieter 1981. Halfway between Quechua and Spanish: the case for relexification. In Arnold Highfield and Albert Valdman (eds.) Historicity and Variation in Creole Studies. Ann Arbor, MI: Karoma, pp. 52–78. Myers‐Scotton, Carol 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Ogilvie, Sarah 2013. Words of the World: A Global History of the Oxford English Dictionary. Cambridge: Cambridge University Press. Poplack, Shana 2004. Code‐switching. In Ulrich Ammon, Norbert Dittmar, Klaus‐Jürgen Mattheier, and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, 2nd edition. Berlin: Walter de Gruyter, 589–96.

Poplack, Shana and Nathalie Dion 2012. Myths and facts about loanword development, Language Variation and Change 24 279–315. Poplack, Shana and Marjory Meechan 1998. How languages fit together in code‐mixing. International Journal of Bilingualism 2: 127–38. Poplack, Shana, David Sankoff, and Chris Miller 1988. The social correlates and linguistic processes of lexical borrowing and assimilation, Linguistics 26: 47–104. Tadmor, Uri 2009. Loanwords in the world’s languages: findings and results. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter Mouton, pp. 55–75. Tadmor, Uri, Martin Haspelmath, and Bradley Taylor 2010. Borrowability and the notion of basic vocabulary, Diachronica 27 226–46. Thomason, Sarah G. 2001. Language Contact: An Introduction. Edinburgh: Edinburgh University Press. Thomason, Sarah G. 2003. Contact as a source of language change. In Brian D. Joseph Richard D. Janda, (eds.) The Handbook of Historical Linguistics. Oxford: Blackwell, pp. 687–712. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Townend, Matthew 2002. Language and History in Viking Age England: Linguistic Relations Between Speakers of Old Norse and Old English. Studies in the Early Middle Ages 6. Turnhout: Brepols. Weinreich, Uriel 1953. Languages in Contact: Findings and Problems. New York: Linguistic Circle of New York. (Reprinted 1964 The Hague: de Gruyter Mouton.) Wichmann, Søren and Anthony Grant 2012. Quantitative Approaches to Linguistic Diversity: Commemorating the Centenary of the birth of Morris Swadesh. Amsterdam: Benjamins (= papers from special issue of Diachronica, 27 (2010)). Winford, Donald 2003. An Introduction to Contact Linguistics. Oxford: Blackwell.

9 Contact and Code‐Switching PENELOPE GARDNER‐CHLOROS 1 Introduction Code‐switching (henceforth CS) is one of the most striking aspects of language contact to the observer. It provides the clearest synchronic evidence that contact is occurring, both in societies and for individuals. In this chapter I will discuss how it fits in with other manifestations and outcomes of contact, including borrowing and translanguaging on the one hand, and language change and shift on the other. CS occurs in contact situations of many types and relates in complex ways both to shift – the displacement of one language/variety by another – and change – meaning changes to the fabric of the varieties themselves. The relationship between CS, borrowing, and translanguaging, discussed in Sections 2 and 3, is an indirect one – as is the relationship between shift and change itself. In Section 3.3, an apparently simple example is given of the way in which languages in contact can be combined through CS. Understanding how and why this occurs involves getting to grips with the context, the motivations, and the contribution of CS to meaning‐making. In this context – that of a Syrian refugee’s Facebook posts – the role of CS in communicating a complex identity takes center stage. It is a fairly safe assumption that some CS will occur in most, if not all, contact situations. It is found among immigrant communities, regional minorities, and native multilingual groups alike. Gumperz and Hernández, who were among the first to study CS systematically, wrote that it could be found “each time minority language groups come into contact with majority language groups under conditions of rapid social change” (1969: 2). Others, e.g. Giacalone Ramat (1995), have on the contrary described it as a feature of stable bilingualism in communities where most speakers can speak both languages. There have been descriptions of contact situations in which CS receives little or no prominence, despite its undoubted presence, e.g. Jones (1998) on Wales, Spolsky and Cooper (1991) on Jerusalem. Other consequences of contact or restructuring may be the primary focus, or the data collection techniques may not center on the informal conversational modes where CS occurs. This is more unusual now, as CS is such a prominent issue in areas of current interest in sociolinguistics: the language of social – and other – media (see p. 192; “superdiversity” and the language of multicultural youth; and the construction and performance of identity/­identities; therefore CS has maintained its profile in much of the recent literature (Androutsopoulos 2013; Bassiouney 2012; Blommaert and Rampton 2011; Chen 2008; Flamenbaum 2014; Stell and Yakpo 2015). At the same time it is now viewed by some as just one sub‐category within a broader framework of language contact, the cover‐term of “translanguaging” having been put forward to encompass a wide range of flexible, personal and often multi‐modal manifestations of multilingualism (Garcia and Li Wei 2014; Li Wei 2018).

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

182  Penelope Gardner‐Chloros CS can embody different types of significance within a contact situation, from accommodation to divergence and from language maintenance to language shift. It reflects social differences and tendencies within the same society and language combination (Bentahila and Davies 1991; Li Wei 1998; Treffers‐Daller 1998), just as it reflects those between different societies and different language combinations (Poplack 1984; McClure 1998). Even before the concept of translanguaging was developed, CS was seen by many researchers as part of a “bigger picture” which includes other crosslinguistic manifestations in the speech of bilinguals (Backus 2015; Gardner‐Chloros 2009). In early research on the topic, CS was overwhelmingly investigated in spoken contexts, being considered typical of speech rather than writing. As a spoken mode, it has occurred throughout history, e.g. German–Latin CS in Luther’s “table‐talk” (Stolt 1964); or the speech of Russian aristocrats in the nineteenth century, as portrayed in Tolstoy’s novels, above all Anna Karenina (Timm 1975). While less has been written about this, it has also always been abundant in written texts, literary or otherwise (Forster 2010; Gardner‐Chloros and Weston 2015; Sebba, Mahootian, and Jonsson 2012). Examples include CS with Latin as one of the languages, in Cicero’s informal letters and other Roman writers (Mullen 2015); French– Italian CS in a thirteenth‐century Coptic phrasebook (Aslanov 2000); German–Latin CS in the works of a seventeenth‐century linguist, Schottelius (McLelland 2004); English–French‐ Latin CS in a variety of Medieval English texts (Trotter 2002; Schendl 2015), etc. Literary examples include Chicano poetry (Valdés‐Fallis 1976), Eco’s The Name of the Rose (where Salvatore speaks a multilingual jargon) and contemporary novels set in bilingual locations such as those by Zadie Smith (English–Creole in White Teeth) and John Markovitch (Spanish– Quechua in The Dancer Upstairs). A wide variety of non‐literary genres include writing in more than one language. Angermeyer (2005) described script‐alternation in contemporary Russian–American advertising as a way of signaling a bilingual identity. A Hassidic newsletter distributed in North London, Hakohol, includes code‐switches from Yiddish and Hebrew, the latter in Hebrew script. The lifestyle magazine Latina, aimed at American women of Hispanic descent, has copious Spanish–English CS. Mahootian (2005) describes this as a reflection of community norms, but also as an implicit challenge to relations of power and dominance between the older Hispanics, the monolingual English‐speaking community, and the younger Hispanic generation. More recently, written CS has been investigated in types of writing which reflect conversational practice, such as personal letters and the email and text messages of young people of mixed heritage (Hinrichs 2006; Androutsopoulos 2007). It is becoming clear that CS is common in written communication in social media even in contexts where it diverges from the spoken practice of the speakers, and is sometimes more frequent there than in the speech of its authors (Gardner‐Chloros, McEntee‐Atalianis, and Ateek, forthcoming; Hong Liu 2018). In the absence of the signals which are provided in a face‐to‐face encounter, CS allows multiple layers of meaning‐making. For example, Lee (2016) argues that CS is used as “a resource for self‐positioning and identity performance.” She describes a Facebook user who deploys Cantonese, English and code‐mixed utterances to alternate between a playful, down‐to‐earth personality and more serious topics on his Facebook page. Warschauer et al. (2007) found that their subjects used English for formal emails while Romanized Egyptian Arabic characterized the same subjects’ relaxed chatting online; and Georgakopoulou (1997) described switching between English and Greek in emails to express solidarity and simultaneously to signal in‐group membership. In some cases the alternation of scripts adds a further dimension to CS between varieties (Bailey 2001; Bianchi 2014; Kosoff 2014; Palfreyman and Al Khalil, 2007; Spilioti 2018; Warschauer, El Said, and Zohry, 2007). This leads to the interesting reflection that written language is intrinsically more resistant to innovation than speech. CS may (still) be considered in some senses nonstandard, but its

Contact and Code‐Switching  183 functions are so diverse – and it is so useful for expressing identity, a topic at the forefront in current sociolinguistics – that it enters the written language in many new and unexpected forms. It remains to be seen what effect these new forms of (written) linguistic co‐existence will have on the languages/varieties involved.

2  Code‐switching, Language Change and Language Shift 2.1  Synchronic and Diachronic Studies of Contact At a pragmatic level, CS has frequently been seen as a compromise or a negotiating strategy in bilingual settings, and thus essentially a mechanism for bringing the languages – and their speakers  –  closer together (e.g. Gal 1988; Han Chung 2006). On the other hand, at a grammatical level, several researchers who have carried out variationist studies have concluded that CS does not lead to convergence between languages (Poplack et al. 2012; Levey et al. 2013; Torres Cacoulos and Travis 2018). So while CS may be the most obvious symptom of bilingual contact, it may not be CS as such which brings about linguistic change. Poplack and Dion (2012) use synchronic variationist methodology to investigate the hypothesis that convergence with English as a result of CS has given rise to the use of phrase‐ final prepositions in Quebec French, as in: (1) J’avais pas personne à parler avec. ‘I had no one to talk to.’ As the authors explain, “Phrase‐final prepositions are prescriptively unacceptable in French, but they are the norm in English, the language with which it has been in intense contact in Canada” (2012, 204). From their analysis of a sizable corpus, they conclude that there is no empirical basis to suggest that it is CS with English that gave rise to this structure. Copious code‐switchers used the form no more frequently than sparse code‐switchers; moreover, linguistic conditioning for stranded prepositions was found to pattern differently to stranded prepositions found in the English spoken in the region, but similarly to another grammatical structure already present in Quebec French, pointing to language‐internal change. Another study, by Levey et al. (2013), came to similar conclusions. As is well‐known from studies of British and US English, be like is used as a quotative verb, and numerous other languages have developed a similar structure (2013). In European French, the closest equivalent are the structures être la or être genre which have followed a similar path of grammaticalization to the structure in English (Cheshire and Secova 2018). But in Canada, where French is in close contact with English, with widespread CS between the two varieties, a quotative structure which parallels the English expression much more closely is found, i.e. être comme. Somewhat counter‐intuitively, Levey et al., like Poplack et al., conclude from their analysis that this too is an internal development, unrelated to the CS and contact between English and French in Canada. A third study, by Torres Cacoulos and Travis (2016, 2018), focuses on clitic pronoun use in New Mexico, where Spanish is in (code‐switched) contact with English. Despite long‐term contact, the patterns of pronoun use by bilinguals continue to align with those of monolinguals in each language; so once again no direct link between CS and language change can be established. However, Torres et al. warn against concluding that this is representative of all cases of language contact: their findings, they specify, may be most typical of cases of long‐ term bilingualism where speakers develop an ability to switch back and forth while keeping the grammars separate. Such results are surprising and should lead us to ask questions about the methodology used – is there a danger of throwing the baby out with the bathwater? CS, as these authors

184  Penelope Gardner‐Chloros demonstrate, may not be directly responsible for specific language changes. However, it does flag up that contact is occurring and, for the speakers involved, increases exposure to the structures of the other language. It therefore acts as a red flag, pinpointing loci where change may occur. In some cases internal processes such as leveling and simplification may, more or less coincidentally, cause a change in language A which goes in the same direction as the structures in language B, and it may be possible to show that contact between the varieties did not cause the change. These are the sort of cases described on p. 183. But looked at over a longer diachronic period and over a large number of changes, it is clear that CS and borrowing are closely related processes, that they represent two different faces of linguistic contact and can result in radical changes. Thomason (2001a) is one of the best‐known authors who has considered this question in a diachronic perspective. She too is of the view that CS itself is not a major mechanism in contact‐induced change, but she does consider that CS is one of the principal mechanisms of borrowing. In support of this position, she points out that “necessary conditions for contact‐ induced change are by no means the same thing as sufficient conditions for change, especially in potential borrowing situations: sometimes, even in the most intense contact situations, with extensive bilingualism in (say) a minority group and with extreme cultural pressure, speakers of the minority language fail, or refuse, to borrow anything much from the dominant language” (2001b: 2). Her evidence, while not specifically concerning CS, shows the significance of contact for change in situations where CS is very likely to have been present. For example, in the case of the Greek community in Asia Minor before 1922, numerous features of Turkish were adopted from the Ottoman overlords. At the same time, native speakers of Turkish  –  i.e. the majority – also played a role in bringing features of Turkish into Asia Minor Greek, because some of them had learned Greek as a second language. Her book Language Contact provides numerous detailed examples of changes which have occurred in different languages as a result of sustained contact between varieties (2001a). This goes beyond the simple borrowing of individual words and can lead to wide‐ranging typological change. For example, at a syntactic level, there may be an increase in the frequency of one of several possible word orders, where one of the languages only has that particular order. Thomason gives the example of the Waikuru an language Kadiweu of Brazil, which allows a variety of word orders – OVS, VOS, SOV, OSV, VSO, and SVO, each of these variants being well attested. But bilingual speakers of Kadiweu and Portuguese show a preference for SVO word order, which corresponds with Portuguese (2001b: 3). Thomason cites similar findings by Pfaff (1979) and Silva‐Corvalán (1994), who also discuss bilinguals’ preference for syntactic patterns which are common to both their languages. Similarly in Middle English, which relexified under the influence of Norman French and Latin, Schendl (2000) argues that bilingualism and CS must have played a major role in the process of lexical borrowing, with mixed‐language texts providing information on the process of widespread relexification of English in the Middle English period. “Generations of educated Englishmen passed daily from English into French and back again in the course of their work, a process which must have led to specific lexical transfers both in the field of technical and of general vocabulary” (2002: 86). Schendl’s examples of CS between English, French, and Latin from a variety of medieval genres (business, religious, legal, and scientific texts) further support this claim. To sum up, rejecting CS as the mechanism of change does not mean rejecting contact, within which CS has a part to play. At the very least, contact plays a role in language‐internal changes due to the presence of the relevant structure in the available “feature pool” (Mufwene 2008). As argued in Gardner‐Chloros and Secova (2018), a structure which is present in this way in the “pool” and which has generally nonstandard connotations, may be taken up for specific sociolinguistic purposes – expressing identity or stance, thus introducing a “change

Contact and Code‐Switching  185 from below” which spreads to non‐code‐switchers. The latter group may end up outnumbering the code‐switchers  –  and lead us to mistakenly conclude that the structure is independent of contact. The advantage of contact‐based explanations is that they provide a motivation for change which internal change‐based explanations often lack; for example, changes do not always follow the path of simplification. Using a comprehensive and nuanced methodology would seem to be the key, with diachronic evidence from cases involving large‐scale restructuring providing the ultimate test.

2.2  The Shift‐change Interface Apart from historical studies such as those in the work of Thomason or Schendl, evidence of the link between CS and language shift may be found in apparent time studies which compare different generations (Chambers 2002; Wagner 2012). This is notably the case when there is a correlation between speakers’ age and the type of CS which characterizes their speech. In an early study of such a case in Greece, Trudgill (1976–7) found that CS from Arvanitika, a dying variety, into Greek was used as a compensatory strategy by younger speakers who were less competent in Arvanitika, making it difficult to identify which aspects of Arvanitika were being lost. Lavandera (1978) reported a similar phenomenon in Buenos Aires where migrants of Italian origin switched between their own variety of Argentinian Spanish, Cocoliche, and Standard Argentinian Spanish to compensate for reduced stylistic options available to them in either variety. Schmidt (1985) contrasted the older, more fluent speakers of Dyirbal, a dying Aboriginal language of Australia, who only used individual words in English, e.g. all right, now, finish (= period), with moderately proficient speakers who use many more imported English words, often adapted to Dyirbal morphology, e.g. ring‐iman ‘she phoned him’, jayil‐gu ‘to jail’, one night, down there, etc. Finally, the younger speakers, who were much less fluent in Dyirbal, code‐switched copiously between the two varieties: (2) We tryin’ to warn ‘We were trying to warn

ban wuigi nomo wurrbay‐gu. her not to speak [Dyirbal].’ (Schmidt 1985)

Apparent time studies such as these are sometimes criticized on the basis that they represent age‐grading (i.e. different linguistic behavior in different generations) rather than actual change over time; but while people definitely can change their pronunciation at different stages in their life (Harrington et  al. 2000), it is relatively unlikely that younger speakers using CS as a compensatory strategy when speaking their ancestral language would become substantially more fluent in it as they grow older. In such cases, therefore, CS would seem to be fairly uncontroversially symptomatic of shift – which is not the same, of course, as being the cause of it. Variations in CS patterns between different subgroups, which are sometimes overlooked in studies of CS communities, may need to be considered where the pattern is less than clear: for example Weston (2012) found clear differences in the code‐switching patterns characterizing younger and the older groups in Gibraltar, which were related to differences in the role played by Spanish and English in their education, background, and individual networks.

2.3  CS and Borrowing In order to make sense of disparate results regarding the role of CS in language change, a closer look at the relationship between CS and borrowing is necessary, since the latter is by definition evidence of language change. This relationship has been widely discussed in the literature (Romaine 1995; Myers‐Scotton 1992; Amuzu 2014).

186  Penelope Gardner‐Chloros As is often pointed out, single words – and in particular common nouns – are both the most frequently borrowed and the most frequently code‐switched items in many corpora (Poplack, Sankoff, and Miller 1988: 62). One explanation of this is provided by Bynon (1977: 231), who says this just reflects the size of the grammatical categories concerned. Another is suggested by Aitchison (1994: 62), who points out that nouns are freer of syntactic restrictions than other word classes. Both these explanations involve language‐internal factors rather than sociolinguistic ones. A third reason for the prevalence of single‐word switches/loans is that these are accessible to bilinguals with even minimal competence in the donor language. At a synchronic level, there is no failsafe method of distinguishing between loans and code‐ switches, as only time can tell if a loanword is more generally adopted over time. Every loan must start life as spontaneous CS, or a calque in the speech of an individual bilingual; some of these switches then spread among speakers of the host language (Haust 1995; Gardner‐Chloros 1995). The diachronic nature of this process is confirmed by the fact that transfers which occur at different historical stages of contact between the languages may go through quite different processes of integration and end up looking quite different in the receiving language. Heath (1989), for example, has shown how some French verbs borrowed in the early French colonial period were adopted in Moroccan Arabic without inflectable verb frames. Others borrowed more recently have been instantly provided with Moroccan Arabic inflectional frames. Their phonemes have also been imported wholesale, and show signs of stabilizing as such (1989: 203). Muysken (2000) shows similar discrepancies between Malay importations into Dutch in the original setting and in the post‐colonial setting of the Netherlands. A comprehensive model which integrates both CS and borrowing was proposed by Auer (1999). In this model, CS is seen as the first point in a chronological progression along a continuum. At the CS stage, the point in the sentence where there is a switch is a significant aspect of the conversation. The next stage is language mixing, where it is not the individual switch points which carry significance, but the use of the overall switching mode – a stage also described by Myers‐Scotton (1993b) as “switching as an unmarked choice.” The third stage is that of fused lects, which are stabilized mixed varieties. This coincides with a loss of variation: the use of elements from one or other variety is no longer a matter of choice, but of grammatical convention. Structures from each variety, which are equivalent in monolingual usage, develop specialized uses. In some cases it may never be completed, as bilingual repertoires may stabilize at any point along the way (as in the case of mixed languages like Warlpiri or Chiac), but it does not allow for any movement in the opposite direction; in other words the fused lects can be expected to spread. This view of borrowing receives support from a study of English borrowing in Japanese. Stanlaw (2004) describes how Japanese has adopted numerous English color terms such as pinku (pink), guree (gray), and orenji (orange), which do not double up for Japanese equivalents but instead replace them, with a shift in meaning or area of application (the Japanese word for gray, for example, has negative connotations and is never applied to clothes). Reflecting upon this, Stanlaw considers that these terms are not direct transfers from English, but “home‐grown,” English‐inspired items created and negotiated within Japanese society. As he points out, the metaphor of “borrowing” is inadequate to describe complexities of cultural and linguistic contact. Given the complexity of differentiating loans and code‐switches in different settings, what are the criteria that might be used to draw a line between the two?

2.3.1  Grammatical Category Nouns may, as mentioned, be the most frequently borrowed – and switched – word class owing to their grammatically self‐contained character, but all grammatical categories are potentially transferable (Boeschoten 1998). The transfer of grammatical structures is characteristic of particular types of sociolinguistic context. Recently examined examples of

Contact and Code‐Switching  187 grammatical change linked to CS include Dorvlo (2014), who describes a “contact‐induced change in progress” in Ghana, where Logba is code‐switched with Ewe, the majority language. Almost all the 7,500 Logba speakers are bilingual, and young speakers are introducing a simplification of the noun class system: a plural suffix borrowed from Ewe is generalizing to all Logba nouns, including those which take a plural prefix in their traditional form. Also in a Ghanaian setting but with different varieties at play, Amuzu (2014) shows how the use of English nominals in English‐Ewe CS brings about changes in Ewe grammar, a phenomenon which he links with the modularity hypothesis. Comparative studies are relatively rare, but can be revealing. A study comparing Punjabi English and Greek‐Cypriot English bilinguals showed that it is not always single nouns which are most frequent transfers. Whereas single‐word switching was the commonest type among a group of London Greek Cypriots, intrasentential switching was almost three times as frequent as single‐word switching among a comparable group of London Punjabi speakers (Cheshire and Gardner‐Chloros 1998). The Punjabi speakers switched massively more than the Greek Cypriots overall, despite the fact that typologically speaking, it should be easier to produce complex switches between Greek and English than between Punjabi and English. This underscores the role of broader sociolinguistic factors in determining the outcome of contact: in this case the relevant factors were, most probably, less normative attitudes towards using the two varieties within the Punjabi‐speaking community, plus the history of contact and CS even in the country of origin.

2.3.2  Necessary and “Luxury” Loans It has also been suggested that a characteristic of loans is that they fill a “lexical gap” in the host language, as in: (3) No me precipitaré en el famoso name‐dropping. ‘I will not throw myself headlong into the famous practice of name‐dropping.’ (McClure 1998: 134) whereas code‐switches tend to substitute for a “native” equivalent. This is negated by the Japanese color‐term loanwords mentioned on p. 186, which show that such so‐called “luxury” loans often enter the language for reasons other than lexical necessity (Campbell 2013). Conversely CS may be a way of “supplementing” the resources of the host language. True translation equivalents are extremely rare, and the use of an expression from another variety, almost by its nature, carries a slightly different meaning/different connotations.

2.3.3  Morphophonemic Integration With the Surrounding Language It is also argued that transfers which are morphophonemically integrated with borrowing languages should be classed as loans – though phonology on its own is considered a less “secure” criterion (Poplack and Dion 2012). In fact, the integration of loan words shows considerable variation at the individual level that can be attributed to factors such as context, interlocutors, orientation towards the language providing the loans, etc. (Furiassi et al. 2012). Both ad hoc code‐switched nouns and verbs, as well as more established loans, can take on the morphophonology of the borrowing variety, as examples from many language combinations have shown (Eliasson 1989; Eppler 1999; Gardner‐Chloros 1991).

2.3.4  A Progression Towards Complementarity As has been noted, words originally introduced via CS acquire a specialized meaning which differs from the synonymous “native” term. Onysko and Winter‐Froemel (2011) found that borrowings from English to German, whilst semantically equivalent to a German word,

188  Penelope Gardner‐Chloros were used in different ways or with a different scope than in English, e.g. shoppen means ‘to shop for leisure’ rather than as a chore, for which speakers still use einkaufen ‘to shop’ (2011: 1557). Complementary meanings for “native” and borrowed discourse markers were reported by Oesch‐Serra (1998), with respect to the adoption of Italian ma ‘but’ in the French of Italian migrants in Switzerland. The French mais continued to be found but each of these discourse markers took on a slightly different function. In a study reminiscent of the Japanese case on p. 186, Lantto (2015) found that Spanish pragmatic markers were used systematically instead of their Basque equivalents by a sample of Basque speakers in Bilbao. However, she did not conclude that these were established borrowings as they retained their “Spanishness,” but instead suggested that the intermediate category of “conventionalized CS” was required to cover such instances. Treffers‐Daller (1994) was led by similar findings to the conclusion that a unified theory of borrowing and CS was needed: in a study of French–Dutch contact in Brussels, she found many words used by her subjects, such as unique and sympathique, belonged equally to either variety (i.e. there was no variation between them and native Dutch equivalents regardless of the language of the conversation). Furthermore, as both French and Dutch have limited morphological marking, the criterion of morphological integration could not be used to determine to which language they “belonged.” Third, both borrowings and code‐switches varied as to whether they were syntactically integrated or unintegrated. In other cases, a distinction between loans and code‐switches may be justifiable due to specific aspects of their use. For example, in Jones (2005), speakers of Jersey French (‘Jèrriais’) tended to flag code‐switched forms (with pauses, self‐corrections, etc.), but not loans; also, those informants with the most positive attitude toward Jèrriais avoided CS entirely, though they did use borrowings. Thus it is less the linguistic phenomena as such than the way in which speakers use them which determines whether “CS” or “loan” is the most suitable characterization. The different possible criteria (e.g. lexical necessity and morphophonemic integration) do not always coincide and the distinction between borrowing and CS ­frequently remains debatable. 2.3.4.1  Mechanisms of Change While the exact mechanisms of language change remain hard to pin down, it seems undeniable that individual bilinguals, exposed to two different structures for expressing a similar meaning, are likely to make choices between them and thereby have the role of catalysts for change, whether they code‐switch abundantly or not. CS can take place in a context where shift is occurring to a greater or lesser extent, but does not constitute shift of itself. Moreover within a given contact situation, convergent and nonconvergent tendencies can co‐exist. This was shown for example in Gumperz’s early study of Hindi–Punjabi CS in Delhi (1964), where CS persisted at a lexical level although the two – closely related – varieties had substantially converged at a grammatical level after decades of contact. Gumperz concluded that the contrast between the varieties remained functional in identity terms, as the languages would otherwise have converged entirely. Rindler‐Schjerve’s (1998) study of CS in Sardinia shed further light on its complexity in particular situations. She emphasized that although the switching occurs in a context of language shift, it “should not be seen as a mechanism which accelerates the shift” (1998: 247). If CS was a straightforward mechanism of change, one would expect the less proficient bilinguals to code‐switch most. But unlike the Trudgill example on p. 185, in Sardinia it is the more balanced bilinguals who switch most, and who also “contribute to the maintenance of Sardinian in that they change the Sardinian language by adapting it to the majority language, thus narrowing the gap between the two closely related codes” (1998: 246). The variations in code‐switching behavior between different groups and different individuals within the same society pose a methodological challenge: they show that CS cannot

Contact and Code‐Switching  189 merely be classified at the community level (Muysken 2000: 221), but requires a more individuated approach. Contact situations occupy widely differing time‐spans, and different sections of the same community may be implicated in very different ways. Shift can take place over several generations or, effectively, within a single generation. It may be difficult to detect, losses being masked by code‐switches (Trudgill, 1976–7) or internal restructuring (Tsitsipis 1998), and it can occur either with heavy linguistic symptoms such as morphological loss or without them (Dorian 1981; Schmidt 1985). That CS can lead to major changes in a community’s linguistic make‐up is demonstrated by the case of so‐called “mixed languages,” for example Warlpiri, spoken in Australia’s Northern Territory (O’Shannessy 2012, 2016) and Chiac, spoken in Moncton, Canada (Perrot 1995, 2014), varieties which are understood to derive from stabilized forms of CS. 2.3.4.2  An Alternative Model: The MLF A different approach to tracing the linguistic pathways through which language change interacts with CS is represented by Carol Myers‐Scotton’s Matrix Language Turnover Hypothesis (1998). This grammatically based approach relies on a framework proposed by Myers‐Scotton in a series of publications (1993a 2000, 2002) where CS is analyzed in terms of the interaction of a Matrix and an Embedded language – the Matrix Language Frame or MLF. The notion of Matrix language, its justifiability on the basis of various criteria and its applicability to all cases of CS has been extensively debated; the reader is referred to discussions in, e.g., Muysken (2000) and Gardner‐Chloros (2009). Briefly, Myers‐Scotton argues that in “standard” CS one variety dictates the grammatical frame of the utterance, with another variety providing “embedded” elements. Recognizing that one is rarely dealing with two discrete, self‐contained “languages,” Myers‐Scotton concedes that, in some cases, the frame itself may be ultimately derived from two contributing varieties, giving a “composite” Matrix. Authors such as Amuzu endorse this composite CS theory (taken from Myers‐Scotton 2002) as “an economical approach to… the bilingual CS phenomenon” (2014: 405). Myers‐Scotton’s model also allows for the roles of the varieties to be reversed within a given community over time, with the old embedded language becoming the new Matrix language. The model’s main weakness is the difficulty of identifying the Matrix conclusively, especially at the level of a whole community, since this implies at least a certain level of stability and uniformity in linguistic practices. In fact, different kinds of CS can co‐exist in the same community at a given time, and it is empirically challenging to distinguish situations where there is a progression from one state to another from those where the different phenomena overlap chronologically. Take, for example, cases where the community includes speakers with different life‐histories, who ostensibly speak the same languages. In Hong Kong, the children of “returnee” migrants who left Hong Kong after Tiananmen Square were brought up in various English‐speaking countries (the UK, Canada, New Zealand, Australia, the US), and have a code‐switching style very different from the local one (see Gibbons 1987). The non‐indigenous style is widely recognized or “enregistered” and is perceived to index a kind of ethnolinguistic betrayal by local HK people (Chen 2008; Weston 2016). In such circumstances contact‐induced change does not proceed smoothly through Auer’s three stages, nor can a clear “turnover” in Myers‐Scotton’s sense be identified. Instead of an old variety being (gradually) abandoned in favor of a newer one, different code‐switched varieties, brought about through contact, co‐exist and assume distinct (identity‐related) functions.

2.4  Stabilization of CS Varieties and Identity CS varieties often stabilize when they assume such identity‐related functions. They may for example become characteristic of young second‐generation immigrant communities which develop a pride in their mixed identity. Over a couple of generations the ability to

190  Penelope Gardner‐Chloros actually code‐switch may fade as a function of shrinking competence in the parental or grand‐parental language, till it subsists only in emblematic form; for example young males of North‐African heritage in the Paris banlieues tend to use Arab words only for swearing, for references to religion and for the occasional discourse marker (Gardner‐Chloros and Secova 2018). In other cases a CS variety can become fully enregistered. Among second‐generation Portuguese in France, a variety known as immigrais ‘immigrese’, by the speakers’ own description a “bastard” Portuguese, including French words and colloquialisms, is understood by the 900,000 strong immigrant population, and is used in writing too. Albatroz, a literary review, is trilingual in Portuguese, French, and immigrais (Muñoz 1999: 31). Hewitt (1986) and Sebba (1993) described types of CS which occur with symbolic and discourse‐ related functions in London, between Jamaican Creole and London English‐speaking young people. Although used sparingly, Creole is said by Sebba to have a we‐code function and CS is described as an “insider activity.” Creole is almost certainly perpetuated by being used in this way, as the majority of young people he studied would not be able to use  it as a means of expression on its own. Zadie Smith’s White Teeth portrays young Afro‐Caribbeans in London who use this type of CS as an intrinsic  –  if somewhat ­inconsistent – aspect of their speech. In any complex linguistic situation, numerous processes operate simultaneously. Recent research has attributed much linguistic innovation in large European cities to the existence of so‐called “multiethnolects” in the speech of young people of immigrant origin (Cheshire and Gardner‐Chloros 2018; Cheshire, Kerswill, Fox, and Torgerson 2011; Nortier and Dorleijn 2013; Wiese 2009). Although the term “multiethnolect” sounds as if it might apply to CS, it in fact designates varieties of the local majority language (English, Swedish, French) used by speakers of many and mixed ethnicities. Code‐switching as such may be present in these varieties only symbolically, or at least, not in the peer‐group interactions which are particularly propitious for the introduction of new structures, though it may have been common in the speech of their parents.

3  A New Variety? 3.1  Vitality and Decline As has been seen, CS can occur both in situations of decline and as a mechanism of vitality. An example of the first is provided by Aikhenvald (2002), who shows how an Amazonian community, the Tariana, have a strong inhibition against any kind of mixing of their language with that of other local tribes, such as the Tucano and the Makú, whom they despise. When such mixing occurs, it evokes ridicule and pity. However, they do increasingly allow CS with Portuguese and even with English, which symbolizes “everything a capitalist paradise could offer” (2002: 211). CS with a more positive identity function is illustrated by Gibbons (1987), who describes a variety called “MIX” spoken by students at Hong Kong University. Unlike a creole, MIX has emerged within a single community and the processes which have occurred have taken place entirely within this relatively homogeneous group. Heller (1988) suggested as a priority for CS research looking at “the extent to which different types of CS are related to different types of boundary maintenance/change processes” and “the generalizability of findings concerning the social conditions under which CS is or is not found” (1988: 268–9). More comparative research is needed for us to realize this aim.

Contact and Code‐Switching  191

3.2  CS and Pidginization/Creolization CS is found alongside pidginization and creolization/decreolization in many parts of the world, and contributes to the process of convergence of different varieties. Gumperz (1964) and Gumperz and Wilson (1971) identified the close relationship between CS and creolization in the early days of CS research; more recent research generally emphasizes the differences. But despite the differences, the relevant processes often co‐occur, derive from similar social factors, and sometimes lead to similar outcomes. Crowley (1990), for example, argues that in Melanesian Pidgin, access to the substratum persists at all levels: stabilization, destabilization, and creolization should all, he claims, be regarded as contributing simultaneously to the gradual evolution of the language (1990: 384–6). As to CS, its presence or absence seems to be largely a function of the prestige attached to the pidgin. Comparing usage on the radio in Vanuatu and the Solomon Islands, Crowley finds more CS in the Solomon Islands, where the pidgin’s prestige is lower than in Vanuatu, to such an extent that with Solomon Islands radio announcers, “It is sometimes difficult to know which language they would claim to be speaking.” In Papua New Guinea, Romaine (1992) showed how CS occurs within a post‐creole continuum which has emerged over a period of 20 years: “In town, Standard English, English spoken as a second language with varying degrees of fluency, highly anglicized Tok Pisin, more rural Tok Pisin of migrants, and the creolized Tok Pisin of the urban‐born co‐exist and loosely reflect the emerging social stratification (1992: 323). Whereas some have been adamant that this situation should be seen in terms of ongoing CS between Tok Pisin and English (Siegel 1994), Romaine believes that there is no principled way for determining to which language many utterances belong (1992: 322). In their classic book Acts of Identity, Le Page and Tabouret‐Keller (1985) gave examples of creole speech from sources in the Caribbean, Belize, and London Jamaican. In these linguistically diffuse situations, CS often involves shifting at particular linguistic levels rather than a wholesale transition from one variety to another. Discrete codes are hard to come by in such unfocused linguistic contexts. As with London Jamaican, the switching between the codes is more symbolic than real: London Jamaican is more a set of norms to be aimed at than an internally coherent and consistent system. Speakers behave as if there were a language called “Jamaican”, but often all they do (perhaps all they know how to do) is to make gestures in the direction of certain tokens associated with Jamaican Creole which have a stereotypical value. (Le Page and Tabouret‐Keller 1985: 180; see also Sebba 1993)

In other cases, CS varieties flourish. Code‐switching between Warlpiri and Aboriginal English/Kriol by parents in baby talk to their children has led to the emergence of a new mixed language, “Light Warlpiri,” by the children who first heard the code‐switched form and subsequently conventionalized and expanded its syntax (O’Shannessy, 2016). The new generation of children have different names for Light Warlpiri and Warlpiri (Lajamanu stail ‘Lajamanu style’ and Yurntumu stail ‘Yuendumu style’), supporting the argument that Light Warlpiri is a separate language system (O’Shannessy, 2012: 91; O’Shannessy and Davidson, this volume). The essential finding is that different types of contact‐induced varieties do not have clear boundaries (Thomason 1997). Heath (1989) gives examples of French borrowings into Arabic and Arabic borrowings into Turkish, which occur in intensive contact zones with alternative, productive adaptation routines. Along with the avoidance of inflections reminiscent of pidgins, some of these importations are also unstable with regard to word class, such as himri in Moroccan Arabic (from English hungry), which fluctuates between noun and

192  Penelope Gardner‐Chloros adjective (1989: 202). The difference lies in the social function of the particular variety. Whereas pidgins and creoles arise as lingua francas, other bilingual mixtures of this type “usually or always serve as salient markers of ethnic‐group identity” (Thomason 1997: 4). Time and again, one finds that understanding the role of CS in language contact – indeed understanding its presence or absence – depends on grasping the interplay of structural and social factors. Treffers‐Daller (1994, 1999) compared CS involving French and Brussels Dutch in Brussels with that involving French and the Alsatian dialect in Strasbourg (Gardner‐ Chloros 1991). In both cases, a Germanic language is in contact with French, and both Alsatian and Brussels Dutch have borrowed extensively from French in their respective contexts. However, in terms of identity, the two situations are very different. Whereas a mixed identity, encompassing both French and Alsatian, is the norm in Strasbourg, in Brussels the two linguistic groups are polarized: almost all the population identify either with French or Dutch, but not with both. Consequently, as Treffers‐Daller points out, Brusselers “no longer consider the mixed code to be an appropriate expression of their identity” and intrasentential CS is accordingly infrequent.

3.3  Intervening Variables: Identity Since the relationship between CS and language contact is complex and indirect, it is necessary to look at CS in context, and in relation to other aspects of the discourse, if its functions in a given case are to be understood. A significant locus of language contact nowadays is that of online writing. Different varieties interact online as well as in speech, and much recent research on CS focuses on its use in social media (Georgakopoulou 1997; Androutsopoulos 2007; Palfreyman and Al Khalil 2007; Tsiplakou 2009; Warschauer, El Said, and Zohry 2007), where it fulfills a variety of functions. Some of these functions are comparable to those of spoken CS: for example, the “inclusive” function of CS, which allows the speaker (or poster) to address several audiences and/or to target comments to specific members of the potential audience, is reproduced in online settings. Other functions have more to do with creating/expressing a particular identity online (for an up‐to‐date summary of the surrounding issues see McEntee‐Atalianis 2019). In a recent study, McEntee-Atalianis, Ateek and Gardner-Chloros (forthcoming) examined over 2,000 posts of a group of Syrian refugees on Facebook. This group of Facebook “friends” consisted of three groups: Syrians settled in the UK, British friends, and family and friends back in Syria. Most of the posts were in Arabic (with some switching between Standard Arabic, Syrian Arabic, and local dialects) and around 10% included CS with English. Some similar functions as in spoken CS were found (e.g. we‐code/they code, addressee specification); but the CS in these posts differs significantly from these speakers’ spoken CS, where it is much less prevalent. The same finding emerges from a systematic study by Hong Liu (2018) which compared online posting by 40 Chinese‐English bilinguals with interview data from the same subjects: both the quantity and the complexity of CS online was greater, with participants reporting scruples towards using complex alternational CS when conversing with fellow Chinese. Thus different motivations and standards may be applied to speech and to writing on social media, just as different standards apply to casual speech and traditional/more formal writing. The following extract from the data in McEntee-Atalianis et al. (forthcoming) provides an example of CS in the online context. A young Syrian settled in the UK constructs his identity through a series of “staged” exchanges between himself and an imagined Shia and Kurd interlocutor. He positions himself as different from other countrymen in that he refuses to “sweep under the carpet” divisions between different groups of Syrians in the UK, which some choose to ignore:1

Contact and Code‐Switching  193 (4) I see the Shia in Britain and (he) tells me you alright Bro. I see the Kurd and (he) tell(s) me Bro. !! As if we do not have any issues in our countries!! ⍰ Brothers !!!! I just want to understand why we are enemies inside and brothers outside!!!? Bros Regards Bros His sense of difference – and of internal divisions being strategically ignored – is conveyed via different “voices” emblematically represented. The CS is an essential aspect of the characterization: whereas the poster himself addresses his audience in Syrian Arabic, the (imagined) interlocutors use young Londoners’ vernacular: “You alright Bro?” He then explains his position, questioning the stance of these fellow Syrians, and their tendency to skate over divisions which would be highly significant in their country of origin. The implication is that they are either insincere, or neglectful of important issues within the community – note his use of the strong term “enemies.” He elaborates on his feelings of discomfort in his final statement, expanding the implications of the casual address term “Bro” to its full form “Brothers.” He implies that the term ‘Bro’ should carry some meaning – one of fellowship or similarity. He signs off using the term himself – after all he too is a young Londoner – but with ironic intent. This brief passage shows that CS can be highly strategic – sometimes even one word in a different variety can be enough to suggest a different “voice.” The boundaries which the poster wishes to convey are clear: he rejects the use of English vernacular address terms implying solidarity between fellow countrymen in a displaced context. For him, the ethnic/ nationalistic/religious differences (Sunnis v Shia/Arab vs Kurd) embody continuing tensions between these groups which cannot be ignored in the UK. At the same time the use of both varieties in a single post underpins a sense of community with those readers who understand both. Like him, they have to navigate the script‐switching which involves writing/reading from left to right and from right to left within the same passage – thus taking part in an “in‐between” type of activity, a “third space,” as described by Bhatt (2008), between speaking and writing, between Syria and the UK, Arabic and English. While this much is shared, other aspects of their identity are not. As to the intended audience, it could include friends still in Syria, showing them the strange transformations undergone by their countrymen in the UK; but also those who are in the UK, including the Shias and Kurds whom he is implicitly criticizing. Such strategic uses of CS are not peculiar to the online context and are often found among young people with a composite sense of belonging. In Paris, young males of Arab heritage but with minimal competence in Arabic use terms from their Islamic religious background, halal or haram to show moral approval or disapproval of young women in their peer‐groups (Gardner‐Chloros and Secova 2018). In the London Greek‐Cypriot community, where convention dictates that women should be less direct and assertive than men, young women switch to Greek‐Cypriot dialect to make jokes or to interrupt, the shared we‐code acting as a way mitigate any suggestion of face‐threatening (Cheshire and Gardner‐Chloros 1998).

194  Penelope Gardner‐Chloros Both written and spoken CS contribute in a multiplicity of ways to meanings beneath the surface message, and the difficulty of characterizing the role of CS in language contact comes from its enormous versatility within the bilingual armory. Because of its ubiquitous and flexible uses it will continue to be present in contact situations, whether they involve convergence/borrowing between the varieties or not.

4 Conclusion In this chapter I have pointed out the difficulties of pinning down the relationship between CS and language contact, specifying that CS is a symptom of contact rather than being identical with it. Contact itself is undeniably a major factor in change, but its relationship with internal change processes, e.g. simplification or leveling, is still not fully understood. Conversely, the study of languages in isolation  –  to the extent that this is a possible scenario – can contribute to our understanding of why certain varieties may remain relatively stable (Schreier 2009). To take an example: a variety of Greek known as Romeika, spoken in present‐day Turkey, has, since Hellenistic times, preserved the infinitive form, which has been lost from Modern Greek spoken elsewhere. As Sitaridou (2014) explains, while other varieties of Greek were in contact with various Balkan languages, Romeika was cut off from medieval Greek varieties between the fourteenth and sixteenth centuries ce. Such a case allows us a glimpse of the range of sociolinguistic factors which need to be considered to explain both change and the lack of it. CS is a good example of the fact that theoretical disagreements between researchers are  sometimes explicable in terms of the different data and circumstances they are ­analyzing – one size does not fit all, as Muysken (2000) demonstrated. In certain bilingual situations varieties alternate without fundamentally altering their essential character; in  other cases CS is rule‐breaking behavior, symptomatic of a shifting linguistic scene. A fuller exploration of these mechanisms is still needed (see the papers in Bilingualism, Language and Cognition 7.2, 2004). At the individual level, Otheguy claims that convergence occurs at the semantic–pragmatic interface, with bilinguals selecting “the most ­parsimonious grammar that serves both languages” (2004: 177). In a similar vein, Bullock and Toribio (2004) distinguish CS from interference and transfer; rather than implying the  imposition of a structural property from one language on another, they see it as an  “enhancement of inherent structural similarities found between two linguistic ­systems” (2004: 91). In many cases, the issue is complicated by the fact that CS coexists and overlaps with other symptoms of contact. Further research on the processes which are considered under the broader umbrella of “translanguaging” may in future help to clarify these issues. To repeat: the key to understanding why CS takes the form it does in any specific case,  and whether it is implicated in convergence, lies in scrutinizing the relevant sociolinguistic factors such factors affect different subgroups and individuals in ­ ­different  ways, and different types of CS often co‐exist within the same community, or  even within the productions of a single individual in different circumstances. At a sociological level, CS has been described as the product of a power struggle between two varieties (Pujolar 2001). At an individual level, it reflects varying bilingual competences and strategies and allows complex identities to be expressed. It serves both to structure discourse and to embody a variety of stances. Like other outcomes of language contact discussed in this volume, CS is not a random and unregulated phenomenon, but is rooted in deliberate, if oftentimes unconscious, choices made by speakers in specific discourse situations.

Contact and Code‐Switching  195

NOTE 1 This chapter draws partly on material from Code‐Switching by the same author (Cambridge University Press, 2009). The permission of Cambridge University Press to reprint extracts from the book is acknowledged with thanks.

REFERENCES Aikhenvald, Aleksandra Y. 2002. Language Contact in Amazonia. Oxford: Oxford University Press. Aitchison, Jean 1994. Words in the Mind, 2nd edition. Oxford: Blackwell. Amuzu, Evershed Kwasi 2014. Producing composite codeswitching: the role of the modularity of language production. International Journal of Bilingualism 18: 384–407. Androutsopoulos, Jannis 2007. Language choice and code‐switching in German‐based diasporic web forums. In Brenda Danet and Susan C. Herring (eds.), The Multilingual Internet. Oxford: Oxford University Press, pp. 340–62. Androutsopoulos, Jannis 2013. Code‐switching in computer‐mediated communication, in Susan C. Herring, Dieter Stein, and Tuija Virtanen (eds.), Handbook of the Pragmatics of Computer‐Mediated Communication. Berlin: de Gruyter Mouton, pp. 659–86. Angermeyer, Philipp S. 2005. Spelling bilingualism: script choice in Russian American classified ads and signage. Language in Society 34.4: 493–531. Aslanov, Cyril 2000. Interpreting the language‐ mixing in terms of codeswitching: the case of the Franco‐Italian interface in the Middle Ages. Journal of Pragmatics 32: 1273–81. Auer, Peter 1999. From codeswitching via language mixing to fused lects: toward a dynamic typology of bilingual speech. International Journal of Bilingualism 3.4: 309–32. Backus, Ad M. 2015. A usage‐based approach to codeswitching: the need for reconciling structure and function. In Gerald Stell, and Kofi Yakpo (eds.), Code‐switching Between Structural and Sociolinguistic Perspectives. Berlin: de Gruyter Mouton, pp. 19–37. Bailey, Benjamin 2001. The language of multiple identities among Dominican Americans. Journal of Linguistic Anthropology 10.2: 190–223.

Bassiouney, Reem 2012. Politicizing identity: Code choice and stance‐taking during the Egyptian revolution. Discourse and Society 23.2: 107–26. Bentahila, Abdelâli and Eirlys Davies 1991. Constraints on code‐switching: a look beyond grammar. Papers for the Symposium on Code‐Switching in Bilingual Studies: Theory, Significance and Perspectives. Barcelona: European Science Foundation, pp. 369–405. Bhatt, Rakesh M. 2008. In other words: language mixing, identity representations, and third space. Journal of Sociolinguistics 12.2: 177–200. Bianchi, Robert 2014. Language and topic. Choice among prolific and non‐prolific posters on an Arabic‐English Website. International Journal of Social Science and Humanity 4.2: 128–31. Blommaert, Jan and Ben Rampton 2011. Language and superdiversity. Diversities 13.2: 1–21. Blommaert, Jan and Ben Rampton 2011. Language and superdiversity. Diversities 13.2: 1–21. Boeschoten, Hendrik 1998. Codeswitching, codemixing, and code alternation: what a difference. In Rodolfo Jacobson (ed.), Codeswitching Worldwide. Berlin: de Gruyter Mouton, pp. 15–24. Bullock, Barbara and Almeida J. Toribio 2004. Introduction: convergence as an emergent property in bilingual speech. Bilingualism, Language and Cognition 7.2: 91–3. Bynon, Theodora 1977. Historical Linguistics. Cambridge: Cambridge University Press. Cameron, Deborah 1992. Feminism and Linguistic Theory. London: Macmillan. Campbell, Lyle 2013. Historical Linguistics: An Introduction, 3rd edition. Edinburgh: Edinburgh University Press. Chambers, Jack K. 2002. Patterns of variation including change. The Handbook of Language Variation and Change, pp. 358–61.

196  Penelope Gardner‐Chloros Chen, Katherine Hoi Ying 2008. Positioning and repositioning: linguistic practices and identity negotiation of overseas returning bilinguals in Hong Kong. Multilingua – Journal of Cross‐ Cultural and Interlanguage Communication 27.1–2: 57–75. Cheshire, Jenny and Penelope Gardner‐Chloros 1998. Code‐switching and the sociolinguistic gender pattern. International Journal of the Sociology of Language 129: 5–34. Cheshire, Jenny and Gardner‐Chloros, Penelope 2018 Introduction: multicultural youth vernaculars in Paris and urban France. Language Innovation and Change in Paris. Special Issue, Journal of French Language Studies, 161–5. Cheshire, Jenny, Paul Kerswill, Susan Fox, and Eivind Torgersen 2011. Contact, the feature pool and the speech community: the emergence of Multicultural London English. Journal of Sociolinguistics 15: 151–96. Cheshire, Jenny and Maria Secova 2018. The origins of new quotative expressions: the case of Paris French. Language Innovation and Change in Paris. Special Issue, Journal of French Language Studies, 209–34. Crowley, Terry 1990. Beach‐la‐Mar to Bislama: Emergence of a National Language in Vanuatu. Oxford: Oxford University Press. Dorvlo, Kofi 2014. Ewe borrowings into Logba. International Journal of Bilingualism 18.4: 428–46. Dorian, Nancy 1981. The Life Cycle of a Scottish Gaelic Dialect. Philadelphia: University of Pennsylvania Press. Eliasson, Stig 1989. English–Maori Language contact: code‐switching and the free morpheme constraint. Reports from Uppsala University Department of Linguistics (RUUL) 18: 1–28. Eppler, Eva 1999. Word order in German–English mixed discourse. University College London Working Papers in Linguistics 11: 285–308. Flamenbaum, Rachel 2014. The pragmatics of codeswitching on Ghanaian talk radio. International Journal of Bilingualism 18.4: 346–62. Forster, Leonard 2010. The Poets’ Tongues: Multilingualism in Literature: The de Carle Lectures at the University of Otago. Cambridge University Press Furiassi, Cristiano, Virginia Pulcini, and Félix Rodríguez González (eds.) 2012. The Anglicization of European Lexis. Amsterdam: Benjamins. Gal, Susan 1988. The political economy of code choice. In Monica Heller (ed.), Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: de Gruyter Mouton, pp. 245–64.

Garcia, Ofelia and Li Wei 2014. Translanguaging: Language, Bilingualism and Education. Basingstoke: Palgrave Macmillan. Gardner‐Chloros, Penelope 1991. Language Selection and Switching in Strasbourg. Oxford: Oxford University Press. Gardner‐Chloros Penelope 1995. Code‐switching in community, regional and national repertoires: the myth of the discreteness of linguistic systems. In Lesley Milroy and Pieter Muysken (eds), One Speaker, Two Languages: Cross‐Disciplinary Perspectives on Code‐ Switching. Cambridge: Cambridge University Press, pp. 68–90. Gardner‐Chloros, Penelope 2009. Code‐Switching. Cambridge: Cambridge University Press. Gardner‐Chloros, Penelope and Maria Secova 2018. Grammatical change in Paris French: in situ question words in embedded contexts. Multicultural Youth Vernaculars in Paris and Urban France. Special issue of Journal of French Language Studies 28.2: July, 181–209. Gardner‐Chloros, Penelope and Daniel Weston (eds.) 2015. Code‐switching in Literature. Special issue of Language and Literature 24.3. Georgakopoulou, Alexandra 1997. Self‐ presentation and interactional alignments in email discourse: the style and code‐switches of Greek messages. International Journal of Applied Linguistics 7.2: 141–64. Georgakopoulou, Alexandra and Tereza Spilioti 2016.The Routledge Handbook of Language and Digital Communication. London: Routledge. Giacalone Ramat, Anna 1995. Code‐switching in the context of dialect‐standard language relations. In Lesley Milroy and Pieter Muysken (eds.), One Speaker, Two Languages: Cross‐ Disciplinary Perspectives on Code‐Switching. Cambridge: Cambridge University Press, pp. 45–67. Gibbons, John 1987. Code‐Mixing and Code‐Choice: A Hong Kong Case Study. Clevedon: Multilingual Matters. Gumperz, John J. 1964. Hindi–Panjabi code‐ switching in Delhi. In Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: de Gruyter Mouton, 1115–24. Gumperz, John J. and Edward Hernández 1969. Cognitive aspects of bilingual communication. Working Papers of the Language Behavior Research Laboratory 28. Berkeley: University of California Press. Gumperz, John J. and Robert D. Wilson 1971. Convergence and creolization: a case from the Indo‐Aryan‐Dravidian border. In Dell Hymes

Contact and Code‐Switching  197 (ed.), Piginization and Creolization of Languages. Cambridge: Cambridge University Press, pp. 151–67. Han Chung, Haesook 2006. Code switching as a communicative strategy: a case study of Korean–English bilinguals. Bilingual Research Journal 30.2: 293–307. Harrington, Jonathan, Sallyanne Palethorpe, and Catherine I. Watson, 2000. Does the Queen speak the Queen’s English? Nature 408: 927–8. Haust, Delia 1995. Codeswitching in Gambia: eine soziolinguistische Untersuchung von Mandinka, Wolof und Englisch in Kontakt. Cologne: Rüdiger Köppe. Heath, Jeffrey 1989. From Code‐switching to Borrowing: Foreign and Diglossic Mixing in Moroccan Arabic. London: Kegan Paul International. Heller, Monica 1988. Where do we go from here? In Monica Heller (ed.), Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: de Gruyter Mouton, pp. 265–72. Hewitt, Roger 1986. White Talk Black Talk. Cambridge: Cambridge University Press. Hinrichs, Lars (ed.) 2006. Codeswitching on the Web: English and Jamaican Creole in E‐mail Communication. Amsterdam: Benjamins. Hong Liu 2018. Intra‐speaker variation in Chinese–English code‐switching: The interaction between cognitive and contextual factors. International Journal of Bilingualism 22.6: 740–62. Jones, Mari C. 1998. Language Obsolescence and Revitalization: Linguistic Change in Two Sociolinguistically Contrasting Welsh Communities. Oxford: Oxford University Press. Jones, Mari C. 2005. Some structural and social correlates of single word intrasentential code‐switching in Jersey Norman French. Journal of French Language Studies 15: 1–23. Kosoff, Zoe 2014. Code‐switching in Egyptian Arabic: a sociolinguistic analysis of Twitter. Al‐‘Arabiyya: Journal of the American Association of Teachers of Arabic 47: 83–99. Lantto, Hanna 2015. Conventionalized code‐ switching: entrenched semantic‐pragmatic patterns of a bilingual Basque‐Spanish speech style. International Journal of Bilingualism 19.6: 753–68. Lavandera, Beatriz 1978. The variable component in bilingual performance. In James E. Alatis (ed.), International Dimensions of Bilingual Education. Washington, DC: Georgetown University Press, pp. 391–409. Le Page, Robert B. and Andrée Tabouret‐Keller 1985. Acts of Identity: Creole‐Based Approaches to

Language and Ethnicity. Cambridge: Cambridge University Press. Lee, Carmen 2016, Multilingual resources and practices in digital communication. In Alexandra Georgakopoulou and Tereza Spilioti (eds.), The Routledge Handbook of Language and Digital Communication. London: Routledge, pp. 118–32. Levey, Stephen, Karen Groulx, and J. Roy 2013. A variationist perspective on discourse‐ pragmatic change in a contact setting. Language Variation and Change 25: 225–51. Li Wei 1998. Banana split? Variations in language choice and code‐switching patterns of two groups of British‐born Chinese in Tyneside. In Rodolfo Jacobson (ed.), Codeswitching Worldwide. Berlin: de Gruyter Mouton, pp. 153–76. Li Wei 2018. Translanguaging as a practical theory of language. Applied Linguistics 39.1: 9–30. Mahootian, Sharzad 2005. Written code: the discourse of social change. Talk presented at The Fifth International Symposium on Bilingualism, Barcelona, March 2005. McClure, Erica 1998. The relationship between form and function in written national language: English codeswitching: evidence from Mexico, Spain and Bulgaria. In Rodolfo Jacobson (ed.), Codeswitching Worldwide. Berlin: de Gruyter Mouton, pp. 125–52. McEntee‐Atalianis, Lisa 2019. Identity in Applied Linguistics Research. London: Bloomsbury. McEntee‐Atalianis, Lisa, Ateek, Mohammed and Gardner‐Chloros, Penelope, forthcoming. Code‐switching, Script‐switching and Identity in Social Media: Syrian refugees on Facebook. McLelland, Nicola 2004. A historical study of codeswitching in writing: German and Latin in Schottelius’ Ausführliche Arbeit von der Teutschen Haubtsprache (1663). International Journal of Bilingualism 8.4, 499–525. Mufwene, Salikoko 2008. Language Evolution: Contact, Competition and Change. London: Continuum. Mullen, Alex 2015. In both our languages: Greek‐Latin code‐switching in Roman literature. In Penelope Gardner‐Chloros and Daniel Weston (eds.), Code‐switching in Literature. Special issue of Language and Literature 24.3: 213–33. Muñoz, Marie‐Claude 1999. Usages du français et double appartenance: le cas des Portugais en France. Education et Sociétés Plurilingues 7, 21–33.

198  Penelope Gardner‐Chloros Muysken, Pieter 2000. Bilingual Speech: A Typology of Code‐Mixing. Cambridge: Cambridge University Press. Myers‐Scotton, Carol 1992. Comparing code‐ switching and borrowing. Journal of Multilingual and Multicultural Development 13.1/2, 19–39. Myers‐Scotton, Carol 1993a Duelling Languages. Oxford: Oxford University Press. Myers‐Scotton, Carol 1993b Social Motivations for Code‐Switching: Evidence from Africa. Oxford: Oxford University Press. Myers‐Scotton, Carol 1998. One way to dusty death: the matrix language turnover hypothesis. In Lenore Grenoble and Lindsay Whaley (eds.), Language Loss and Community Response. Cambridge: Cambridge University Press, pp. 289–316. Myers‐Scotton, Carol 2000. Testing the 4‐M model: an introduction. International Journal of Bilingualis, 4.1: 1–8. Myers‐Scotton, Carol 2002 Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Nortier, Jacomine and Dorleijn, Margreet 2013. Multi‐ethnolects: Kebabnorsk, Perkerdansk, Verlan, Kanakensprache, Straattaal, etc. In Peter Bakker and Yaron Matras (eds.), Contact Languages: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 229–71. Oesch‐Serra, Cecilia 1998. Discourse connectives in bilingual conversation: the case of an emerging Italian–French mixed code. In Peter Auer (ed.), Code‐switching in Conversation: Language, Interaction and Identity. London: Routledge, pp. 101–24. Onysko, Alexander and Esme Winter‐Froemel 2011. Necessary loans‐luxury loans? Exploring the pragmatic dimension of borrowing. Journal of Pragmatics 43: 1550–67. O’Shannessy, Carmel 2012. The role of codeswitched input to children in the origin of a new mixed language. Linguistics 50.2: 305–40. O’Shannessy, Carmel, 2016. Distributions of case allomorphy by multilingual children: Speaking Warlpiri and Light Warlpiri. Linguistic Variation 16: 68–102. Otheguy, Ricardo 1995. When contact speakers talk, linguistic theory listens. In Ellen Contini‐ Mrava and Barbara Sussman Goldberg (eds.), Advances in Linguistic Sign Theory. Berlin: de Gruyter Mouton, pp. 213–42. Otheguy, Ricardo 2004. Single‐language and code‐switching strategies in immigrant and heritage varieties: Spanish subject personal pronouns in Toribio’s cross‐modal hypothesis. Bilingualism, Language and Cognition 7.2: 175–7.

Palfreyman, David and Muhamed Al Khalil 2007. ‘A funky language for teenzz to use’: representing Gulf Arabic in instant messaging. In Brenda Danet and Susan C. Herring (eds.) The Multilingual Internet. Oxford: Oxford University Press, pp. 43–64. Perrot, Marie‐Eve 1995. Tu worries about ça, toi? Métissage et restructurations dans le chiac de Moncton. Linx 33: 79–85. Perrot, Marie‐Eve 2014. Le trajet linguistique des emprunts dans le chiac de Moncton: quelques observations. Minorités Linguistiques et Société 4: 200–28. Pfaff, Carol W. 1979. Constraints on language mixing: intrasentential code‐switching and borrowing in Spanish/English. Language 55: 291–318. Poplack, Shana 1984. Variable concord and sentential plural marking in Puerto Rican Spanish. The Hispanic Review 52.2: 205–22. Poplack, Shana and Nathalie Dion 2012. Myths and facts about loanword development. Language Variation and Change 2403: 279–315. Poplack, Shana, David Sankoff, and Christopher Miller 1988. The social correlates and linguistic processes of lexical borrowing and assimilation. Linguistics 26.1: 47–100. Pujolar, Joan 2001. Gender, Heteroglossia and Power: A Sociolinguistic Study of Youth Culture. Berlin: de Gruyter Mouton. Rindler‐Schjerve, Rosita 1998. Code‐switching as an indicator for language shift? Evidence from Sardinian–Italian bilingualism. In Rodolfo Jacobson (ed.), Codeswitching Worldwide. Berlin: de Gruyter Mouton, pp. 221–48. Romaine, Suzanne 1992. Language, Education and Development: Urban and Rural Tok Pisin in Papua New Guinea. Oxford: Oxford University Press. Romaine, Suzanne 1995. Bilingualism, 2nd edition. Oxford: Blackwell. Schendl, Herbert 2000 Linguistic aspects of code‐switching in medieval English texts. In David A. Trotter (ed.), Multilingualism in Later Medieval Britain. Cambridge: D. S. Brewer, 77–92. Schendl, Herbert 2015. Code‐switching in Early English Literature. In P. Gardner‐Chloros and D. Weston (eds.), Special Issue, Language and Literature, Code‐switching in Literature 24.3: 233–49. Schmidt, Annette 1985. Young People’s Djirbal. Cambridge: Cambridge University Press. Schreier, Daniel 2009. Language in isolation, and its implications for variation and change. Language and Linguistics Compass 3.2: 682–99.

Contact and Code‐Switching  199 Sebba, Mark 1993. London Jamaican: Language Systems in Interaction. London: Longman. Sebba, Mark, Sharzad Mahootian, and Carla Jonsson 2012. Language Mixing and Code‐ switching in Writing. London: Routledge. Siegel, Jeff 1994. Review of Suzanne Romaine Language, Education and Development. Language in Society 23: 144–9. Silva‐Corvalán, Carmen 1994. Language Contact and Change: Spanish in Los Angeles. Oxford: Oxford University Press. Sitaridou, Ioanna 2014. The Romeyka infinitive: continuity, contact and change in the Hellenic varieties of Pontus. Diachronica 31.1: 23–73. Spilioti, Tereza 2018. The language of social media. In Philip Seargeant, Ann Hewings and Stephen Pihlaja (eds.), The Routledge Handbook of English Language Studies. London: Routledge, pp. 310–23. Spolsky, Bernard and Robert L. Cooper 1991. The Languages of Jerusalem. Oxford: Oxford University Press. Stanlaw, James 2004. Japanese English: Language and Culture Contact. Hong Kong: Hong Kong University Press. Stell, Gerald and Kofi Yakpo (eds.) 2015. Code‐ switching Between Structural and Sociolinguistic Perspectives. Berlin: de Gruyter Mouton. Stolt, Birgit 1964. Die Sprachmischung in Luthers Tischreden: Studien zum Problem der Zweisprachigkeit. Stockholm: Almquist and Wiksell. Thomason, Sarah G. 1997 Introduction. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 1–8. Thomason, Sarah G. 2001a. Language Contact: An Introduction. Edinburgh: Edinburgh University Press. Thomason, Sarah G. 2001b Contact‐induced typological change. In Martin Haspelmath, Ekkehard Koenig, Wulf Oesterreicher, and Wolfgang Raible, (eds.), Language Typology and Language Universals, Sprachtypologie und Sprachliche Universalien: An International Handbook Berlin: Walter de Gruyter, 1640–8. Timm, Leonora 1975. Spanish–English code‐ switching: el porqueì y how‐not‐to. Romance Philology 28.4: 473–82. Torres Cacoulos, Rena and Catherine E. Travis 2016. Two languages, one effect: structural priming in spontaneous code‐switching. Bilingualism: Language and Cognition 19.4: 733–53. Torres Cacoulos, Rena and Catherine E. Travis 2018. Bilingualism in the Community: Code‐switching

and Grammars in Contact. Cambridge: Cambridge University Press. Treffers‐Daller, Jeanine 1994. Mixing Two Languages, French–Dutch Contact in a Comparative Perspective. Berlin: de Gruyter Mouton. Treffers‐Daller, Jeanine 1998. Variability in code‐switching styles: Turkish–German code‐ switching patterns. In Rodolfo Jacobson (ed.), Codeswitching Worldwide. Berlin: de Gruyter Mouton, pp. 177–98. Treffers‐Daller, Jeanine 1999. Borrowing and shift‐induced interference: contrasting patterns in French–Germanic contact in Brussels and Strasbourg. Bilingualism, Language and Cognition 2: 77–80. Trotter, David A. (ed.) 2002. Multilingualism in Later Medieval Britain. Cambridge: D. S. Brewer. Trudgill, Peter 1976–7. Creolization in reverse: reduction and simplification in the Albanian Dialects of Greece. Transactions of the Philological Society 7, 32–50. Tsiplakou Stavroula 2009. Doing (bi)lingualism. Language alternation as performative construction of online identities. Pragmatics 19.3: 361–91. Tsitsipis, Lukas D. 1998. A Linguistic Anthropology of Praxis and Language Shift: Arvanitika (Albanian) and Greek in Contact. Oxford: Oxford University Press. Valdés‐Fallis, Guadalupe 1976. Code‐switching among bilingual Mexican–American women: towards an understanding of sex‐related language alternation. International Journal of the Sociology of Language 7: 65–72. Wagner, Suzanne Evans 2012. Age grading in sociolinguistic theory. Wiley Online Library, pp. 71–82. Warschauer, Mark, Ghada R. El Said, and Ayman G. Zohry 2007. Language choice online: globalization and identity in Egypt. In Brenda Danet and Susan C. Herring (eds.), The Multilingual Internet. Oxford: Oxford University Press, pp. 303–19. Weston, Daniel 2012. Code‐switching variation in Gibraltar. International Journal of Bilingualism 17.1: 3–22. Weston, Daniel 2016. Bits, chunks and channel‐ switching: perceptions of Cantonese‐English code‐switching. Journal of Chinese Linguistics 44: 384–414. Wiese, Heike 2009. Grammatical innovation in multiethnic urban Europe: new linguistic practices among adolescents. Lingua 119: 782–806.

10 Contact and Mixed Languages PETER BAKKER 1 Introduction Mixed languages can be said to be the most extreme result of language contact. Languages that have been called mixed languages are of such a nature that they are no longer classifiable in family trees: when applying the comparative method in historical linguistics, the results points in two different directions as regards genealogical affiliations. In a less strict definition of mixed languages, one can also consider languages that, for instance, inherited one component of a source language. Creoles and pidgin are sometimes called mixed, as can be seen where the vocabulary was inherited from e.g. English, Arabic or French (the lexifier), but the grammatical system was largely recreated. In these cases, there is only one genealogical source, however. The delineation of mixed languages versus non‐mixed languages was very well illustrated by the two opening quotes in Thomason and Kaufman (1988: 1). Max Müller had written: “Es gibt keine Mischsprachen” (mixed languages do not exist) in 1871 and Hugo Schuchardt wrote in 1884: “Es gibt keine völlig ungemischte Sprache” (all languages show signs of mixture). Now, more than a century later, all linguists agree that there are no languages without lexical or grammatical borrowings  –  a statement that is also true for language isolates, languages spoken more or less in isolation on islands or mountainous areas, and small languages as well as huge languages in the sense of numbers of speakers. Does this universal presence of influence from other languages then mean that the concept of mixed languages is void? Or is there in fact a clear border between mixed and non‐mixed languages? Schuchardt was a prolific author, who is most famous for his work on Romance languages and his work on creoles and pidgin languages. In the century following Schuchardt’s claim, there was a discussion in which the consensus was that true mixed languages do not exist. Scholars agreed that it would always be possible to comfortably place every language into a family tree. This view changed, however, as the result of a number of publications. First and foremost, the pioneering study by Thomason and Kaufman (1988), who synthesized research on the results of different types of language contact. In their landmark book, they included a number of case studies on languages that were considered “mixed,” such as Ma’a/Mbugu, Mednyj Aleut, and Michif, all of which continue to play an important role in the debate on mixed languages. A few years later, the results of a workshop devoted to mixed languages were published (Bakker and Mous 1994), and this book included chapters on additional mixed languages. These two books and the studies included in it, contributed to a turn of the tide, to the general acceptance of the existence of mixed languages, and to the identification

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

202  Peter Bakker of social‐historical origins and structural patterns common between some of them. Some highlights in the debate include Thomason (1995), Matras (2000), Matras and Bakker (2003). Since then, additional mixed languages have been identified, and thanks to a larger set of data, it has become possible to make (provisional) classifications based on the nature of the mixtures. Many of the cases show systematic dichotomies between fundamental parts of a language, such as grammar and lexicon, or nouns and verbs, or verbal inflection versus the rest. Those are parts of a language that are normally taken to be indicative of the genealogical affiliation of languages, but in mixed languages such fundamental components point in different directions. Within the area of pidgin and creole studies and the study of language contact, one can notice a gradual acceptance of the existence of mixed languages of types like Ma’a and Mednyj Aleut. The Arends, Muysken, and Smith textbook (1995) on pidgins and creoles was the first to include a chapter on mixed languages beyond creoles and pidgins. The recent textbook by Velupillai (2015) even has “mixed languages” in the title, beside creoles and pidgins. The Atlas of Pidgin and Creole Language Structures (APiCS) (Michaelis et al. 2013) also included a handful of mixed languages. Similarly, the name of the Creole Language Library was changed into Contact Languages Library in 2017. For that reason, we deal with pidgins and creoles here as well, even though they can and should be distinguished from mixed languages.

2  Pidgins, Pidgincreoles, Creoles, Mixed Languages As a thought experiment, let us create fictitious specimens of pidgins, creoles and mixed languages. Here we provide examples, and in Section 3 more information is given about these language types. As our point of departure, we take English, and the influencing language is Spanish. The message one wants to transmit is (1a) (1b)

We would not (have) walk(ed) to the sea if it rained. (English) No í‐ba‐mos a camin‐ar hasta el mar si llov‐iera. (Spanish) NEG go‐IMPERF‐1PL to walk-INF towards DEF.M.SG sea IF rain‐SUBJ.PST.31

In this sentence, we encounter habitual/imperfective aspect, unreal/irrealis mood, negative, past, and conditional. A pidgin is used between people who do not share a language but they do need to communicate, and language norms have developed. In a pidgin, this same statement could be expressed as in (2): (2)

No go playa bikos rreen NEG go beach because rain

(Spanish Pidgin English)

One Spanish word, playa, commonly known, is used. There is Spanish influence in the sound system, e.g. the trilled /r/ in rain and the vowel would not be diphthongized. The hearer has to interpret whether this deals with a past, present, or future action, and a real or hypothetical action. It could be even shorter, like (3): (3)

Reen yes, mar no Rain yes sea no

(Spanish Pidgin English)

A pidgincreole is a former pidgin that has become a default language, maybe even a native language for some, and both language use and structure are expanded to a full communicative language. In a pidgincreole, the same thing could be expressed as in (4): (4)

Os no vaka tuwa mar spos dattaim reen bin fol (Pidgincreole English) 1PL NEG walk TO sea suppose/if then rain HB fall

Contact and Mixed Languages  203 Again, the phonology is close to Spanish, but now the habitual (bin < been) and conditional (spos < suppose) have been grammaticalized. A preposition tuwa (from emphatic towards rather than nonsalient to) is also present. A more salient or emphatic form of the pronoun is used (os < us, not we). A creole has become a native language to all, shows some more grammaticalization of tense‐mood‐aspect with bin o from been go, and some more phonological developments, as in (5): (5)

Oso no bin o vaka tuwa mar spo reen bin fol (Creole English) 1PL NEG PAST FUT walk suppose/if then rain HB fall

Now we provide two examples of two types of mixed languages in the strict sense. We first give the phrase in Spanish, and then in an intertwined form: (6a)

No í‐ba‐mos a camin‐ar hasta el mar si llov‐iera. (Spanish) NEG go‐PST.IRR‐1PL INF walk‐INF to the sea if rain‐IRR.3

(6b)

donit go‐ba‐mos a vok‐ar til el sisi si rren‐ería (intertwined mixed Spanish‐English) NEG go‐PST.IRR‐1PL INF walk‐INF to the sea if rain-IRR.3

In (6b), donit is from don’t, go‐ from go, and the word for sea is doubled to disambiguate it from the conditional conjunction si. Finally, there are metatypized languages (Ross 2006, 2007) in which, at least in theory, all the morphemes are replaced by morphemes from another language, preserving the structural traits of one language but both lexical and grammatical morphemes are from another language. (7)

We would not have walked to the sea if it rained. (English) Nosotros haría no haber caminado hasta el mar si el lluvia‐do. (Spanish in English frame)

(8)

No í‐ba‐mos a camin‐ar hasta el mar si llov‐ería. Not go‐ed‐us to walk‐ing until the sea if rain‐ed (English in Spanish frame)

(Spanish)

Sentences (2) to (8) are fictitious examples to illustrate results of language contact, as results of different socio‐historical processes. They may seem contrived, but on pp. 206–211 and pp. 213–215 we will p ­ resent examples from the real world. All these types have been shown, or assumed, to have emerged within a short time span, between one generation and less than a century. Clearly, these types are all languages that do not fit the historical‐comparative model, in which ­lexicon, inflection, and other morphology, and syntactic structures are slowly and concurrently adjusted through time.

3 The Historical‐comparative Method and Mixed Languages One of the goals of the widely successful comparative method in historical linguistics is to establish whether two (or more) languages are related or not. This is done using three criteria: (i) many everyday roots should be cognates, that is, demonstrably go back to the same roots in an ancestral language; (ii) the cognate words must be relatable to each other through

204  Peter Bakker systematic and regular sound changes in order to exclude borrowings and (iii) the grammatical systems have to show similarities as well, notably in their morphological paradigms – if present. Even though the method can be considered universally accepted, there are challenges as well, such as proven regularity of sound changes even in borrowings or lexical innovations, and the lack of overt morphology in some language families (e.g. Sinitic). Such challenges, however, are marginal, compared to the enormous success and predictive power of the method, in establishing language families, suggesting internal relationships and reconstructing ancestral forms of the language. The method has led to many successful predictions (or retrodictions) in historical linguistics. However, mixed languages defy the historical‐comparative method. Mixed languages could be defined as those languages in which the different basic chunks of a language point to different ancestries. That is, they are languages for which the comparative method does not work. It is important to note that their existence does not falsify the comparative method, or otherwise cast doubt on that method: mixed languages are the exceptions to the patterns, that confirm the rule. Owing to their special history of genesis, as shown on pp. 211–215, they have special characteristics. They defy classification within the patterns of what has been called “normal transmission” (cf. Thomason and Kaufman 1988), i.e. the near‐perfect learning of languages by children from parents and peers. Mixed languages have special scenarios in that they presumably came about suddenly, and not through intergenerational transmission, and therefore they have distinct properties, just like volcanic rock is ostensibly different from other types of rock due to its sudden rather than slow genesis, due to volcanic eruption rather than long‐term pressure. Following Thomason and Kaufman (1988), the vast majority of the languages of the world are the result of “normal transmission.” Some of these languages may have undergone considerable influence from other languages (e.g. English underwent considerable French influence, or Turkish has a considerable number of Persian/Arabic elements), but their genetic affiliation with one language family is not obscured, despite lexical and grammatical influence from other languages. We discuss these exceptions to regular transmission here, and provide examples below. First, pidgins had an interrupted transmission. Pidgins emerge in situations of sudden contact between people who need to improvise an oral (or signed) communication system for trade, labor relations, and social contacts. Content words are taken from one, rarely two or more, languages, but very little of the grammar is transmitted, thus fluking the condition of the historical‐comparative method that the lexicon and grammar must be transmitted: only the lexicon is transmitted, and, in the case of pidgins, only a small part. Juvonen (2008) counted pidgin lexicons and found between 500 and 2,000 words, but some pidgins with even fewer words seemed to have been sufficient in the pertinent communicative situations. Second, creoles have undergone an interrupted transmission. The lexicon tends to be almost exclusively from one language, aptly called the lexifier. As we will see on pp. 208–211, creoles show a similar lack of grammatical and lexical transmission as pidgins, but in contrast to pidgins, the creators of the creoles have innovated a full‐fledged grammatical system. Even though some influence from other languages than the lexifier can be identified, many of the newly developed grammatical systems cannot be related to any of the substrate or adstrate languages either – those are languages other than the lexifier, spoken when the language emerged or spoken at a later stage in its development, parallel to the creole. Since the grammatical system was not taken over from the lexifier along with the lexical roots, much of the transmission was interrupted, and most of the grammatical system was newly created through grammaticalization. Specialists have traditionally also recognized a class of expanded pidgins or pidgincreoles (Bakker 2008). The latter label is to be preferred, as “expanded pidgins” is

Contact and Mixed Languages  205 ambiguous between a pidgin with structural expansion (more complex) or one with social expansion (wider use). Pidgincreoles are often called “pidgin” by the speakers, but they are to be distinguished from pidgins on social grounds: they have some native speakers; they may be an official language, or a language unconnected with a specific ethnic group, or a default language used with strangers. It appears that these pidgincreoles are structurally indistinguishable from creoles, but distinct from pidgins, and for that reason they can be subsumed under creoles for the study of linguistic properties of creole languages, also in this chapter. This group includes languages like Tok Pisin and other forms of Melanesian Pidgin, West African Pidgin English, Sango in Central Africa, and Naga Pidgin or Nagamese in Northeast India. Fourth, the label semicreoles is used for a small number of languages that underwent some clear morphological reduction in relation to a lexifier, but not much expansion in the form of the development of a new grammatical system (Holm 2003). A probably exhaustive list of known cases consists of Kituba and Lingala in Southwestern Africa, Hiri Motu in Papua New Guinea, Afrikaans in South Africa, vernacular Brazilian Portuguese and, in the view of some, Reunion Creole in the Indian Ocean as well. Fifth, mixed languages in the narrow sense had an interrupted transmission. For instance, languages that combine the lexicon from one language with the grammatical system of another, or the verbs from one language and the nouns from another, show a major discontinuity of a major part of a source language. As a rule, and in contrast with pidgins and creoles, the preserved components are transferred more or less intact into the new, mixed languages.

4 A Historical Perspective on Research on Mixed Languages The problem in classification of mixed languages can be illustrated with the way two ­linguists who do not believe in the concept of mixed languages, classify three concrete mixed languages, Mednyj/Copper Island Aleut (Russian pronouns and verbal morphology, the rest Aleut), Ma’a (Bantu grammatical frame, a lexicon of mostly Southern Cushitic origin) and Berbice Creole (lexicon Dutch and Eastern Ijo, morphology Eastern Ijo). Greenberg (1999: 632) classifies Copper Island Aleut (CIA) as Aleut. Indeed, the bulk of the lexical and grammatical roots are from Aleut. On the other hand, it is “evident” for Van Driem (2001: 167) that “the genetic core of CIA is Russian.” Van Driem has especially the verbal inflection in mind, which is taken as indicative of its grammatical affiliation, and which is Russian. Obviously, they cannot both be right. Ma’a (Mous 1994, 2003) is another example. The language is sometimes described, somewhat simplistically, as a language with the grammatical system of a Bantu language, with its agglutinative morphology, and a (mostly) Southern Cushitic lexicon. For Greenberg (1999), Ma’a is genetically Cushitic, since many basic and everyday language roots have clear cognates in Cushitic languages. For Dimmendaal (1995: 367), however, Ma’a is a Bantu language, because the morphology is Bantu. Obviously, both of them cannot be right, if one thinks of family trees. A final example of contested genetic affiliation is Berbice Creole of Guyana, a creole language with lexical roots from Dutch (hence the common name “Berbice Dutch”; Kouwenberg 1994), Eastern Ijo, a language of Nigeria, and Lokono, an Amerindian language of the Arawakan family. If we abstract away from the fact that it is a creole language, which, according to some, is outside the frame of genetic classification because of the lack of transmission of the grammatical system, the controversy here is about the origin of the lexical and

206  Peter Bakker grammatical roots. In Berbice Creole, all inflectional morphology is Ijo, but most of its l­ exicon is Dutch. For Dimmendaal (1995), it is a Germanic language, because so many roots are from Dutch. But here, his position in fact contradicts his classification of Ma’a: for Ma’a, he takes the morphological system as his criterion for classification, and not the lexicon, whereas for Berbice Creole, he takes the lexicon as his criterion, and not the fact that the plural suffix, the tense suffix and the aspectual suffix, the nominalization suffix and the object clitic (in fact all the morphology in the language) are all from Ijo rather than Dutch. Of course, both positions are equally defendable and rejectable, and therefore the category of mixed language is useful in these cases. Such deviant and incompatible positions by eminent historical linguists are easily solved if one accepts the existence of mixed languages with vocabularies and morphosyntactic properties derived from different languages. One has to recognize these mixed languages as exceptions to the one parent‐one daughter language genetic model, and the mixed languages can be linked to exceptional socio‐historical events. If any of the three criteria (basic lexicon, morphology and grammatical traits) point massively in two different directions, the languages should be considered mixed, rather than using ultimately arbitrary c­ riteria in each case.

5 Pidgins, Creoles, and Semicreoles and Their (Un)mixed Nature One question to be discussed is whether pidgins, creoles, and semicreoles are mixed languages or not, as opinions diverge on this.

5.1  Pidgin Languages: Mixed and Unmixed Pidgin languages (Parkvall 2017; Parkvall and Bakker 2013) are lexically and grammatically reduced makeshift languages. By definition, there are no mother tongue speakers, and ­neither is the language an official language or a lingua franca. Pidgins come about in situations of sudden contact in which a spoken or signed communication system is needed, and the parties involved want to keep a social distance from one another. The French sentence (9a) is expressed quite differently in Pidgin French (9b): (9a)

Si tu dis un mot, tu fin‐ir‐as mort. IF you say one word you end‐FUT‐2SG dead

(French)

(9b)

twa parle œ̃ mo, twa fini mor (New Caledonia Pidgin French; Hollyman 1964: 60; 1932) 2SG speak one word 2SG finish dead ‘If you say one word, you will end up dead.’

We can notice the following: emphatic pronouns are used for the French unstressed ones (tu > toi); verbal inflection is dropped and verbs are replaced with uninflected forms, for instance infinitives (fin‐ir‐as > fini); the subordinate clause including the conditional marker (si) is replaced with two juxtaposed main clauses. The rather indistinct verb di‐ ‘to say’ is replaced with more salient parl‐ ‘to speak’. Also typical for pidgins is the lack of tense and aspect marking, lack of definite and indefinite articles, lack of all forms of gender marking, and absence of plural marking.

Contact and Mixed Languages  207 Here is an example from Chinese Pidgin English (Pernel 1938): (10)

spose my catchee one length table big too much ‘What if I take a length for the table that is bigger.’

(Chinese Pidgin English)

Here we see, again, emphatic pronouns, uninflected verbs, and further the use of “too much” for “more” or “very” and a more salient form from “suppose” to indicate a conditional as well as a more salient verb “to catch” for “to take.” There is some phonological adjustment with a paragogic vowel in catchee. A copula is lacking. The previous examples are pidginized forms or languages with not much morphology. A morphologically rich language may pidginize as well. The morphology may be dropped, or remain as a meaningless part of a stem, as in Hudson Bay Eskimo Pidgin. Instead of an inflected verb like oqaluttuup‐pakkit ‘I told you’ a simpler verb ukak‐tuk ‘he talks’ is used, combined with overt personal pronouns, so the complex verbal sentence is rendered in a way in which three meaning components are presented in three different words rather than one complex one (cf. Canadian Inuktitut uvanga ‘I’, ilvit ‘you’, and uqaqtuq ‘he talks’) (Bakker and van der Voort 2017: 415). (11)

Awonga igbik ukaktûk. (Pidgin Eskimo) I you  talk ‘I told you.’ (etymologically ‘I you he talks’)

A final example is from Icelandic Pidgin Basque, as found in a seventeenth‐century Icelandic‐Basque word list. We give the same sentence in Basque and in Pidgin Basque: (12a)

Kristo eta Maria‐k bale‐a ema‐ten d‐i‐da‐te‐n‐ean eman‐go d‐i‐zu‐t buztan‐a (Basque) Christ and Maria‐ERG whale‐DEF give‐PRES it‐to‐me-they‐SUBORD‐when give‐FUT it‐to‐you‐I tail‐DEF

(12b)

Christ Maria presenta for mi balia, for mi presenta for ju bustana. (Pidgin Basque) Christ Maria give to me whale, to me give to you whale ‘If Christ and Maria provide me with (give) a whale, I will give you the tail.’

Case marking and the complex verbal morphology are dropped; nouns are used with the definite articles, which is also the citation form. An international word presenta is used, and emphatic pronouns with a preposition (for ju, for mi) from Zealandic Dutch or English (Bakker 1987). Pidgins are typically very analytic (Roberts and Bresnan 2008), and also languages with complex morphologies like Inuktitut, Basque, Algonquian languages and others can be ­simplified to languages with no or very little morphology. Constituent ordering always seems to be like one or more of the contact languages. In pidgins, there is no grammatical component that can be linked to another language; the grammar is just altered through simplification. This implies that pidgins are not in themselves mixed languages. Because of their simplified, almost grammarless nature, pidgins can by definition only show mixture in the lexicon. In the first three pidgins discussed on pp. 206–207, virtually no lexical admixture is found, in that all roots are from French, English and Inuktitut respectively. Only the last example shows traces from Basque, Germanic, and Romance languages. Some pidgins, however, are lexically mixed in the sense that a similar quantity of roots seems to come from two different source languages, usually with no systematic semantic patterns. These pidgins have been called 50‐50 pidgins. The following pidgins can be considered 50‐50 pidgins: Chinese‐Russian Pidgin of Harbin, China (Lyovin 1968); Russenorsk of the

208  Peter Bakker North Cape, Norway; Yimas‐Arafundi Pidgin of Sepik, Papua New Guinea (Foley 2013), Trio‐Ndyuka Pidgin of Suriname and French Guiana (Huttar 1996), and early New South Wales Pidgin of Australia (Troy 1994). Others may be as well, but often not sufficient material is available to draw conclusions about the admixture. This is an example of the Trio‐Ndyuka Pidgin, with Ndyuka (English‐lexifier creole) words in bold (Crevaux 1880: 404–5): (13) Panakiri ouani oua. Ala pikinini alele. Nono poti. Echimeu ouaca. Cassava mia oua. White.person want no all child dead hole put quickly walk cassava eat no ‘We don’t want white people. All the children are dead. We have put them in a hole. Leave quickly. We have no cassava to eat.’ This is uttered by an Amerindian woman, speaking to a French explorer, his Ndyuka guide/friend and their company. Seven words are from Ndyuka, the creole language, and six are from Trio. The roughly equal proportion of roots from both languages is representative of the pidgin as a whole, and permeates all the material, and their distribution seems to be largely arbitrary (Huttar and Velantie 1997, Meira and Muysken 2017). Such arbitrary lexical mixtures in the basic lexicon seem to be limited to pidgins among the languages of the world. No non‐pidginized languages are known with similar unpredictable mixtures in the basic vocabulary. Calling such pidgins mixed languages is justified because of the mixture of the basic lexicon, but its grammatical structures cannot play a role in the characterization. In most of the ca. 100 known pidgins (cf. Parkvall 2017; Parkvall and Bakker 2013), however, the bulk of the lexicon is from one language only. For those other pidgins, two reasons have been brought up to call them mixed. The first is that they have lost the grammatical system without having replaced it by a new system. The other reason may be that one would assume that the pidgin now has grammatical traits from one or more other languages, the substrate languages. The new pidgin structures are clearly only the result of a process of reduction, and not of addition of new features from other languages (even though there may be a few, e.g. the use of a classifier piecee in Chinese Pidgin English under the influence from Chinese). In the realm of pidgins, in my view only the lexically mixed pidgins are part of the ­category “mixed languages.”

5.2  Creole Languages: Mixed and Unmixed The nature of creoles can illustrate how the structures of a lexifier are altered radically after the creolization process, and how lexical items are grammaticalized to create a full‐fledged language. This French sentence (14a) has become (14b) after a process of creolization, most likely after a process of pidginization. Below the creole, the origin of each of the morphemes is indicated. (14a)

Elle la fait bouillir pour qu’elle soit molle/douce. 3SG.F.SUB 3SG.F.OBJ make.3SG boil.INF for that‐she BE.SUBJ.3SG soft.F ‘She is boiling it (“her”) so that it (“she”) will be soft.’

(14b)

Li ka bui‐l pu li mu (Karipuna Creole French, Brazil) [lui ? bouillir‐lui pour lui moux] 3SG PROG boil‐3SG for 3SG soft ‘She is boiling it so that it will be soft.’

Contact and Mixed Languages  209 All lexical and grammatical morphemes in (14b) are from French, but a number of fairly radical changes have made the creole language virtually unintelligible for French speakers. Gender in the pronouns has been dropped (li ‘he, she, her, him’ < lui ‘him’), and only one form (originally masculine) is used for females/feminine and males/masculine gender. There is no gender in the nouns either (not visible here), and the French articles (le, la, les, un, une, des) have disappeared. There is no gender distinction in the adjectives anymore (French moux M/molle F, doux M/douce F). The overt causative marker fait ‘(s)he makes’ has disappeared and the creole verb bui < bouillir can be used both as transitive and intransitive. The preverbal object clitic la (feminine) has disappeared, and instead a postverbal clitic li (gender‐ neutral) is used. The complementizer que has disappeared and the preposition pour/pu now functions as a complementizer (which happens in many creoles, also with other lexifiers; see Bickerton 1984). This type of radical changes between the lexifier and the creole are quite common results of creolization. All creoles have dropped all manifestations of grammatical gender (adjectival inflection, definite and indefinite articles) and, with very few exceptions, gender distinctions in personal pronouns have also disappeared. English is the only Germanic language without grammatical gender, but it has preserved the gender distinction in the singular pronouns (he/ she, him/her). In the 6,000–8,000 year history of the 450 languages that are descendants of Proto‐Indo‐European, virtually all preserved gender (except Armenian, Bengali, Persian and a few other Iranian languages, and a Danish and a Dalmatian dialect), whereas all 100 creoles within only a few centuries time lost gender. In a study of grammatical stability, Wichmann and Holman (2009) found that all four gender‐connected features are temporally “very stable.” Thus, pace Mufwene (e.g. 2015) and DeGraff (e.g. 2009), creolization is not an ordinary process of change. Even if all reductions observable in creoles compared to their lexifiers can be encountered in change in non‐creoles, the mere quantity of changes, including rare ones like loss of gender, leave no doubt that creolization differs from change in uninterrupted generational transmission of languages. Are creoles mixed languages? No, as no mixture in the lexicon can be spotted in creoles, with only a few exceptions: almost all creoles have just one conspicuous lexifier. Non‐lexifier roots tend to be limited to local flora and fauna and cultural inheritance (Bartens and Baker 2012). There are, however, a few creoles with a significant portion of their vocabulary from other languages: the Portuguese Creole Angolar has many Bantu words; Spanish‐lexifier creole Palenquero in Colombia as well, English‐lexifier creole Saramaccan has many Portuguese words, and that is probably to do with their social history, as these are all spoken by maroons. Berbice Dutch is probably the creole with the most pervasive influence from a language other than the lexifier, as can be seen in example (15). The creole is very different from the lexifier, even taking into consideration a dialectal Dutch origin. (15a) Ze hadd‐en bom‐en ge‐plant, maar de koe‐ien hebb‐en ze beschadig‐d. (Dutch) 3PL have.PST‐PL tree‐PL PARTIC‐plant but DEF.ART cow‐PL have‐PL 3PL damage‐PARTIC ‘They had planted trees, but the cows have damaged it/them.’ (15b)

(Berbice Creole; Kouwenberg 2013: 28–90) endi wa plan‐tɛ bom bat di kuj‐ap sei‐t‐o2 3PL.DEM PST plant‐PFV tree but DEM cow‐PL damage‐PFV‐3SG ‘They had planted trees, but the cows destroyed it.’

Compared to the lexifier Dutch, we can observe the following changes. The non‐emphatic third person plural ze has been replaced by an emphatic form endi, possibly related with

210  Peter Bakker dialectal forms like hunnie or hullie (www.meertens.knaw.nl/ewnd/boeken/woord/110928). The irregular past form of the tense‐aspect auxiliary hebben (had) is not transmitted. A tense particle wa has a similar function. A perfective marker ‐tɛ from Ijo is added to both verb stems, as is the Ijo object clitic ‐o. There is no plural marking on the word for ‘tree’ in the creole, as plural marking is optional and not used in generic meanings. But the word for ‘cows’ has a plural marker ‐ap from Ijo rather than from Dutch. The conjunction bat is obviously from English. The pervasiveness of both Ijo and Dutch in the basic lexicon and structure, makes Berbice Creole one of a few mixed creoles. Creoles are normally not mixed languages, however, with the above‐named exceptions because of their lexical admixture. Creoles have a lexicon from one language, but no significant component from another language can be identified.

5.3 Semicreoles The term semicreole was perhaps first used by Révah (1963) in connection with Popular Brazilian Portuguese. Compared to European Portuguese, Brazilian Portuguese, in its most vernacularized forms, has lost some of the verbal and nominal inflection of European Portuguese, and pronouns are used more frequently in Brazilian Portuguese. Semicreoles are mother tongues. They have their origin in another language, but the grammatical system of that language has been simplified within a short period, due to ­language contact. One could say that they are nativized varieties that have been significantly simplified after being acquired by more second language speakers than first language speakers. Afrikaans (Roberge 2002; den Besten 2012) is an example, spoken as different varieties in South Africa by descendants of people mostly from the south of the present‐day Netherlands, France (Huguenots), and people formerly designated as Coloreds (under the Apartheid regime), the latter forming the majority of Afrikaans speakers today. (16a)

Mijn moeder is uit Kaapstad terug‐ge‐kom‐en.  (Dutch) 1SG.POSS mother BE.3SG from Cape.Town back‐PTCP‐come‐PTCP

(16b)

My ma het van Kaapstad (af) terug gekom. (Afrikaans) 1SG.POSS mother PST from Kaapstad back PTCP‐come ‘My mother has returned from Cape Town.’ (den Besten and Biberauer 2013: 29–163)

Note the following differences between Dutch and Afrikaans. Whereas Dutch distinguishes between object pronouns and possessive pronouns for first person (mijn, mij), in Afrikaans my is used for both. Afrikaans ma is from a colloquial Dutch form ma(ma) “moeder.” The suppletive paradigm of the Dutch verb ‘to be’ is reduced to just the forms wees (imperative and infinitive) and is (present) in the present tense. This tense/aspect auxiliary has been replaced by a form of the auxiliary het/had ‘to have’, also with paradigm reduction compared to its source, Dutch hebben. The participle has been preserved except that the ending is dropped. Afrikaans word order in this sentence is fairly close to Dutch, with the auxiliary/inflected verb in second place, and the main verb at the end (Dutch also allows other orders). The ­differences between Dutch and Afrikaans are thus primarily in the realm of grammatical reduction. There are also (not visible here) some grammatical innovations, such as double negation and reduplication. Kituba is another example of a semicreole: (17a)

(Yándi) ka‐ku‐zól‐elé. (Kikóngo; Mufwene 1997: 176) (He/she) AGR‐you‐like‐NEAR.PERFECT ‘He/she likes you.’

Contact and Mixed Languages  211 (17b)

Yándi zól‐a ngé. He/she like‐ENDING you ‘He/she likes you.’

(Kituba; Mufwene 1997: 176)

Mufwene (1997) calls the Kituba language “less agglutinating” compared to its lexifier Kikóngo. “Kitúba has simplified Kikóngo’s tense/aspect morphological system” (p. 179). He describes changes from the perspective of the lexifier as follows: “Most obvious are the absence of a subject‐agreement prefix on the verb, the lack of object pronominal prefixes, and the absolute reliance on independent and morphologically invariant pronouns as well as on their syntactic positions to determine the subject and objects in a sentence. Aside from derivational extensions, the verb form varies only according to tense (…)” (1997: 175). For a study of the history of Kituba, see Samarin (1991). Other languages that have been called semicreoles are Reunionnais, because it lost many of the French grammatical properties, but on the one hand it preserved gender in the articles, unlike all other creoles. Thus, semicreoles are languages that are not as different from their lexifiers as creole languages are. They have undergone reduction of paradigmatic structures, but no significant grammatical innovations, unlike in creoles.

5.4  Are Pidgins, Creoles, and Semicreoles Mixed Languages? After the discussion on pp. 207–211, we can ask whether pidgins, creoles, and semicreoles are mixed languages. Some people may defend the view that these three types of languages are mixed, because their structures have been significantly altered through processes of loss and, in the case of creoles, of expansion as well. Other people may demand that a language has to have components from two identifiable sources, if one wants to call a language mixed. For them, neither of the three types are mixed in themselves. I side personally with the latter view. In my view, pidgins, creoles, and semicreoles are not mixed languages. The only mixed pidgins and creoles are those with extreme lexical admixture in the basic lexicon, as discussed on pp. 207–210 – a small minority. As far as I have been able to establish, only pidgins may display an arbitrary blend of the basic lexicon. This would imply that, if it was a language with different origins of the basic lexicon, it was a lexically mixed pidgin at some point in the past. In Table 10.1, we distinguish five types of contact languages, where creoles and pidgincreoles are combined into one category. They are classified based on the degree and type of influences of the source language(s), using parameters of reduction, the number of languages involved and the presence and absence of grammatical innovations. The category of “mixed languages” differs in most parameters from the others, and I will show why in the next section.

6 Mixed Languages Proper: Structural Types and Links with Genesis Meakins (2016; see also 2013a) and Bakker (2017) have classified mixed languages into different types. Their classifications should be considered preliminary because so few (less than 30) mixed languages have been documented. Both of them do not include creoles in their taxonomic overviews. Mixed languages proper are rare among the world’s languages, and certainly not all of them have been documented before they fell into disuse. Most are, or were, spoken by small ethnic groups, and they appear to have emerged under special social circumstances. In our discussion, we only include languages in which major basic chunks of the language point to different source languages.

Table 10.1  Differences between pidgins, creoles, semicreoles, and mixed languages proper. Pidgins

Creoles and pidgincreoles

Semicreoles

Mixed languages

Lexical reduction Grammatical reduction Number of source languages of most forms Reason for emergence

Much Much One, sometimes two

Much Much One, rarely two

? Significant One

Little hardly Two

Sudden need for communication Mostly not identifiable

Large groups of L2 learners; imperfect learning of target Most are identifiable

Act of identity by bilinguals

Origin of grammatical features identifiable Newly developed grammatical features Full‐fledged grammatical system

Hardly any

Need for a common language; shift from a pidgin Mostly grammaticalized from lexifier Many

Few

Virtually all are identifiable None or hardly any

No

Yes

Yes

Yes

Contact and Mixed Languages  213

6.1  G‐L Mixed Languages To the extent that one can speak of frequency in a class that only contains a few dozen ­languages, the G‐L (Grammar‐Lexicon) type is the most common. These languages combine the lexicon of one language with the grammatical system of another. It was not yet identified as a type in Thomason and Kaufman (1988), although a few were discussed, such as Angloromani (English grammar, Romani lexicon), Ma’a/Mbugu (Cushitic lexicon, Bantu grammar). A few dozen of them are known, among them Media Lengua (Kichwa and Spanish; Muysken 1981, 1997a) and Callahuaya (Quechua, Puquina, and Spanish; Muysken 1997b). I will give an example of Chindo or Chinese Indonesian, spoken by an ethnic group whose members descend from Chinese male traders who spoke a contact variety of Malay and local Javanese‐speaking women. Dreyfuss and Oka described the language as follows: “The majority of the content words are Indonesian, while the majority of grammatical affixes are Javanese” (1979: 250). When Dreyfuss and Oka discussed the language, the type was unknown, and they wondered whether it was “A new kind of language hybrid.” This example shows the four affixes and function words, in italics, of Javanese origin and Indonesian content words. (18)

O, neʔ di-dateng-no pigi rumah s-ribu Oh if PA-come-BEN go house one-thousand ‘Oh, if he’s called to the house, a thousand?’

This is not an isolated, contrived or specially selected example, but the combination of the two languages is systematic: in their calculations, 88% of the grammatical affixes are from Javanese, and 90% of the lexical roots are from Malay. A less well‐known example is the language of the Spanish‐Philippine mestizos from Vigan, Luzon, who mixed Spanish content words (bold) with an Ilocano grammatical system. This example is from a letter written before 1883 (Steinkrüger 2008: 226–7): (19) i‐participar‐co qca TH‐inform‐I 2SG.OBL

a na‐nombrar‐anac LNK PRF‐nominate‐1SG

a cabo del barrio LNK chief of. the neighborhood

ditoy qt sentir‐e‐c unay to ca‐asan‐mo3 and feel‐TR‐1SG very the NOM‐absence‐your here ‘I inform you that they made me chief of the “barrio” (neighborhood) and I regret a lot your absence here.’

6.2  V‐N Mixed Languages V‐N mixed languages combine verbal roots and structures from one language and nominal roots and structures from another. The most famous examples are Michif (Rhodes 197; Bakker 1997) and Gurindji Kriol (Meakins 2013b). In example (20), the second position auxiliary is from local Creole English (im‐in < him been), as is najawan (< another one) and the rest is from the Gurindji Aboriginal language. Sixty‐four percent of verbs and 67% of nouns are from Kriol (Meakins 2013c). (20)

najawan‐tu im‐in jawurra karu another.one‐ERG him‐been steal kid ‘Someone else, he has stolen a child.’

(Gurindji Kriol)

214  Peter Bakker Michif combines Cree (Algonquian) verbs, with all their polysynthetic complexity, with French nouns, including French definite and indefinite articles (French elements italics). (21) lii gros dans kii‐ushihtaaw‐ak, enn gros gros silibraasyon thePL big.F dance PAST‐make‐3PL a.F big.F big.F celebration mihceet kii‐nipah‐eewak lii vash, lii koshon, lii pul, many PAST‐kill‐they.them the.PL cow the.PL pig the.PL chicken ‘They organized big dances, a really big celebration. They slaughtered many cows, pigs and chickens.’ (Norman Fleury, speaker) Michif and Gurindji Kriol differ from one another in that Michif only has French nouns, with very few exceptions, and Gurindji Kriol can have Gurindji and Kriol nouns, but with Kriol case marking (Meakins and McConvell 2005). Bakker (2017) therefore distinguishes between V‐N and V‐NN mixed languages.

6.3  F‐R Mixed Languages F‐R mixed languages are languages in which the grammatical frame is from one language, but all of the roots are from another. Similar phenomena have been described as grammatical convergence (resulting in a Sprachbund or linguistic area like the Balkans or South India, when multiple languages are involved), and metatypy. No examples in which both the noun phrase and the verbs have been metatypized are known. Sri Lankan Portuguese (SLP) is derived from a creolized version of Portuguese that became an agglutinative language under the influence of the neighboring languages. It also changed from SVO to SOV, and developed case marking (including experiencer dative) and prefixing verbal morphology (Smith 1979: 212). (22)

əwərə e:w wi:tu poli:s‐su sirwi:s‐ntu ippa na:n vantu poli:s‐fa ve:lay‐la now I TOP police‐GEN work‐LOC Agora, quanto a mim, estive no trabalho policial. ‘Now as for me, [I] was in police work.’

tiñə (SL Portuguese) ifu‐nt‐an (SL Tamil) be‐PAS‐CNC (Portuguese)

In SLP, Portuguese possessive adjective sua has become a possessive suffix –su in SLP. The SLP locative suffix ‐ntu is from the P. preposition junto ‘near’. Note that Tamil (and Sinhala) verbs are suffixing, but the SLP verbs are prefixing. Portuguese is SVO, but SLP has SOV constituent order. In other words, a rather radical typological change, sometimes called conversion, has taken place (Bakker 2000). In the noun phrase, syntax has been affected, but in the verb only a semantic match (Bakker 2006).

6.4  Other Types: INFL‐rest The final type is one in which the verbal inflection, often taken to be characteristic for the genealogical affiliation of a language, is from a language different from the rest of the language. There are only two known examples of such languages, Copper Island/Mednyj Aleut (Russian INFL, otherwise mostly Aleut) and Hubner Mischsprache (Slovenian INFL, otherwise mostly German). In both cases, not only INFL but also personal pronouns, possessive pronouns, and deictic terms are from a Slavonic language, as in example (23):

Contact and Mixed Languages  215 (23) cwå nāj‐e hājʒ‐r drrixt‐a moj prō(ə)dr tam hintr nāʃme gôrte (Mischsprache; Slovenian italics) two new‐PL house‐PL build‐3 my.M brother there behind our garden ‘My brother built two new houses there behind our garden.’ (Lessiak 1959: 219) Copper Island Aleut is spoken by people of mixed Aleut and Russian descent (Sekerina 1994) and Hubner Mischsprache by German‐speaking immigrants in rural Slovenia (Pohl 1995).

7 Socio‐historical Scenarios for the Genesis of Mixed Languages Proper Bakker (1997) suggested two main socio‐historical scenarios for the genesis of mixed languages proper. One is the situation in which women who are speakers of language W, and men who are speakers of language M, intermarry, and their descendants are sufficient in number to form a new ethnic group. The members of the new ethnic group will call themselves “metis,” “half‐child,” or a similar term, and their language appears, in a number of cases, as a mixed language, with the women providing the grammatical system, and the men the lexicon. Alternatively, the locals (in the cases discussed here always women) provide the grammatical system, and the immigrants the lexicon. The second scenario is that of nomads speaking language N who settle in a region where language S is spoken. They may develop a language in which the lexicon is from the N language and the grammatical system from language S. Both events are acts of identity (LePage and Tabouret‐Keller 1985). It should be stressed, however, that not all descendants of settled nomads and mixed marriages speak mixed languages. This model has some predictive, or retrodictive power (i.e. if we have a mixed language, then one of these scenarios must have taken place): new mixed languages have indeed been discovered afterwards that follow this pattern. There are, however, also exceptions. Media Lengua of Ecuador and Gurindji Kriol of Australia, for instance, are spoken by indigenous populations that do not fit either of these scenarios. For the Hubner Mischsprache, we have no data to reconstruct a socio‐historical scenario.

8  Mixed Languages and Mixture of Fluids: A Metaphor Many phenomena have been subsumed under mixed languages, including code‐switching, the spontaneous blending of two or more languages by bilinguals, languages with more than occasional borrowings – including English with French and Norse, Maltese with Italian and English, Chamorro with Spanish, or the languages in a linguistic area or Sprachbund that have taken over structural traits from neighboring languages, e.g. in South Asia, the Balkans, or the Northwest Coast of North America. I will try to make it clear how “mixture” could be interpreted with a metaphor of fluids and other substances in an aquarium. If one pours one carton of milk into an empty aquarium and then another carton of milk, is the result a mixture? The result is milk, a mixture of two virtually identical components. If one mixes a red fluid with a yellow fluid in the aquarium, the result will be a fluid with an orange color. Is it then mixed, because the new homogenous fluid was originally two different fluids? The original components have disappeared, and a new mixture has appeared. Can one call it mixed? If one mixes identical quantities of water

Table 10.2  Mixture of substances as metaphors of language mixture.

Are components distinguishable?

Can components be returned to original state?

Have components changed visual appearance?

Do components belong to different domains in matter physics (liquid, solid, gas)

Milk and milk

no

no

no

no

Red and yellow Water and oil Water and pebbles Water and air

no yes yes yes

no yes yes no

yes no no yes

no no yes yes

Language contact equivalent Normally transmitted languages No equivalent Code‐switching Mixed language Creole

Contact and Mixed Languages  217 and oil, and the oil floats on top of the water, is the result mixed? Yes, there are two distinct components in the same container. But the two components have not merged into one. If one throws a quantity of pebbles in an aquarium, so the mass is equally water and stones, are the contents of the aquarium mixed? Yes, as one can observe mixed contents. No, because the stones and water are separate, and have hugely different physical properties. Finally, if one has an aquarium filled with water, and after some time, only half of the water is left due to evaporation, is the result mixed? Yes, if one would claim that half of the water has become air, then the result is a mixture of water and air. No, if one considers only what is left. This illustrates the problem of mixed languages in that in all cases one can accept the result is mixed, or dispute that it is mixed (Table 10.2). The differences boil down to whether the components, after the mixture, have become indistinguishable or not, and whether or not the result of the two components have new properties. Two cartons of milk have mixed to an identical result; two colored fluids have mixed to become a new fluid; two fluids have mixed because the oil floats on the water in the same space and the aquarium shows a mixture of two physically different entities, and in the last case only part of the original entity is maintained, whereas one can discuss whether the rest is nothing, or air.

9 Conclusions In this chapter, I have discussed different types of languages that have resulted from contact: mixed languages proper have emerged where people have merged the two languages they speak at a fairly fluent level. Pidgins and creoles are not mixed languages, as they have emerged where people did not have a language in common, and needed to create one, and there are no two components from clearly identifiable sources. Some people remain skeptical, however, about the existence of mixed languages (Versteegh 2017).

NOTES 1 Abbreviations used in the glosses: AGR agreement; ART article; BEN benefactive, causative; CNR concord; DEF definite; DEM demonstrative; ERG ergative; F feminine; FUT future; GEN genitive; IMPERF imperfective; HB habitual; INF infinitive; IRR irrealis; LNK linker; LOC locative, NEG negation; OBJ object; NOM nominalizer; PA passive transitive; PARTIC participle; PAS past; PFV perfective; PL plural; POSS possessive; PRES present; PRF perfect; PROG progressive; PST past; PTCP participle; SG singular; SUB subject; SUBJ subjunctive; SUBORD subordinator; TH theme; TOP topic; TR transitive; TRANS transitive. 2 Bold elements are from Ijo. 3 Perhaps asan is derived from Spanish ausencia ‘absence’ or ausente ‘absent’.

REFERENCES Aboh, Enoch Oladé 2015. The Emergence of Hybrid Grammars. Language Contact and Change (Cambridge Approaches to Language Contact). Cambridge: University Press. Arends, Jacques, Pieter Muysken, Norval Smith 1995. Pidgins and Creoles: An Introduction. Amsterdam: Benjamins.

Bakker, Peter 1987. A Basque nautical pidgin: a missing link in the history of fu. Journal of Pidgin and Creole Languages 2.1: 1–30. Bakker, Peter 1997. “A Language of our Own.” The Genesis of Michif – the Mixed Cree‐French Language of the Canadian Métis. New York: Oxford University Press.

218  Peter Bakker Bakker, Peter 2000. Rapid language change: creolization, intertwining, convergence. In C. Renfrew, A. McMahon, R. L. Trask (eds.), Time Depth in Historical Linguistics (Papers in the Prehistory of Languages). Cambridge: McDonald Institute for Archaeological Research, pp. 585–620. Bakker, Peter 2006. The Sri Lanka Sprachbund: the newcomers Portuguese and Malay. In Yaron Matras, April McMahon, and Nigel Vincent (eds.) Linguistic Areas. Convergence in Historical and Typological Perspective. Basingstoke: Palgrave Macmillan, 135–59. Bakker, Peter 2008. Pidgins versus creoles and pidgincreoles. In Silvia Kouwenberg and John Singler (eds.), Handbook of Pidgin and Creole Studies. Oxford: Wiley‐Blackwell, pp. 130–57. Bakker, Peter 2017. Typology of mixed languages. In A. Y. Aikhenvald and R. M. W. Dixon (eds.), The Cambridge Handbook of Linguistic Typology. Cambridge: Cambridge University Press, pp. 217–53. Bakker, Peter and Maarten Mous (eds.) 1994. Mixed Languages. 15 Case Studies in Language Intertwining. Amsterdam: IFOTT. Bakker, Peter and Hein van der Voort 2017. Polysynthesis and language contact. In M. Fortescue, M. Mithun, and N. Evans (eds.), The Oxford Handbook of Polysynthesis. Oxford: Oxford University Press, pp. 408–27. Bartens, Angela and Philip Baker (eds.) 2012. Black Through White: African Words and Calques which Survived Slavery in Creoles and Transplanted European Languages. London: Battlebridge. Besten, Hans den 2012. Roots of Afrikaans. Selected writings of Hans den Besten, edited by Ton van der Wouden. Amsterdam: Benjamins. Besten, Hans den and Theresa Biberauer 2013. Afrikaans structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐online.info/contributions/29 Bickerton, Derek 1984. The language bioprogram hypothesis. The Behavioral and Brain Sciences 7: 173–88. Crevaux, Jules 1880. De Cayenne aux Andes par l’Oyapock, le Yary, le Parou, l’Amazone et l’Iça, retour par le Yapura. Bulletin de la Société de Géographie, mai 1880, pp. 385–416.

DeGraff, Michel 2009. Language acquisition in creolization and, thus, language change: some Cartesian‐Uniformitarian boundary conditions. Language and Linguistics Compass 3.4: 888–971. Dimmendaal, Gerrit J. 1995. Do some languages have a multi‐genetic or non‐genetic origin? An exercise in taxonomy. In Robert Nicolai and Franz Rottland (eds.), Proceedings of the Fifth Nilo‐Saharan Conference, Nice, 1992. Cologne: Rüdiger Köppe, 354–69. Dreyfuss, G. R. and Djoehana Oka 1979. Chinese Indonesian: a new kind of language hybrid? Pacific Linguistics A‐57: 247–74. Foley, William A. 2013. Yimas‐Arafundi Pidgin. In Susanne Maria Michaelis, Philippe Maurer, and Martin Haspelmath and Magnus Huber (eds.), Contact Languages Based on Languages from Africa, Asia, Australia, and the Americas. Oxford: Oxford University Press, pp. 105–13. Greenberg, Joseph H. 1999. Are there mixed languages? In L. Fleishman, M. Gasparov, T. Nikolaeva, A. Ospovat, V. Toporov, A. Vigazin, R. Vroon, and A. Zalizniak (eds.), Essays in Poetics, Literary History and Linguistics Presented to Viacheslav Vsevolodovich Ivanov on the Occasion of his Seventieth Birthday. Moscow: OGI, pp. 626–33. Hollyman Jim K. 1964. L’ancien pidgin français parlé en Nouvelle‐Calédonie. Journal de la Société des Océanistes 20: 57–64. doi: https:// doi.org/10.3406/jso.1964.1901 Holm, John A. 2003. Languages in Contact: The Partial Restructuring of Vernaculars. Cambridge: Cambridge University Press. Huttar, George and Frank J. Velantie 1997. Ndyuka‐Trio Pidgin. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 99–124. Juvonen, Päivi 2008. Complexity and simplicity in minimal lexica: the lexicon of Chinook Jargon. In Matti Miestamo, Kaius Sinnemäki and Fred Karlsson (eds.), Language Complexity: Typology, Contact, Change. Amsterdam: Benjamins, pp. 321–40. Kouwenberg, Silvia 1994. A Grammar of Berbice Dutch. Berlin: de Gruyter Mouton. Kouwenberg, Silvia 2013. Berbice Dutch. In Susanne Maria Michaelis, Philippe Maurer and Martin Haspelmath and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures. English‐based and Dutch‐based Languages. Oxford: Oxford University Press, pp. 275–84. LePage, Robert and Andrée Tabouret‐Keller. 1985. Acts of Identity: Creole‐Based Approaches to

Contact and Mixed Languages  219 Language and Ethnicity. Cambridge: Cambridge University Press. Lessiak, Primus 1959. Die deutsche Mundart von Zarz in Oberkrain. A. Grammatik. Marburg: N. G. Elwert Verlag. Lyovin, Anatole 1968. The Sino‐Russian mixed language in Manchuria. Working Papers in Linguistics 3: 135–64 (Department of Linguistics, University of Hawaii). Matras, Yaron 2000. Mixed languages: a functional‐communicative approach. Bilingualism: Language and Cognition 3.2: 79–99. Matras, Yaron and Peter Bakker (eds.). 2003. The Mixed Language Debate: Theoretical and Empirical Advances. Berlin: de Gruyter Mouton. Meakins, Felicity 2013a. Mixed languages. In Peter Bakker and Yaron Matras (ed.), Contact Languages: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 159–228. Meakins, Felicity 2013b. Gurindji Kriol structure dataset. In Susanne Maria Michaelis, Philiippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available online at http://apics‐online.info/ contributions/72 Meakins, Felicity 2013c. Gurindji Kriol. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 3: Contact Languages Based on Languages from Africa, Asia, Australia, and the Americas. Oxford: Oxford University Press. Meakins, Felicity 2016. Mixed languages. Oxford Research Encyclopedia of Linguistics. Oxford: Oxford University Press. doi:10.1093/ acrefore/9780199384655.013.151 Meakins, Felicity and Patrick McConvell 2005. Gurindji Kriol: A mixed language emerges from code‐switching. Australian Journal of Linguistics 25.1: 9–30. Meira, Sergio and Pieter Muysken 2017. Cariban in contact: new perspectives on Tiriyo‐Ndyuka Pidgin. In Kofi Yakpo and Pieter Muysken (eds.), Boundaries and Bridges. Language Contact in Multilingual Ecologies. Berlin: de Gruyter Mouton, pp. 197–227. Michaelis, Susanne Maria, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) 2013. The Atlas of Pidgin and Creole Language Structures. Oxford: Oxford University Press. Mous, Maarten 1994. Ma’a or Mbugu. In Peter Bakker and Maarten Mous (eds.), Mixed

Languages. 15 Case Studies in Language Intertwining. Amsterdam: IFOTT. Mous, Maarten 2003. The Making of a Mixed Language: The Case of Ma’a/Mbugu. Amsterdam: Benjamins. Mufwene, Salikoko S. 1997. Kituba. In Sarah G. Thomason, (ed.), Contact languages. A Wider Perspective. Amsterdam: Benjamins, pp. 173–208. Mufwene, Salikoko S. 2015. Pidgin and creole languages. In James D. Wright (ed.), International Encyclopedia of the Social and Behavioral Sciences, 2nd edition, vol. 18: 133–145. Oxford: Elsevier. Muysken, Pieter 1981. Halfway between Quechua and Spanish: the case for relexification. In Arnold Highfield and Albert Valdman (eds.), Historicity and Variation in Creole Studies. Ann Arbor, MI: Karoma, pp. 52–78. Muysken, Pieter 1997a. Media Lengua. In Sarah G. Thomason (ed.), Contact languages. A Wider Perspective. Amsterdam: Benjamins, pp. 365–426. Muysken, Pieter 1997b. Callahuaya. In Sarah G. Thomason (ed.), Contact Languages. A Wider Perspective. Amsterdam: Benjamins, pp. 427–47. Parkvall, Mikael 2017. Pidgin languages. Oxford Research Encyclopedias. Oxford: Oxford University Press. Available at http:// linguistics.oxfordre.com/view/10.1093/ acrefore/9780199384655.001.0001/ acrefore‐9780199384655‐e‐58 Parkvall, Mikael and Peter Bakker 2013. Pidgins. In Yaron Matras and Peter Bakker (eds.), Contact Languages. Berlin: de Gruyter Mouton, pp. 15–64. Pernel, Pim 1938. Article about Penang, Malaysia around 1900 in Dutch newspaper Het Vaderland, November 6, 1938. Pohl, Heinz‐Dieter 1995. Slowenischer‐ Deutscher Sprachkontakt in Krain. Bemerkungen zur “Hubner Mischsprache.” Klaus Sornig et al. (eds.), Festschrift Norman Denison. In Grazer Linguistische Monographien 10: 315–22. Révah, Israël Salvatore 1963. La question des substrats et superstrats dans le domaine linguistique brésilien. Les parlers populaires brésiliens doivent‐ils être considérés comme des parlers “créoles” ou “semi‐créoles?” Romania 84: 336: 433–50. doi:https://doi. org/10.3406/roma.1963.2916 Rhodes, Richard 1977. French Cree – a case of borrowing. In William Cowan (ed.), Actes du

220  Peter Bakker huitième congrès des algonquinistes. Ottawa. Carleton University, pp. 6–25. Roberge, Paul 2002. Afrikaans: considering origins. In Rajend Mesthrie (ed.), Language in South Africa. Cambridge: Cambridge University Press, pp. 79–103. Roberts, Sarah and Joan Bresnan 2008. Retained inflectional morphology in pidgins: a typological study. Linguistic Typology 12: 269–302. Ross, Malcolm D. 2006. Metatypy. In Keith Brown (ed.), Encyclopedia of Language and Linguistics, vol. 8. Oxford: Elsevier, pp. 95–9. Ross, Malcolm D. 2007. Calquing and metatypy. Journal of Language Contact: Thema 1: 116–43. Samarin, William 1991. The origins of Kituba and Lingala. Journal of African Languages and Linguistics 12.1: 47–77. Sekerina, Irina A. 1994. Copper Island (Mednyj) Aleut (CIA): A mixed language. Languages of the World 8: 14–31. Smith, Ian. 1979. Convergence in Sri Lanka: A Creole Example. Lingua 48: 193–222. Steinkrüger, Patrick. 2008. Hispanisation processes in the Philippines, in Thomas Stolz, Dik Bakker and Rosa Salas Palomo (eds.), Hispanisation. The Impact of Spanish on the Lexicon and Grammar of the Indigenous Languages of Austronesia and the Americas. Berlin: de Gruyter Mouton, pp. 203–36.

Thomason, Sarah G. 1995. Language mixture: ordinary processes, extraordinary results. In Carmen Silva‐Corvalán (ed.), Spanish in Four Continents: Studies in Language Contact and Bilingualism. Washington, DC: Georgetown University Press, pp. 15–33. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of Californian Press. Troy, Jakelin Fleur 1994. Melaleuka: a history and description of New South Wales pidgin (Australian National University Ph.D. dissertation). Van Driem, George 2001. Languages of the Himalayas. An Ethnolinguistic Handbook of the Greater Himalayan Region Containing an Introduction to the Symbiotic Theory of Language, 2 vols. Leiden: Brill. Velupillai, Viveka 2015. Pidgins, Creoles and Mixed Languages: An Introduction. Amsterdam: Benjamins. Versteegh, Kees 2017. The myth of the mixed languages. In Benjamin Saade and Mauro Tosco (eds.), Advances in Maltese Linguistics. Berlin: de Gruyter Mouton, pp. 245–66. Wichmann, Søren and Eric W. Holman 2009. Assessing Temporal Stability for Linguistic Typological Features. Munich: LINCOM Europa.

11 Contact and Sociolinguistic Variation MAYA RAVINDRANATH ABTAHIAN AND JONATHAN KASSTAN 1 Introduction In a seminal paper addressing the study of language variation and change, Weinreich, Labov, and Herzog laid out the principles for the study of language as an object possessing variation that demonstrates “orderly heterogeneity” (1968: 100). One of the guiding principles of this research paradigm has been the “use of the present to explain the past” (also known as the Uniformitarian Principle, Labov 1978), a doctrine borrowed from geology which proposes that the forces that act on language today are the same as those that acted on language in the past. Assuming that all change in language must be preceded by some period of variability, the adoption of the Uniformitarian Principle allows for the examination of language change synchronically (in “apparent time”), through close investigation and quantification of linguistic variables that are correlated with social characteristics of speakers and interactions between speakers. This approach to the study of language in society is known as variationist sociolinguistics (e.g. Tagliamonte 2012). In this chapter we focus on research in the variationist paradigm that intersects with the field of language contact. We predominantly focus on sound change, which forms the bulk of the work at this interface, as well as a significant part of the tradition of variationist sociolinguistics. However, we also make reference to particular case studies at other levels of linguistic description. We focus on some of the current discussions of the types of changes that are possible in the context of different language‐contact scenarios, as well as some of the distinctions that are made between different types of change and different types of language‐ contact scenarios. In particular, we highlight how variationist methods have contributed to our understanding of contact‐induced change in such contexts. An ongoing area of inquiry that provides a further theme for this chapter is whether we expect the principles of language variation and change as we understand them today, and as largely laid out by Labov (1994, 2001, 2010), to be applicable in communities that do not speak a dominant standardized language, and are not Western, urban, and similarly socially stratified, which broadly characterizes the sites of study in much of the work in the field (see e.g. Nagy and Meyerhoff 2008; Stanford 2016; Hall et al. 2019; Meyerhoff 2019). So, how do the principles of language variation and change apply in disparate rural communities, in very small language communities, in highly multilingual communities, in communities undergoing language shift, or in communities where a clear standard is not readily

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

222  Maya Ravindranath Abtahian and Jonathan Kasstan identifiable? These last three scenarios in particular are of interest here, as these are the types of language communities where language contact is “part of the social fabric of everyday life” (Sankoff 2001: 638) and thus part of the social context for the types of language variation and change that we will discuss here. Moreover, many of these are also communities where indigenous minority languages are being lost, as the social processes that promote language shift and endangerment affect some 50–90% of these languages (see Lee and Van Way 2018). As indigenous languages are used in fewer domains and their speaker communities change, we lose opportunities to observe language variation and change in situations that are different from those more commonly studied (Stanford and Preston 2009: 4). In Sections 3 and 4 we discuss linguistic and social factors in the study of language contact and sociolinguistic variation, drawing on a number of examples from Stanford and Preston’s (2009) volume, and from Winford (2003), as well as from overviews in Sankoff (2001), Winford (2013), and Ravindranath (2015). In Section 5 we make use of our own more recent empirical research to discuss a number of distinctions that have been made with respect to different types of changes in the context of language contact. First, to orient the reader, we offer a brief overview on the origins of the variationist paradigm. As with most overviews of language variation and contact, we begin with Uriel Weinreich.

2  Language Contact and Variationist Sociolinguistics Uriel Weinreich was a central figure in arguing for orderly heterogeneity as a basic function of language, that is, for the systematicity of what had previously been termed “free variation,” (Weinreich et al. 1968). Any two ways of saying the same thing had – in the structuralist tradition  –  been seen as outside of the scope of linguistic theory, and attributed to issues of “performance” (contra “competence” in Chomsky’s classic dichotomy). The use of the term “free variation” served to exclude variation from the scope of inquiry to begin with: it said nothing about why speakers alternate competing forms, or where such forms originate (Fischer 1958: 47). The notion that competing forms can alternate freely remained a thorny issue, however, because it augured poorly with historical linguistic theory that sought categorical and predictable rules. The label “orderly heterogeneity” then served to highlight that human behavior can be variable, but nonetheless conform to systematic and predictable rules. To adequately capture this systematicity, variationists schematize variation as a linguistic variable: an abstract unit representing two competing variants that convey the same referential meaning. Labov (1963)’s seminal study on the island of Martha’s Vineyard demonstrated that linguistic variables could be used to track language change in apparent time (i.e. across age groups). Fundamentally, the apparent‐time construct hinges on the postulation that the grammar stabilizes in adulthood. Therefore, any progressive change in competing variants could be identified in principle as a possible change in progress. Thus, to paraphrase Bailey’s (1973) oft‐quoted maxim, it is generally accepted that where change emerges, variation preceded it, but not all variation necessarily leads to change. One goal of variationist sociolinguistics is to understand and model the factors that promote change. Weinreich et al. also considered what they saw as key empirical challenges to building a socially accountable theory of linguistic change. We outline these challenges in some detail in Table 11.1, as they provide a roadmap for how variationist sociolinguistics has engaged with contact linguistics. The constraints problem mandates establishing the envelope of variation: that is, delineating not only all possible conditions under which a given linguistic variable is said to vary, but also the possible variants that alternate for the variable. The transition problem requires mapping how a given variable transitions in a given community: is the change gradual? If  so, is this common across all variables? Do changes (such as phonetic changes) move

Contact and Sociolinguistic Variation  223 Table 11.1  Empirical principles for a theory of language change (Weinreich et al. 1968: 101–2). The constraints problem The transition problem The embedding problem The evaluation problem The actuation problem

What are the set of possible changes and possible conditions for changes which can take place in a structure of a given type? What intervening stages can be observed, or must be posited, between any two forms of a language defined for a language community at different times? What other linguistic changes are associated with the given change, or how might a particular change trigger other changes? How can the observed changes be evaluated – in terms of their effects upon the linguistic structure, upon community efficiency, and on the wide range of nonrepresentational factors involved in speaking? Why do certain instances of variation become changes while others do not?

regularly or at different rates in a lexically diffuse manner? The embedding problem considers not only the social embedding of a particular variable, but also the extent to which it is embedded in the linguistic system. Observing patterns in this way allows for a prediction for how language change progresses through a given community. Understanding how variation is embedded necessarily requires understanding how said variation is evaluated by speakers (the evaluation problem). Does a particular change lead to misunderstanding? Is the incoming variant evaluated negatively? Do patterns of evaluation lend themselves to testable hypotheses about all changes? Lastly, the actuation problem requires the researcher to consider how and why changes take off to begin with in some cases, but not in others. The actuation problem is a particular challenge in variationist research because, as Kiesling (2011: 11) among others points out, changes likely begin unobserved, and so identifying the beginning of a change becomes something of a quandary. More recent work at the interface of formal and variationist sociolinguistics has instead attempted to posit the conditions under which variation is likely to emerge in the grammar, as a means to predicting the actuation of variation (e.g. Cornips and Corrigan 2005; Cheshire et al. 2013). Nonetheless the bulk of variationist research thus far has tended to focus on transition, embedding, and evaluation (Table 11.1). We see variationist sociolinguistics as an important and appropriate paradigm for addressing key research questions in language‐contact studies more broadly, principally: “what types of contact‐induced change will occur and when” (Thomason and Kaufman 1988: 13). In fact, Weinreich’s thinking stemmed from his work in language contact. In a seminal study, Weinreich (1953 [1964]) considers the sociocultural and linguistic mechanisms of contact‐induced language change, where he foregrounds the bilingual speaker as the locus of the contact: “two or more languages will be said to be in contact if they are used alternately by the same person” (1964: 1).1 Under this model, any points of difference between their linguistic systems can be considered vulnerable and therefore potentially subject to contact‐induced change. Studies at the intersection of variationist sociolinguistics and language contact that address most clearly embedding and evaluation have for the most part tended to focus on transfer effects in some or all of the contact varieties. Typically, where varieties are in contact at the community level, there tends to be an unbalanced relationship in that at least one variety will exert at least some influence on at least one other variety. In the contact literature, Thomason (2001: 60) presents a useful typology of contact‐induced mechanisms and processes, the most common types being transfer (e.g. of phonology), replication (of structural or grammatical forms), and borrowing (e.g. of lexis, morphosyntactic material such as conjunctions, pragmatic markers, or new phonemes or prosodic features, etc.). In variationist research, less attention has been dedicated to the lexicon, as lexical variation

224  Maya Ravindranath Abtahian and Jonathan Kasstan tends to be more recalcitrant to systematic quantitative analysis by comparison with phonological variation (principally because of issues of form ~ function asymmetry; Sankoff 2001).2 In general, the intensity of contact, the typological fit of the two varieties in contact, and the particular sociolinguistic ecology will dictate the level of contact‐induced change. This continuum can extend from borrowing and transfer on one end to complete language shift on the other. For instance, in one classic study, Gal (1978) demonstrated that bilingual women in Oberwart were progressively shifting from Hungarian to German at a macro (or community)‐ level in order to improve their socio‐economic conditions. Gal was able to describe this change by mapping the extent to which younger speakers were abandoning Hungarian in key domains of usage in apparent time. In later work, Gal (1984) identified the (micro‐level) linguistic correlates of this shift. She reports that, while her older Oberwart speakers consistently style‐shift in traditional sociolinguistic interviews for three Hungarian variables, her younger speakers acquire the variants for these linguistic variables, but not the associated interactional constraints, and so were found to use the variants interchangeably irrespective of the level of formality. What is less clear is the extent to which this linguistic outcome of contact (i.e. the contraction of the socio‐stylistic continuum) is a result of the younger speakers’ imperfect learning, or whether this results from social‐network pressures, or a lack of standard norms, all of which are arguments that Gal (1984) appeals to. The case studies thus serve to highlight the complexity of differentiating so‐called internally motivated change from externally motivated change, a dichotomy that we address further on p. 227. One central question that has driven research at the interface of language contact and variationist sociolinguistics has been the extent to which we can be sure that a change is contact‐induced. Following Thomason (2001: 93–4), Poplack and Levey (2010: 410) have argued that, to determine whether contact‐induced change is occurring, researchers need to: a. establish that the proposed incoming (or “interference”) features were not present in the variety prior to contact; b. establish that the proposed incoming features were present in the source variety prior to contact; c. account for internal motivations. What is considered to be an interference feature can be broadly defined: e.g. “direct importations from a source language” that otherwise would not have arisen (Thomason 2001: 62), such as morphemes or higher‐level structures, with or without structural change to the source features in the process of transfer. Similarly, later changes can take place in the replica variety as a result of the features borrowed earlier on in that variety’s history (a “snowball effect”). Although in all cases it can be difficult to distinguish internal vs external motivations for the change taking place, in the language‐contact literature a distinction is often made between the relative importance of linguistic or social factors in predicting different types of contact effects. In Section 3 and 4 of this chapter we consider some of these linguistic and social factors from a variationist perspective, all the while making it clear that a nuanced interpretation and a clear balance of evidence from constraints hierarchies remains critical in the investigation of contact effects.

3 Linguistic Factors in Language Contact and Sociolinguistic Variation In studies that have focused on transfer, Thomason (2001: 76–7) proposes at least two linguistic factors that are generally agreed to affect the systemic outcome of the contact‐ induced change: (i) the markedness of the feature (i.e. “simpler” or “transparent” unmarked

Contact and Sociolinguistic Variation  225 features transfer over marked features), and (ii) the typological distance, or “congruency,” between the contact varieties (as similarity between varieties can promote more transfer), and these factors are discussed most clearly in relation to the transfer of categorical features from one variety into another (for a discussion of these factors, see Thomason 2001).3 Where most disputes arise, however, is in the true source of features described as resulting from transfer; this debate is perhaps most pronounced in the creole literature. In an attempt to navigate this concern, comparative sociolinguistic analysis (e.g. Poplack and Tagliamonte 2001) has offered important methodological and theoretical input on approaches to language contact. Where contact studies have looked for instance at the transfer of categorical features, variationist approaches to language contact have not only looked at transfer of variable features, but have also made use of statistical modeling and multivariate analysis (such as multiple regression) to enable researchers to focus on the transfer of constraints (per Table 11.1) operating on preference of forms, too. As this research has shown, such an approach to language contact provides a nuanced research design for identifying the source variety by according analytical weight to the variable rules operating on the transferred feature. Meyerhoff (2009) adopts such comparative quantitative methods to evaluate claims that variable null subjects and objects in vernacular Bislama (Malo Island, Northern Vanuatu) originate in the variable production of the same feature in Tamambo (the L1 of most speakers on the island). In Bislama, subjects and objects can be realized as phonetically overt full NPs, as pronouns, or as null arguments, and prior work has suggested that animacy as a semantic feature can and does constrain argument realization across the substrate languages of the wider region (see e.g. Crowley 2002 on Tok Pisin). This work thus presents a falsifiable claim that can be tested empirically using variationist techniques. Meyerhoff compares data taken from a corpus of Bislama and Tamambo oral speech samples, testing the hypothesis that the constraints on null subject and object realization are probabilistically and distributionally similar across varieties. This is an important step in comparative sociolinguistic research if it is to be argued that transfer has indeed taken place from Tamambo into Bislama: if this were the case, we would expect some parity in terms of the most important constraints operating on the variation in both varieties. However, Meyerhoff instead finds conflicting evidence for transfer: “the same (or similar) factor groups emerge as significant constraints on null subject and null object realisation in Tamambo and Bislama, but which factors constrain the variable differ in the two languages” (2009: 391). In other words, while the same feature appears to be operative in Bislama and Tamambo, the variable rules differ. In the aggregate, the evidence suggests that animacy as an internal‐linguistic factor does not predict null object realization in Tamambo in the same way that it is predicted in Bislama, but there is parity in terms of probabilistic distribution of constraints for some (i.e. principally third person) subjects across varieties, “where a robust distinction [for Bislama] seems to be going in the same direction as some substrate languages” (2009: 391). In short, animacy in Bislama appears to be grammaticalizing in the same direction as Tamambo. Further, Meyerhoff marshals these observations to make a broader theoretical claim concerning the extent to which features can be “transformed” in the course of transfer from one variety to another (for a fuller discussion see Meyerhoff 2003). The case study highlights how the adoption of variationist methods in a context of language contact can illuminate which internal‐linguistic constraints carry greater weight in the variable realization of the transferred feature, which in turn can provide evidence for identifying the origin of the proposed transfer feature. In subsequent work, Meyerhoff has also demonstrated how linguistic‐external factors such as speaker age can impact upon this variation. We turn next to the role of social factors such as speaker age in change.

226  Maya Ravindranath Abtahian and Jonathan Kasstan

4 Social Factors: Sociolinguistic Ecology of Language‐Contact Scenarios We rely on a broad understanding of “social” here to ask: how do different types of ­language‐contact scenarios affect the structural outcomes of contact? Even if we begin with Weinreich’s most basic definition that languages in contact are those that are spoken by bilinguals (1968: 1), there is a range of social contexts that may produce communities of bilingual speakers, and which may influence the linguistic outcomes of contact (Thomason and Kaufman 1988; Sankoff 2001; Winford 2003). These social contexts, arising generally from the movement of people from one place to another, may be broadly grouped into three categories: immigration, conquest, and migration – that result in two general outcomes, namely stable multilingualism or language shift. In both cases, inequitable distribution in either the domains of usage of those languages or the relative prestige of the languages in particular domains is common, and this is likely to affect the linguistic outcomes. Thus, some of the social factors that need to be considered include the presence or absence of imperfect learning on the part of the speech community, the history and the length of contact, the power relationships between the groups, the relative mobility of the speakers, patterns of social interactions between the groups, and attitudes towards the languages in question (Winford 2013). In the context of immigration, linguistic assimilation of the immigrant group to the dominant local language is generally rapid enough that the effect of learning the new dominant language as an L2 rarely persists as interference or substrate effects beyond the first‐generation immigrants, although in some cases features may persist as socio‐indexical identity markers of the group for some generations thereafter (see for instance Dubois and Horvath 1998 on the persistence of interdental stops in English speakers of Cajun heritage in Louisiana). Sankoff (2001) lists a few studies that have demonstrated phonological effects of substratum influence among English‐dominant Hispanic immigrants in the United States, including Santa Ana (1996) and Fought (1999). Prince (1988) has also demonstrated a substrate effect of Yiddish in the English of descendants of Yiddish speakers with respect to the discourse function of so‐called Yiddish‐movement as a type of syntactic Focus Movement. In the context of conquest, where a longer period of contact may persist before shift occurs to the dominant language, and where the initial majority of speakers are speakers of the non‐ dominant language, there may be more contact effects on the (minority) dominant language. One possible outcome here is what are sometimes termed indigenous varieties, such as World Englishes (Kachru 1985) like Nigerian English and Indian English; vernacular Irish English is also frequently cited as an example of substantial substratum influence on an incoming, socially dominant, minority language in language shift, which was largely completed by the late nineteenth century (Hickey 2010, this volume; Odlin 1997). Taiwanese Mandarin, as another case study, provides at least two examples of what seem to be substratum influence of Southern Min on the Mandarin spoken in Taiwan, one syntactic/semantic and one phonetic (Kubler 1985; Baran 2014). These two examples also serve to demonstrate a few generalizations and questions about the study of language variation in the context of shift. The first example demonstrates the transfer of the semantics of a Southern Min verb to an analogous verb in Mandarin. Both Southern Min and Mandarin have verbs for ‘have’ and ‘not have’ (ū, bóu; and yŏu, méiyŏu, respectively). In both languages these are main verbs, but in Southern Min they may also be auxiliary verbs that indicate the completion or non‐completion of an action and the affirmation or negation of existence. Unlike in Standard (Beijing) Mandarin, Taiwanese Mandarin speakers can use the verbs yŏu ‘have’ and méiyŏu ‘not have’ as auxiliary verbs (Kubler 1985), and he argues that this is due to contact with Southern Min, a majority (but less socially dominant) language in Taiwan following the imposition of standard Mandarin as the primary language of education post‐World War II. These analogous

Contact and Sociolinguistic Variation  227 structures in the substrate and contact variety demonstrate a phenomenon that Silva‐ Corvalán refers to as a lexico‐syntactic calque, whereby a word in the receiving language incorporates some (new) semantic components from the donor language in a word that already shares some semantic component (Silva‐Corvalán 2008: 218). Here and elsewhere she argues that this is only possible when there is some shared element to begin with, or possibly some ongoing change that is further promoted by analogy with the contact language, giving several examples of structures that have been transferred from English to Los Angeles Spanish not through direct borrowing of the structure, per se, but through a gradual loss of discourse‐pragmatic constraints, register constraints, and selection restrictions on the analogous structure in Spanish (2008: 220). A phonetic distinction between Taiwanese and Standard Mandarin provides a second example of an enduring effect of L1 transfer that has become a socio‐indexical marker. A highly salient feature of Taiwanese Mandarin (TM) is the de‐retroflection of fricative [ʂ] and affricates [tʂ] and [dʐ], also argued to be due to contact with Southern Min. Kubler cites this as a simple L2 acquisition effect, that the feature of retroflection is absent in Southern Min and thus difficult for L2 Mandarin speakers to acquire (1985: 157). Yet although this stigmatized feature is described as being a feature of older L2 speakers of Mandarin, it persists in the Mandarin spoken natively by young Taiwanese speakers. In a variationist analysis of [ʂ]‐retroflection, Baran (2014) demonstrates that phonetic variables that are associated with TM have become imbued with sociocultural meanings that may be used variably by young speakers as they negotiate conflicting identities and stances. In other words, what may once have been a transfer effect of language contact now demonstrates sociolinguistically meaningful variation for native speakers of Taiwanese Mandarin. Baran’s treatment of these linguistic variables in the context of the social lives of her speakers serves to demonstrate the value of a variationist approach to language‐contact effects, i.e. considering variation in linguistic structure situated in its social context. In particular, this type of approach allows researchers to focus on what Winford calls the “micro‐level” social factors of interpersonal relationships and interaction (2013: 367).

5  Four Dichotomies In this section we highlight and expand on a few notable cases of variation in the context of contact with respect to some key (and dichotomous) notions: internally‐ vs externally motivated changes; phonetically natural vs. unnatural changes; and socially “predictable” vs “unpredictable” effects in contact settings. In what follows we discuss each one of these, using examples from the literature and from our own work on Garifuna (an Arawak ­language of Belize), and Francoprovençal (a Romance language spoken in France, Italy, and Switzerland); both languages are endangered in at least some of the communities in which they are used (Bonner 2001 and Zulato et al. 2018 respectively). We hope that these illustrations also serve to highlight the contributions of the study of sociolinguistic variation to the study of language contact, based on the primary assumptions of variationist sociolinguistics: (i) that we can quantify variation in linguistic form and its relationship with the linguistic environment and social context; and (ii) that no single linguistic or social factor can fully explain linguistic variation: instead it is better to think of these factors as part of a larger constraints hierarchy.

5.1  Internal vs External The distinction between internal and external motivations for change arises from an old Boas‐Sapir controversy over the possibility of distinguishing between inherited similarities between two languages (i.e. due to genetic affiliation) and similarity due to language contact

228  Maya Ravindranath Abtahian and Jonathan Kasstan (Thomason and Kaufman 1988: 5). In the field, however, the distinction between what counts as external and what counts as internal is not always defined consistently, with some authors using the term “external” exclusively for those changes that arose through contact with another language variety, and others using “external” to refer to all socially‐motivated changes, with language contact as a sub‐type of these (e.g. Hickey 2012). Under the first definition, socially‐motivated variation (i.e. where variants are linked to different micro‐ or macro‐ level social categories, stances, or meanings) would not necessarily be considered externally motivated variation, as social motivation for competition between variants of one form may occur with or without language contact. We illustrate these distinctions further here with a description of variation in phonemic r in Garifuna. Although Aikhenvald (1999: 72) identifies Garifuna as one of the “few healthy Arawak languages,” noting that the overwhelming majority of the 40 living Arawak languages are now endangered, even Garifuna’s status is not stable, and there is widespread language shift in those parts of Central America where it is spoken. English and Belizean Creole (Kriol) both contact languages in all of the Garifuna‐speaking communities in Belize, and even in previously Garifuna‐dominant domains many speakers are shifting to use of English or Kriol. The phoneme inventory of Garifuna includes 17 consonants, with one lateral /l/ and one rhotic (Cayetano 1992; Taylor 1955). In the Belizean variety of Garifuna, the rhotic may exhibit two phonetic variants: an alveolar tap [ɾ] or an alveolar approximant [ɹ]. Contact seems a likely (if not confirmed) motivation for the variation between the tap and the approximant, as the varieties that are in contact with Spanish and English are not the same with respect to this feature. In Honduras, where Garifuna is in contact with Spanish (and where alveolar rhotics include a tap and a trill), the tap is the only reported variant (Haurholm‐Larsen 2016; Suazo 1991). In Belize, where Garifuna is primarily in contact with varieties of English and Kriol that use the approximant, the approximant is the more common variant (Ravindranath 2008, 2009), demonstrating a real‐time change from Taylor’s (1955) description of the rhotic as varying between an apical flap and a trill (based on fieldwork with older speakers in the late 1940s). In fact, the change from tap to approximant seems to be an ongoing change: of the 61 speakers interviewed for Ravindranath (2009) only six of them produced any tokens of /ɾ/, and all of these speakers were born before 1955. Speakers born after that time were categorical users of the approximant. Thus in addition to the aforementioned differences between Honduran and Belizean varieties and the real‐time comparison with Taylor (1955), an apparent‐time analysis of the behavior of the speech community ­demonstrates an ongoing change in progress toward use of the approximant that provides additional support for a contact‐motivated (i.e. external) explanation for the change. However, a second type of variation simultaneously affects the rhotic. In both Honduran and Belizean varieties of Garifuna the rhotic may be variably deleted, as in the variable deletion of r between vowels in a word like ‘see’ ariha ~ eiha. With respect to this variation, none of the traditional contact explanations would seem to apply, as this is not a feature of any of the contact languages and there is no particular reason to see r‐deletion intervocalically as a more unmarked form. Moreover, Ravindranath (2008, 2009) demonstrates that deletion of r also exhibits change in apparent time, with younger speakers and women leading the change. And social motivations cannot be ruled out as a factor in the variation, as all speakers exhibit more deletion in interviews than in the more formal task of telling a story from a picture book. More broadly, these observations are significant because they suggest that principles of linguistic change  –  in this case, a change from below (Labov 1990)  –  posited in monolingual Western societies can be applied in the types of speech communities that are less frequently discussed in the variationist literature (cf. Chirkova et al. 2018; Kasstan 2019). Moreover, this variation demonstrates what seems to be two distinct changes in progress, both affecting the same phonemic category and for speakers in the same speech community.

Contact and Sociolinguistic Variation  229 It serves as an illustration of the fact that a clear‐cut distinction between internal and external is not only impossible, but it also may be inaccurate to try to identify only one source of change in a particular phonemic category. As Dorian (1993: 132) points out in a rejection of the internal‐external distinction, “dichotomies have the effect of nudging us in the direction of an either/or discrimination,” and, similarly, Mougeon and Beniak (2002: 10) point out that “[what] linguists would label as internal or external according to their theoretical persuasion are in all probability both.” We find at least some support for this assertion in the data described here.

5.2  Natural vs Unnatural Most treatments of sound change, particularly from a historical perspective, present a classification of the types of sound changes most likely to be encountered in the world’s languages (Campbell 2013: 24), and a distinction between those changes that are considered natural, due to common articulatory and perceptual processes, and those that are considered unnatural, in that their results are counter to those same processes (e.g. Kiparsky 2003, but see also Hale 2003). Naturalness may refer to both the type of change (e.g. lenition, fortition, etc.) as well as the environment for the change (e.g. word‐finally, between vowels, etc.). Categories of sound changes that are considered natural include assimilation (especially between two adjacent sounds) and lenition (including deletion). The alternate processes of dissimilation and fortition, as well as more sporadic processes such as metathesis, are considered natural, but less common, while changes that do not fall into any of the categories would be considered unnatural. While the distinction cannot be made on the same articulatory and perceptual grounds for morphological, syntactic, or semantic change, similar attempts are made to identify common cross‐linguistic processes. Grammaticalization, for example, is thought to be a natural direction of change (Hopper and Traugott 1993); Traugott has further argued that semantic changes often demonstrate the same directionality (Traugott 1982). One question that is relevant to language contact with respect to natural changes is the role that language contact may play, for example, in promoting unnatural or unusual changes, or conversely, in promoting the emergence of unmarked categories. As we have outlined already, the latter seems to be more likely, and this is confirmed to some extent in the case that Palosaari and Campbell (2011) for example describe, of a merger of velar and uvular consonants in the Mam language of Tuxtla Chico in contact with Spanish, which lacks that distinction. However, the former is also attested. For instance, in an example of dialect contact, Larsen (1917, cited in Trudgill 1986) describes the emergence of a mid‐rounded variant in the vernacular of the inner Sogn area of Norway (see Table 11.2). This variant is described as “unusual, unexpected, and impossible to explain” when situated in its historical linguistic context (Trudgill 1986: 69). In fact, in both Sogn and Standard Norwegian, /ɔ/ is a common form in a large number of words, but not in the neighboring Halling dialect, which has /ø/ as the regular form. Thus, while the regional Norwegian Table 11.2  Neighbor opposition in Norwegian Sogn (adapted from Trudgill 1986: 69–70). Sogn

Halling

Standard Norwegian

English gloss

/bjɔrk/ /çɔt/ /smɔr/

/bjørk/ /çøt/ /smør/

/bjørk/ /çøt/ /smør/

‘birch’ ‘meat’ ‘butter’

230  Maya Ravindranath Abtahian and Jonathan Kasstan standard has focused on /ø/, Sogn has diverged from the neighboring variety (Halling) with /ɔ/, which is described as a “salient” marker of Sogn speech, and which has been extended to Sogn lexical items which historically had /ø/. As linguistic‐internal reasons cannot account for the dominance of /ɔ/ in Sogn, Larsen invokes the notion of nabo‐opposisjon (“neighbor opposition”), whereby speakers deliberately distantiate their speech in the face of a competing form for socio‐psychological reasons. This example also serves to illustrate the importance of language attitudes in language contact, which we turn to in section 5.3.

5.3  Predictable vs Unpredictable Effects It is an axiom that language change is itself unpredictable. As Thomason, among others, has already pointed out, “even the most ‘natural’ structural changes  –  common changes that occur frequently in diverse languages all over the world – often do not happen” (2001: 77). It is therefore not possible to predict with certainty that change will take place, even where variability is present, and above we have outlined some general tendencies in language contact that can promote change. However, there is one overriding factor that, even in the most favorable of environments, can impede or even reverse linguistic change, and that is the effect of language attitudes. As Table 11.1 outlines, Weinreich et al. understood the social embedding of linguistic variation to be connected to the social evaluation of variation, and it is for this reason that the embedding and evaluation problems they outlined have figured most prominently in variationist research. Indeed, Labov has gone as far as to suggest that “language change may simply reflect changes in interlocutor frequencies which are in turn the result of changes in social preferences and attitudes” (2001: 191). It is therefore paramount that language ­attitudes be made a central component of any research design investigating any possible linguistic change in progress. It is also important to highlight that language ­attitudes have been studied from the perspective of many diverse methodological and ­theoretical frameworks. In what follows, we focus specifically on the contribution of Communication Accommodation Theory (CAT) (e.g. Giles et  al. 1973), given that this is where the intersection of variationist methods and language‐contact research has been most fruitful. CAT, like other theories grounded in the social psychology of similarity attraction (e.g. Social Identity Theory, Tajfel 1978), broadly holds that a speaker (as a rational agent) can be more positively evaluated by a hearer if the speaker maneuvers to reduce dissimilarities between themselves and the hearer (i.e. minimizing social distance through linguistic accommodation). In the short‐term, then, accommodation can lead to linguistic convergence (in turn minimizing social distance) or linguistic divergence (increasing social distance), depending on the interactional goals of the speaker. Trudgill (1986) and Kerswill and Williams (2000) have shown in detailed studies on contact‐induced change how, in dialect contact (i.e. of two typologically similar varieties), accommodation is the central social mechanism underpinning broader linguistic processes of koinéization (for a discussion, see Trudgill 1986). These works make clear that contact leading to change can follow a principled path, and that, crucially, all other things being equal, it is in the second‐generation peer‐group networks where variable input is focused, and subsequently incremented.4 In focusing, it is typically the most frequent or unmarked variants (as we have seen already) that win out over rarer more marked forms performing the same work.5 These processes are mediated by CAT in that speakers accommodate (i.e. converge) on one another. In Kerswill and Williams’ (2000) quantitative study of new‐dialect formation in Milton Keynes, for instance, focusing was observed to be most prominent in second‐generation adolescent peer groups, where pressure to reduce social distance was most acute, which means eliminating linguistic differences distinguishing speakers. This was demonstrated for instance with the goat and goose

Contact and Sociolinguistic Variation  231 lexical sets, where the authors observe that focusing of the most fronted variants emerges in the adolescent 12‐year old females, who produce the most advanced forms. However, this is not always the case, and speakers can also diverge linguistically, under certain social conditions: where divergence takes place, linguistic hard boundaries are maintained or intensified (for a discussion, see Giles et al. 1978). Relatedly, Mufwene (2001: 4–6) describes cases of both dialect and language contact as producing feature pools, made up of variants contributed to by the input varieties. Under the “feature pool” model, speakers select from combinations of features, sometimes modifying them into new structures in the output varieties. Where the feature pool hypothesis differs from much of the earlier dialect contact work however is in the focusing of features of low‐frequency among the input varieties. As Mufwene (2001: 6) suggests, language contact can result in more complex competition in the feature pool, and in some cases, under focusing, a feature can emerge that was “statistically too insignificant to produce the same output under different ecological conditions.” While Trudgill (2008) has recently eschewed the role of social factors in the emergence of new varieties (a debate that questions the extent to which accommodation is an inherently automatic process, cf. Babel 2010), language attitudes undeniably play a pivotal role in mediating which features win out (see also Rodríguez‐Ordóñez 2019). The literature on dialect and language contact therefore indicates that language attitudes can either promote linguistic change, or serve as an overriding or inhibiting factor.

5.4  Contact‐induced Change in Stable vs Obsolescing Languages We have suggested on p. 224 using examples taken from Gal (1978, 1984) that cases of language shift (a macro‐level change) can precipitate micro‐level structural changes in the grammar of individuals and across the speech community. The label used to refer to these interrelated phenomena is “language obsolescence”: the progressive reduction and loss in the competence or in the use of a language, leading to language death (Dorian 1989: 2). It is important to consider language obsolescence as an area of inquiry here because it provides a unique opportunity to understand how linguistic change can progress with little time depth, by comparison with change observed in comparatively “healthy” languages. This occurs because the prototypical case of language obsolescence, often termed “gradual language death” (Campbell and Muntzel 1989: 182), is characterized by an unstable diglossic context, that is, the contraction in domains of usage available to the obsolescing variety, which leads in most cases to periods of bilingualism, and thus a decrease in linguistic competence in the threatened variety. A corollary of this, identified in most cases of language obsolescence that we are aware of, is the emergence in a speech community of a “proficiency continuum of speakers” (Dorian 1981: 114). Dorian distinguishes principally between three types of speaker along her continuum: older fluent speakers, younger fluent speakers, and semi‐speakers (1977: 23–32, 1981: 114–7), though other typologies exist that have built on Dorian’s work (e.g. Sasse 1992: 61; Grinevald Craig 1997:259–60). As linguistic attrition (or “decay”) is often characteristic of semi‐ speakers, they have been referred to as “potential harbingers of language death” (Jaffe 2015: 23). However, we contend that a greater focus should be placed on the linguistic production of such speakers given that the apparent‐time construct (see Section  1) is underpinned by the assumption that a speaker’s grammar stabilizes in adulthood (this postulation is crucial in modeling communal change). Contexts most clearly associated with language obsolescence therefore challenge “the belief in the automatic and complete competence of ‘native speakers’ in their ‘native language’” (Doerr 2009: 36). Accordingly, we agree with Dorian that “semi‐speakers […] merit special study for the light they may shed on our understanding of the human language facility in general, to the benefit of linguistic description and theory” (1981: 156).6

232  Maya Ravindranath Abtahian and Jonathan Kasstan The types of linguistic changes that have been documented in language obsolescence are by no means unique to threatened languages; indeed, many of them are commonly observed in healthy varieties, too. For instance, many studies dedicated to the linguistic consequences of language obsolescence have tended to focus on grammatical simplification resulting from the disuse of the threatened variety (Mougeon and Beniak 1989: 299). However, ­simplification, as we have seen, is a very common process in contact‐induced change. For example, Jones and Singh (2005: 88) compare moribund East Sutherland Gaelic and Norwegian, highlighting that commentators have identified a loss of case differentiation in both languages. However, as they also point out, it is the amount and the rate of change that characterizes a context of language obsolescence, given the precarious issues associated with the threatened variety. Other common types of contact‐induced change cited in Jones and Singh’s (2005: 87–92) typology include: reduction (or loss) of some component of the grammar (such as the loss of grammatical gender in the Elmolo third person paradigm); a change from more synthetic to more analytical morphology; and a generalization of unmarked categories where historically marked features are normatively required. Using a case study on Francoprovençal, below, we briefly build on this typology with the addition of two further changes that we see as being typically contact‐induced in language obsolescence, but which are accorded comparatively much less attention: (i) the overgeneralization of salient (even stereotyped) forms with strong positive valorization, and (ii) range in the socio‐stylistic continuum. Francoprovençal (also known as “Arpitan” or just “patois”) is a severely endangered and fragmented Romance language spoken transnationally across French, Italian, and Swiss borders (see Zulato et al. 2018). Speakers of Francoprovençal have long been dwindling as language shift has progressively been taking place: the language is no longer transmitted intergenerationally in France, where French has long been entrenched in the critical domain (the home), and the decline is well advanced in Switzerland, too. In terms of its phonological inventory, Francoprovençal shows many similarities with its closest dominant contact language, French (see Kasstan 2015a, 2015b), but there are important allophonic differences that form the basis of this discussion. To illustrate processes (i) and (ii), we focus on one phonological feature common in Francoprovençal (and other Romance languages), that does not exist in Modern French: the palatalization of lateral approximants in obstruent + lateral onset clusters /kl, gl, fl, pl, bl/. In the Francoprovençal dialects spoken in France, historical atlas recordings illustrate that this categorical allophonic contrast has been maintained for the monts du Lyonnais region over half a century, whereby /kl, gl/ palatalize to [kj] and [gj], but the labial sets do not palatalize (cf. Gilliéron and Edmont 1902–10; Gardette 1950–6). However, Kasstan and Müller (2018) and Kasstan (2019) illustrate using variationist techniques that /l/‐palatalization has not only become variable, but is progressively being leveled among the oldest and most fluent speakers in that they are able to maintain the allophonic contrast exclusively with velar consonants (they thus maintain internal‐linguistic constraints operating on the variable), but they only clearly maintain the contrast in the most formal speech styles; the contrast is all but lost in much more informal conversational styles (see Figure 11.1, p < .001). Much like Gal (1984), we therefore observe a loss of interactional (i.e. social) constraints. Therefore, as more attention is paid to speech, speakers orient to and begin to over‐ generalize the “most typical” or “salient” Francoprovençal‐like target, whereas in casual speech (in this case group interviews directed by the speakers), the lateral approximant tends to categoriality, given that all speakers are now dominant in their L2, French, which does not have this feature. This evidence problematizes the traditional view of linearly ordered style variation (as conceived by Labov 1966) and the primacy of the first‐learned variety: those speakers who acquired Francoprovençal as an L1 only produce the most vernacular variants when most attention is paid to speech. However, contra the traditional literature in language obsolescence (e.g. Dressler 1972 :454; Dressler and Wodak‐Leodolter 1977: 36–7),

Contact and Sociolinguistic Variation  233

***

80

Count n = 152

60

Median approximant

40

Lateral approximant

20

0 Casual

Speech style

Wordlist

Figure 11.1  Distribution of palatal and non‐palatal variants according to speech style in a sample of 16 native speakers of Francoprovençal (taken from Kasstan 2019: 18).

the data still show that speakers are able to style‐shift according to their interactional needs in spite of the heavy contraction of usage domains. Monostylism is therefore not a foregone conclusion in gradual language death, as is commonly reported (for an extended discussion, see Kasstan 2019). Instead, natural languages can still be spoken in ways appropriate to speakers’ various stylistic needs, in spite of very limited access to occasions and contexts with which to practice (Dorian 1994: 326).

6  Future Trajectories 6.1  Sociolinguistically Informed Language Documentation To make a distinction between variation and change in vital vs. obsolescing languages requires thinking both about potential structural differences between vital and obsolescing languages as well as differences in the social contexts of their use. Because language obsolescence occurs entirely in the context of multilingual communities, this distinction is highly relevant to the topic of sociolinguistic variation and language contact, as is the topic of language endangerment more generally. Indeed, highly multilingual communities are fertile ground for the intersection of sociolinguistic variation and language contact. Moreover, as attention to the issue of language endangerment to the field of language description and documentation (LDD) has grown, the issue of documenting sociolinguistic variation in under‐studied, multilingual speech communities is a burgeoning area of inquiry (Hildebrandt, Jany, and Silva 2017; Meyerhoff 2017).

234  Maya Ravindranath Abtahian and Jonathan Kasstan LDD emerged as a growing subfield of linguistics following increased awareness of the loss of languages worldwide (e.g. in Hale 1992) and the perceived importance of recording languages and language practices while living speakers remain. In large part this enterprise was primarily focused on severely endangered languages with few living speakers, with the necessary outcome of working with few fluent consultants to produce dictionaries, texts, and grammars, and the perhaps unintended consequence of prioritizing the speech of a few fluent consultants (much like the NORMs of early dialectology, cf. Chambers and Trudgill 1998). Himmelman later argued for a distinction between language description, and describing “the linguistic practices and traditions of a speech community,” i.e. language documentation (1998: 166). Others have advocated for language documentation that is ethnographically grounded. Harrison for example argues that “there can be no adequate description of grammar in isolation from the cultural knowledge that underlies it” (2005: 28), giving the example of directional verbs in Tuvan, in which knowledge of which verb to use depends upon knowledge of the physical and cultural landscape of the area of Siberia. By building on the observation that language documentation should involve the documentation of linguistic practices and behaviors as well as linguistic description, more recent work in LDD has recognized the importance of documenting variation in un(der)‐ documented and minority languages. Thus, one future trajectory that we see situated at the intersection of language contact and sociolinguistic variation is in sociolinguistically informed language documentation. That is, within the domain of the increased focus on language documentation (including better documentation of cultural practices and the contexts of language use), there is also room for better description of linguistic variation in that context, including analyses of synchronic sociolinguistic variation as well as variationist analyses of the types of language‐contact outcomes we have discussed here (see Childs, Good, and Mitchell 2014; Hildebrandt, Jany, and Wilson 2017; Meyerhoff 2017; Nagy 2009). Sociolinguists have methods for eliciting, describing, and theorizing linguistic variation, and these can complement existing methods in language documentation. At the same time, as Meyerhoff (2017) points out, documentary linguistics has the advantage of a whole‐language perspective that allows for the consideration of mutual dependence or independence of variables in a single linguistic system, what Meyerhoff terms a “linguistic symphony,” in favor of the “linguistic sonatas” of previous variationist work. Finally, Meyerhoff also notes that the selection of variables in un(der)described varieties, as it happens alongside documentation, may “throw up insights that would not be predicted by existing theory” (Labov 2015 in Meyerhoff 2017: 545), which relates to the second trajectory that we expect will bear fruit in the intersection of language contact and sociolinguistics.

6.2 Principles of Linguistic Change Beyond Western Urban Communities Loh and Harmon (2014: 23) suggest that 0.1% of the world’s population (or about 8 million people, about the size of London) are responsible for keeping one half of the world’s estimated 7,000 languages alive. It is therefore surprising that, leaving aside some seminal work in the variationist paradigm that has focused on lesser‐studied languages (including Eckert 1980 on Gascon, Sankoff 1980 on multilingualism in Papua New Guinea, and Rickford 1979 on Guyanese Creole), and a handful of more recent studies such as those mentioned here (among others, e.g. Chirkova et al. 2018 on Ganluo Ersu), variationist studies continue to focus on dominant Western urban (and predominantly monolingual) societies (see Nagy and Meyerhoff 2008; Stanford 2016; Hall et al. 2019, as per p.221 for discussion).

Contact and Sociolinguistic Variation  235 As we have written in the introduction to this chapter, we may still ask the question of how the principles of language variation and change apply in highly multilingual communities, or in communities where a clear standard language is not prescribed. We see this as another future trajectory that lies at the intersection of contact linguistics and variationist sociolinguistics, as most of these are communities where multilingualism and language contact are the norm. With respect to the dichotomies that we discuss in this chapter, and the unanswered questions that remain with respect to the nature of language variation in the context of contact, we find that the answers are likely to be found in under‐studied language communities that we have attempted to highlight in this chapter.

7 Conclusion In this chapter we have focused on research at the intersection of the fields of language variation and change (or “variationist sociolinguistics”), and language contact. We have discussed a number of distinctions that have been made with respect to different types of changes in the context of language contact, including the internal and external dichotomy, the categorization of changes as natural or unnatural, and predictable or unpredictable changes, as well as the characterization of languages as stable or obsolescing, with concurrent assumptions about their patterns of variation and change. In each of these cases we argue that the study of language variation and change has something to add to these discussions, with its focus on linguistic structure and social ecology, and emphasis on the fact that both social and linguistic factors may be implicated in the outcomes of language contact in a more continuous and nuanced manner than is traditionally assumed in the literature.

NOTES 1 Weinreich (1968) adopts the label “language contact” to imply bilingualism, and he reserves “interference” for the effects of language contact on linguistic structure, i.e. the influence of one linguistic system on another. We make use of slightly different terminology here, per the variationist tradition: principally, “contact induced change,” “language change” will be used in the sense adopted by Weinreich for “interference.” 2 The linguistic variable as originally conceived hinges on a change in linguistic form, but not a change in referential meaning (i.e. “the variants are identical in reference or truth value, but opposed in their social and/or stylistic significance,” Labov 1972: 271). This constraint is most clearly reflected at the phonological or phonetic level (e.g. an alternation between [t] and [ʔ]), but much less so at the level of e.g. syntax and, say, an alternation between some active and passive constructions (for a discussion, see Lavandera 1978). 3 It should be pointed out that the generalizations made by Thomason (2001) have not always been supported by empirical evidence (cf. e.g. Doğruöz and Backus 2009 on Heritage Turkish morphosyntax). 4 Incrementation in variationist sociolinguistics refers to the mechanism by which linguistic changes advance in a step‐by‐step fashion generationally through a speech community (Labov 2001: 446) 5 Other factors have been cited however, including regularity of paradigms, transparency (form‐ function symmetry), and salience. For a fuller discussion, see Siegel (1997). 6 It is important to highlight that any linguistic consequences of such contexts may owe nothing to the influence of the dominant contact variety, but, as Mougeon and Beniak point out, instead may “constitute autonomous developments triggered by system‐internal tendencies to restructure, the result most often being structural simplification” (2002: 2).

236  Maya Ravindranath Abtahian and Jonathan Kasstan

REFERENCES Aikhenvald, Alexandra Y. 1999. The Arawak language family. In Robert M. W. Dixon and Alexandra Aikhenvald (eds.), The Amazonian Languages. Cambridge: Cambridge University Press, pp. 65–106. Babel, Molly 2010. Dialect divergence and convergence in New Zealand English. Language in Society, 39.5: 437–56. Bailey, Charles‐James N. 1973. Variation and Linguistic Theory. Arlington, VA: Center for Applied Linguistics. Baran, Dominika 2014. Linguistic practice and identity work: variation in Taiwan Mandarin at a Taipei County high school. Journal of Sociolinguistics 18.1: 32–59. Bonner, Donna 2001. Garifuna children’s language shame: ethnic stereotypes, national affiliation and transnational immigration as factors in language choice in southern Belize. Language in Society 30.1: 81–96. Campbell, Lyle 2013. Historical Linguistics: An Introduction, 3rd edition. Cambridge, MA: MIT Press. Campbell, Lyle and Martha C. Muntzel 1989. The structural consequences of language death. In Nancy C. Dorian (ed.), Investigating Obsolescence. Cambridge: Cambridge University Press, pp. 189–96. Cayetano, Roy 1992. Towards a common Garifuna orthography. Dangriga: National Garifuna Council of Belize. Available at www. ngcbelize.org/the‐culture/language/garifuna‐ orthography/ Chambers, Jack and Peter Trudgill 1998. Dialectology. Cambridge: Cambridge University Press. Cheshire, Jenny, David Adger, and Sue Fox 2013. Relative who and the actuation problem. Lingua 126: 51–77. Childs, Tucker, Jeff Good, and Alice Mitchell 2014. Beyond the ancestral code: towards and model for sociolinguistic language documentation. Language Documentation and Conservation 8: 168–91. Chirkova, Katia, James N. Stanford, and Dehe Wang 2018. A long way from New York City: socially stratified contact‐induced phonological convergence in Ganluo Ersu (Sichuan China). Language Variation and Change 30.1: 109–45. Cornips, Leonie and Karen P. Corrigan 2005. Syntax and Variation: Reconciling the Biological and the Social. Amsterdam: Benjamins.

Crowley, Terry 2002. Serial Verbs in the Oceanic: A Descriptive Typology. Oxford: Oxford University Press. Doerr, Neriko Musha 2009. Investigating native speaker effects: towards a new model of analyzing native speaker ideologies. In Neriko Doerr (ed.), The Native Speaker Concept. Berlin: de Gruyter Mouton, pp. 15–43. Doğruöz, A. Seza and Ad Backus. 2009. Innovative constructions in Dutch Turkish: an assessment of ongoing contact‐induced change. Bilingualism: Language and Cognition 12.1: 41–63. Dorian, Nancy C. 1977. The problem of the semi‐speaker in language death. International Journal of the Sociology of Language 12: 23–32. Dorian, Nancy C. 1981. Language Death: The Life Cycle of a Scottish Gaelic Dialect. Philadelphia: University of Pennsylvania Press. Dorian, Nancy C. (ed.) 1989. Investigating Obsolescence. Cambridge: Cambridge University Press. Dorian, Nancy C. 1993. Internally and externally motivated change in language contact settings: doubts about the dichotomy. In Charles Jones (ed.), Historical Linguistics: Problems and Perspectives. London: Longman, pp. 131–55. Dorian, Nancy C. 1994. Stylistic variation in a language restricted to private‐sphere use. In Douglas Biber and Edward Finegan (eds.), Sociolinguistic Perspectives on Register. Oxford: Oxford University Press, pp. 217–32. Dressler, Wolfgang U. 1972. On the phonology of language death. Chicago Linguistic Society Papers 8: 448–57. Dressler, Wolfgang U. and Ruth Wodak‐ Leodolter 1977. Language preservation and language death in Brittany. International Journal of the Sociology of Language 12: 33–44. Dubois, Sylvie and Barbara M. Horvath 1998. Let’s tink about dat: interdental fricatives in Cajun English. Language Variation and Change 10.3: 245–61. Eckert, Penelope 1980. The structure of a long‐ term phonological process: the back vowel chain shift in Soulatan Gascon. In William Labov (ed.), Locating Language in Time and Space. New York: Academic Press, pp. 179–219. Fischer, John 1958. Social influences on the choice of a linguistic variant. Word 14.1: 47–56.

Contact and Sociolinguistic Variation  237 Fought, Carmen 1999. A majority sound change in a minority community: /u/‐fronting in Chicano English. Journal of Sociolinguistics 3.1: 5–23. Gal, Susan 1978. Peasant men can’t get wives: language change and sex roles in a bilingual community. Language in Society 7: 1–16. Gal, Susan 1984. Phonological style in bilingualism. In Deborah Schiffrin (ed.), Meaning, Form and Use in Context: Linguistic Applications. Washington DC: Georgetown University Press, pp. 290–302. Gardette, Pierre 1950–6. Atlas linguistique et ethnographique du Lyonnais. Paris: Éditions du CNRS. Giles, Howard, Donald M. Taylor, and Richard Bourhis.1973. Towards a theory of interpersonal accommodation through language: some Canadian data. Language in Society 2.2: 177–92. Gilliéron, Jules and Edmond Edmont 1902–10. Atlas linguistique de la France. Paris: Honoré Champion. Grinevald Craig, Colette 1997. Language contact and language degeneration. In Florian Coulmas (ed.), The Handbook of Sociolinguistics. Oxford: Blackwell, pp. 257–70. Hale, Mark 2003. Neogrammarian sound change. In Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics. Oxford: Blackwell Publishing, pp. 343–68. Hale, Ken, Michael Krauss, Lucille J. Watahomigie, Akira Y. Yamamoto, Colette Craig, LaVerne Masayesva Jeanne, Nora C. England. 1992. Endangered languages. Language 68.1: 1–42. Hall, Damien, Jonathan R. Kasstan, and David Hornsby. 2019. Beyond obsolescence: a twenty‐first century research agenda for the langues régionales. Journal of French Language Studies 29.2: 155–68. Harrison, K. David 2005. Ethnographically informed language documentation. In Peter K. Austin (ed.), Language Documentation and Description, vol. 3. London: SOAS, pp. 22–41. Haurholm‐Larsen, Steffen 2016. A Ggrammar of Garifuna (University of Bern Ph.D. dissertation). Hickey, Raymond 2010. Contact and language shift. In Raymond Hickey (ed.), Handbook of Language Contact. Wiley‐Blackwell, pp. 151–69. Hickey, Raymond 2012. Internally‐ and externally‐motivated language change. In Juan Manuel Hernández‐Campoy and Juan Camilo Conde‐Silvestre (eds.), The Handbook of Historical Sociolinguistics. Oxford: Blackwell, pp. 388–407.

Hildebrandt, Kristine A., Carmen Jany, and Wilson Silva 2017. Introduction: documenting variation in endangered languages. Language Documentation and Conservation, Vol. 13: 1–5 Himmelman, Nikolaus P. 1998. Documentary and descriptive linguistics. Linguistics 36: 161–95. Hopper, Paul J. and Elizabeth C. Traugott 1993. Grammaticalization. Cambridge: Cambridge University Press. Jaffe, Alexandra 2015. Defining the new speaker: theoretical perspectives and learner trajectories. International Journal of the Sociology of Language 231: 21–44. Jones, Mari C. and Ishtla Singh 2005. Exploring Language Change. Abingdon: Routledge. Kachru, Braj B. 1985. Standards, codification and sociolinguistic realism: the English language in the outer circle. In Randolph Quirk and Henry G. Widdowson (eds.), English in the World: Teaching and Learning the Language and Literatures. Cambridge: Cambridge University Press, pp. 11–30. Kasstan, Jonathan R. 2015a. Variation and change in Francoprovençal: a study of an emerging linguistic norm (University of Kent Ph.D. dissertation). Kasstan, Jonathan R. 2015b. Illustrations of the IPA: Lyonnais (Francoprovençal), Journal of the International Phonetic Association 45.3: 340–55. Kasstan, Jonathan R. 2019. Emergent sociolinguistic variation in severe language endangerment. Language in Society. FirstView. Kasstan, Jonathan R. and Daniela Müller 2018. (l) as a sociolinguistic variable in Francoprovençal. International Journal of the Sociology of Language 249: 99–118. Kerswill, Paul and Ann Williams 2000. Creating a new town koine: Children and language change in Milton Keynes. Language in Society, 29.1: 65–115. Kiesling, Scott 2011. Linguistic Variation and Change. Edinburgh: Edinburgh University Press. Kiparsky, Paul 2003. The phonological basis of sound change. In Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historial Linguistics. Oxford: Blackwell Publishing, pp. 313–42. Kubler, Cornelius C. 1985. The influence of Southern Min on the Mandarin of Taiwan. Anthropological Linguistics 27.2: 156–76. Labov, William 1963. The social motivations of sound change. Word 19: 274–309. Labov, William 1966. The Social Stratification of English in New York City. Washington. DC: Center for Applied Linguistics.

238  Maya Ravindranath Abtahian and Jonathan Kasstan Labov, William 1972. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William 1978. On the use of the present to explain the past. In Philip Baldi and Ronald Werth (eds.), Readings in Historical Phonology. Pennsylvania: Pennsylvania State University Press, pp. 275–312. Labov, William 1990. The intersection of sex and social class in the course of linguistic change. Language Variation and Change 2.2: 205–54. Labov, William 1994. Principles of Linguistic Change, vol. I: Internal Factors. Oxford: Wiley‐Blackwell. Labov, William 2001. Principles of Linguistic Change, vol. II: Social Factors. Oxford: Wiley‐Blackwell. Labov, William 2010. Principles of Linguistic Change, vol. III: Cognitive and Cultural Factors. Oxford: Wiley‐Blackwell. Labov, William 2015. The discovery of the unexpected. Asia‐Pacific Language Variation 1.1: 7–22 Lavandera, Beatriz R. 1978. Where does the sociolinguistic variable stop? Language in Society 7.2: 171–82. Lee, Nala H. and John R. Van Way 2018. Assessing degress of language endangerment. In Kenneth L. Rehg and Lyle Campbell (eds.), The Oxford Handbook of Endangered Languages. Oxford: Oxford University Press, pp. 48–65. Loh, Jonathan and David Harmon. 2014. Biocultural Diversity: threatened species, endangered languages. Zeist: WWF Netherlands. Odlin, Terrence 1997. Hiberno‐English: pidgin, creole, or neither? Center for Communication and Language Studies Occasional Paper No. 49. Available at https://eric.ed. gov/?id=ED412742. Meyerhoff, Miriam 2003. Formal and cultural constraints on optional objects in Bislama. Language Variation and Change 14.3: 323–46. Meyerhoff, Miriam 2009. Animacy in Bislama? Using quantitative methods to evaluate transfer of a substrate feature. In James Stanford and Dennis Preston (eds.), Variation in Indigenous Minority Languages. Amsterdam: Benjamins, pp. 369–96. Meyerhoff, Miriam 2017. Writing a linguistic symphony: analyzing variation while doing language documentation. Canadian Journal of Linguistics 62.4: 525–49. Meyerhoff, Miriam 2019. Unnatural bedfellows? The sociolinguistic analysis of variation and language documentation, Journal of the Royal Society of New Zealand 49.2: 229–41.

Mougeon, Raymond and Édouard Beniak 1989. Language contraction and linguistic change: the case of Welland French. In Nancy C. Dorian (ed.), Investigating Obsolescence: Studies in Language Contraction and Death. Cambridge: Cambridge University Press, pp. 287–312. Mougeon, Raymond and Édouard Beniak 2002. Linguistic Consequences of Language Contact and Restriction: The Case of French in Ontario, Canada, 2nd edition. Oxford: Oxford University Press. Mufwene, Salikoko 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Nagy, Naomi 2009. The challenges of less commonly studied languages: writing a sociogrammar of Faetar. In James Stanford and Dennis Preston (eds.), Variation in Indigenous Minority Languages. Amsterdam: Benjamins, pp. 397–417. Nagy, Naomi and Miriam Meyerhoff 2008. The social lives of language. In Miriam Meyerhoff and Naomi Nagy (eds.), Social Lives of Language – Sociolinguistics and Multilingual Speech Communities. Amsterdam: Benjamins, pp. 1–17. Palosaari, Naomi and Lyle Campbell 2011. Structural aspects of language endangerment. In Peter K. Austin and Julia Sallabank (eds.), The Cambridge Handbook of Endangered Languages. Cambridge: Cambridge University Press, pp. 100–19. Poplack, Shana and Stephen Levey 2010. Contact‐induced grammatical change: a cautionary tale. In Peter Auer and Jürgen Erich Schmidt (eds.), Language and Space: An International Handbook of Linguistic Variation. Vol. 1. Theories and Methods. Berlin: de Gruyter Mouton, pp. 391–419. Poplack, Shana and Sali Tagliamonte 2001. African American English in the Diaspora. Oxford: Blackwell. Poplack, Shana, Lauren Zentz, and Nathalie Dion 2012. What counts as (contact‐induced) change. Bilingualism: Language and Cognition 15.2: 247–54. Ravindranath, Maya 2008. The effect of language shift on a sound change in progress. University of Pennsylvania Working Papers in Linguistics 14.2: 17. Available at: https://repository. upenn.edu/cgi/viewcontent.cgi?article=1064 &context=pwpl Ravindranath, Maya 2009. Language shift and the speech community. Sociolinguistic change in a Garifuna community in Belize. University of (Pennsylvania Ph.D. dissertation).

Contact and Sociolinguistic Variation  239 Ravindranath, Maya 2015. Sociolinguistic variation and language contact. Lang & Ling Compass 9.6: 243–55. Rickford, John R. 1979. Variation in a Creole continuum: quantitative and implicational approaches (University of Pennsylvania Ph.D. dissertation). Rodríguez‐Ordóñez, Itxaso 2019. The role of linguistic ideologies in language contact situations. Lang & Ling Compass 11.8: e12351. Sankoff, Gillian 1980. The Social Life of Language. Philadelphia: University of Pennsylvania Press. Sankoff, Gillian 2001. Linguistic outcomes of language contact. In Peter Trudgill, J. K. Chambers, and Natalie Schilling‐Estes (eds.), Handbook of Sociolinguistics. Oxford: Basil Blackwell, pp. 638–68. Santa Ana, Otto 1996. Sonority and syllable structure in Chicano English. Language Variation and Change 8: 63–89. Sasse, Hans‐Jürgen 1992. Language decay and contact‐induced change: similarities and differences. In Matthias Brenzinger (ed.), Language Death: Factual and Theoretical Explorations with Special Reference to East Africa. Berlin: de Gruyter Mouton, pp. 59–80. Siegel, Jeff. 1997. Mixing, leveling and pidgin/ creole development. In A. K. Spears and Donald Winford (eds.), The Structure and Status of Pidgins and Creoles. Amsterdam: John Benjamins, pp. 111–49. Silva‐Corvalán, Carmen. 2008. The limits of convergence in language contact. Journal of Language Contact. (Thema 2), pp. 213–24. Stanford, James N. 2008. Child dialect acquisition: new perspectives on parent/peer influence. Journal of Sociolinguistics 12.5: 567–96. Stanford, James N. 2016. A call for more diverse sources of data: variationist approaches in non‐English contexts. Journal of Sociolinguistics 20.4: 525–41. Stanford, James N. and Dennis Preston 2009. The lure of a distant horizon: variation in indigenous minority languages. In James Stanford and Dennis Preston (eds.), Variation in Indigenous Minority Languages. Philadelphia: Benjamins, pp. 1–22. Suazo, Salvador 1991. Conversamos en Garifuna: grámatica y manual de conversación. Tegucigalpa, Honduras: COPRODIEM.

Tagliamonte, Sali 2012. Variationist Sociolinguistics: Change, Observation, Interpretation. Oxford: Wiley‐Blackwell. Tajfel, Henri 1978. Interindividual behaviour and intergroup behaviour. In Henri Tajfel (ed.), Differentiation between Social Groups: Studies in the Social Psychology of Intergroup Relations. London: Academic Press, pp. 289–323. Taylor, Douglas M. 1955. Phonemes of the Hopkins (British Honduras) dialect of Island Carib. International Journal of American Linguistics 21.3: 233–41. Thomason, Sarah G. 2001. Language Contact. Edinburgh: Edinburgh University Press. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Traugott, Elizabeth. 1982. From propositional to textual and expressive meanings: some semantic‐pragmatic aspects of grammaticalization. In Winfred P. Lehmann, and Yakov Malkiel (eds.), Perspectives on Historical Linguistics. Amsterdam: Benjamins, pp. 245–71. Trudgill, Peter 1986. Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 2008. Colonial dialect contact in the history of European languages: on the irrelevance of identity to new‐dialect formation. Language in Society 37.2: 241–80. Weinreich, Uriel 1953 [1964]. Language Contact. Findings and Problems. The Hague: de Gruyter Mouton. Weinreich, Uriel, Labov, William and Marvin Herzog 1968. Empirical foundations for a theory of language change. In Winfried P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics. Austin: University of Texas Press, pp. 95–195. Winford, Donald 2003. An Introduction to Contact Linguistics. Oxford: Blackwell. Winford, Donald 2013. Social factors in contact languages. In Peter Bakker and Yaron Matras (eds.), Contact Languages: A Comprehensive Guide. Berlin: Walter de Gruyter, pp. 363–416. Zulato, Alessia, Jonathan R. Kasstan, and Naomi Nagy 2018. An overview of Francoprovençal vitality in Europe and North America. International Journal of the Sociology of Language 249: 11–29.

12 Contact and New Varieties PAUL KERSWILL 1  Introduction: “New” Varieties The term “new variety” implies the convergence, by a population of speakers, on a set of linguistic norms which are different from previous norms. There are two epistemological issues here. The first is defining what is meant by a “population.” Trivially, this can refer to some set of people who, in our case, stop speaking Variety A and start speaking Variety B. More usefully, the term “population” can be applied in the sociolinguistic sense of “speech community”  –  individuals having some affinity with others through sharing linguistic norms, both in terms of linguistic structure and in terms of patterns of variation and ­evaluation. This essentially Labovian view (Labov 1989: 2; Patrick 2002: 584–8) places the focus on collective linguistic behavior, with subjective ideas about identity and community subordinated to that behavior (Coupland 2010: 101). Linguistic change, then, constitutes change in this collective behavior. The second epistemological issue is how to set criteria for a “new” variety  –  just how much change is needed for a variety to qualify for this label? In speech communities where there are low‐to‐moderate rates of in‐ and out‐migration, language change is gradual and the concept of a “new” variety is not relevant. In terms of the Labovian model, local speech norms are passed on by inter‐generational transmission, with incremental changes in dialect features being introduced by children and adolescents (D’Arcy 2015). The formation of a new variety (which may be a language or a dialect) involves much more than just changes in norms. We need to allow for a preceding period of relative absence of norms, or diffusion, followed by focusing (Le Page and Tabouret‐Keller 1985). “Focusing” refers to the reduction in the number of variant forms and the increase in sociolinguistically predictable variation; this process represents the (re‐)emergence of norms. Importantly, new varieties lack the inherent inter‐generational continuity of gradually changing speech communities (Kerswill 2002: 695–8). A new variety is likely to emerge in two types of scenario, both involving migration. The first involves the large‐scale migration of newcomers to an existing community, while the second sees the migration of individuals from different speech communities to a new location where they, as (voluntary or involuntary) settlers, form an entirely new community. In the former scenario, an existing dialect or dialects may be “swamped” by the dialects or languages of the incomers (Lass 1990, 2004: 367–8), or at least heavily modified. The latter scenario may give rise to rapid creolization in cases where migrants’ languages are not mutually intelligible, where no language is shared by a large enough minority, and where an economically or politically powerful language is a remote target (see Sippola and Cardoso, this volume). Under either scenario, if the varieties in contact are mutually intelligible,

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

242  Paul Kerswill koineization may occur; this leads to the formation of a new dialect (Trudgill 1986, 2004). New varieties may also be formed where contact is less obvious: this is often the case for language shift, where a population adopts a different language, in the process forming a new variety of that language marked by substrate influence (see Durkin, this volume). In  this chapter, I will deal with dialect leveling (Section  3), new‐dialect formation in tabula rasa territories (Section 4), koineization in new towns (Section 5), and ethnolects and multiethnolects (Section 6).

2 Accommodation To understand the dynamics of face‐to‐face dialect and language contact, we should explore the social psychological mechanism which underlies it: accommodation. New varieties emerge, initially, through countless individual acts of linguistic accommodation (adjustments) performed by speakers of every age interacting with others, the adjustments being variously conscious (strategic, in a manner modeled by Communication Accommodation Theory – Giles 1973) and unconscious (through interactive alignment – Pickering and Garrod 2004, cited in Tuten 2008; Wachsmuth, de Ruiter, Jaecks, and Kopp 2013). Accommodation may be responsive to context, or it may be long‐term (Trudgill 1986), leading to what Labov (2001: 415–17) calls “vernacular reorganization.” Accommodation does not, however, entail convergence to an interlocutor’s production of a particular linguistic item. Accommodatory behavior also reflects the speaker’s intentions and motivations, as well as what they believe they are doing linguistically; in either case, the speaker may well end up diverging, not converging. This is most likely to occur when there is a power asymmetry (Gallois, Ogay, and Giles 2005: 126–7). Accommodation takes place in a wider, but community‐specific social context. For our purposes, broad demographic and social factors are crucial to the outcomes of any kind of dialect (or language) contact. The following must count among the principal non‐linguistic factors: 1. The proportion of incomers to existing inhabitants. This calculation allows us to include both entirely new settlements and locations with existing populations. 2. The proportions of children to adults in the initial stage of the contact (Kerswill and Williams 2000: 90). 3. Relations between social groups (e.g., class, ethnicity, gender), and effects on opportunities for interaction. 4. Wider linguistic ideologies (Milroy 1999). 5. Personal and social identity formation. Against this backdrop, much of the outline of the new variety is determined by the relative frequency of linguistic forms heard among the population in the early stages, some of those forms being the outcome of accommodation, others belonging to a speaker’s quasi‐permanent vernacular. This deterministic mechanism, proposed by Trudgill, Gordon, Lewis, and Maclagan (2000), will be central to our discussion. The outcome is modified by the social and demographic factors just mentioned, some early in the koineization process, some later. In particular, if in the early stages of contact social divisions are relatively absent (as we shall see), then accommodation, leading to convergence, will prevail, especially among the crucial first‐child generation (cf. Kerswill and Williams 2000, and following sections). When social divisions emerge, whether they are new or reinforcements of divisions inherited from the place of origin, then dissociation – the opposite of accommodation – may additionally occur (cf. Hickey 2000, 2005).

Contact and New Varieties  243

3  (Dialect) Leveling Contact between varieties typically leads to leveling, which Britain (2018: 149) defines as: … the eradication of marked linguistic features, marked (a) in the sense of being in a minority in the ambient feature pool after the contact “event,” (b) in the sense of being over(t)ly stereotyped, or (c) marked in the sense of being found rarely in the world’s languages and/or learned late in child language acquisition.

Leveling is a central process in dialect contact. The term is often used in the collocation “dialect leveling,” which refers to leveling at the regional level; here, large‐scale, long‐term migration is not central; instead, routine contacts between people in neighboring towns and villages, for example through commuting, are the driving forces (see Britain 2013, 2018). This kind of regional dialect leveling (Kerswill 2003), or supralocalization (Milroy 2002), is found, for example, in southern England, where differences between dialects/accents are being reduced (Torgersen and Kerswill 2004; see Auer 2018 for discussion of this and related phenomena in Europe). Any kind of change through external contact is distinct from the inter‐generational transmission found in low‐contact communities (Labov 2007). However, changes resulting from regional dialect leveling are often similar to those found in the migration‐driven changes which lead to new varieties. We turn to these next.

4  New‐dialect Formation New dialects come about as a result of the migration of speakers of different, but mutually intelligible varieties of a single language to a new location  –  a process Trudgill calls new‐ dialect formation (NDF; Trudgill 1986, 2004; Trudgill et al. 2000). As already mentioned, central to NDF is koineization: According to Trudgill (1986: 107), koineization begins with the mixing of features from the different varieties, giving rise to a high degree of variability. This is followed by the reduction in the number of forms present in the mix through the leveling of variant forms (especially in phonology and morphology), and simplification – the reduction of phonological and morphological complexity. Provided social conditions are right, this results in a new, distinct, partly hybrid, partly innovative version of the parent varieties; this version is by definition more focused (Le Page and Tabouret‐Keller 1985) than the initial mixed stage. The result is a new dialect, or koine. In this chapter, we will be concerned with what Siegel (1985: 364) calls immigrant koines, as opposed to the Koine, or lingua franca Greek, which formed in the Eastern Mediterranean in the Hellenistic period (Siegel 1985: 358). We will also cover the other scenario already sketched out: that leading to rapid changes when an existing population is linguistically overwhelmed by a critical mass of incomers, typically in the formation of new towns. Koineization is a component of change in this case, too.

4.1 Tabulae Rasae: South African Bhojpuri and New Zealand English In an imaginary experimental scenario, one might observe the formation of a new dialect ab initio by depositing a population of people, carefully chosen to be speakers of different dialects and screened to avoid reflecting existing sociolinguistic divisions, on a desert island, and then return after a generation or two to observe the outcome. This scenario is, of course, impossible on ethical and practical grounds. There are, however, a number of locations

244  Paul Kerswill around the world where new dialects, isolated from their ancestral homelands, have been investigated in what Trudgill (2004) calls tabula rasa conditions, where speakers of the language concerned had not previously lived. Among other similar cases are German‐language islands in Eastern Europe and the USA (Rosenberg 2005) and Bhojpuri (Hindi) dialects in Fiji (Siegel 1987, 1997) and South Africa (Mesthrie 1991) (and elsewhere  –  see Trudgill 1986: 99–102). Historical work in this tradition is represented by Tuten (2003), Lodge (2004), Dollinger (2008), and McColl Millar (2016; on Scots in Orkney and Shetland). Bhojpuri speakers from India arrived in South Africa (KwaZulu‐Natal) as indentured laborers from the middle of the nineteenth century to around 1910 (Mesthrie 1991: 72), and many remained, forming distinct communities. South African Bhojpuri is a highly mixed variety, as Mesthrie points out: “SB [South African Bhojpuri] does not accord with any single language or dialect of North India, displaying – rather – a blend of features from several sources” (1991: 104). The Bhojpuri studies show that, in relation to many of the input dialects, there is simplification in morphology and morphophonemics, with a reduction in the number of morphological categories which are marked, as well as simpler paradigms. Thus, South African Bhojpuri shows no trace of the present negative copula (Mesthrie 1991: 98), and lacks “respectful” forms in pronouns and verbs (1991: 97, 100). However, these two features show a pattern which can be directly linked to the demographic and social make‐up of the Bhojpuri communities. The present negative copula is a minority feature that would have been brought by speakers of Eastern Bhojpuri, its absence being a consequence of the fact that these speakers were in a minority (1991: 104). Here, demography played a crucial role. Second, “[i]t seems that this feature [respect marking] did not survive in the koine formation process in Natal, no doubt because of the leveling of social distinctions among [South African Bhojpuri] speakers” (1991: 100). Again, this is a critical observation: we will argue that the social upheaval caused by migration, and the need to establish a living in a new environment, causes old social distinctions to be lost, and with them the linguistic marking of those distinctions. Finally, we look at the demography of the early Bhojpuri‐speaking settlers in South Africa, to see if we can find an explanation for this simplification. Mesthrie (1991: 6) cites statistics indicating an overwhelmingly adult, predominantly male stream of migrants. Trudgill (2011: 237) suggests that “in fact simplification is most likely to occur in situations involving language learning by adults, who are typically poor second‐language learners as compared to small children, particularly so far as informal acquisition in short‐term contact situations is concerned.” South African Bhojpuri fits this account well. The Bhojpuri studies do not give us clear information on the original settlers. A study which comes close to this is an investigation of the early stages of New Zealand English reported in Trudgill et al. (2000), Gordon, Campbell, Hay, Maclagan, Sudbury and Trudgill (2004), and Trudgill (2004). The study was based on oral history recordings of elderly New Zealanders made in 1946–8 (which we refer to as the Mobile Unit recordings).1 Crucially, those recorded included people born in the period when New Zealand English was, according to Trudgill (2004: xi), being formed: 1850–90. The recordings thus afford a window on the speech of the earliest children growing up in colonial New Zealand, albeit sampled in their late adulthood. Gordon et al. (2004) present a comprehensive report and discussion. Trudgill (2004), however, presents the argument that new‐dialect formation is almost entirely deterministic in the sense that the shape of a new dialect can be predicted with some precision from a knowledge of the input varieties – the dialects of the original immigrants. This theory relies on careful consideration of three stages of new‐dialect formation and argues that the types of accommodatory behavior and selection of linguistic variants at each stage proceed in a predictable way which is largely unconnected with social factors  –  other than demography. Before critiquing his position, I will outline Trudgill’s stages. These correspond roughly to the first three generations of speakers (Trudgill et  al. 2000). At Stage I, adult

Contact and New Varieties  245 migrants will level away features which are in a small minority in the mix of dialects they encounter, subject to the individual’s ability to do so (Trudgill 2004: 89–93). In the New Zealand case, these would have included traditional dialect features. There would have been a great deal of inter‐individual variability at this stage, and people would themselves have been inconsistent in their usage (intra‐individual variability). At Stage II, the demographic distribution of features begins to determine the shape of the focused variety, which is still to appear. The Mobile Unit recordings represent this stage. According to Trudgill, the absence of a stable adult norm, or even a peer‐group dialect, means that children pick features to some extent “at will,” “from a kind of supermarket” (2004: 103, 108). The reason for this is that, in this tabula rasa context, they are not influenced by prestige or identity‐marking functions (pp. 151–7). Trudgill states that these young speakers did not “indulge in long‐term accommodation to one another” (p. 108), but rather that they selected features based on their frequency, always allowing for a “threshold rider” which filters out relatively infrequent features (p. 110). As at Stage I, speakers show considerable inter‐ and intra‐individual variability, much more so than in a community where the transmission of norms has been inter‐generational and without substantial external contact. Examples of this variability in two small communities are as follows (from Kerswill and Trudgill 2005: 209–10): (1) Inter‐individual variability in Arrowtown: GOAT: [oˑ] [oˑʊ] [oʊ] [ɔʊ] [ɵʉ] [ɐˑʊ] (2) Intra‐individual variability in the speech of Mr Riddle, Palmerston: /eɪ/ and /ǝʊ/ as in face and goat alternate between Scottish‐sounding monophthongal pronunciations with [e] and [o] and very un‐Scottish pronunciations with the wide diphthongs [æɪ] and [ɐʊ]. Stage III represents the focusing of the new variety, with alternate realizations leveled out, leaving only one, or two in the case of reallocation where variants are “reallocated” to a linguistic or sociolinguistic function (Britain and Trudgill 1999: 245). In New Zealand, this resulted in a very homogeneous variety apparently by 1900 (with the exception of the preservation of rhoticity in the far south), though with considerable social variation in terms of accent. Trudgill (2004: 115–28) argues strongly for a purely quantitative explanation of the features adopted by the new variety, citing several vowels and consonants including the retention of /h/, the maintenance of the /ʍ/  –  /w/ distinction (as in which/witch), the merger of unstressed /ǝ/ and /ɪ/ on /ǝ/ (as in rabbit), and broad diphthongs in words of the GOAT, FACE, MOUTH, and PRICE sets. In all but one of these, Trudgill adduces a simple majority principle: in the Mobile Unit recordings, the presence of these features is more common than their absence. In the case of the merger of /ǝ/ and /ɪ/, /ǝ/ was not in a majority, with only 32% of tokens having this vowel. Here, in a rather post hoc manner (which he admits – Trudgill 2004: 120), he appeals to a markedness principle: /ǝ/ is less “marked” in this position than /ɪ/, and this helped guarantee its survival. An important component of Stage III is the notion of “drift,” to explain apparently parallel developments in early New Zealand English and English in England. Drift, according to Trudgill, refers either to the continuation of a change in the “home” country or a tendency or propensity for a change (Trudgill 2004: 132–3). Hickey (2003: 229–35) agrees with the quantitative facts, but prefers to see changes in terms of individual speakers, especially children, detecting what is innovative and what conservative, or reinterpreting small phonetic changes as a shift of phonemic status. The latter would be the case in the parallel shift in both England and New Zealand from lax /ɪ/ to tense /iː/ in words like coffee (HAPPY‐tensing).

246  Paul Kerswill Hickey agrees with Trudgill’s point about the dynamic relationship between the emergent dialect and its British antecedents. We still need to explain the almost complete lack of regional variation in New Zealand English. Trudgill provides a partial answer. In accounting for the similarities between Australian and New Zealand English, he quotes a mixing‐bowl metaphor from Bernard (1981): “[T]he ingredients of the mixing bowl were very much the same, and at different times and different places the same process was carried out and the same end point achieved” (Trudgill 2004: 161). Trudgill points out that the dialect mixes in the different settlements in New Zealand were not the same (a point we will return to), so other factors must be brought to bear. New Zealand was a mobile society, and Trudgill agrees with Britain, who writes, “settlement isolation, mobility, transience and individualism led to the emergence of an atomistic society freeing people both from subservience and from the need to conform that tight‐knit local communities often engender” (2005: 164–5, referring to Fairburn 1982). Mobility in such a society, with a lack of local speech norms, would, it is argued, quickly lead to uniformity. At the same time, social prestige played no part, as Trudgill argues for both Stage II and Stage III, because children would not have been exposed to a standard ideology in New Zealand at that time – and in any case, children align themselves linguistically with local speech, especially that of their peers. Britain (2005: 165–6) points out that, in the early decades of European settlement, there was no compulsory education, so that overt pressures from prestige varieties could scarcely have had any effect. Both literacy and English‐style social stratification came a little later, and in any case after the earliest settlers’ children had begun the process of koineization. Moreover, New Zealand was engaged in what Belich (1996: 330) calls “custom shedding … Highly overt class differences … were leading candidates for the discard pile” – echoing the Bhojpuri communities’ leveling of social distinctions. Together, social structure, lack of overt norms, the seemingly random choice of forms at Stage II, and a choice of forms at Stage III based solely on relative frequencies of forms used by the Stage II speakers all contributed to a rapid focusing on a single set of features (Trudgill 2004: 114–15). Even so, the homogeneity is much greater than would be expected, and we need to account for this. First, we will take a closer look at the “social factors” rejected by Trudgill.

4.2 Social Factors vs Determinism in Tabula Rasa New‐dialect Formation We can now relate Trudgill’s findings and arguments to the general model of dialect formation I presented at the end of Section 2. That model allowed for “social factors” to affect the process at different stages. Trudgill states that such factors did not affect children’s dialect development at either Stage II or III in tabula rasa new‐dialect formation. His argument is that the social set‐up of early colonial New Zealand meant that “prestige,” “stigma,” “identity,” and “ideology” counted for nothing in children’s adoption of features. Instead, he appeals to the notion of “behavioral co‐ordination” (citing Cappella 1997). This motivates people to subconsciously converge as the default, following the principles outlined in the discussion of accommodation in Section  2; this, Trudgill posits, then leads to the newly focused New Zealand English of the end of the nineteenth century. At face value this is easy to accept – but only if we accept all of Trudgill’s assumptions about the nature of early anglophone New Zealand and about the nature of childhood language acquisition. The problem lies with the conceptualization of the tabula rasa. The Stage II scenario as outlined by Trudgill was most likely fairly rare. Clearly it existed in the very first English‐speaking settlements where children were present. After this, it would have existed only in small settlements populated by immigrants arriving at roughly the same time. Conditions for Trudgill’s Stage II

Contact and New Varieties  247 would not have existed in the larger settlements, except briefly on their founding; a complex social make‐up, including institutionalized ethnic divides  –  much of it aimed at Irish ­immigrants, who needed a permit of stay in Christchurch (Hickey, p.c.)  –  quite quickly emerged in the new urban communities. According to Belich (1996: 405), in Christchurch, social stratification, some of it imported from Britain, was present very early. The city was founded as a Church of England settlement, and its ethos is shown by early complaints about Australian and “half‐breed” incomers (Belich 1996: 339; see also Kerswill 2007). Meyerhoff (2006) argues strongly for a more complex and nuanced view of the early stage. She points out that several of the variables Trudgill investigates in fact show significant effects of parents’ origins for these same Stage II speakers – the model predicts the absence of such effects. It is likely that, in a diffuse speech community, children will be more, rather than less, influenced by parental varieties, as has been observed in two studies of new‐dialect formation in Norway (Kerswill and Williams 2000: 75). Meyerhoff also points out that children were born to large families with a great age range among the children, and that parents and older siblings would, like everywhere else, use a range of contextual styles which would be detected by children. In such communities, children at both Stages II and III would have coexisted, even within the same family, a fact which strongly curtails the time span of the putative social‐factor‐free Stage II. Where does this leave the deterministic account? Almost all of Trudgill’s features support his model, or at least do not contradict it. Trudgill’s view is as follows. The outcome, according to Trudgill (2004: 157), is a composite in which Stage III children selected “upper‐class H Retention, lower‐class Diphthong Shift, urban nonrhoticity, and the Rural Weak Vowel merger, from all the features available to them.” This, in his view, guarantees that they were not motivated by any prestige or identity‐based factors; elsewhere, he argues: “But this kind of baggage is not relevant to 7‐year‐old children in the colonial situation – which is precisely why the mixing of variants from different dialects of different regional origin and different degrees of social status … always takes place” (Trudgill 2008: 279). While I accept that normative pressures from British ideas of acceptable usage would have had a lesser effect in early New Zealand than in Britain, it is impossible to claim that the children grew up in what amounted to a social vacuum, devoid of adult intervention and isolated from adult norms (Holmes and Kerswill 2008; see Gordon 2010 on prescriptive comments from inspectors). Here, one would have to envisage an even more extreme experiment: the depositing of a population of pre‐adolescents on a desert island, Lord of the Flies‐like; even then, there would already have been social differentiation, which (as the film shows) rapidly increased. A tabula rasa quickly ceases to be the clean slate its name suggests (if it ever was, since adults in a new community are never entirely deracinated). We cannot therefore claim that social factors had no effect in the early period. The question is, which social factors? Prestige and social identity formation cannot be ruled out even at the start of the koineization process, since parents’ ideas, brought from the home country, about good and bad behavior and acceptable linguistic practices including politeness and, even, correctness, would have been brought to bear. Very few New Zealand children at Stage II and certainly Stage III would have been growing up in the highly idealized situation Trudgill envisages; in any case, the Mobile Unit recordings are heavily biased toward small communities, including farmsteads and a gold‐rush settlement, so are not especially representative of how the early settlers lived. We would also expect (as Trudgill suggests) that different koines emerged in some of the early settlements, while mobility later led to a leveling of the differences between them. We must conclude that Trudgill’s three‐stage deterministic model may be applicable, but only to settlements which remained relatively isolated, and which come closest to the imaginary experiment mentioned earlier. However, taken as a whole New Zealand English is pretty much as predicted. Trudgill’s idealized route, with its absence of identity factors,

248  Paul Kerswill doubtless pertained in a limited number of places. But it existed alongside a considerable majority of socially complex, more “normal” situations. Thus, while the deterministic model has validity in a rather idealized situation, it is impossible to accept its detailed implementation “on the ground.” A different approach is taken by Baxter, Blythe, Croft, and McKane (2009), who develop a mathematical model based on population genetics, applying it to accommodation and token frequency in the New Zealand data. Their conclusion is that the determinism model needs to be supplemented by a recognition that interlocutors (“interactors”) are “weighted,” meaning that speakers allocate different evaluations (“weightings”) to the features the interlocutors produce. In other words, linguistic variants are not neutral, because their users are not evaluated neutrally (Baxter et al. 2009: 271). The authors conclude: “The prevalence of the majority variants means that they were more likely to be propagated, for the reason that Trudgill advocated; but the mechanisms by which they are propagated are probably social.”

4.3  Homogenization in New Varieties All this said, there is still a gap in our understanding. The geographical homogeneity of the English of New Zealand is not merely a relative matter (compared to, say, Great Britain), but (barring Southland rhoticity) virtually total. Phonetic divergence has only recently emerged (by all accounts) in the form of distinct Maori and Pasifika Englishes (Hay, Maclagan, and Gordon 2008: 105–9). The mixing‐bowl metaphor does not in itself predict total homogeneity; instead, it predicts differences resulting from variations in the ingredients. Thus, in a highly mobile and atomistic society, we might expect regional varieties to have emerged around Auckland and the other main centers, in the manner of the regional dialect leveling in southern England today (Kerswill 2003; Britain 2018). However, this does not seem to have been the case. Hickey proposes a mechanism of supraregionalization to account for this: “[D]ialect speakers progressively adopt more and more features of a nonregional variety which they are in contact with. There does not have to be direct speaker contact …” (2003: 236; see also Hickey 2013). In Trudgill’s Stage III, the new variety can “be seen as a product of unconscious choices made across a broad front in a new society to create a distinct linguistic identity” (Hickey 2003: 215). This supraregional variety, according to Hickey, would have emerged in the “melting‐pot” settlements, which had mixed populations of relatively high density and size but with similar mixes, and then spread to the much more dialectally distinctive rural settlements. This process can be observed in contemporary Europe, notably in Denmark, where regional dialects have all but given way to a non‐conventionally prestigious but “modern” Copenhagen‐based variety (Pedersen 2005; Kristiansen and Jørgensen 2005). But for New Zealand English, in spite of high mobility, it is difficult to see how sufficient familiarity with the new variety could have come about to act as a reliable model. We cannot be certain how or exactly when the entirely nonregional variety appeared, but it is necessarily distinct from the focusing at local or regional speech community level which presumably preceded it. We should therefore introduce a Stage IV, at which new‐ dialect formation is already complete at the local/regional level, and at which supraregionalization is about to set in. Demographic and other social factors  –  especially gender and class  –  mentioned by Gordon (2008) guided the spread of New Zealand English, either promoting it or hindering it. Crucially, they did not actually determine its form. Instead, we have an image of the fully formed variety spreading inexorably, meeting varying degrees of resistance. Interestingly, the gold‐rush town of Arrowtown (a focus of the Gordon–Trudgill studies) is one of the pockets where Trudgill’s model seems to have applied; but in the context of New Zealand it was too small and peripheral to have been a center of influence in itself. What Gordon

Contact and New Varieties  249 describes matches rather closely what Hickey means by supraregionalization. We have no access to its mechanism in those early days; fortunately, we have the modern sociolinguistic record of countries like Denmark and (to a more limited extent) Ireland, where the same process has been documented in recent times.

5  Koineization in New Towns The second type of new‐dialect formation we will discuss is that of the new town, which we take as representative of new settlements where there are prior speakers of the language concerned, and where, therefore, there is face‐to‐face contact with existing speakers, whether in the new location or the “home” location. Importantly, the proportion of residents who are incomers vs locals becomes relevant to the outcome of the contact. The term “new town” is an official UK designation for a new, planned urban settlement placed on previously more or less unoccupied land, built throughout the twentieth century but particularly after World War II. The earliest new towns described in the sociolinguistic literature are, however, in Norway, where several were established at the heads of fjords in the period 1910–20 to harness hydroelectric power for the smelting of various ores. These included Odda and Tyssedal (east of Bergen), Sauda (southeast of Bergen), as well as Høyanger (in the Sognefjord, north of Bergen). The first two of these were investigated by Sandve (1976), and an extensive report and interpretation of his findings appear in Kerswill (2013: 523–6), with a focus on dialect mixing, inter‐dialect (intermediate, compromise) forms, simplification, and the marked effect of differences in the geographical origins of the in‐migrants to the two towns, which are located only 5 kilometers apart. Høyanger was discussed in some detail by Trudgill (1986) as a prototypical case of new‐dialect formation. Basing himself on Omdal (1977), he established the stages which he later applied to New Zealand. Omdal noted a transition from the traditional rural dialect of the oldest speakers, who grew up before the establishment of the town, through the extreme linguistic heterogeneity of the children who grew up in the new town, to a more stable new‐dialect spoken by the third generation – who were young adults in the 1970s. So far, the scenario is similar to New Zealand, with the important difference that some of the children growing up were the offspring of the original population. These people’s speech was mixed, but was closer to the old dialect than was that of the descendants of incomers. In 2001, Randi Solheim conducted sociolinguistic interviews in Høyanger, as well as obtaining archive recordings and previous descriptions, giving her a real‐time window of 45 years (Solheim 2006, 2009). As expected, the new Høyanger dialect includes features from both major input varieties, West Norwegian (the dialect area in which the town is located, and from where the vast majority of in‐migrants arrived) and East Norwegian (including the capital, Oslo, from where many of the managers and engineers came). However, despite the relatively small number of people from the east, a large number of high‐frequency words have an eastern form. As in most koines, there are inter‐dialect forms (Trudgill 1986: 62–5). In Høyanger, these are a compromise between the two main dialect sources, often involving the blending of a western stem with an eastern inflection – or vice versa. The eastern forms which also represent simplifications (particularly loss of velar–palatal alternation and umlaut in verbs) are now widespread in western Norway, as part of regional dialect leveling (see Kerswill 2013: 525 for details) – a point I will return to in a discussion of the new dialect of Milton Keynes. However, most of the eastern forms, including the pronominal forms dere ‘you (pl.),’ de (/diː/) ‘they’, and noen ‘some’, are rare elsewhere in the region, thus demonstrating that the dialect is a koine. The Høyanger study gives us a direct window on social influences on the outcome of new‐dialect formation. We have already seen that the number of high‐frequency eastern

250  Paul Kerswill Table 12.1  Evolution of Two Salient Høyanger Variables. East Norwegian and bokmål

West Norwegian and nynorsk

Generation I (rural dialect)

Generation II

Generation III

Gloss

ikke /ɪkǝ/ jeg /jæɪ/

ikkje /ɪçǝ/ eg /eːɡ/

ikkje eg

ikke and ikkje jeg and eg

ikkje eg

negator ‘I’

Note: Usage in Generation II was variable, as expected in the early stage of a mixed dialect. By Generation III, the eastern forms had all but disappeared.

forms is disproportionate to the number of in‐migrants from that region: Solheim attributes this to the high social prestige of those people. However, a number of items initially took an eastern form, only to be replaced by the original western realization. Table 12.1 shows two such instances, where there has been a shift from the original dialect forms, through East Norwegian forms, back to the original variants. Solheim uses a terminology similar to that of Trudgill, labeling the stages “generations” in order to link individual life experiences more directly with developments in the town itself. “Generation I,” however, refers not to the in‐ migrants, but to the existing dialect‐speaking population. Solheim ascribes this shift to the notion that the East Norwegian forms, being associated both with the former managerial class and Standard Norwegian in its prestigious bokmål instantiation, are too strong as markers to be acceptable in a West Norwegian dialect. This point of view is indeed expressed by some of the informants themselves. It is worth noting that these two items, the negator and the form for ‘I’, are regularly cited by Norwegians as regional and social dialect markers, doubtless partly because they are also associated with the two versions of Standard Norwegian – nynorsk (mainly western and rural) and bokmål (eastern, northern and urban). This is an indication that language ideology (in this case in favor of the less‐prestigious, western forms) may be a direct motivation for a wholesale change in linguistic usage. A somewhat different argument is put forward by Neteland (2017) in relation to the new dialect of Sauda, where the outcomes are broadly similar: She finds that the forms that ‘win out’ are those that were not only in a majority in the original mix, but also those that are spreading by a process of regional dialect leveling coupled with diffusion from urban centers. She explains: “The empirical data from Sauda show how the historical sociocultural context of language use, norms and change on the national and regional level provides a ‘target variety’ for the koine formation process on the local level [my emphases]” (Neteland 2017: 52). In addition to new dialects following national leveling trends, there is also evidence of local identity formation guiding changes in usage. Two diphthongs, /æɪ/ and /øy/, originally had local realizations in Høyanger, [aɪ] and [ɔy], respectively, setting them apart from much of the rest of the country, which typically has [ɛɪ] and [øy]. These supralocal variants of both diphthongs were in a large majority in Generation II, only to be largely replaced by the local form in Generation III. Meanwhile, Solheim has real‐time evidence that Generation III speakers in fact increased their use of the supralocal variant during their lifetimes (in line with national trends), while today’s youngest speakers, Generation IV, seem to be increasing their use of the local variant of /øy/. They do this most markedly in the place name Høyanger, insisting explicitly that [hɔyˈɑŋǝr] is the correct pronunciation (Solheim 2008: 6). The advantage of the Høyanger studies is that it is possible to show how local social factors affected outcomes, thus confirming some of the criticisms of Trudgill’s deterministic

Contact and New Varieties  251 model that it does not pay attention to the mechanism of transmission at Stages II and III. Solheim notes that a central observation emerging from my work on more recent data from Generation II, not least from my encounters with the speakers themselves, is that individuals’ personalities and life worlds are, to a great extent, decisive influences on their language use. It is likely that background factors such as these are particularly significant for this generation, since, in their formative years, there was no stable, local linguistic norm on which they could rely. (Solheim 2006: 237, my translation)

As we have seen, Neteland emphasizes regional and national trends, driven by contact, over local identity, personalities, and life worlds, and she criticizes Solheim in this respect; my view is that both pertain. There is every indication that the outcomes in places like Høyanger, Odda, Tyssedal, and Sauda are predictable using a quantitative model so long as social, ideological, and demographic factors are taken into account. Since I put forward a similar argument for the “tabula rasa” koines, the difference between them and the new towns discussed here is not a fundamental one. We turn now to Milton Keynes, which is by far the most recent of the new‐dialect situations we review here. It is the only one where it has been possible to log changes more or less as they happen, rather than at a distance of 50–150 years (Kerswill and Williams 2000, 2005; Williams and Kerswill 1999; Kerswill and Trudgill 2005). It is the newest, and largest, of Britain’s new towns, designated in 1967, and is situated some 80 kilometers northwest of London, in an already extensively leveled dialect area. We have already seen how the dialects of Høyanger and Sauda partly adumbrate leveling changes in their own dialect regions. What is the situation with a much more highly connected new town in an already leveled region? Population statistics for Milton Keynes reveal a rapid increase, particularly from the mid‐ 1970s to the late 1980s when the population increased by a factor of 2.41, with a tailing off of the rate of increase after that, as follows: Year: Population:

1967 60,000

1971 66,800

1977 96,300

1987 161,500

1997 196,920

2007 227,796

2017 266,800

(Sources: www.mkiobservatory.org.uk; http://ukpopulation2018.com/population‐of‐milton‐ keynes‐2018.html)

Sociolinguistic recordings were made in 1991 and 1992, some 24 years after designation and 14 years after large‐scale in‐migration had got under way. Children and young people at that time were representatives of Stage II, their parents Stage I, an assumption corroborated by the fact that only 10 of the 48 adults recorded were born in Milton Keynes. The primary interest was to observe new‐dialect formation as it was actually happening, rather than two or more generations later. The project could not, of course, observe the fully focused outcome, since Stage III had not been reached, though as it turned out focusing was already well under way at Stage II. The sample was composed of eight girls and eight boys in each of three age groups of 4, 8, and 12 years old, in addition to the principal caregiver (46 mothers, 1 father, 1 aunt). The families were selected to be classifiable as “working class.” The project was informed by a number of “Principles of koineization”: Outcomes in post‐contact varieties: 1. Majority forms found in the mix, rather than minority forms, win out. 2. Marked regional forms are disfavored. 3. Phonologically and lexically simple features are more often adopted than complex ones.

252  Paul Kerswill The migrants and the first generation of native‐born children: 4. Adults, adolescents, and children influence the outcome of dialect contact differently. 5. The adoption of features by a speaker depends on his or her network characteristics. The time scale of koineization: 6. There is little or no historical continuity with the locality, either socially or linguistically. Most first‐ and second‐generation speakers are oriented towards language varieties that originate elsewhere. 7. From initial diffusion, focusing takes place over one or two generations. 8. Because of sociolinguistic maturation, the structure of the new speech community is first discernible in the speech of native‐born adolescents, not young children. Not surprisingly on the basis of our previous discussion, Principles 1 and 2 can be shown to apply in Milton Keynes. Little can be said about Principle 3, because most of the input dialects were already leveled southeastern ones (see further, Kerswill and Williams 2000: 85–9). Principle 4 could also not be fully addressed, though it was clear that the high proportion of children to adults relative to the general population in the 1990s would favor early focusing (26.1% under‐16s compared to 20.1% for England and Wales). However, Trudgill (2002, 2011) has reasoned that in cases where dialect contact and language contact are relatively sustained and involve children, complex features may be learned more easily than when contact mainly involves adults. Principles 5 and 8 can both be demonstrated by an examination of index scores on a phonetic variable, the fronting of the vowel of GOAT. Table 12.2 shows the values for this variable. Kerswill and Williams (2000: 93–4) show, first, that it is the 12‐year‐olds who have the highest scores, the girls exceeding the boys. Second, the older children have greater fronting than their caregivers, the conclusion being that this age group is propelling the change. Third, there appeared to be a group of low scorers and another of high scorers in this age group. The low scorers (mean score 1.3) are two boys and two girls, and appear to be socially quite isolated individuals. The high scorers are four girls (mean 2.1), forming a group of friends who are sociable and well integrated at school. Given that the fronting of this vowel is an ongoing change, the obvious conclusion is that it is female‐led. What this means for new‐dialect formation is that “integrated” children with broad social contacts are in a position to focus the new‐dialect forms; in the sample, all the other children – be they 4‐year‐olds or the less‐integrated older children – are linguistically more heterogeneous, showing features of their parents’ accents more strongly than the integrated children, and are clearly not in the lead in the focusing process. Interestingly, this shows that focusing can start at Stage II. It also confirms Principle 8 in showing older children to be in the vanguard (see Kerswill and Williams 2005: 1026–32 for a fuller discussion). We turn now to Principle 6. A marker of a new dialect is the absence of any stable, locally based dialect to serve as a model for acquisition. This implies that there is a break in Table 12.2  Key to GOAT‐fronting. Degree of fronting

Value

Location/type

(ou) – 0: [oː], [oʊ] (ou) – 1: [ǝʊ], [əʊ̟] (ou) – 2: [ǝʏ] (ou) – 3: [ǝɪ]

score: 0 score: 1 score: 2 score: 3

(Northern and Scottish realization) (older Buckinghamshire and London) (fronting) (fronting and unrounding)

Contact and New Varieties  253 continuity between generations – not of language acquisition, which would lead to pidginization or creolization, but of the transmission of local dialect features. This is as true in new towns as in tabula rasa situations. The “new‐dialect” status of Milton Keynes English can be confirmed by examining the diphthong /aʊ/ of MOUTH. There are a number of variants of this vowel, which appear to be converging on a Received Pronunciation‐like [aʊ] in southeast England, moving away from local pronunciations such as [ɛɪ] and [ɛʊ̟]. The striking point is that there is an abrupt and complete disjunction between the variants, first between pre‐Stage I (the original inhabitants of the pre‐new town district) and Stage I (the adult in‐migrants), and then again between Stage I and Stage II: none of the Stage I or Stage II speakers uses the two favored, conservative variants of the pre‐Stage I people. A parallel study in Reading, a well‐established town in the same region, showed a similar trend, but with the conservative tokens still in use, albeit at a low level, by the youngest speakers. This demonstrates continuity, absent from Milton Keynes. It also shows that Milton Keynes is ahead of Reading in this feature, which is spreading as part of regional dialect leveling in the south‐east of England. Principle 7 can be illustrated, again using the GOAT‐fronting data. Figure 12.1 shows the index score for all the 48 children, ranked from high to low. Against each child the caregiver’s score has been plotted. Two patterns stand out. The first is that there is no obvious link between the adults and children, with seven adults having scores close to zero (representing Scottish and northern English pronunciations). Second, the average score for the children is higher than for the caregivers, suggesting change (as we have seen). Further to this is the fact that two children have a score close to zero. This comes as no surprise when we realize that these are 4‐year‐olds, matching their caregivers. We assume from the data for the older children that these children will align themselves with the other children as they grow older; for the one child re‐recorded 18 months later, this is indeed the case, on this feature as on others. The result is greater homogeneity among older children, a shift which is necessarily much greater in a new town than elsewhere, because almost all of the adults are from elsewhere. This also, of course, supports Principle 8. We saw that the population of Milton Keynes more than doubled in the period equivalent to Stage I. Kerswill (2018) applies insights from present‐day new dialects to the Industrial Revolution in Britain in order to find candidates for “new dialects.” Only one town/city

3 (Most fronted) 2.5

(ou) index

2 caregivers children

1.5 1 0.5

(Least fronted) 0 Subjects

Figure 12.1  Association of Milton Keynes children’s scores for GOAT‐fronting with those of their caregivers (from Kerswill and Williams 2000: 102).

254  Paul Kerswill appeared to fulfill the criterion of having a majority of incomers at any stage of its development: this is Middlesbrough, founded from scratch in 1830, which shows circumstantial evidence of new‐dialect formation.

6  Ethnolects and Multiethnolects Language shift following immigration may lead to new varieties: in the UK, the clearest examples are the South Asian Englishes of cities like Bradford, which has a large population with a Pakistani, mainly Panjabi‐speaking heritage. The speech of many of this group (Wormald 2016) can be characterized as an ethnolect (Clyne 2000). Large‐scale immigration of people with a much wider range of linguistic backgrounds to northwest European cities in the last 20–40 years has given rise to new versions of urban vernaculars. These are often referred to as multiethnolects, or urban contact dialects (see Kerswill and Wiese forthcoming). Recent work on these varieties has been carried out in Copenhagen (e.g. Quist 2008), Stockholm (Bijvoet and Fraurud 2006), Oslo (e.g. Aarsæther et  al. 2008), the Netherlands (Hinskens 2015), Germany (Freywald, Mayr, Özçelik, and Wiese 2011), and elsewhere (Nortier and Svendsen (eds.) 2015.). We will take the example of London, a city with a long history of immigration. Various incoming groups are said to have influenced London English, but it is only with large‐scale immigration post‐World War II and especially since the late 1950s that we see verifiable change – at least in the speech of working‐class young people. The new variety has been labeled Multicultural London English (Cheshire, Kerswill, Fox, and Torgersen 2011). We will take the example of diphthongs in London English, which are usually said to be subject to Diphthong Shift (Wells 1982: 306–10). This means that the vowels of face, mouth, price, and goat have long trajectories, a development which has been taking place over a century or more and is a further development of the Great Vowel Shift. Figure  12.2 shows vowel plots for an elderly working‐class Londoner (Kerswill, Torgersen, and Fox 2008: 457). If this is read as a traditional vowel diagram, these characteristics can be clearly seen. The lines represent the trajectories of the diphthongs. F2 2500 2300 2100 1900 1700 1500 1300 1100 900

700

500 200

300 CHOICE FACE

400

GOAT

500

F1

MOUTH PRICE DRESS TRAP

600

STRUT START FOOT

700

800

Figure 12.2  Vowel plots for an elderly male speaker from Hackney, born 1938.

Contact and New Varieties  255 F2 2500 2300 2100 1900 1700 1500 1300 1100 900

700

500 200

300 CHOICE FACE

400

GOAT

500

F1

MOUTH PRICE DRESS TRAP

600

STRUT START FOOT

700

800

Figure 12.3  Vowel plots for Brian, aged 17, Afro‐Caribbean, Hackney.

Figure 12.3 shows the same vowels for a young inner‐city Londoner, who is representative at least of speakers of Afro‐Caribbean and West African ethnicities. The diphthongs, especially face, price, and goat, have much shorter trajectories, and face and goat are now high peripheral vowels. goat in fact shows a development which is the opposite of that found in Milton Keynes (and elsewhere in the southeast of England). The study showed that, in fact, all ethnicities, including white Anglos, could variably use this system, and that the strongest predictors were residence in the inner city (rather than the suburbs) and a friendship network which was highly multiethnic. Contact, in this case, is in the first instance with the languages of the immigrants (cf. Stage I in Trudgill’s terminology). However, the changes leading to Multicultural London English are not driven directly by second‐language speakers, but by the L1 English of their immediate offspring (Stage II), which may well contain copies of L2 features such as near‐monophthongal face and goat. It is, in other words, another case of dialect contact. One may ask why this has happened now, and not at an earlier period. The rate of immigration in recent years has been such that there are working‐class areas of London where languages other than English are used by a large proportion of residents – sometimes being collectively a majority. In these inner‐city communities, children grow up acquiring English to a high level, but the model is not the local, Cockney dialect. Instead, their models may be L2‐speaking adults (especially caregivers) and older children who themselves did not have an L1 model, as well as non‐British varieties of English. Children from (London) English‐ speaking families also grow up in this environment, acquiring many of the features and using them, at least variably. This outline suggests that a deterministic model could be applicable, though research remains to be done in this area; in any case, the same provisos are likely to apply as in the other cases discussed in this chapter. These varieties are strong markers of social identity, and are often considered “youth styles” (e.g. Rampton 2008). Finally we should mention a new line of research which is likely to become important with the current massive urbanization of Sub‐Saharan Africa. African countries are almost all highly multilingual, and this means that the continent’s cities are themselves becoming more multilingual still. The capital of Ghana, Accra, is a particular case in point. The

256  Paul Kerswill country’s most widely spoken language, Akan, is not the original language of Accra, but the in‐migration of Akan speakers means that it has become the dominant language there. This has two consequences: there are now many L2 speakers of Akan in the city, while, at the same time, as Yankson (2018) has found out, a new Accra variety of Akan is emerging, spoken by the Akan migrants and their children.

7 Conclusion In this chapter, I have attempted to show that new‐dialect, or koine formation is a process which can be divided into several stages, at each of which speakers engage in certain ­behaviors and where social factors always play a part, but in different ways. The overall mechanism is in part deterministic. However, even if outcomes are highly predictable, many different factors, both social and identity‐related, come into play before the outcome is achieved.

NOTE 1 The recordings were made by the Mobile Disc Recording Unit of the National Broadcasting Corporation of New Zealand (see Gordon et al. 2004: 3–5 and Trudgill 2004: x).

REFERENCES Aarsæther, Finn, Ingvild Nistov, Toril Opsahl, Unn Røyneland, and Bente Ailin Svendsen 2008. On some formal and functional aspects of Norwegian spoken among youth in multicultural Oslo. Paper given at UPUS‐ seminar at Centre de cooperation FrancoNorvegienne en Sciences Sociales et Humaines, 26 September. Auer, Peter 2018. Dialect change in Europe – leveling and convergence. In Charles Boberg, John Nerbonne, and Dominic Watt (eds.), The Handbook of Dialectology. Oxford: Wiley‐Blackwell, pp. 159–76. Baxter, Gareth J., Richard A. Blythe, William Croft, and Alan J. McKane 2009. Modeling language change: an evaluation of Trudgill’s theory of the emergence of New Zealand English. Language Variation and Change 21.2: 257–96. Belich, James 1996. Making Peoples: A History of the New Zealanders. Honolulu: University of Hawai‘i Press. Bernard, John R. 1981. Australian pronunciation. In The Macquarie Dictionary. Sydney: Macquarie Library, pp. 18–27.

Bijvoet, Ellen and Kari Fraurud 2006. Svenska med något utländskt [Swedish with a little foreign language]. Språkvård 3: 4–10. Britain, David 2005. Where did New Zealand English come from? In Allan Bell, Ray Harlow, and Donna Starks (eds.), Languages of New Zealand. Wellington: Victoria University Press, pp. 156–93. Britain, David. 2013a. The role of mundane mobility and contact in dialect death and dialect birth. In Daniel Schreier and Marianne Hundt (eds.), English as a Contact Language. Cambridge: Cambridge University Press, pp. 165–81. Britain, David 2013b. Space, diffusion and mobility. In J. K. Chambers and Natalie Schilling (eds.), The Handbook of Language Variation and Change, 2nd edition. Oxford: Wiley‐Blackwell, pp. 471–500. Britain, David 2018. Dialect contact and new dialect formation. In Charles Boberg, John Nerbonne, and Dominic Watt (eds.), The Handbook of Dialectology. Oxford: Wiley‐ Blackwell, pp. 143–58. Britain, David and Peter Trudgill 1999. Migration, new‐dialect formation and

Contact and New Varieties  257 sociolinguistic refunctionalisation: reallocation as an outcome of dialect contact. Transactions of the Philological Society 97: 245–56. Cappella, Joseph N. 1997. Behavioral and judged coordination in adult informal social interaction: vocal and kinesic indications. Journal of Personality and Social Psychology 72: 119–31. Cheshire, Jenny, Paul Kerswill, Susan Fox, and Eivind Torgersen 2011. Contact, the feature pool and the speech community: the emergence of Multicultural London English. Journal of Sociolinguistics 15.2: 151–96. Clyne, Michael 2000. Lingua franca and ethnolects in Europe and Beyond. Sociolinguistica 14: 83–9. Coupland, Nikolas 2010. The authentic speaker and the speech community. In Carmen Llamas and Dominic Watt (eds.), Language and Identities. Edinburgh: Edinburgh University Press, pp. 99–112. D’Arcy, Alexandra 2015. Variation, transmission, incrementation. In Patrick Honeybone and Joseph Salmons (eds.), The Oxford Handbook of Historical Phonology. Oxford: Oxford University Press, pp. 583–602. Dollinger, Stefan 2008. New‐Dialect Formation in Canada: Evidence from the English Modal Auxiliaries. Amsterdam: Benjamins. Fairburn, Miles 1982. Local community or atomised society? New Zealand Journal of History 16: 146–67. Freywald, Ulrike, Katharina Mayr, Tiner Özçelik, and Heike Wiese 2011. Kiezdeutsch as a multiethnolect. In Friederike Kern and Margret Selting (eds.), Ethnic Styles of Speaking in European Metropolitan Areas. Amsterdam: Benjamins [Studies in Language Variation 8], pp. 45–73. Gallois, Cindy, Tania Ogay, and Howard Giles (2005). Communication accommodation theory: a look back and a look ahead. In William Gudykunst (ed.), Theorizing about Intercultural Communication. Thousand Oaks, CA: Sage, pp. 121–48. Giles, Howard 1973. Accent mobility: a model and some data. Anthropological Linguistics 15: 87–109. Gordon, Elizabeth 2008. Ian Gordon Fellowship lectures. Victoria University of Wellington, September. Gordon, Elizabeth 2010. Written evidence of early New Zealand English pronunciation. In Raymond Hickey (ed.), Varieties of English in Writing. The Written Word as Linguistic Evidence. Amsterdam: John Benjamins, pp. 349–64,

Gordon, Elizabeth, Lyle Campbell, Jennifer Hay, Margaret Maclagan, Andrea Sudbury, and Peter Trudgill 2004. New Zealand English: Its Origins and Evolution. Cambridge: Cambridge University Press. Hay, Jennifer, Margaret Maclagan, and Elizabeth Gordon 2008. New Zealand English. Edinburgh: Edinburgh University Press. Hickey, Raymond 2000. Dissociation as a form of language change. European Journal of English Studies 4.3: 303–15. Hickey, Raymond 2003. On the process of new dialect formation. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 213–39. Hickey, Raymond 2005. Dublin English: Evolution and Change. Amsterdam: Benjamins. Hickey, Raymond 2013. Supraregionalisation and dissociation. In J. K. Chambers and Natalie Schilling (eds.), The Handbook of Language Variation and Change, 2nd edition. Oxford: Wiley‐Blackwell, pp. 537–54. Hinskens, Frans 2015. Wijdvertakte wortels: over etnolectisch Nederlands. Nieuwjaarsboekje Meertens Instituut. Amsterdam: Amsterdam University Press. Holmes, Janet and Paul Kerswill 2008. Contact is not enough: a response to Trudgill. (Invited response to Peter Trudgill 2008. Colonial dialect contact in the history of European languages: on the irrelevance of identity to new‐dialect formation. Language in Society 37.2: 241–54). Language in Society 37.2: 273–7. Kerswill, Paul 2002. Koineization and accommodation. In J. K. Chambers, Peter Trudgill, and Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. Oxford: Blackwell, pp. 669–702. Kerswill, Paul 2003. Dialect levelling and geographical diffusion in British English. In David Britain and Jenny Cheshire (eds.), Social Dialectology: In Honour of Peter Trudgill. Amsterdam. Benjamins, pp. 223–43. Kerswill, Paul 2007. Review of Peter Trudgill, 2004: Dialect Contact and New‐Dialect Formation: The Inevitability of Colonial Englishes. Language 83: 657–61. Kerswill, Paul 2013. Koineization. In J. K. Chambers and Natalie Schilling (eds.), The Handbook of Language Variation and Change, 2nd edition. Oxford: Wiley‐Blackwell, pp. 519–36. Kerswill, Paul 2018. Dialect formation and dialect change in the Industrial Revolution: British vernacular English in the nineteenth century. In Laura Wright (ed.) Southern English. Varieties

258  Paul Kerswill Then and Now. Berlin: de Gruyter Mouton, pp. 8–38. Kerswill, Paul, Eivind Torgersen, and Sue Fox, 2008. Reversing “drift”: innovation and diffusion in the London diphthong system. Language Variation and Change 20: 451–91. Kerswill, Paul and Peter Trudgill 2005. The birth of new dialects. In Peter Auer, Frans Hinskens, and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press, pp. 196–220. Kerswill, Paul and Heike Wiese forthcoming. Urban Contact Dialects and Language Change: Insights from the Global North and South. London: Routledge. Kerswill, Paul and Ann Williams 2000. Creating a new town koine: children and language change in Milton Keynes. Language in Society 29.1: 65–115. Kerswill, Paul and Ann Williams 2005. New towns and koineisation: linguistic and social correlates. Linguistics 43.5: 1023–48. Kristiansen, Tore and Jens Normann Jørgensen 2005. Subjective factors in dialect convergence and divergence. In Peter Auer, Frans Hinskens, and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press, pp. 287–302. Labov, William 1989. Exact description of the speech community: short a in Philadelphia. In Ralph Fasold and Deborah Schiffrin (eds.), Language Change and Variation. Amsterdam: Benjamins, pp. 1–57. Labov, William 2001. Principles of Linguistic Change: Social Factors. Oxford: Blackwell. Labov, William 2007. Transmission and diffusion. Language 83: 344–87. Lass, Roger 1990. Where do extraterritorial Englishes come from? Dialect input and recodification in transported Englishes. In Sylvia Adamson et al. (eds.) Papers from the 5th International Conference on English Historical Linguistics. Amsterdam: Benjamins, 245–80. Lass, Roger 2004. South African English. In Raymond Hickey (ed.), Legacies of colonial Englishes. Cambridge: Cambridge University Press, pp. 363–86. Le Page, Robert and Andrée Tabouret‐Keller 1985. Acts of Identity: Creole‐Based Approaches to Language and Ethnicity. Cambridge: Cambridge University Press. Lodge, R. Anthony 2004. A Sociolinguistic History of Parisian French. Cambridge: Cambridge University Press.

McColl Millar, Robert 2016. Contact: The Interaction of Closely Related Linguistic Varieties and the History of English. Edinburgh: Edinburgh University Press. Mesthrie, Rajend 1991. Language in Indenture: A Sociolinguistic History of Bhojpuri – Hindi in South Africa. Witwatersrand: Witwatersrand University Press. Meyerhoff, Miriam 2006. Linguistic change, sociohistorical context and theory‐building in variationist linguistics: new dialect formation in New Zealand. English Language and Linguistics 10.1: 173–94. Milroy, Lesley 1999. Standard English and language ideology in Britain and the United States. In Tony Bex and Richard J. Watts (eds.), Standard English: The Widening Debate. London: Routledge, pp. 173–206. Milroy, Lesley 2002. Introduction: mobility, contact and language change – working with contemporary speech communities. Journal of Sociolinguistics 6.1: 3–15. Neteland, Randi 2017. Koine formation in context. Journal of Historical Sociolinguistics 2017: 37–54. Nortier, Jacomine and Bente A. Svendsen (eds.) 2015. Language, Youth and Identity in the 21st Century: Linguistic Practices across Urban Spaces. Cambridge: Cambridge University Press. Omdal, Helge 1977. Høyangermålet – en ny dialekt [The speech of Høyanger – a new dialect]. Språklig Samling 18: 7–9. Patrick, Peter L. 2002. The speech community. In J. K. Chambers, Peter Trudgill, and Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. Oxford: Blackwell, pp. 573–97. Pedersen, Inge Lise 2005. Processes of standardisation in Scandinavia. In Peter Auer, Frans Hinskens, and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press, pp. 171–95. Pickering, Martin and Simon Garrod 2004. Toward a mechanistic psychology of dialogue. Behavioral and Brain Sciences 27: 169–226. Quist, Pia 2008. Sociolinguistic approaches to multiethnolect: language variety and stylistic practice. International Journal of Bilingualism 12.1–2: 43–61. Rampton, Ben 2008. Crossing into class: language, ethnicities and class sensibility in England. In Carmen Llamas and Dominic Watt (eds.), Language and Identities. Edinburgh: Edinburgh University Press, pp. 134–43.

Contact and New Varieties  259 Rosenberg, Peter 2005. Dialect convergence in the German language islands. In Peter Auer, Frans Hinskens, and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press, pp. 221–35. Sandve, Bjørn Harald 1976. Om talemålet i industristadene Odda og Tyssedal: generasjonsskilnad og tilnærming mellom dei to målføra [On the spoken language in the industrial towns Odda and Tyssedal: generational differences and convergence between the two dialects]. Unpublished Cand. philol. thesis, University of Bergen. Siegel, Jeff 1985. Koines and koineization. Language in Society 14: 357–78. Siegel, Jeff 1987. Language Contact in a Plantation Environment. Cambridge: Cambridge University Press. Siegel, Jeff 1997. Mixing, leveling, and pidgin/ creole development. In Arthur K. Spears and Donald Winford (eds.), The Structure and Status of Pidgins and Creoles. Amsterdam: John Benjamins, pp. 111–49. Solheim, Randi 2006. Språket i smeltegryta: sosiolingvistiske utviklingsliner i industrisamfunnet Høyanger [Language in a melting pot: sociolinguistic lines of development in the Høyanger industrial community] (Norwegian University of Science and Technology, Throndheim Ph.D. thesis). Solheim, Randi 2008. Språket i smeltegryta: presentasjon av doktoravhandlinga [Language in a melting pot: presentation of a doctoral thesis]. Maal og Minne 1: 1–8. Solheim, Randi 2009. Dialect development in a melting pot: the formation of a new culture and a new dialect in the industrial town of Høyanger. In Frans Gregersen and Unn Røyneland (eds.), Sociolinguistics, special issue of Nordic Journal of Linguistics 32.2. Torgersen, Eivind and Paul Kerswill. 2004. Internal and external motivation in phonetic change: dialect levelling outcomes for an English vowel shift. Journal of Sociolinguistics, 8: 24–53. Trudgill, Peter 1986. Dialect in Contact. Oxford: Blackwell. Trudgill, Peter 2002. Linguistic and social typology. In J. K. Chambers, Peter Trudgill, and

Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. Oxford: Blackwell, pp. 707–28. Trudgill, Peter 2004. New‐Dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press. Trudgill, Peter 2008. On the role of children, and the mechanical view: a rejoinder. Language in Society 37.2: 277–80. Trudgill, Peter 2011. Social structure, language contact and language change. In Ruth Wodak, Barbara Johnstone, and Paul Kerswill (eds.), The Sage Handbook of Sociolinguistics. London: Sage, pp. 236–48. Trudgill, Peter, Elizabeth Gordon, Gillian Lewis, and Margaret Maclagan 2000. Determinism in new‐dialect formation and the genesis of New Zealand English. Journal of Linguistics 36, 299–318. Tuten, Donald N. 2003. Koineization in Medieval Spanish. Berlin: de Gruyter Mouton. Tuten, Donald N. 2008. Identity formation and accommodation: sequential and simultaneous relations (response to Peter Trudgill, 2008, Colonial dialect contact in the history of European languages: on the irrelevance of identity to new‐dialect formation, Language in Society 37: 2, 241–80). Language in Society 37.2: 259–62. Wachsmuth, Ipke, Jan de Ruiter, Petra Jaecks, and Stefan Kopp (eds.) 2013. Alignment in Communication. Towards a New Theory of Communication. Amsterdam: Benjamins. Williams, Ann and Paul Kerswill 1999. Dialect levelling: change and continuity in Milton Keynes, Reading and Hull. In Paul Foulkes and Gerard Docherty (eds.), Urban voices. Accent studies in the British Isles. London: Arnold, pp, 141–62. Wells, John C. 1982. Accents of English. Cambridge: Cambridge University Press. Wormald, Jessica 2016. Regional Variation in Panjabi‐English (University of York Ph.D. thesis). Yankson, Solace Ago 2018. Language Contact and Change in Linguistically Heterogeneous Urban Communities: The Case of Akan in Accra. Utrecht: LOT.

13 Contact in the City1 HEIKE WIESE

1 Background: An Historic View on Urban Language Contact Throughout history, cities have been a magnet for immigration, taking in a constant influx of new residents, both from more rural areas of the same region or country and from regions and countries further afar, who would bring with them a wealth of different dialects and languages (see Mackey 2005 for an overview). Well‐known historic examples for highly multilingual cities include such early cosmopolitan cities as Bombay, Dar es Salaam, Alexandria, or Constantinople, many of which still retain their multicultural and multilingual character today (Gupta 2000). In an account from the early seventeenth century, the English travel writer Thomas Coryat relates that in Venice one could “heare all the languages of Christendome, besides those that are spoken by the barbarous Ethnickes” (Dursteler 2012: 47), and in an often‐cited letter from the early eighteenth century, another English author, Lady Mary Wortley Montagu, describes Constantinople  –  where she lived when her husband was the English ambassador there – as a “Tower of Babel” (O’Quinn and Heffernan 2012: Letter 41). The large trading cities of the Levant were polyglot and cosmopolitan, with language mixing in the streets and in families as well as in formal contexts (Strauss 2011; Mansel 2014), and cities in the Early Modern Mediterranean in general, including not only the Ottoman, but also the Habsburg and Venetian empires, were characterized by a large cultural and linguistic diversity and fluidity (Dursteler 2012; cf. also Lucassen and Lucassen 2013 on multiethnic empires in Medieval Europe). The strong dominance of a single national language and the related monolingual perspectives we see in Europe today, are a historically comparably recent phenomenon: a legacy from a period of nation‐state building that involved what Bommes and Maas (2005: 182) called the “counter‐factual ideological construction” of “one country, one people, one language,” which is still very much part of the self‐image in most modern European states (cf. Vogl 2012 for an historical overview). In contrast to this, in Medieval and Early Modern Europe, multilingualism was a normal part of life, an unremarkable fact of everyday linguistic experiences and practices (cf. Putzo 2011). Classen (2013) shows in an analysis of a large range of early texts including, among others, Beowulf (between ca. 800 and ca.1100), the Nibelungenlied (ca. 1200), and El cantar de mio Cid (ca. 1200), that even though there are numerous descriptions of encounters during travels or war that will have involved speakers

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

262  Heike Wiese of different linguistic backgrounds, in the stories this never incited any special ­comments: multilingualism was so much the norm that at no point, this seems to have been regarded as a challenge. In urban areas, different local dialects came together with a range of foreign languages, supplemented by Latin as a language of religion and education. In particular in the cities, widespread multilingualism was further supported by the fact that merchants’ sons and (male) students usually spent longer periods abroad for training and education. Accordingly, students’ argots as predecessors of modern youth languages were often influenced by ­language contact at least in the form of lexical borrowing (Mihm 2001).

2  Contact in Modern Cities The heterogeneous and fluid character of urban speech communities makes them particularly open to language variation and change, and they can accordingly play a pioneering role in linguistic innovation (cf. Vanderkerckhove 2010). According to the UN report on World Urbanization Prospects, more than half of the world’s population lived in urban areas in 2014, with an upward trend (UN 2015), so cities are an increasingly important domain for the investigation of language use in general, and language contact in particular, with urban language putting a spotlight on developments at both sociolinguistic and structural levels. Traditionally, the linguistic investigation of dialects concentrated on rural areas. The focus was on a group of speakers whom Chambers and Trudgill (1980) described as “non‐mobile old rural males,” short NORMs: in order to identify a particularly pronounced, “authentic” dialect, traditional dialectology tended to favor older, male speakers who had spent most of their life in the same rural area. Accordingly, a dialect atlas resulting from such a survey, rather than reflecting the breadth of linguistic practices within different regions, might more fittingly be called an “atlas of old men’s village speech.” Urban speech communities were not so much the center of attention since they were seen as less likely to provide homogeneous, historically grown dialects, given the high degree of contact and fluidity at social and linguistic levels. However, it is just this dynamic that makes urban centers interesting for investigations of linguistic and social differentiation. Accordingly, modern sociolinguistics has had a strong focus on urban language since Labov’s seminal works in the 1960s and 1970s.2 A comparably new line of research into urban multilingualism targets the linguistic landscape, that is, language that is visible in the public domain and can provide insights into local linguistic ecologies and power relations, and (contact‐)linguistic practices.3

2.1  Dialect Leveling and Change Urban language contact often leads to the formation of new dialects. From a general perspective, we can distinguish two main kinds of urban dialects that have been in the focus of contact‐linguistic studies: the development of koines as a result of dialect leveling based on different regional dialects, and the emergence of new versions of existing local dialects under the influence of other languages or dialects. From the perspective of language contact, koines represent the formation of a new mixed variety and can hence be regarded as a counterpart to new contact languages such as pidgins and creoles, and mixed languages. In contrast, the second kind of dialect reflects developments within an existing variety that were triggered or facilitated by language contact. An example for the first case has been described by Kerswill (2002) for Milton Keynes, a “new town” in the UK that was founded in 1967. Internal immigration supported here a koineization based on contact between the region’s original dialect and several other dialects

Contact in the City  263 brought in by new residents from other regions (cf. also Kerswill, this volume).4 An example of the second kind of dialect change has been described by Wölck (2002) for New York State, where immigration from Germany, Italy, and Poland supported new linguistic developments not just in minority, heritage language use, but also in the majority language of the receiving country, that is, English.

2.2  New Urban Contact Dialects Over the last decades, a major focus of research on urban language contact has been on new ways of speaking that emerged as peer‐group vernaculars among adolescents in multiethnic urban neighborhoods. Following Wiese (forthcoming), such new ways of speaking can be captured as urban contact dialects, defined as “urban vernaculars that emerged in ­contexts of migration‐based linguistic diversity among locally born young people, marking their speakers as belonging to a multiethnic peer group.” While such contact dialects can be observed in different cities across the world, two geographical areas have been particularly in the center of interest lately: Northwestern Europe and Sub‐Saharan Africa.5

2.2.1  The Linguistic Dynamics of Multiethnic Urban Youth An important sociolinguistic characteristic of urban contact dialects is their association with multiethnic neighborhoods, where “ethnicity” should be understood as a social category, constructing groups that are believed to share a common descent – typically geographically associated  –  and culture (e.g. Moran 2014; Fought 2002). These dialects typically emerge among young people who grow up in a mixed urban neighborhood where a substantial part of the older generation (e.g. the speakers’ parents or grandparents) have immigrated to the city from rural areas and/or other countries. These locally born speakers, then, find themselves as part of a new, multiethnic urban generation that has access to a wealth of different dialects and languages as part of the broader cultural heritage in their peer group. This large range  –  sometimes described as “superdiversity”6  –  supports the creative combination and integration of a multitude of linguistic resources, in linguistic practices captured, e.g. by such concepts as “translanguaging” (cf. García 2009). Over time, such practices can lead to new contact dialects as markers of a new, multiethnic urban generation. Findings on the distribution of these dialects so far suggest that they tend to emerge first in peer‐group situations among young people, and can later spread to other age groups, becoming more general markers of social class, multiethnicity, or urbanness, cf. Wiese (forthcoming) for an overview. That these dialects have their roots in youth language practices is not surprising from the perspective of language variation and change. Young adolescent speakers are often regarded as the main agents of change since they are a social group that is forging new identities for themselves and strongly orients towards peers. Accordingly, Kerswill (1996) and Eckert (2000) identify adolescents as a core group in linguistic innovation, and Tagliamonte and d’Arcy (2009) discuss converging quantitative data that points to an “adolescent peak” in the usage of new phenomena, which might be a requirement for language change (cf. also Labov 2001). In both Africa and Europe, urban youth is a particularly vital sector of society. Africa shows not only a high increase of urbanization, but also the highest level of population growth globally, and young people make up a large proportion of the population, with 60% of the population being under 25 years old (UN 2015). While Europe, in sharp contrast to Africa, faces population declines and is challenged by an aging population in general (UN 2015), this is not true for its urban population with a migrant background, which make up a substantial part of the speech community for urban contact dialects: this community is vital

264  Heike Wiese and grows faster than the rest  –  for instance, in Germany, the average age of the overall population is 45.3 years, but the population with migration background shows an average age of only 35.4 years, and over a third of children under 10 years have a migration background.7 According to the 2017 census, in cities with over 50,000 inhabitants, nearly half of residents under 18 (48%) were from a family with an immigration history. As mentioned on p. 261, cities in general are a particularly vibrant site of language contact, supported by a large influx of new residents. In present‐day Europe, this is primarily due to political and labor immigration from other countries, while in Africa, domestic immigration from other, more rural parts plays an important role as well. As the historic examples of students’ language have already shown, urban youth language as a rule participates in such language contact, thus bringing together two sources of dynamics, which make them particularly interesting, providing a domain where, as Nortier and Dorleijn (2013) put it, “language contact can be caught, as it were, ‘red handed.’”

2.2.2  Multilingual Mixed Languages vs New Majority Language Vernaculars The multiethnolectal urban contact dialects that can emerge under such favorable conditions, can take on two different forms: they can constitute multilingual mixed languages that integrate different contact languages, or new vernaculars of a majority language of the larger society. This two‐fold distinction of contact‐linguistic codes is related to the sociolinguistic make‐up of the larger societal context, underlining the relevance of such factors for the outcome of language contact.8 In the context of widespread societal multilingualism, multilingual practices such as code‐switching and mixing are a normal part of everyday encounters. In such contexts, urban contact dialects often take on the form of mixed languages in the sense of Thomason (2001). According to the contact‐linguistic taxonomy developed there, Mixed Languages combine grammatical and lexical subsystems from two source languages, thus indicating a mixed ancestry (cf. also Matras 2009: ch. 10.3; Meakins 2013; Velupillai 2015: ch. 3; Bakker, this volume). In the core type of “intertwined” Mixed Languages, grammatical features dominantly go back to one language, while the lexicon is largely contributed by a second language. Urban contact dialects with their much larger range of contact languages as potential sources typically draw on more than two contact languages, especially for their lexicon, making them a special kind of “Multilingual Mixed Language” (cf. Wiese, forthcoming). Examples for urban contact dialects that have the form of multilingual mixed languages are found, e.g. in many African countries today, where linguistic developments can benefit from the larger context of societal multilingualism and a long tradition of embracing the normality of multilingual practices (cf. Mufwene 2008: ch. 13 on multilingualism in African history). (1) illustrates this type of contact‐linguistic code with an example from Camfranglais, an urban contact dialect from Cameroon (Kießling and Mous 2004): in this example, a grammatical frame provided by Cameroonian French integrates lexical borrowings from two contact languages, ‘kick’ from English and ‘agogo’ from Hausa. (1)

On a   kick mon agogo. GENERIC.PRON.3PS. has steal   my watch ‘They stole my watch / Someone has stolen my watch.’

Historically, there is also an example from Europe for an urban contact dialect that takes on the form of a Mixed Language, namely Old Helsinki Slang. This dialect emerged in early‐ twentieth century Helsinki, when monolingual national ideologies were still restricted to the upper classes. Like modern examples, Old Helsinki Slang is associated with the formation of a new, mixed urban group; in this case a bilingual working‐class culture that emerged as the

Contact in the City  265 result of Finnish‐speaking immigrants mixing with the initially predominantly Swedish‐ speaking local population. Structurally, this new dialect resembled modern African examples, showing such characteristic patterns of language mixing as a Finnish morphosyntactic frame combined with a dominantly Swedish lexicon (cf. de Smit 2010). In contrast to this, in present‐day Europe, a societal macro context that is characterized by a strong monolingual habitus (e.g. Hüning et al. 2012) supports urban contact dialects that constitute new vernaculars of the majority language. In modern European nation states, we typically find a strong ideological association tying language to space, where a single language is constructed as belonging to a distinctive geographical region (often the country as a whole, or, in the case of multilingual European states, a delimited region within a country). This language constitutes a strong majority language whose dominance is further supported through such institutions as education, bureaucracy, or the legal system. This widespread monolingual habitus in modern Europe is at odds with the local reality of urban settings, where multilingualism is an unremarkable everyday reality, in many ways reminiscent of the multilingual normality of earlier periods (see Section 1). In these urban settings, many speakers grow up with one or more heritage language(s) in addition to the majority language, and children and adolescents from mono‐ and multilingual homes alike typically acquire at least some words or short routines from other languages spoken in their peers’ families, leading to a horizontal multilingualism that is in evident contrast to the “one language  –  one country” ideology that dominates the societal macro context.9 Nevertheless, this macro context is powerful enough to preserve the national language’s dominance even in multilingual local settings, given that this language is not only the language of educational (and other) institutions, but also typically functions as a widespread lingua franca, among multilingual speakers with different repertoires as well as monolingual majority language speakers. Under such conditions, then, urban contact dialects, rather than constituting mixed languages, form new vernaculars of the majority language. Such new vernaculars do not even‐handedly integrate two or more contact languages, but remain close to this single, dominant majority language, with contact‐linguistic transfer mostly restricted to a number of lexical items and short routines from heritage languages. Structural transfer is rare: rather than genuine contact‐induced change, what we most often find at the grammatical level is variation and change that takes up internal tendencies of the majority language. Accordingly, crosslinguistically one can find similar developments in countries with related majority languages, even though the contact‐linguistic settings are characterized by typologically different heritage languages. Developments get a special boost in the dynamic setting of extensive language contact, but they reflect, to a large part, majority language patterns. An example are “verb‐third” (V3) word order options that have been observed in urban contact dialects across countries with related Germanic majority languages, namely Sweden, Denmark, Germany, the Netherlands, and Norway, but in the context of such typologically diverse heritage languages as, e.g. Turkish, Arabic, Kurdish, Greek, and Serbian or Croatian (cf. Wiese 2009). The Germanic languages on which these contact dialects are based are usually described as “verb‐second” (V2), capturing a pattern that puts the finite verb in second position in main declaratives (and some other sentence types), with another major sentence constituent in front of the verb; this can be the subject, but also, e.g. an object or an adverbial. For urban contact dialects of such Germanic languages, findings from different studies point to additional V3 options, where two constituents rather than one can be placed before the finite verb (see Section 2.2.2 for a further discussion of V3). (2) through (4) give three examples, from Denmark, Germany, and Norway (data from Quist 2000, Wiese 2013, and Freywald et al. 2015, respectively).

266  Heike Wiese (2)

Normalt man går på ungdomsskolen.    (Danish) usually one goes to supplementary school ‘Usually one goes to the supplementary school.’

(3)

danach  isch muss zu mein vater  (German; from KiDKo/MuH28MK10) afterwards I must to my father ‘Afterwards, I’ll have to go to my father.’

(4)

Nå  de   får     betale. now they get/must  pay ‘Now, they must pay.’

(Norwegian)

The fact that such similar V3 patterns can be observed across contact settings suggests that these patterns relate to internal options of the Germanic V2 languages involved here, rather than to transfer from other languages. This kind of contact‐linguistic dynamics can be described as “contact‐facilitated” variation and change: it is contact‐related in the sense of being supported by a dynamic linguistic setting that generally favors language variation and change, but it is not contact‐induced in that it does not involve specific patterns of another language (Wiese, forthcoming). We can hence distinguish two kinds of contact‐ related dynamics for urban contact dialects: contact‐induced versus contact‐facilitated variation and change. Contact‐facilitated patterns occur particularly often in urban contact dialects that represent new vernaculars of a majority language. The local setting with its large diversity of multilingual practices supports a higher tolerance of linguistic variation, allowing a more liberal linguistic system. Given a dominant monolingual habitus within the larger societal context, this might not lead to the emergence of new mixed varieties, but such a loosening of grammatical restrictions can still give these new urban dialects an innovative potential that makes them pioneers within the general dialectal landscape of the majority language (Wiese 2013; cf. also Hinskens 2011 on the special dynamics of new urban contact dialects compared to historically older ones; Matras 2011 on innovations and elaborations in contexts of widespread bilingualism and laxer normative control). Contact‐induced and contact‐facilitated variation and change for a particular urban contact dialect can cover a wide spectrum, with gradual rather than categorical differences. At one end, we find direct transfers of specific elements or patterns from another language, at the other end, phenomena that are shared with other, more monolingually based varieties, but might not be as far developed there. We will tease this apart in a case study from Germany, in Section 3.2 where we take a closer look at V3 and some other pertinent examples.

2.3  Metrolinguism and Market Jargons An interesting setting of particularly intense contact and a high degree of linguistic fluidity are urban markets or bazaars, which have lately attracted more and more attention from linguistic studies. In such linguistically rich settings, sellers and customers act as multicompetent speakers (cf. Matras 2013) accessing a diverse pool of linguistic resources in their market interactions that allows them to creatively add new elements to their repertoire. Accordingly, contact usually does not so much lead to new dialects, but to less stable, highly fluid practices of linguistic mixing and integration. Such settings have been in the

Contact in the City  267 focus of sociolinguistic approaches to “metrolingualism” (cf. Pennycook and Otsuji 2015), targeting spatial repertoires that are determined by linguistically highly diverse local contexts (a bazaar, but also, e.g. a restaurant or a shop) and challenge traditional concepts like code‐switching or code‐mixing. From a contact‐linguistic point of view, such multilingual practices could be seen as something like the “jargons” described in creole studies; cf. Velupillai (2015: 534), who defines this as “individual ad hoc solutions in individual contact situations leading to a highly variable and unstable contact language.” If taken literally, there is no such thing as a jargon, then, since these solutions will differ from speaker to speaker and emerge differently in each communicative situation. However, as we will see in the next section for the example of two urban markets in Berlin, one can often find some recurring patterns in these settings, and perhaps the best way to think of an urban market’s jargon is as an integrative linguistic practice characterized by (i) access to a spatially determined, but principally open range of linguistic resources and (ii) a liberal use of individual ad hoc solutions, but at the same time (iii) guided by local customs of language choice and language dominance and (iv) centering around a shared core of recurring patterns.

3  A Case in Point: Language Contact in Berlin In our third and last section, let us have a look at the city of Berlin. Even though Germany’s largest city by some margin, with slightly more than 3.6 million inhabitants (2019) Berlin is not exactly huge in international terms. However, as a vibrant urban metropolis, it offers a broad range of language contact settings that allow us to explore and illustrate some of the different facets of urban language contact we discussed in the first two sections.

3.1  Language Contact in the Traditional Berlin Dialect Berlin has always been a site of language and dialect contact, right from the start, with the two thirteenth‐century towns Berlin and Cölln  –  the predecessors of present‐day Berlin – based on the immigration of Low German and Dutch speakers from the lower Rhine region to an area further characterized by a Slavic language, Sorbian, especially in rural settlements (cf. Buts 1988). Today, Sorbian is a protected minority language in Germany, spoken, e.g. in some of the Spreewald region south of Berlin in the state of Brandenburg. As typical for urban areas, immigration‐based dialect contact led to dialect leveling in Berlin. Initially, the spoken language was Low German, and writing was in Latin. In North German cities including Berlin, Low German was later used as a written language as well, complementing Latin. However, in the fifteenth and sixteenth centuries, writing was conducted more and more in High German, and in the seventeenth and eighteenth centuries, spoken and written French gained in importance as an additional contact language, fashionable through language practices of the nobility. This was further boosted by French‐speaking Huguenots, and similarly, other languages also gained through immigration, e.g. Yiddish. Further immigration adding to the city’s linguistic diversity came from Low German speaking rural areas, with large numbers of immigrants, for instance in the aftermath of the Thirty Years’ War (1618–48), and, in the nineteenth century, as a result of industrialization which led to an immense growth of the city and to a linguistic setting characterized by intense dialect and language contact.11 This kind of setting supported an urban dialect whose High German character reflects broad dialect leveling, but which also integrates lexical transfers from a range of contact languages, as well as retaining some remnants of its initially Low German basis. For instance, Bulette ‘meat ball’, referring to a staple of traditional Berlin cuisine, goes back to French ‘boulette’, the particle dalli ‘quick/snappy’ has a Slavic origin (> Polish dalej), and the verb

268  Heike Wiese zocken came in via Yiddish, meaning ‘to play’, ‘to play cards’. As a colloquial term, zocken used to be associated primarily with gambling, but has meanwhile gained additional currency among young people, who use it to refer to computer games (while the morphologically related, prefixed abzocken is a colloquial term meaning ‘to rip off / take someone for a ride’). A Low German influence is visible, e.g. morphologically in deviations from the High German accusative–dative distinction, or phonologically by the use of [t] instead of [s] in some lexical items, e.g. wat instead of was ‘what’. The twentieth century saw separate developments in East and West Berlin as a result of the Berlin Wall. This social and linguistic segregation led to differences that were still noticeable a decade after the wall came down in the late 1990s, especially in domains of lexicon and phonology (cf., for instance, Dittmar and Bredel 1999). In East Berlin, during the period of the German Democratic Republic (1949–1990), state contract workers, e.g. from North Vietnam, lived in separate boarding houses, and informal interactions with the local German‐speaking population were not encouraged, hence this did not have a noticeable contact‐linguistic effect on the Berlin dialect. West Berlin participated in the intake of large labour immigration since the 1960s, and these immigrants contributed to a vibrant contact‐linguistic setting. This was particularly true for traditional working‐class neighborhoods, e.g. Wedding and Neukölln, and in areas that had become particularly disadvantaged because of the Wall, e.g. parts of Kreuzberg that were cut off on three sides from surrounding areas. Today, these neighborhoods (and similarly multilingual inner‐city neighborhoods in other German cities) support an urban contact dialect of the kind we discussed in Section 2.2, namely ‘Kiezdeutsch’, lit. ‘(neighbor‐)hood German’, a new German vernacular (see Section 3.2 below). Since the 1990s, immigration from Eastern Europe and increased domestic immigration (in particular from South Germany) further contributed to the linguistic melting pot of Berlin as a whole, as did the influx of international companies and institutions and their workforce and refugees forced into emigration by political and economic crises in the Middle East and the Global South. Ironically, today we can observe some ethnic and social linguistic stratification where only the older urban dialect is constructed as authentic local language use, indexical for an imagined monolingual German in‐group of genuine, ethnically homogeneous “Berliners” – even though this dialect has always been contact‐based, with a mixed, immigration‐born speech community. In contrast to this, a new urban dialect like Kiezdeutsch is often perceived as alien to the local linguistic landscape and a potential threat to it, and its speakers, even though they are typically at least second‐generation Germans, are excluded from this in‐group.12

3.2  Contact‐related Variation and Change in Kiezdeutsch Kiezdeutsch is an instance of the second kind of contact‐linguistic code we discussed earlier, namely a new vernacular of the majority language (= in this case, of German) – which is not surprising given the strong monolingual habitus we find in present‐day Germany, and the accordingly strong dominance of German as a majority language (cf. Gogolin 2002; Wiese 2015). Kiezdeutsch hence provides us with a good showcase for teasing apart contact‐ induced transfer and contact‐facilitated internal dynamics, two kinds of contact‐related language variation and change we distinguished for such urban contact dialects in Section 2.2. As mentioned there, this distinction is not categorical, but gradual, and we can often observe both kinds of dynamics at work. In the present section, let us have a look at some clear‐cut examples as well as a number of in‐between cases for Kiezdeutsch. For the purpose of this illustration, we will concentrate on phenomena where Turkish might play a role as a source language.

Contact in the City  269

Figure 13.1  Lexical integration in a love message (Berlin‐Kreuzberg).

A clear case of direct influence is lexical transfer. In the context of urban contact dialects, such transfer leads to integrations at grammatical, and sometimes also graphemic levels, and new elements can be used across speakers, independent of linguistic backgrounds. Figure 13.1 gives an illustration:13 this is from a message on a playground in Berlin celebrating the love between Ingo + Inga. In this message, the writer (presumably Ingo) uses a number of Turkish loan words within a German text. One of them, rendered in Figure 13.1, is Aşkim ‘my love’, as part of a passage ’Liebe sie mein Aşkim, Du and Ich!”, lit. ‘Love her my Aşkim, you and I!’ The Turkish source, ‘aşkım’, is a morphologically complex item c­ onsisting of a nominal base ‘aşk’ and a 1sg possessive suffix ‘‐ım’. In the Kiezdeutsch example, ‘Aşkim’ is combined with the German first person possessive pronoun mein ‘my’, which suggests that it is treated as a mono‐morphematic noun, that is, its integration into German has led to an erasure of the internal Turkish structure. Graphematic integration is visible in two aspects: the replacement of Turkish ‘ı’, which is not a grapheme in German, by ‘i’, which is; and the capitalization of Aşkim in accordance with German spelling rules for nouns. Note, though, that the Turkish ‘ş’, which is not a grapheme in German either, is preserved, indicating that the integration is not categorical. An example for the transfer of a pattern above the level of individual words is the m‐reduplication described for Kiezdeutsch,14 cf. (5): (5)

er sagt zu m=meiner cousine so fettsack METTsack  (KiDKo, MuH27WT) he says to m=my cousin mp fatbag matbag ‘He says, like, “Fatso, Matso!” to my cousin.’

In Turkish, m‐reduplication is well established in spoken language; it is syntactically fully integrated and can be used with bases of different syntactic categories (e.g. Stolz 2008). At the semantic or pragmatic level, it expresses vagueness and can also have pejorative effects. In German, m‐reduplication seems to be a novel phenomenon, so far restricted to the multilingual contexts characteristic of Kiezdeutsch. In these contexts, though, it is not simply taken over as is, but develops its own dynamics and acquires language‐specific characteristics. Syntactically, m‐reduplication seems to be less flexible and largely (although not exclusively) restricted to nominal bases in German. Phonological integration into German is achieved by replacing the whole onset by [m] (while in Turkish, only the first consonant is replaced). Pragmatically, Kiezdeutsch adds a youth language aspect to the construction: by using m‐reduplication, speakers present themselves as “cool” or “chilled.”15 Hence, this is an example for contact‐induced change since contact triggers the development of a new pattern, however, one that takes on a life of its own when it is integrated into the receiving language. A further step in the direction of internal change can be seen in developments that have parallels in a heritage language, but also a clear internal motivation from within the

270  Heike Wiese grammatical system of the majority language. A candidate for such a two‐fold source in Kiezdeutsch is the use of gib(t)s lit. ‘gives.it’ as an existential particle, cf. (6): (6)

WEIßte doch, die die in verschiedene FARben gibs? (KiDKo, MuH9WT) thatpl in different colors     gibs know youcl PART those ‘You know them – those that come in different colors?’

The form ‘gib(t)s’ derives from existential ‘gibt es’ lit. ‘gives it’. In the original pattern, a 3sg verb ‘gibt’ is combined with an expletive pronominal subject ‘es’ and an accusative object for the Theme, i.e. ‘Es gibt NPacc’, with a meaning similar to English “There is NP.” In spoken language, ‘es’ is cliticized, yielding ‘gibts’ or phonologically reduced ‘gibs’ when we have the order ‘gibt es’. This is in accordance with a general rule in German that pronouns in that position (after the finite verb) are cliticized, which frequently happens in main declaratives with a sentence‐initial adverbial. This generally yields ‘gibs’ in such sentences as, e.g. ‘Hier gibts …’, lit. ‘Here gives‐it …’ (‘Here, there are …’). However, this rule cannot capture the novel use of ‘gibs’ in subordinate sentences requiring SOV order, as in (6). And there is some evidence for the Kiezdeutsch construction moving even further away from standard German, with the Theme NP as a subject rather than an accusative object (cf. Wiese 2013). This suggests that ‘gibs’ is treated as a mono‐morphematic form, making the subject position vacant, which can then be appropriated by the Theme. Unlike standard German, Turkish possesses a construction like this, based on the existential particle ‘var’ (negated ‘yok’), which thus makes a plausible source for this use of ‘gibs’ in Kiezdeutsch. However, a closer look at German shows that the new use of ‘gibs’ gets also a solid motivation from within, since the conventional pattern sits rather uneasily in the general layout of the syntax‐semantics interface. With the expletive subject ‘es’ it contains a syntactic argument that does not have a semantic counterpart, and  –  related to this  –  the highest thematic role (the Theme) does not correspond to the subject, as it normally should, but to the object. Seen from this point, the development in Kiezdeutsch can be motivated by syntax‐semantics alignment: mono‐morphematization of ‘gibs’ gets rid of the semantically empty element ‘es’, and reinterpretation of the accusative as a nominative allows an association of the highest thematic role with the subject. Such a regularization can draw on two lines of support within German (cf. Wiese 2013 for a discussion). For one, there is a general tendency in spoken German towards a univerbation of existential gibt es to gib(t)s, based on frequent cliticization. Second, in modern German, accusative and nominative forms are often identical at the surface, which provides a basis for reinterpretation (e.g. in (6) above, the expression for the Theme, third person singular feminine die, could be nominative as well as accusative). This suggests a strong internal appeal for the development of gibs. This German‐internal motivation is further supported by such evidence as in (7), which comes from German in Namibia (Wiese et al. 2014): (7)

da gibs auch n berühmter SÄNger hier in namibia famous singer here in Namibia there gibs also acl ‘There is also a famous singer here in Namibia.’

This data is from informal conversations in a speech community where German is spoken as a first language and is also embedded in a multilingual context. Unlike in Germany, Turkish does not play a significant role here. Nevertheless, we do find the gibs construction and, since the expression for the Theme (n berühmter Sänger) is a masculine singular, this is one of the cases where we can even identify a distinct nominative. This suggests that the motivation from within the grammatical system of German is a crucial force behind the development,

Contact in the City  271 pointing to a primarily contact‐facilitated, rather than contact‐induced pattern. At the same time, this does not rule out that the Turkish existential construction gives the development of gibs an additional boost in Kiezdeutsch in Germany, especially since Turkish competences are widespread in its community. An example that is even further towards the end of the scale where internal trends are dominant (rather than influences of specific contact languages), is the V3 word order in main declaratives that we discussed in Section 2.2.2, that is, the option to place two constituents, rather than just one, before the finite verb. As argued already, the fact that there is converging evidence for this option from a range of Germanic “verb‐second” languages, points to an internal motivation. At the surface, this pattern shows some similarity with SVO orders known from early stages of language acquisition, and some earlier accounts from Germany who associated it primarily with young speakers of a Turkish heritage background described it as a transformation of V2 to SVO (e.g. Auer 2003). However, subsequent studies have revealed structural features that indicate the integration of this pattern into the topology of German sentences (including sentential brackets that delineate the so‐called “forefield” in the left periphery and the “middle field” following it).16 Furthermore, cross‐linguistic evidence points to a general information‐structural motivation for the pattern: V3 options allow speakers to place both framesetters and topics in the left periphery, a linearization that might reflect general, language‐independent preferences for ordering information.17 In Kiezdeutsch contexts, learner SVO can be found, e.g. in the German of the original immigrants, often the parents or grandparents of some of the speakers, and speakers will hence be familiar with it. Accordingly, its surface similarity might add a further source, but this presumably plays a minor role in the development of V3. Further support for this comes from V3 evidence in more monolingual contexts of ­present‐day German (Wiese 2013; Wiese and Müller 2018). In these contexts, V3 is not, as traditional wisdom has it, ruled out by a strict V2 constraint, but can in fact be observed in informal speech, where it follows the same pattern found for Kiezdeutsch contexts, but might be quantitatively less frequent (Wiese and Rehbein 2015). This is hence a case where the multilingual context of Kiezdeutsch does not so much provide a novel grammatical ­blueprint, but rather contributes to the general linguistic dynamics, giving urban contact dialects a quantitative advantage over language use from more monolingual settings. Such outcomes are characteristic of urban contact dialects like Kiezdeutsch that take the form of new majority language vernaculars under the pressure of a monolingual societal habitus. However, even under such societal conditions, we also find urban settings of such intense contact that practices of substantial language mixing are the rule rather than an exception. A prime example for such settings are urban markets, where speakers can integrate elements from a rich pool of linguistic resources. For an illustration, let us now have a look at two such markets in Berlin.

3.3  Two Urban Markets as Sites of Intense Language Contact Urban markets with their rich pool of diverse linguistic resources have been in the focus of approaches to “metrolingual practices” (see Section  3.2) investigating the specific spatial repertoires that emerge at such places. Two examples from (socio‐)linguistically different parts of Berlin are the Maybachufer market in Berlin‐Neukölln and the Dong Xuan Center in Berlin‐Lichtenberg. The Maybachufer market, also known as the “Turkish market,” is a street market in a vibrant multiethnic and multilingual neighborhood, with a linguistic ecology that includes Kiezdeutsch and the traditional Berlin dialect as well as heritage varieties of Turkish, Arabic, Kurdish, and a large range of other languages and dialects. The Dong Xuan Center, on the other hand, is set in the more monolingually German neighborhood of

272  Heike Wiese Berlin‐Lichtenberg. The center is a covered wholesale and retail market founded by a Vietnamese immigrant who had initially come to East Berlin as a contract worker during the DDR. In terms of products, stall owners’ backgrounds, and customer focus, the Maybachufer market has a more middle‐Eastern and Mediterranean orientation, compared to the Dong Xuan Center’s more East‐Asian character. However, both markets cater to multi‐ and monolingual locals, broader immigrant communities, and tourists alike, and both support a rich contact‐linguistic setting. While German is still a frequently used lingua franca in interethnic encounters among locals, it is complemented by Turkish or Vietnamese as a salient market language on the Maybachufer and Dong Xuan market, respectively; by English as a global language (e.g. in encounters with tourists), plus a range of other languages, including heritage languages of stall owners (some of them first‐generation immigrants) and a range of additional languages used by different groups of customers; which all form part of the markets’ linguistic ecologies and can enter speakers’ repertoires. In this setting, German does not exert its usually overwhelming dominance as a majority language anymore, but is integrated into a more diverse setting, as reflected, e.g. by an abundance of multilingual signs. The photos in Figure 13.2 illustrate these typologically different, but similarly diverse linguistic landscapes of the two markets. The sign on the top is by a Vietnamese‐heritage shop owner in a Dong Xuan market hall and brings together four languages and their respective scripts, namely Vietnamese (on top, “Asia Hoàng Dúc  –  identifying the owner’s name), German (below, “Asian groceries”),

Figure 13.2  Multilingual (and multi‐script) signs at Dong Xuan and Maybachufer markets.

Contact in the City  273 Mandarin (bottom left, “Asian food wholesale self‐service market”) and Thai (bottom right, “Asian food wholesale market”). The signs on the bottom are from a Maybachufer perfume stall owned by a seller of monolingual German background, who asks customers not to put bottles directly to their noses, using handwritten signs in three languages, Arabic, German, and Turkish, together with printed signs informing customers that perfumes are for women and men, in German, French, Italian, and English. Note that in both cases, German is part of the linguistic selection, but unlike, e.g. on multilingual official signs in Berlin, it is not distinguished by a prominent top or left‐most position, a larger font size, etc. As the following transcript from communications at a vegetable stall on Maybachufer illustrates, such a setting favors a creative integration of linguistic resources. The seller is a first‐generation immigrant from Turkey with Turkish and Kurdish as his heritage languages; the customer speaking here is a local with a monolingually German background. Turkish elements are marked by bold script, English ones by underlining, and Kurdish ones are in small caps.18 (8)

seller

(to passers‐by): (to Italian tourist):

customer (to seller): seller

(to colleague): (to customer):

Karpuz! Melone, Gurken! water.melon melon cucumbers One Stück – fünfzig Cent. Fifty Cent. one cl fifty cent fifty cent İki tane Aubergine, bitte. two cl eggplant please Badıncan! eggplant Bittschön, zwei Auberginen. Alles? – Ein Öro. please two eggplants everything one euro

In this communication, both seller and customer spontaneously integrate elements from a range of languages in their utterances. As discussed in Section 3.2, such multilingual practices could be characterized as a contact‐linguistic “jargon.” On the Maybachufer market, this jargon includes some recurring patterns, such as the frequent use of numeral classifiers (Stück, tane), similar to what we would find in Turkish, and the absence of plural marking for measure nouns (Cent), which is in accordance with both German and Turkish grammar, while English would require plural forms (e.g. for the example in (8), it would be fifty cents, not cent). Some recurring lexical characteristics of this market jargon are sellers’ use of Madame as a term of address, which often serves as an attention‐getting device towards (female) customers, thus closing a lexical‐pragmatic gap in standard German, which is currently lacking a term conventionally used for this. The traditional Berlin dialect has ‘junge Frau’, lit. ‘young woman’, for such cases, with a counterpart junger Mann, lit. ‘young man’ (since this is used independently of the addressee’s age, it often causes irritation if used towards non‐locals). Other recurring characteristics of the market jargon are some phonetic/phonological modifications of German words, in particular [Ɂø͜r͜ o] (transcribed as Öro in (8)) for Euro (standard German [Ɂɔi͜ro]), and recurring bittschön ‘please’ / ‘here you are’ as a variant of standard German bitteschön. A semantic innovation on a Turkish basis is the use of çıtır, illustrated on the market sign in Figure  13.3. This sign advertises cucumber (German Gurke) as ‘crunchy’, using an adjective çıtır in an interesting combination: in standard Turkish, çıtır would not be applied to vegetables, but rather to flat dry foods (e.g. potato crisps), whereas vegetables would be characterized with kütür.

274  Heike Wiese

Figure 13.3  Crunchy cucumbers: use of çıtır on the Maybachufer market.

If such noncanonical patterns stabilize and get further established within the market’s linguistic practices, they might eventually form the basis of a shared, spatially‐bound market grammar; in the words of a Turkish‐German seller at a Maybachufer mocca stall: (9)

One also learns a lot of languages in addition, after all, from customers, neighbors at the other stalls; sometimes just a few words, for instance, “uno, dos, tres,” but one creates one’s own grammar then.19

The creative integration and further development of different linguistic resources we can thus observe on urban markets brings back some multilingual normality to European cities: a linguistic ecology where language contact is embraced as an enrichment rather than a challenge, and the engagement with diverse linguistic resources is accepted as a normal and positive condition of human communication.

4 Conclusion In this chapter, we have encountered cities as sites of rich language contact, characterized by a pronounced social and linguistic fluidity, a constant intake of new speakers and speaker groups, and the wealth of linguistic resources and multilingual competences that comes with this. From a contact‐linguistic point of view, we distinguished two kinds of new linguistic forms that often emerge in such settings: new mixed varieties and new versions of existing ones. Instances of the first kind have been described as scenarios of dialect leveling, where koines emerge as new mixed dialects. These new varieties draw on a range of different regional dialects contributed by new arrivals to the city, including from the surrounding countryside. Other instances we discussed are multilingual mixed languages, which draw on linguistically more distant varieties and languages. The local linguistic ecology that supports multilingual mixed languages are multiethnic urban neighborhoods characterized by a range of heritage languages as a result of immigration. Their primary speaker base are not the immigrants themselves, but locally born young people, often descendants of immigrants, who grow up in such neighborhoods and forge a new, urban multiethnic identity. Such

Contact in the City  275 languages typically emerge in urban areas under the umbrella of a societal macro context where multilingual linguistic practices are accepted as normal and prevail in everyday encounters, as, for instance, in many African countries today. Instances of the second kind of contact‐linguistic outcome are cases of dialect change where an existing dialect acquires new features as a result of immigration, as has been the case, for instance, for the traditional Berlin dialect. Other instances we discussed are new majority language vernaculars that emerge in multiethnic urban neighborhoods similar to the ones supporting multilingual mixed languages. Unlike the latter, they do not integrate different languages alike, but rather stay within the realm of the language that has the status of a majority language in the broader society. The societal macro setting of such new vernaculars is typically characterized by a strong monolingual habitus and a correspondingly pronounced dominance of this majority language, which impedes the formation of fully‐blown mixed languages. As we have seen for the example of German Kiezdeutsch, the contact‐ related variation and change we see in such dialects can be located on a continuum from contact‐induced phenomena drawing on other‐language patterns to contact‐facilitated phenomena that reflect internal tendencies of the majority language. The example of multilingual mixed languages and such new majority language vernaculars highlighted the impact of the societal macro context on contact‐linguistic outcomes. To emphasize the similarity of their speaker base in multiethnic urban neighborhoods, I brought them together under the concept of “urban contact dialects.” Interestingly, we saw that even under conditions of a societal monolingual bias, one can also find systematic language mixing. Pertinent examples are multilingual mixed market jargons that emerge on urban markets with a linguistically highly diverse base of sellers and customers. As we saw for the example of two Berlin markets, such jargons can be described as integrative linguistic practices that access a broad range of spatially determined linguistic resources and allow individual ad hoc solutions, however they do not follow some “anything goes” policy, but are guided by local customs of language choice and dominance and contain a core of recurring patterns. Taken together, our exploration of cities as sites of language contact has shed a light on a number of threads that are interesting for contact‐linguistic research, among them the status of urban contact dialects in multiethnic areas, the relationship of mixed languages and new urban vernaculars, the impact of a multilingual vs monolingual societal habitus on contact‐linguistic developments, and the emergence of multilingual market practices that seem to defy the strong dominance of a single majority language, at least at a local level within a spatially determined (and restricted) linguistic setting. For future research, it would be particularly interesting to further investigate such market practices from a contact‐linguistic perspective, and also from a broader perspective of linguistic practices and language structure. If we embrace multilingualism as the normal condition of human language, it is in such settings, where the restrictions of codified monolingual norms are loosened, that we can expect a more natural language use, something we might think of as “free‐range language.”

NOTES 1 The research for this chapter was supported by funding from the German Research Foundation (DFG) for Collaborative Research Centre 1278 / project A01, and Research Unit FOR 2537 / projects P6 and Pd. 2 Cf. Labov (1972). For recent sociolinguistic perspectives on urban language see, e.g. contributions in Heinrich and Smakman (2018), and Coulmas (2018: ch. 7) for an overview and a discussion of ­language profiles as a basis of research on urban multilingualism.

276  Heike Wiese 3 Compare, e.g. contributions in Shohamy and Gorter (2009) for an overview; in Blackwood (2017) for a recent methodological discussion. 4 For similar developments in another geographical region, cf. Miller (2004) on historic and present examples of dialect leveling in Arabic cities. 5 For an overview of recent research, compare, e.g. such collected volumes as Quist and Svendsen (2010), Källström and Lindberg (2011), Kern and Selting (2011), Nortier and Svendsen (2015) for Europe; Hurst (2014); Nassenstein and Hollington (2015); Mensah (2016) for Africa; Kerswill and Wiese (forthcoming) for an integration. 6 E.g. Vertovec (2007), Blommaert et al. (2011); for criticism of this concept, see Pavlenko (2018). 7 The Federal Statistical Office in Germany defines “migration background” as covering persons who immigrated to Germany after 1949, have a foreign nationality and/or have at least one parent for whom that holds. 8 Cf. also Lim and Ansaldo (2016: Ch.8) on the relevance of sociolinguistic aspects for contact‐ linguistic dynamics in general. 9 Cf. also Quist (2010) on untying the “language‐body‐place connection” for the example of a Copenhagen school setting. 10 KiDKo (“KiezDeutsch‐Korpus) is an annotated corpus with data from spontaneous conversations among adolescents in Berlin, Germany. The main subcorpus, KiDKo/Mu, is from a multilingual and multiethnic neighborhood; in addition, there is a complementary subcorpus, KiDKo/Mo, which provides comparable data from a more monolingually German setting. The first two letters in speaker codes identify the subcorpus (Mu or Mo), the last letter identifies speaker’s family language repertoires (additional family/heritage language for bilingual speakers: e.g. K‐Kurdish, T‐ Turkish, A‐Arabic, or only German for monolingual speakers: D). The corpus (including also some other subcorpora) is freely accessible at www.kiezdeutschkorpus.de. 11 For an historic overview, cf. Schildt and Schmidt (1992). 12 See also Wiese (2015). For similar patterns in other European countries, compare, e.g. Kerswill (2014) on concerns in the public debate in the UK that Multicultural London English might replace traditional Cockney. 13 From KiDKo/LL: “From the ‘hood’ with love,” a linguistic‐landscape subcorpus of KiDKo (see also Note 10 above), accessible via www.kiezdeutschkorpus.de. 14 Wiese (2013); Wiese and Polat (2016); cf. also Şimşek (2012: ch. 4.2) on examples for m‐reduplication in German‐Turkish language contact. 15 Cf. Wiese and Polat (2016) for a detailed analysis of the conceptual domains and pragmatic networks involved here. 16 E.g. Wiese (2013), te Velde (2017); cf. also Selting and Kern (2009) for deviations from V2 in bilingual (Turkish‐German) settings. 17 Cf. Wiese (2011, 2013) for Kiezdeutsch; Freywald et  al. (2015) for a cross‐linguistic confirmation (from urban contact dialects based on Germanic majority languages); Walkden (2017) for a ­diachronic perspective. 18 From Maybachufer fieldnotes (H. Wiese); ‘cl’ – numeral classifier. 19 From Maybachufer fieldnotes; German original, my translation (H. Wiese).

REFERENCES Auer, Peter 2003. ’Türkenslang’: Ein jugendsprachlicher Ethnolekt des Deutschen und seine Transformationen. In Annalies Häcki Buhofer (eds.), Spracherwerb und Lebensalter. Tübingen: Francke, pp. 255–64. Blackwood, Robert (ed.) 2017. Methodology in linguistic landscape research. Linguistic Landscape (special issue) 3.3.

Blommaert, Jan, Ben Rampton, and Massimiliano Spotti (eds.) 2011. Language and superdiversities. Diversities (special issue) 13.2. Bommes, Michael and Utz Mass 2005. Interdisciplinarity in migration research: on the relation between sociology and linguistics. In Michael Bommes and Ewa Morawska (eds.), International Migration Research. Constructions, Omissions and the

Contact in the City  277 Promises of Interdisciplinarity. Utrecht: Ashgate, pp. 179–202. Butz, Georg 1988. Grundriß der Sprachgeschichte Berlins. In Norbert Dittmar and Peter Schlobinski (eds.), Wandlungen einer Stadtsprache. Berlinisch in Vergangenheit und Gegenwart. Berlin: Colloquium, pp. 1–40. Chambers, Jack K. and Trudgill, Peter 1980. Dialectology. Cambridge: Cambridge University Press. Classen, Albrecht 2013. Multilingualism in the Middle Ages and the Early Modern Age: The literary‐historical evidence. Neophilologus 97:131–45. Coulmas, Florian 2018. An Introduction to Multilingualism. Language in a Changing World. Oxford: Oxford University Press. de Smit, Merlijn 2010. Modelling mixed languages: some remarks on the case of Old Helsinki Slang. Journal of Language Contact – VARIA 3: 1–19. Dittmar, Norbert and Ursula Bredel 1999. Die Sprachmauer. Die Verarbeitung der Wende und ihrer Folgen in Gesprächen mit Ost‐ und WestberlinerInnen. Berlin: Weidler. Dursteler, Eric R. 2012. Speaking in tongues: language and communication in the early modern Mediterranean. Past and Present 217: 47–77. Eckert, Penny 2000. Linguistic Variation as Social Practice. Oxford: Blackwell. Fought, Carmen 2002. Ethnicity. In Jack K. Chambers, Peter Trudgill and Nathalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. London: Blackwell, pp. 444–72. Freywald, Ulrike, Leonie Cornips, Natalia Ganuza, Ingvild Nistov, and Toril Opsahl 2015. Beyond verb second – a matter of novel information‐structural effects? Evidence from Norwegian, Swedish, German and Dutch. In Jacomine Nortier and Bente A. Svendsen (eds.) 2015. Language, Youth, and Identity in the 21st Century. Oxford: Oxford University Press, pp. 73–92. García, Ofelia 2009. Education, multilingualism and translanguaging in the 21st century. In Ajit K. Mohanty, Minati Panda, Richard Phillipson, and Tove Skutnabb‐Kangas (eds.), Multilingual Education for Social Justice: Globalising the Local. New Delhi: Orient Blackswan, pp. 140–58. Gogolin, Ingrid 2002. Linguistic and cultural diversity in Europe: a challenge for educational research and practice. European Education and Research Journal 1.1: 123–38.

Gupta, Anthear Fraser 2000. Bilingualism in the cosmopolis. International Journal of the Sociology of Language 143: 107–19. Heinrich, Patrick and Dick Smakman 2018. (eds.), Urban Sociolinguistics – The City as a Linguistic Process and Experience. New York: Routledge Hinskens, Frans 2011. Emerging Moroccan and Turkish varieties of Dutch: ethnolects or ethnic styles? In Friederike Kern and Margret Selting (eds.), Ethnic Styles of Speaking in European Metropolitan Areas. Amsterdam: Benjamins, pp. 101–39. Hüning, Matthias, Ulrike Vogl, and Olivier Moliner (eds.) 2012. Standard Languages and Multilingualism in European History. Amsterdam: Benjamins. Hurst, Ellen (ed.) 2014. South African Linguistics and Applied Language Studies 32.2 (special issue) Tsotsitaal Studies: Urban Youth Language Practices in South Africa. Källström, Roger and Inger Lindberg (eds.) 2011. Young Urban Swedish. Variation and Change in Multilingual Settings. Gothenburg: University of Gothenburg. Kern, Friederike and Margret Selting (eds.) 2011. Ethnic Styles of Speaking in European Metropolitan Areas. Amsterdam: Benjamins Kerswill, Paul 1996. Children, adolescents, and language change. Language Variation and Change 8: 177–202. Kerswill, Paul 2002. Koineization and accomodation. In Jack K. Chambers, Peter Trudgill, and Nathalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. London: Blackwell, pp. 669–702. Kerswill, Paul 2014. The objectification of ‘Jafaican’. The discoursal embedding of Multicultural London English in the British media. In Jannis Androutsopoulos (ed.), Mediatization and Sociolinguistic Change. Berlin, New York: de Gruyter, pp. 427–56. Kießling, Roland and Maarten Mous 2004. Urban youth languages in Africa. Anthropological Linguistics 46.3: 303–41. Labov, William 1972. Language in the Inner City. Studies in Black English Vernacular. Philadelphia: University of Pennsylvania Press. Labov, William 2001. Principles of Linguistic Change. Vol.2: Social Factors. Oxford: Blackwell. Lim, Lisa and Umberto Ansaldo 2016. Languages in Contact. Cambridge University Press. Lucassen, Jan and Leo Lucassen 2013. European migration history. In Steven J. Gold and Stephanie J. Nawyn (eds.), Routledge

278  Heike Wiese International Handbook of Migration Studies. London: Routledge, pp. 52–63. Mackey, William F. 2005. Multilingual cities. In Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier, and Peter Trudgill (eds.), Sociolinguistics. An International Handbook of the Science of Language and Society, vol. 2/1. Berlin: de Gruyter, pp. 1304–1312. Mansel, Philip 2014. Cities of the Levant – the past for the future? Asian Affairs 45.2: 220–42. Matras, Yaron 2009. Language Contact. Cambridge: Cambridge University Press. Matras, Yaron 2013. Languages in contact in a world marked by change and mobility. Revue Française de Linguistique Appliquée: Langues en contact 18.2: 7–13. Matras, Yatron 2011. Universals of structural borrowing. In Peter Siemund (ed.), Linguistic Universals and Language Variation. Berlin: de Gruyter, pp. 204–33. Meakins, Felicity 2013. Mixed languages. In Peter Bakker and Yaron Matras (eds.), Contact Languages: A Comprehensive Guide. Berlin, New York: de Gruyter, pp. 159–228. Mensah, Eyo (ed.) 2016. Sociolinguistic Studies 10: 1–2 (special issue) The Dynamics of Youth Language in Africa. Mihm, Arend 2001. Oberschichtliche Mehrsprachigkeit und ’Language Shift’ in den mitteleuropäischen Städten des 16. Jahrhunderts. Zeitschrift für Dialektologie und Linguistik 68.3: 257–87. Miller, Catherine 2004. Variation and Changes in Arabic Urban Vernaculars. In Martine Haak, Kees Versteegh, and Rudolf Dejong (eds.), Approaches to Arabic Dialects: Collection of Articles presented to Manfred Woidich on the Occasion of his Sixtieth Birthday. Amsterdam: Brill. 177–206. Moran, Anthony 2014. Identity, race, and ethnicity. In Anthony Elliott (ed.), Routledge Handbook of Identity Studies. London: Routledge, pp. 170–85. Mufwene, Salikoko 2008. Language Evolution: Contact, Competition, and Change. London: Bloomsbury. Nassenstein, Nico and Andrea Hollington (eds.) 2015. Youth Language Practices in Africa and Beyond. Berlin: de Gruyter. Nortier, Jacomine and Dorleijn, Margreet 2013. Multi‐ethnolects: Kebabnorsk, Perkerdansk, Verlan, Kanakensprache, Straattaal, etc. In Peter Bakker and Yaron Matras (eds.), Contact Languages: a Comprehensive Guide. Berlin, New York: de Gruyter Mouton, pp. 229–71.

Nortier, Jacomine and Bente A. Svendsen (eds.) 2015. Language, Youth, and Identity in the 21st Century. Oxford: Oxford University Press. O’Quinn, Daniel and Teresa Heffernan (eds.) 2012. Lady Mary Wortley Montague: The Turkish Embassy Letters. Peterborough, ON: Broadview Press. Pavlenko, Aneta 2018. Superdiversity and why it isn’t. Reflections on terminological innovation and academic branding. In Stephan Breidbach, Lutz Küster, and Barbara Schmenk (eds.), Sloganizations in Language Education Discourse. Clevedon: Multilingual Matters. Pennycook, Alastair and Emi Otsuji, 2015. Metrolingualism. Language in the City. London: Routledge. Putzo, Christine 2011. Mehrsprachigkeit im europäischen Kontext. Zu einem vernachlässigten Forschungsfeld interdisziplinärer Mediävistik. In Michael Baldzuhn and Christine Putzo (eds.), Mehrsprachigkeit im Mittelalter. Kulturelle, literarische, sprachliche und didaktische Konstellationen in europäischer Perspektive. Mit Fallstudien zu den “Disticha Catonis.” Berlin, New York: de Gruyter, pp. 3–34. Quist, Pia 2000. Ny kobenhavnsk “multietnolekt.” Om sprogbrug blandt unge i sprogligt og kulturelt heterogene miljoer. Danske Talesprog, vol. 1. Institut for Dansk Dialektforskning. Copenhagen: C. A. Reitzels Forlag, pp. 143–212. Quist, Pia 2010. Untying the language‐body‐ place connection: a study on linguistic variation and social style in a Copenhagen community of practice. In Peter Auer and Jürgen Erich Schmidt (eds.), Language and Space: Theories and Methods. Berlin, New York: de Gruyter, pp. 632–648. Quist, Pia and Bente A. Svendsen (eds.) 2010. Multilingual Urban Scandinavia: New Linguistic Practices. Clevedon: Multilingual Matters. Rehbein, Ines, Sören Schalowski, and Heike Wiese, 2014. The KiezDeutsch Korpus (KiDKo) Release 1.0. In Nicoletta Calzolari, Khalid Choukri, Thierry Declerck, Hrafn Loftsson, Bente Maegaard, Joseph Mariani, Asunción Moreno, Jan Odijk, Stelios Piperidis (eds). Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14). Reykjavik, Iceland, pp. 3927–34. Schildt, Joachim and Harald Schmidt 1992. Berlinisch. Geschichtliche Einführung in die Sprache einer Stadt. Berlin: Akademie Verlag.

Contact in the City  279 Selting, Margret and Friederike Kern 2009. On some syntactic and prosodic structures of Turkish German in talk‐in‐interaction. Journal of Pragmatics 41: 2496–514. Shohamy, Elana, and Durk Gorter (eds.) 2009. Linguistic Landscape. Expanding the Scenery. London: Routledge. Şimşek, Yazgül 2012. Türkendeutsch: Formen und strukturelle Merkmale. In Barbara Jańczak, Konstanze Jungbluth, and Harald Weydt (eds.), Mehrsprachigkeit aus deutscher Perspektive. Tübingen: Narr, pp. 155–79. Stolz, Thomas. 2008. Total reduplication vs. echo‐word formation in language contact situations. In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages. Amsterdam: Benjamins, pp. 107–32. Strauss, Johann 2011. Linguistic diversity and everyday life in the Ottoman cities of the Eastern Mediterranean and the Balkans (late 19th–early 20th century). The History of the Family 16.2: 126–41. Tagliamonte, Sali A. and Alexandra D’Arcy 2009. Peaks beyond phonology: adolescence, incrementation, and language change. Language 85.1: 58–108. te Velde, John 2017. Temporal adverbs in the Kiezdeutsch left periphery: combining late merge with deaccentuation for V3. Studia Linguistica 71. 3: 301–36 Thomason, Sarah G. 2001. Language Contact: An Introduction. Washington, DC: Georgetown University Press. UN 2015. World Urbanization Prospects, The 2014 Revision. New York: United Nations. Vanderkerckhove, Reinhild 2010. Urban and rural language. In Peter Auer and Jürgen Erich Schmidt (eds.), Language and Space: Theories and Methods. Berlin: de Gruyter, pp. 315–32. Velupillai, Viveka 2015. Pidgins, Creoles and Mixed Languages. An Introduction. Amsterdam: Benjamins. Vertovec, Steven 2007. Super‐diversity and its implications. Ethnic and Racial Studies 30.6: 1024–54. Vogl, Ulrike 2012. Multilingualism in a standard language culture. In Matthias Hüning, Ulrike Vogl, and Olivier Moliner (eds.), Standard languages and multilingualism in European history. Amsterdam, Philadelphia: Benjamins, pp. 1–42.

Walkden, George 2017. Language contact and V3 in Germanic varieties new and old. Journal of Comparative Germanic Linguistics 20.1: 49–81. Wiese, Heike 2009. Grammatical innovation in multiethnic urban Europe: new linguistic practices among adolescents. Lingua 119: 782–806. Wiese, Heike 2011. The role of information structure in linguistic variation: evidence from a German multiethnolect. In Frans Gregersen, Jeffrey Parrott, and Pia Quist (eds.), Language Variation – European Perspectives III. Proceedings of ICLaVE 5. Amsterdam: Benjamins, pp. 83–95. Wiese, Heike 2013. What can new urban dialects tell us about internal language dynamics? The power of language diversity. Linguistische Berichte 19: 208–45. Wiese, Heike 2015. “‘This migrants’ babble is not a German dialect!” – the interaction of standard language ideology and “us”/ “them”‐dichotomies in the public discourse on a multiethnolect. Language in Society 44.3: 341–68. Wiese, Heike forthcoming. Urban contact dialects. In Salikoko Mufwene and Anna María Escobar (eds.), The Cambridge Handbook of Language Contact. Cambridge: Cambridge University Press. Wiese, Heike and Hans G. Müller 2018. The hidden life of V3: an overlooked word order variant on verb‐second. Linguistische Berichte: 202–23 (special issue 25) Non‐Canonical Verb Positioning in Main Clauses, Mailin Antomo and Sonja Müller (eds.). Wiese, Heike, and Nilgin Tanış Polat 2016. Pejoration in contact: m‐reduplication and other examples from urban German. In Rita Finkbeiner, Jörg Meibauer, and Heike Wiese (eds.), Pejoration. Amsterdam: Benjamins, pp. 243–68. Wiese, Heike and Ines Rehbein 2015. Coherence in new urban dialects: a case study. Lingua 172/3: 45–61 (special issue) Coherence, Covariation and Bricolage, Gregory Guy and Frans Hinskens (eds.). Wölck, Wolfgang (2002). Ethnolects – between bilingualism and urban dialect. In Li Wei, Jean‐Marc Dewaele, and Alex Housen (eds.), Opportunities and Challenges of Bilingualism. Berlin, New York: de Gruyter, pp. 157–70.

14 Linguistic Landscapes and Language Contact KINGSLEY BOLTON, WERNER BOTHA, AND SIU‐LUN LEE 1 Introduction Somewhere between the mid 1960s and late 1990s, the appearance and texture of international ­cities changed noticeably, in response to the growth of international capitalist economies, international airline travel on a previously unprecedented scale, the related growth of international tourism, and a number of other dynamics closely related to what came to be recognized as late‐twentieth century “globalization.” One material and very visible way in which cities around the world began to change was at the level of signage, which included advertising billboards, shop names, and public signage of all kinds. An obvious linguistic intruder in this process was the English language, which, as the language of such globally‐promoted brands as Coca‐Cola, KFC, Nike, and Starbucks, began to adorn the cityscapes of the vast majority of the world’s international cities, as well as a surprisingly large number of the world’s less important international cities (Bolton 2012). In this context, it was perhaps not surprising that a number of the early articles dealing with what has come to be called “linguistic landscape” (LL) research focused on the spread of English in the public signage of urban settings. Early studies of English signage in international settings included a number of articles in English Today, including Ross (1997) on shop names in Milan, McArthur (2000) on signage in Sweden and Switzerland, Schlick (2002) on Austria and Slovenia, MacGregor (2003) on Tokyo, and Dimova (2007) on Macedonia. These studies were perhaps significant in identifying and commenting on the noticeable Anglicization of public signage in various settings around the world, but typically lacked a clear theorization in explaining the dynamics of such phenomena, relying instead on explanations related to “creativity,” “prestige, style and modernity,” and “western consumerism.” Although the field of linguistic landscapes (LL) can be traced back to Spolsky and Cooper’s (1991) study of signage in Jerusalem, and to even earlier sources (Coulmas 2009), the popularity of linguistic landscapes as a field of inquiry in applied linguistics, multilingualism, and sociolinguistics is of very recent origin. Modern linguistic landscape studies are often dated from an article by Landry and Bourhis (1997), who defined the field of study as: “The language of public road signs, advertising billboards, street names, place names, commercial shop signs, and public signs on government buildings combines to form the linguistic landscape of a given territory, region, or urban agglomeration” (p. 25). Today, work on linguistic landscapes has become increasingly important in a number of overlapping fields, including applied linguistics, multilingualism, and sociolinguistics, and is not

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

282  Kingsley Bolton, Werner Botha, and Siu-Lun Lee predominantly concerned with English alone, but, rather, with the totality of linguistic and extra‐linguistic semiotics at play in public space in (typically) urban communities. Emergent themes in linguistic landscape research include globalization, language policy, multilingualism, multimodality, world Englishes and much else (Gorter 2012). Over the last 20 years, linguistic landscape research has become a very popular field for investigation, one that now has its own dedicated journal and a very large body of publications. Whether or not the current proliferation of research in the field will yield a coherent collection of sociolinguistic insights may be debatable, but the relevance of language contact to the field of linguistic landscape is indisputable. Since the first modern linguistic landscape studies of the early 2000s, approaches to research in the field have evolved from their early beginnings, which, for the most part, were focused on the quantitative analysis of signage and other texts on display in public spaces. For the purposes of this discussion, the development of the field might then be understood in terms of “first‐wave” LL studies as opposed to “second‐wave” LL studies, which attempt to move away from quantitative approaches towards qualitative approaches grounded in social theory, cultural studies, human geography and political analysis. In addition, recent commentators have also pointed to a “third wave” of LL studies, primarily concerned with the ways in which issues of identity and political conflict are played out in public spaces. This chapter sets out to survey the international frontline of research on the topic of linguistic landscapes, and complements this theoretical discussion with a case study from contemporary Hong Kong. As should be clear from the review of previous research, language contact is central to the whole enterprise of linguistic landscape research, whether of a quantitative, qualitative, or critical perspective. In its essence, LL research focuses on the multilingual and translingual intermeshing of languages in public spaces throughout numerous contexts, not least in the world’s global cities, as well as more local contexts experiencing globalization. In many contexts, this involves language contact between English and the national and local languages of particular nations, although LL research is by no means confined to language contact phenomena involving English. Evidence of this can be seen in the Benjamins journal Linguistic Landscape. The range of languages, and languages in contact, in research articles published by the journal is impressively diverse. Of these, English, predictably, receives most coverage (as a major topic of 34 articles), followed by Spanish (12), Arabic (11), Hebrew (6), Chinese (5), French (5), German (3), Russian (3), Amharic (2), Basque (2), Bengali (2), Italian (2), Malay (2), and Turkish (2). Other languages which were the subject of single articles included Albanian, Armenian, Azerbaijani, Bahasa Indonesia, Bangla, Bosnian, Breton, Catalan, Croatian, Estonian, Farsi, Finnish, Flemish, Galician, Greek, Hindi, Hokkien, Hungarian, Japanese, Latin, Maori, Persian, Polish, Quichua/ Kichwa, Serbian, Tamasheq, Tamil, Thai, Ukrainian, and Urdu (these counts were taken from the Linguistic Landscape Volume 1, issue 1–2, up to and including Volume 4, issue 3). It is hoped that this current chapter will serve to highlight the importance of language contact in linguistic landscape research, not only through the critical review of LL literature presented here, but also through a detailed case study of the linguistic landscape of a Hong Kong society, conflicted by both political and linguistic loyalties.

2  First‐wave Linguistic Landscape Studies Even though Landry and Bourhis (1997) was primarily focused on the social‐psychological dimensions of the macro‐sociolinguistic analysis of bilingual Canada, its primary definition of linguistic landscape as concerned with the analysis of “public road signs, advertising ­billboards, street names, place names, commercial shop signs, and public signs on government buildings” influenced a whole range of what might be called “first‐wave” linguistic

Linguistic Landscapes and Language Contact  283 landscape studies, the focus of which was primarily on the language of such signage, and the analysis of the linguistic landscape with reference to such dichotomies as governmental versus private signage and top‐down versus bottom‐up signage. In many of these early studies, the methodology was typically focused on the quantitative analysis of patterns of language use in such texts, although various first‐wave studies also tackle issues related to language policies. Arguably the first detailed study of linguistic landscape from a broadly‐based sociolinguistic perspective was the volume edited by Durk Gorter (2006), entitled Linguistic Landscape: A New Approach to Multilingualism. This volume comprised four substantive chapters reporting on LL studies in Israeli cities (Ben‐Rafael, Shohamy, Amara, and Trumper‐Hecht), Bangkok (Huebner), Backhaus (Tokyo), and Friesland and the Basque country (Cenoz and Gorter), together with a scholarly introduction and conclusion from the editor. The chapters in this volume might be collectively labeled first‐wave studies because they share a number of similarities in terms of approach and methodology. First, these studies largely adopted a quantitative approach, which involved the counting of particular units of linguistic signs; second, such studies focused on the description and analysis of linguistic signs/features rather than other semiotic units; third, they had a shared concern with multilingualism in public spaces; and finally they typically adopted a framework based on the distinction between top‐down (government) versus bottom‐up (non‐ government) signage. For example, Backhaus’ article on “Multilingualism in Tokyo: A look into the linguistic landscape” investigated official and nonofficial signage in the city of Tokyo, which, despite its reputation for monolingualism, revealed “an impressive diversity of languages other than Japanese” (Backhaus 2006: 52). In this study Backhaus surveyed the environs of 28 stations on the Yamanote railway line in the center of Tokyo. Within these 28 survey areas a total of 11,834 signs were identified. Of these, 19.6% were categorized (according to rather strict criteria) as “multilingual,” where very often Japanese was used together with foreign languages such as English (which was on 97.6% of all multilingual signs), Chinese (2.7%), Korean (1.7%), as well as 11 other languages with less than 1%. These included French, Portuguese, Spanish, Latin, Thai, Italian, Persian, Tagalog, German, Arabic, and Russian. Backhaus’ (2007) monograph on Linguistic Landscapes: A Comparative Study of Urban Multilingualism in Tokyo was based on his Ph.D. thesis on the same topic, and essentially constituted a longer and far more detailed account of the research discussed in his (2006) article in the Gorter volume. A third major “first‐wave” publication in establishing LL as a distinct field of study was the volume edited by Elana Shohamy and Durk Gorter (2009) entitled Linguistic Landscape: Expanding the Scenery. This landmark collection comprised 20 chapters on various aspects of linguistic landscape research, and incorporated multimodal theories to include also sounds, images, and graffiti’ and that “LL has a major role to take in activism in the domains of education and critical thinking” (Shohamy and Gorter 2009: 4). A fourth first‐wave study was the edited volume from Shohamy, Ben‐Rafael, and Barni on Linguistic Landscape in the City (2010), which covered similar ground to the three earlier works. Finally, the volume edited by Gorter, Marten and Van Mensel (2012) on Minority Languages in the Linguistic Landscape may also be regarded as first‐wave in orientation, in the sense that they focus on individual languages in the wider linguistic landscape.

3  Second‐wave Linguistic Landscape Studies The term “second‐wave” LL studies is something of an artificial construct as the break between earlier studies and the newer perspectives that emerge around 2010 is not necessarily clear‐cut, but this classification is used here in order to emphasize that a number of studies

284  Kingsley Bolton, Werner Botha, and Siu-Lun Lee were published after 2010 claiming to adopt a much wider view of LL as a field of inquiry, and challenging the earlier research paradigm. A typical feature of second‐wave studies is a preference for qualitative (including ethnographic) rather than quantitative methodologies, as well as an emphasis on the semiotics and multimodality of the linguistic landscape. In many senses, this break can be said to begin with Jaworski and Thurlow (2010). Jaworski and Thurlow’s edited volume entitled Semiotic Landscapes: Language, Image, Space (2010) attempted to make a conscious break from previous frameworks for LL by eschewing the term “linguistic landscapes,” in favor of a distinctly different approach, “to emphasize the way written discourse interacts with other discursive modalities: visual images, nonverbal communication, architecture and the built environment,” adding that, “we thus take semiotic landscape to mean, in the most general sense, any (public) space with visible inscription made through deliberate human intervention and meaning making” (Jaworski and Thurlow 2010: 2). The editors emphasize that their goal in this volume “is to move on from the predominantly survey‐based, quantitative approaches […] and also to complicate some of the taken‐for‐granted dichotomies in favor of more nuanced, genre‐ and context‐specific analyses of language in ‘landscape texts’” (14). Ultimately, they argue that: The city itself can be read as a text, as a festival of signs – an ‘iconosphere’, […] in which the tensions between the globalizing and localizing displays of words and images manifest the aggressive ideology and dominance of global capitalism and often struggling, local ­identities of communities rooted in ‘real’ or ‘imagined’ places. (Jaworski and Thurlow 2010: 31–2)

The volume includes a total of 13 chapters in addition to the introduction, with settings of investigation including Australia, Germany, Hong Kong, Ireland, the Isle of Man, Israel, Jamaica, Poland, Russia, South Africa, Switzerland, the UK, the US, and Wales. What makes this collection ‘second wave’ in orientation is arguably the foregrounding of extra‐linguistic and non‐linguistic (or non‐language) aspects of the semiotics of public spaces, with chapters on graffiti, the spatial constraints and orientations of Hong Kong school children, the commodification of silence in elite tourism, and the sign systems of national monuments and architecture. Nevertheless, there is also noticeable leakage with earlier research paradigms, and despite the editorial preference, the phrase “linguistic landscape” pops up irrepressibly throughout the whole book. A second study of this type is Blommaert’s (2013) monograph on Ethnography, Superdiversity and Linguistic Landscapes, where the author takes an ethnographic approach to the investigation of linguistic landscapes with particular reference to immigrant languages in the context of Antwerp, Belgium. Here, Blommaert’s discussion of LL is informed by Vertovec’s (2007) notion of “superdiversity,” which Blommaert explains by reference to (i) the end of the Cold War, and migration and movement from the former Soviet Union and Eastern Europe into Western Europe, as well as larger patterns of migration; and (ii) the emergence of the internet and mobile phones as new technologies, leading to “more complex forms of communication and knowledge circulation” (Blommaert 2013: 5). A third example of second‐wave study is the co‐authored book by Blackwood and Tufi (2015) on The Linguistic Landscape of the Mediterranean, which includes six substantive chapters on the Mediterranean context. Another book that might qualify for categorization as a second‐wave study is the volume co‐authored by Pennycook and Otsuji (2015) on Metrolingualism: Language in the City. Although this volume is only partly concerned with LL in any mainstream sense, the volume does make the attempt to expand the discoursal space on urban multilingualism, and the “spatial repertoires” of

Linguistic Landscapes and Language Contact  285 individuals and groups, with individual chapters dealing with such topics as ethnic businesses, linguistic diversity and contestation, mobility, food, restaurants, and street markets. Other edited collections of this type worth mentioning are Stroud, Peck, and Willams’ (2018) Making Sense of People and place in Linguistic Landscapes, and Pütz and Mundt (2019) on Expanding the linguistic landscape. Finally, a very recent second‐wave study is the co‐authored work by Eliezer Ben‐Rafael and Miriam Ben‐Rafael, on Multiple Globalizations: Linguistic Landscapes in World‐Cities (2019). As in earlier writings from the Ben‐Rafaels, the broad framework for their research is derived from the sociology of language, which, in this volume, is concerned with a number of related issues, including world‐cities and globalization, multiculturalism, and the national principle in language policies. Second‐wave studies of the kind discussed already, drawing inspiration from semiotics and recent explorations in the context of superdiversity, have also been complemented in recent years by linguistic landscape studies with a political orientation, explicitly concerned with LL and the politics of conflict, exclusion, and protest. Such studies may thus be regarded as constituting a “third‐wave” approach to this field.

4  Third‐wave Linguistic Landscape Studies The volume edited by Rubdy and Said (2015) entitled Conflict, Exclusion and Dissent in the Linguistic Landscape brings together a total of 13 substantive chapters dealing with LL issues in the context of political conflict and contestation. In her introduction to the volume, Rubdy states that the book “is anchored within current issues and debates in the field of linguistic landscape research […] and focuses on the dynamics of the linguistic landscape as a site of conflict, exclusion and dissent often arising from mechanisms of language policies, language politics, language hierarchies and the ethnolinguistic struggles engendered by them” (p. 1). Rubdy also goes on to explain the concern of the volume with “exclusion,” later noting that “[a] crucial dimension of the notion of exclusion in this book […] refers to the way lack of visibility of languages on street signs disfavours language/ethnic minorities” (p. 3). A second example of a “third‐wave” approach is the edited volume by Blackwood, Lanza, and Woldemariam (2016) entitled Negotiating and Contesting Identities in Linguistic Landscapes, which presents a wide‐ranging collection of essays on LL across multiple geographical ­contexts. Interestingly, in their discussion of theory, they describe three distinct approaches (or “waves”) of LL studies: In what may be referred to as the first wave of LL studies, building directly on the work of Landry and Bourhis (1997), quantitative methods dominated in the assessment of the ethnolinguistic vitality of minority languages […] This approach draws specifically on the social anthropology and sociocultural perspectives to the study of language, culture and identity. Increasingly in recent years, more qualitative approaches have in many ways taken the lead methodologically, although a combination of methods often prevails in newer studies, as illustrated in this volume. With the qualitative wave of studies, the third epistemological tradition, the participatory/relational perspective tradition, has enlightened many studies investigating how individuals construct, negotiate and contest identities in the public space. (Blackwood, Lanza and Woldemariam 2016: xvii)

Here, they also add that linguistic landscapes are also invoked by individuals and groups in acts of “belonging,” “engagement,” and “resistance” (xvii). The volume comprises 15 chapters, dealing with such locations as Belarus, Egypt, England, Ethiopia, France, Germany, Israel, Italy, Korea, Moldova, South Africa, Tanzania, and Tunisia.

286  Kingsley Bolton, Werner Botha, and Siu-Lun Lee

5  A Singular Contribution: Scollon and Scollon (2003) One important work that has had an immense impact on linguistic landscape studies, without actually invoking the rubric “linguistic landscape,” has been the monograph co‐ authored some 16 years ago by Ron and Suzie Scollon, entitled Discourses in Place: Language in the Material World. The aim of their volume was stated as presenting “the first systematic analysis of the ways we interpret language as it is materially placed in the world,” adding that their research is informed by perspectives from linguistic anthropology and social psychology, as well as “intercultural communication, sociolinguistics, cultural studies, semiotics, visual anthropology and sociology, and cultural geography” (xi). Interestingly, the term “linguistic landscape” does not occur once in the Scollons’ book, despite the fact that this has been one of the most impactful and insightful studies of LL ever to have been published. Instead of “linguistic landscape,” the Scollons prefer to talk about “geosemiotics,” defining the field thus: Geosemiotics is the study of the meaning systems by which language is located in the material world. This includes not just the location of the words on the page you are reading now but also the location of the book in your hands and your location as you stand or sit reading this. It includes the urban planning designs by which the streets in your city are laid out as well as the signs placed on those streets. It also includes the many forms of what Erving Goffman calls the interaction order – the ways we organize ourselves as single individuals or as conversational partners out for a lunchtime walk – through which we interact with these many semiotic systems of the world around us. (Scollon and Scollon 2003: x)

They later explain that the intellectual inspiration for this approach has been drawn from a number of sources, including Erving Goffman on interaction order, Peirce and Silverstein on indexicality, and Kress and van Leeuwen on semiotic structures. They draw numerous examples from the China and Hong Kong context, as well as many others from the US and across Europe. In a key section of the volume, the Scollons provide an anthropologically informed and layered explanation of how the semiotics of place contribute to the creation of “semiotic aggregates” in particular contexts. The key insight here is that there are very few places where single semiotic systems are in operation, and in most social contexts, multiple ­discourses and multiple semiotic systems co‐exist. Consider, they suggest, the semiotic aggregate of a crowded shopping street on a Saturday afternoon: The many different kinds of social interactions that can take place in a crowded Saturday afternoon shopping area are supported or actually encouraged by the wide public walkways, providing passage spaces, directly in conjunction with a high concentration of big department stores and other shopping venues with their own configurations of display spaces, passage spaces, special use spaces, and backstage areas. […] The many discourses which fall together in a single place such as this shopping area produce such a semiotic aggregate. We know we are in a shopping district, for example, because we see signs of different shops, or by the traffic directions placed there by the city, or by the kinds of pedestrians we see. Each of these are separate realizations of different semiotic actions, but together they form a place that we can read, and which tells us, for example, what sort of shopping might be done in that area of the city. (Scollon and Scollon 2003: 175)

The significance of the Scollon and Scollon (2003) volume on Discourses in Place is substantial, as their work represents a detailed and comprehensive attempt to construct a comprehensive theory as well as a practical methodology for investigating many aspects of (what later came to be called) linguistic landscape research, and has informed multiple waves of LL research from the early 2000s to the present, including Lou (2016, 2017).

Linguistic Landscapes and Language Contact  287

6 Linguistic Landscapes Applied: The Case of Hong Kong’s “Lennon Walls” In this section, we report on the ways in which research on linguistic landscape in a particular sociolinguistic context – that of contemporary Hong Kong – can inform our understanding of language change, language contact and language conflict. As is widely known, Hong Kong was one of the last British colonies to gain its independence in the final years of the twentieth century, when, in 1997, the celebrated “handover” of Hong Kong to mainland China took place. Linguistically, Hong Kong was very different from most other cities in mainland China, given its history of separate development during the colonial period, which, to some extent, promoted a high degree of autonomy in language matters, as well as in other issues of political and social importance. More specifically, on its return to the motherland, Hong Kong was a city where the dominant variety of Chinese Putonghua, the “common language,” or standardized Chinese of mainland China was not spoken, but Cantonese, originally the regional language of Guangdong Province, but, in modern Hong Kong, the dominant variety of “Chinese” used in a wide swathe of high domains throughout the community. In addition, by 1997, Hong Kong was a quasi‐city‐state where the government officially recognized two written languages, Chinese and English, and three spoken languages, Cantonese, Putonghua, and English, in a language policy that valorized “biliteracy and trilingualism” (Bolton, Bacon‐Shone, and Lee 2019). In present‐day Hong Kong, the results of the latest census indicate that Cantonese is the usual home language of 88.9% of the population (and spoken by an additional 5.7%). This compares with 4.3% who claim English as a home language (and 48.9% as “another” or second language), and 1.9% who claim Putonghua as their home language (and 46.7% stating that this is an additional language for them). Thus, the current totals for self‐reported knowledge of the three major languages as a “usual” and “another” language are 94.6% for Cantonese, 53.2% for English, and 48.6% for Putonghua. These figures give at least a broad indication of the extent of trilingualism in Hong Kong society, notwithstanding the fact that there are obvious limitations to a reliance on self‐reported levels of language proficiency in censuses and language surveys (Bacon‐ Shone, Bolton and Luke 2015).

6.1  Hong Kong’s Linguistic Landscape Hong Kong provides a fascinating place for studies of the linguistic landscape, given the presence of multiple varieties of Chinese, as well as other Asian and European languages. However, the one linguistic fact that distinguishes Hong Kong from every other city in mainland China (including Guangzhou, i.e. Canton) is that the vast majority of people regard Cantonese as their first language, and that Hong Kong is “the greatest Cantonese city that the world has ever seen” (Harrison and So 1997: 12). The widespread use of Cantonese in Hong Kong society, in high domains as well as low, is very different from the official policy in mainland China, which promotes Putonghua (the “common language”), together with simplified Chinese characters instead of the “full characters” used in Hong Kong and Taiwan. From an official perspective in mainland China, Cantonese is regarded as a “regional dialect,” but, in Hong Kong, the language is seen by the vast majority of the community as the essential language of the society. Given the wide use of the language in education, religion, the print and broadcast media, and government, “the status of Cantonese is much higher than is normally thought and cannot be simply brushed aside as the ‘vernacular’” (Sin and Roebuck 1996: 252). Since the 1990s, the Hong Kong government has promoted a policy of “biliteracy and trilingualism,” which refers to literacy in written Chinese and English, and a spoken

288  Kingsley Bolton, Werner Botha, and Siu-Lun Lee command of Cantonese, Putonghua, and English. In stark contrast, China’s official language policy since the 1950s has aimed at the spread of Putonghua, the removal of illiteracy, the promotion of simplified rather than “full” Chinese characters and the romanization system of pinyin (Bolton and Lam 2006: 350). Language policy in the People’s Republic of China was enshrined in a new language law in October 2000, “The Law of the National Commonly Used Language and Script of the People’s Republic of China,” which stipulates that: “Schools and other educational organizations will take Putonghua and standard Chinese characters as the basic language and characters to be used in teaching and study” (Rohsenow 2004: 41; Zhang and Yang 2004: 154). This law has now been promoted in most other Chinese cities, including Guangzhou, which is currently experiencing a language shift from Cantonese to Putonghua (Lai 2009). In Hong Kong, too, since the 1997 handover, moves have been made to promote the national language in local schools, and in 1998, Putonghua became a compulsory subject in all Hong Kong schools, and today more and more children are learning Putonghua as a second (or third) language. However, despite initiatives to promote Putonghua in schools and in other domains, it is evident that the vast majority of Hong Kong people have a strong attachment to Cantonese, and that the spoken language, combined with the use of traditional Chinese characters continue to serve as crucial markers of Hong Kong identity. Published research on linguistic landscapes in Hong Kong have been relatively sparse since Scollon and Scollon’s (2003) pioneering work. Those that have occurred have focused on a wide range of topics related to linguistic landscapes and the semiotics of public spaces in the Hong Kong community. Jaworski and Yeung (2010) studied residential signage and naming practices in the local property market; Lai (2013) surveyed bilingual/multilingual signage throughout the city; and Luk (2013) examined language play and linguistic creativity. The two articles from Danielewicz‐Betz and Graddol (2014) and Graddol and Danielewicz‐Betz (2015) both examine linguistic variation along the border between Hong Kong and mainland China, while Lou’s (2016, 2017) research adopts a geosemiotic approach to restaurant signage, and public markets. Wong and Chan (2018) construct a linguistic landscape by studying bilingual signage through the filter of photographs in the government’s Hong Kong Year Book, from 1957 until 2014.

6.2  Hong Kong’s Lennon Walls It is widely believed that the first “Lennon Wall” was created in Prague, after the murder of John Lennon in New York in 1980. Although the wall was ostensibly an open forum for Lennon fans to express their grief at the loss of the singer, the Prague wall also became a space for expressing dissent with the communist government of the day, through lyrics, painted graffiti, poetry and other forms of art (Kwok, 2019). Thirty‐four years later, Hong Kong’s protestors created a large Lennon Wall of their own during the Occupy Movement of 2014, situated near to the Legislative Council (the Hong Kong equivalent of Parliament). The Occupy Movement (or “Umbrella Movement”) at that time arose in response to a number of factors, including dissatisfaction with economic and social conditions for young people, as well as increasing frustration with the Hong Kong government’s reluctance to expand democratic voting rights in the community (as was originally promised at the time of the handover to China in 1997). For three months in 2014, thousands of young people occupied major roads in the center of the city, and created sites of civil disobedience resembling rock festivals. The Lennon Wall that protestors created at that time did not use the spray paint and colored chalk of the Prague wall, but instead utilized that most mundane item of office stationery, the Post‐it note. Suddenly the large wall close to the legislature was full of thousands of Post‐it messages, as large numbers of people, reveled in this new way to express their concerns and hopes, and to make their voices heard (see Figure 14.1 below).1

Linguistic Landscapes and Language Contact  289

Figure 14.1  The 2014 Lennon Wall in Hong Kong.

Eventually, in December 2014, the Occupy protestors lost the sympathy of the public, and the demonstrators withdrew. Over the four and a half years that followed, the protest movement in Hong Kong was subdued, and a number of the leaders of the movement were arrested and imprisoned. By early 2019, it appeared that the territory’s pro‐Beijing government was in the ascendancy, and that the previously rebellious Hong Kong public had been cowed into submission. In February, the government of the China‐appointed Chief Executive, Carrie Lam, announced the introduction of new legislation (an ‘Extradition Bill’) that would allow the police to send alleged criminals in Hong Kong to mainland China to face trial (in a criminal justice system less responsive to an internationally‐ recognizable form of “rule of law”). As a result of this, protests were rekindled in Hong Kong, with various street demonstrations against the bill, including a march to government headquarters of an estimated 1 million people on June 9, followed by a second huge demonstration of 2 million people on June 16. Numerous demonstrations followed throughout the turbulent summer months, with many violent clashes between police and demonstrators (BBC 2019). On September 4, Carrie Lam announced the withdrawal of the extradition bill, but despite that, at the time of writing (mid September 2019) street protests by young people are continuing. During the protests of 2019, the Lennon Walls reappeared in Hong Kong, but this time with scores of Lennon Walls scattered throughout the city, in universities and colleges, libraries, mass transit railway (MTR) stations, pedestrian underpasses, and a wide range of other public spaces (see Figures  14.2 and 14.3). By mid July, there were an estimated 120 walls across the 18 administrative districts of the city. At first sight, the Hong Kong Lennon Walls present an incongruously‐familiar disruption to an architecture of public spaces that relies on the usual modernist configuration of urban concrete and steel girders. The colored and fragile paper notes are familiar accessories for the city’s enlisted office workers and over‐burdened students, and are now re‐deployed from their usual uses as aides‐memoire, to provide a means of expression for the voices of Hong Kong people. Yu (2019) describes the texture of the Lennon Wall close to the University of Hong Kong, in Shek Tong Tsui, in the following way: The Lennon Wall [here] encircles a pillar supporting a flyover that runs down a hill between apartment buildings. From the bottom of the narrow street, the wall looks like a neon‐ coloured sock on the foot of an elephant. When you get up close, you find hundreds of

290  Kingsley Bolton, Werner Botha, and Siu-Lun Lee

Figure 14.2  Taipo MTR station (July 2019).

Figure 14.3  Shatin bus terminal (July 2019).

sticky notes secured with tape, bearing messages that exhort (‘Stand Together Hong Kong’), curse (the administration; Chief Executive Carrie Lam, pro‐Beijing politicians, the police); that address mainlanders with solidarity (‘Democracy is a Good Thing’), with xenophobia (middle fingers, ‘Go Home’); that appeal (‘Five Demands for the government’), or pay homage (‘I love Denise Ho’  –  a well‐known Canto‐pop singer). Clear plastic wrap is wound around the notes to protect them against the elements, including human tampering. When I visited it earlier this month, a small group of college‐aged volunteers were keeping watch over the wall while handing out memo pads to passersby, inviting them to add their thoughts. (Yu 2019: 13)

The vast majority of messages and slogans written on the Lennon Walls are written in Chinese characters, in particular the Hong Kong version of Modern Standard Chinese traditional (full)

Linguistic Landscapes and Language Contact  291 characters, although these have also been supplemented by the use of colloquial Cantonese characters, as well as a smattering of English slogans (“Be with us,” “Fight for freedom,” “God Bless HK,” “HK add oil!,” “Stop shooting us,” “Together we stand”). Yu suggests that the walls serve a range of functions, as “part artwork, part bulletin,” but also to “raise awareness and for emotional support,” providing “a chance to pause and heal” (Yu 2019).

6.3  Reading Hong Kong’s Lennon Walls In August 2019, the authors of this chapter set out to document the use of languages by conducting a focused research project on the Lennon Walls of the city. The overarching aim of this study was to investigate how the use of languages in such contexts reflected the dynamics of multilingualism in Hong Kong society today. More specifically, our study set out (i) to describe patterns of language use on the Lennon Walls during the 2019 Hong Kong ­anti‐extradition bill protests (which, at the time of writing this chapter, are still an ongoing event); and (ii) to explain such patterns with reference to the dynamics of language contact and language conflict in contemporary Hong Kong society.

6.3.1 Methodology As discussed already in connection with Lennon Walls appearing in numerous locations in Hong Kong, with Post‐it notes, artwork, pieces of paper, and posters, one assumption was that the “language” of the walls, their images, messages and slogans, would resonate with the wider sociolinguistic realities of the Hong Kong community. For a period of two weeks in late‐August, the researchers then began compiling a visual database of the Lennon Walls across Hong Kong. In preparation for this report on the project, we initially took 338 photographs of walls in various locations of Hong Kong, including Causeway Bay (on Hong Kong Island), Kwun Tong (the MTR station), and Shatin (the Chinese University of Hong Kong and New Town Plaza). Given the time‐scale and the unfolding series of events, this was not designed as a first‐wave quantitative project, but envisioned very much as an exploratory second‐/third‐wave study, interested in both the semiotics of multilingual diversity as well as issues of language contact and language conflict. What follows is that our discussion of results adopts a selective approach to analyzing what might be judged as salient features of linguistic and semiotic communication on the Lennon Walls that we photographed. Here, examples of language use are taken from our own database of photographs, supplemented by other illustrative material from the Creative Commons and newspaper sources, which serve to situate our research more broadly in the context of the Hong Kong protest movement of 2019.

6.3.2 Results Overall, it was clear that the most frequently used language was the Hong Kong version of Standard Written Chinese, with traditional (or full) Chinese characters dominating the linguistic landscape of Post‐its and posters. The dominance of this writing system can be seen in Figure 14.4 below, which was taken at the Lennon Wall in the Kwun Tong MTR (Mass Transit Railway) station. In Figure 14.4, three of the Post‐it messages contain the expression 加油 gā yáu (‘add oil’). ‘Add oil’ is explained in the OED online “as an exclamation expressing encouragement or support,” broadly equivalent to “Come on!,” or “Go for it!” We have an example of this in (1) below, which can be seen slightly right of center in Figure 14.4 on p. 292. (1)

香 港 加 油! Hēung góng gā yáu! ‘Hong Kong, add oil!’

292  Kingsley Bolton, Werner Botha, and Siu-Lun Lee

Figure 14.4  Lennon Wall at the Kwun Tong MTR (1).

Some Post‐its in Figure 14.4 have somewhat longer messages, as in the four‐line message close to the lower right‐hand corner of the photograph, which is set out below: (2)

香 港 人 加 Hēung góng yàhn gā ‘Hong Kong people, add oil!’ 團 結 才 能 tyùhn git chòih nàhng ‘Unite together and break.’ 高 牆 gōu chèuhng ‘the high wall’ 永 不 放 棄! wíhng bāt fong hei!

油! yáu! 打 dá

破 po!

‘Never give up!’ Example (2) above is a message encouraging the protestors to unite and to stand up against injustice (‘the high wall’), which might refer to the extradition bill itself, but also in addition has the potential to index a range of protestors’ grievances, including police brutality, Beijing’s interference in Hong Kong affairs, the lack of real democracy, and so on. Although less frequent, the English language was also visible in a number of the Lennon Walls, as in Figure 14.5 below, which again is a photograph of a Lennon Wall near Kwun Tong MTR station, where we have the phrases “We have nothing to lose,” and “Let’s fight!!” These latter examples express encouragement to the protestors, while simultaneously indicating the bilingual abilities of the protestors, many of whom are young college and university students who have at least a basic command of English. Other messages, however, need somewhat more contextualization and explanation, as with the four‐line English message in the bottom right‐hand corner of the photograph, which states HongKongers/Be water/Be Great/Add Oil. In fact, the admonition Be water, which became an important slogan in the 2019 protests, draws on a celebrated quotation from Bruce Lee,

Linguistic Landscapes and Language Contact  293

Figure 14.5  Lennon Wall at the Kwun Tong MTR (2).

who in a 1971 televised interview explained his personal martial arts style of 截拳道 jiht kyùhn douh in the following way: I said empty your mind. Be formless, shapeless, like water. Now you put water into a cup, it becomes the cup. You put water into a bottle, it becomes the bottle. You put it in a teapot, it becomes the teapot. Now water can flow or it can crash. Be water, my friend. (Lee 1971)2

Bruce Lee’s suggestion of “being water,” may be understood, in martial arts, to refer to the need of being adaptable and flexible in any fighting situation, and in the summer of 2019, demonstrators took this advice to heart in planning any executing highly mobile and ­shifting demonstrations throughout numerous locations in the city. Thus these four lines simultaneously index Hong Kong culture, history, martial arts, and society, as well as (via translation) the Chinese language. In other examples, we also observed that the posters, Post‐its, and slogans are bilingual, containing both Chinese characters and written English, often English translations of the Chinese texts. In some cases, the translations from the original Chinese texts might be regarded as an irreverent of “Hong Kong English” or at least “Hong Kong word play,” utilizing deliberate mistakes in translation, as can be seen in Figure 14.6, and Example (3). Even though the English text in Figure 14.6 is a direct translation from the Chinese text below it, the English version makes no sense to anyone unfamiliar with Chinese (traditional) characters, and the meaning potential of the text. (3)

  STOP CHARGING    OR WE KISS YOUR SPRING POCKET 停 止 衝 tìhng jí chū ng 否 則 錫 你 fáu jāk sek néih

擊 gı̄k 春 chēun

袋 doih

294  Kingsley Bolton, Werner Botha, and Siu-Lun Lee

Figure 14.6  Poster on a Lennon Wall at the Chinese University of Hong Kong (1).

In order to understand the text in (3) above, it is necessary to know that the phrase 春袋, chēun doih (‘spring pocket’) is a euphemism for ‘balls’ or testicles’. However, in an online discussion of this phrase, a netizen commentator used the wrong Chinese character for ‘kick’, which should have been 踢 tek, but instead wrote 錫 sek, which translates as ‘kiss’. Other netizens then gleefully adopted this mistake and translated the phrase directly into English, thus creating an interesting example of bilingual banter (Apple Daily 2019). Another example of language choice and mixing is illustrated in Figure 14.7, where the text accompanying the cartoon policeman holding a gun in the photograph reads as: (4)

開 槍 so what? Hōi chēung so what? ‘I shoot you so what?’ 洗 唔 洗 wtsapp sái m sái WhatsApp ‘Do I need to WhatsApp’ 你 néih

先? sı̄n?

‘You first?’ What is interesting here is that the written form of Chinese here deploys colloquial Cantonese characters 洗唔洗, sái m sái (‘need not need’). In Standard Written Chinese, this phrase would be written as 要不要 yiu bāt yiu. The expression 洗唔洗 (sái m sái) would be incomprehensible to a Putonghua Chinese speaker, given that the character 洗 (sai) translates as ‘to wash’ in Putonghua, and the 唔 (m) character may be understood as ‘oh’ in Putonghua. Thus, in Putonghua, the literal translation of the phrase 洗唔洗 would correspond to ‘wash oh wash’, and appear nonsensical to a Putonghua speaker. In addition, the insertion of the English

Linguistic Landscapes and Language Contact  295

Figure 14.7  Close up of the Lennon Wall at the Chinese University of Hong Kong (2).

phrases, so what and wtsapp (as a verb) into the two questions in Cantonese expressions serves to mimic the routinely code‐switched speech of young people in Hong Kong. The examples of language contact and language choice in the above examples, underline Scollon and Scollon’s (2003: 2) view that “the study of the social meaning of the material placement of signs” are an important consideration when interpreting the meaning of language in public display. The linguistic landscapes of Hong Kong’s Lennon Walls during the current protests illustrate how the language used on these walls reflects and indexes the broader sociolinguistic realities and conflicts of the wider society. What is particularly striking in this context is that the language used in the messages, posters, Post‐its, and slogans of the Lennon Walls is overwhelmingly local language, reflecting the long‐established norms of a modern Hong Kong that developed in the late colonial period through to its troubled postcolonial present. Thus, the local linguistic repertoire dominating the language world of Hong Kong people has a variety of linguistic resources at its disposal, which are very different from the linguistic codes of contemporary communist China. These local linguistic resources include (i) spoken Cantonese, (ii) the full (or “traditional”) characters of the Hong Kong version of Modern Standard Chinese (MSC), (iii) locally‐used “Cantonese characters” which have a long tradition of use in Guangdong and Hong Kong, (iv) code‐ mixed Cantonese and English, and (v) romanized Cantonese, the use of which again has its own history in southern China and Hong Kong.3 The use of romanized script appears to have been given a boost in the recent protests, where romanized slogans have been inserted into announcements, messages, and posters, allegedly in order that this strategy would make it difficult for any non‐Cantonese speaker to understand the meanings of such messages. It is claimed that this became popular after a protest website called LIHKG began using romanization in order to baffle various parties in mainland China who were monitoring communications on the website. One example of this can be seen in Figure 14.8, where the poster reads gwong fuk heung gong, si doi gak ming (‘take back [or liberate] Hong Kong, revolution of our times’), which has become a popular slogan during recent protests. In the same photograph, you can also see the phrase wai dor lei ah gung yuen, which is a partly phonological rendering of ‘Victoria Park’, followed in turn by

296  Kingsley Bolton, Werner Botha, and Siu-Lun Lee

Figure 14.8  Poster announcing protest.

the phrase heung gong yan dou si gin! ‘Hong Kong people arrive there meet’ with the numbers 8.18, in other words, announcing a planned protest in Victoria Park on August 18. This call to protest was also splashed across the Apple Daily newspaper’s front page on 18 August, where the lead headline is 今日 wai yuen gin (‘Today Vic. Park meet’, while further down the page, the protestors are reminded to be water (Figure 14.9).

7 Conclusion In this chapter, we have set out a survey of linguistic landscape research from the late 1990s up to the recent past. As is evident from this review of the field, linguistic landscapes are currently a major topic of research in such fields as applied linguistics, language policy, ­multilingualism, and sociolinguistics. Since the early 2000s, a large number of book‐length studies of linguistic landscapes have been published, and from 2015 this field of study now has its own dedicated journal, Linguistic Landscape (Benjamins). Given the scope of contemporary linguistic landscape studies, the relevance of the LL field to the issue of language contact is undeniable. However, relatively little linguistic landscape research has been focused on linguistic variation from a traditional sociolinguistic perspective, but instead

Linguistic Landscapes and Language Contact  297

Figure 14.9  Front page of Apple Daily newspaper.

adopted approaches owing more to discourse analysis, multimodality, semiotics, and much else. The different approaches to research that have emerged in the field have reflected a wide range of diverse interests. First‐wave studies typically adopted a quantitative approach, and were often concerned with the distinction between official, top‐down and private, ­bottom‐up signage. Second‐wave studies attempted to move away from a quantitative methodology, and to draw more widely on cultural and social theory, as well as the methodologies and lexicon of semiotic research. Third‐wave studies, by contrast, have been typically concerned with questions of political and social conflict, identity, and progress, as well as issues of social justice. It is thus apparent that, at a theoretical level, there is not a single unified theory, but rather a number of overlapping and sometimes competing theories within the field. Even this categorization of approaches (or waves) is not unproblematic, given the frequent overlap and leakage between first and second, and second and third waves in the research literature. At the same time, when it comes to methodology, there is a similar diversity, if not diffuseness, in relation to materials and method. One of the major reasons for the growing popularity of LL research over the last 20 years has been technological innovation, or more

298  Kingsley Bolton, Werner Botha, and Siu-Lun Lee specifically, the development of digital techniques and the availability and ease of capturing images through smartphone photography. Despite this (or indeed because of this), there is an obvious need for a rigorous methodology in this field. As early as 2006, Gorter raised a number of methodological issues, including the problems of sampling, the identification of signs, the categorization of signs, as well as the issue of sampling, and asking: “what constitutes such an object or sign?” and “what constitutes the unit of analysis?” (Gorter 2006: 3–4). More recently, Gorter has highlighted the importance of technological innovations to LL studies, suggesting that the growth of LL studies has been directly linked to digital technology, and the increasing mediatization of society so that “[d]igital screens can be observed everywhere and all those screens contain continuously changing ‘language.’” The distinction between offline and online worlds is blurring and both are part of the “‘linguistic landscape’ that its users are immersed in’” (Gorter 2019: 50–1). Whether or not Gorter’s view of future LL research dominated by digital technology is accurate may be debatable, but evidence for that vision is available in the continuing and shifting urban semiotic landscape of the Hong Kong community right now. Today, in the ongoing campaign for democracy, the political debates of the day are spelled out in varieties of localized language across a range of media, including Facebook and YouTube, Bridgefy, FireChat, and Telegram, and other forms of social messaging, as well as the posters, Post‐its, and street art of Hong Kong’s colorful Lennon Walls. The Lennon Walls of 2019 are papered with the hopes and dreams of a society dealing with political and linguistic conflict, as well as the shifting realities of a liberal society facing authoritarian rule and the perceived antagonism towards its vernacular language, Cantonese, in a situation of multilingual contact.

NOTES 1 Figures  14.1–143 are from the Wikipedia Creative Commons archives. Figure  14.1 is from 2014, where the photographer’s name is given as Ceeseven. Available at https://commons.wikimedia. org/wiki/File:Close_view_of_Hong_Kong_Lennon_Wall_on_2014‐11‐08.JPG. The photographer for Figure 14.2 is identified as Studio Incendo. Available at https://commons.wikimedia.org/wiki/ File:Hong_Kong_IMG_ 20190712_230551_ (48266553157).jpg. The photographer for Figure  14.3 is identified as Wpcpey. Available at https://commons.wikimedia.org/wiki/File:Shatin_Station_ Bus_Terminal_Lennon_Wall_201907.jpg. 2 The interview with Bruce Lee can be accessed on YouTube at the following URL: www.youtube. com/watch?v=uk1lzkH‐e4U 3 An illuminating YouTube film was made by Professor David Clarke of the University of Hong Kong on the Lennon Wall in Tai Po. This can be accessed at the following link: www.youtube.com/ watch?v=ex3ElMlVo8o

REFERENCES Apple Daily 2019. 大大力錫佢春袋 – 邁克 [Strongly kiss spring pocket – Mike]. Apple Daily. Available at https://hk.lifestyle.appledaily.com/lifestyle/ columnist/%E9%82%81%E5%85%8B/daily/ article/20190808/20744819 Backhaus, Peter 2006. Multilingualism in Tokyo: a look into the linguistic landscape. In Durk

Gorter (ed.), Linguistic Landscape: A New Approach to Multilingualism. Clevedon: Multilingual Matters, pp. 52–66. Backhaus, Peter 2007. Linguistic Landscapes: A Comparative Study of Urban Multilingualism in Tokyo. Clevedon: Multilingual Matters.

Linguistic Landscapes and Language Contact  299 Bacon‐Shone, John, Kingsley Bolton, and Kang Kwong Luke 2015. Language Use, Proficiency and Attitudes in Hong Kong. Hong Kong: Social Sciences Research Centre, University of Hong Kong. BBC (2019). Hong Kong: Timeline of extradition protests. BBC News, September 4. Available at www.bbc.com/news/world‐asia‐china‐ 49340717 Ben‐Rafael, Eliezer and Miriam Ben‐Rafael 2019. Multiple Globalizations: Linguistic Landscapes in World‐Cities. Leiden: Brill. Ben‐Rafael, Eliezer, Elana Shohamy, and Monica Barni (eds.) 2010. Introduction: an approach to “ordered disorder.” In Elana Shohamy, Eliezer Ben‐Rafael, and Monica Barni (eds.), Linguistic Landscape in the City. Clevedon: Multilingual Matters, pp. xi–xxviii. Blackwood, Robert J., Elizabeth Lanza and Hirut Woldemariam (eds.) 2016. Negotiating and Contesting Identities in Linguistic Landscapes. London: Bloomsbury Academic. Blackwood, Robert J. and Stefania Tufi 2015. The Linguistic Landscape of the Mediterranean: French and Italian Coastal Cities. Basingstoke: Palgrave Macmillan. Blommaert, Jan 2013. Ethnography, Superdiversity and Linguistic Landscapes: Chronicles of Complexity. Bristol: Multilingual Matters. Bolton, Kingsley 2012. World Englishes and linguistic landscapes. World Englishes 31.1: 30–3. Bolton, Kingsley, John Bacon‐Shone and Siu‐lun Lee 2019. Societal multilingualism in Hong Kong. In Peter Siemund and Jakob Leimgruber (eds.), Multilingual Global Cities: Dubai, Hong Kong and Singapore. London: Routledge. Bolton, Kingsley and Agnes S. L. Lam 2006. Applied linguistics in China. In Kevin Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edition, vol. 1. Oxford: Elsevier, pp. 350–6. Coulmas, Florian 2009. Linguistic landscaping and the seed of the public sphere. In Elana Shohamy and Durk Gorter (eds.), Linguistic Landscapes: Expanding the Scenery. New York: Routledge, pp. 13–24. Danielewicz‐Betz, Anna and David Graddol 2014. Varieties of English in the urban landscapes of Hong Kong and Shenzhen. English Today 30.3: 22–32. Dimova, Slobodanka 2007. English shop signs in Macedonia. English Today 23: 18–24. Gorter, Durk 2006. Linguistic Landscape: A New Approach to Multilingualism. Clevedon: Multilingual Matters.

Gorter, Durk 2012. Linguistic landscape. In Carol A. Chapelle (ed.), The Encyclopedia of Applied Linguistics. Online edition. Oxford: Blackwell, pp. 1–5. Gorter, Durk 2019. Methods and techniques for linguistic landscape research: about definitions, core issues and technological innovations. In Martin Pütz and Neele Mundt (eds.), Expanding the Linguistic Landscape: Linguistic Diversity, Multimodality and the Use of Space as a Semiotic Resource. Bristol: Multilingual Matters, pp. 38–57. Gorter, Durk, Heiko F. Marten, and Luk Van Mensel (eds.) 2012. Minority Languages in the Linguistic Landscape. Basingstoke: Palgrave Macmillan. Graddol, David and Anna Danielewicz‐Betz 2015. Borderland English: signs of transition across the expiring China‐Hong Kong border. Asian Englishes 17.1: 3–28. Harrison, Godfrey and Lydia K. H. So 1997. The background to language change in Hong Kong. In Sue Wright and Helen Kelly‐Holmes (eds.), One Country, Two Systems, Three Languages: A Survey of Changing Language Use in Hong Kong. Clevedon: Multilingual Matters, pp. 8–17. Jaworski, Adam and Crispin Thurlow (eds.) 2010. Semiotic Landscapes: Language, Image, Space. London: Continuum. Jaworski, Adam and Simone Yeung 2010. Life in the Garden of Eden: the naming and imagery of residential Hong Kong. In Elana Shohamy, Eliezer Ben‐Rafael, and Monica Barni (eds.), Linguistic Landscape in the City. Bristol: Multilingual Matters, pp. 153–81. Kwok, Ben 2019. The spirit of Lennon Wall never dies. EJInsight. Available at www.ejinsight. com/20190820‐the‐spirit‐of‐lennon‐wall‐ never‐dies/ Lai, Chloe 2009. Linguistic heritage in peril. South China Morning Post, October 11, p. 12. Lai, Mee Ling 2013. The linguistic landscape of Hong Kong after the change of sovereignty. International Journal of Multilingualism 10.3: 251–72. Landry, Rodrigue and Richard Y. Bourhis 1997. Linguistic landscape and ethnolinguistic vitality: an empirical study. Journal of Language and Social Psychology 16: 23–49. Lee, Bruce 1971. Interview with Pierre Burton. YouTube. Available at www.youtube.com/ watch?v=uk1lzkH‐e4U Lou, Jackie Jia 2016. Shop signs as monument: the discursive recontextualization of a neon sign. Linguistic Landscape 2.3: 211–22.

300  Kingsley Bolton, Werner Botha, and Siu-Lun Lee Lou, Jackie Jia 2017. Spaces of consumption and senses of place: a geosemiotic analysis of three markets in Hong Kong. Social Semiotics 27.4: 513–31. Luk, Jasmine 2013. Bilingual language play and local creativity in Hong Kong. International Journal of Multilingualism 10.3: 236–50. MacGregor, Laura 2003. The language of shop signs in Tokyo. English Today 19: 18–23. Marten, Heiko F., Luk Van Mensel, and Durk Gorter 2012. Studying minority languages in the linguistic landscape. In Durk Gorter, Heiko F. Marten, and Luk Van Mensel (eds.), Minority Languages in the Linguistic Landscape. Basingstoke: Palgrave Macmillan, pp. 1–15. McArthur, Tom 2000. Interanto: the global language of signs. English Today 16: 33–43. Peck, Amiena, Christopher Stroud, and Quentin Willams (eds.) 2018. Making Sense of People and Place in Linguistic Landscapes. London: Bloomsbury Academic. Pennycook, Alastair and Emi Otsuji 2015. Metrolingualism: Language in the City. Abingdon: Routledge. Pütz, Martin and Neele Mundt (eds.) 2019. Expanding the Linguistic Landscape: Linguistic Diversity, Multimodality and the Use of Space as a Semiotic Resource. Bristol: Multilingual Matters. Rohsenow, John S. 2004. Fifty years of script and written language reform in the P. R. C. In Minglang Zhou and Hongkai Sun (eds.), Language Policy in the People’s Republic of China: Theory and Practice since 1949. Boston: Kluwer Academic, pp. 21–43. Ross, Nigel J. 1997. Signs of international English. English Today 13: 29–33. Rubdy, Rani and Selim Ben Said (eds.) 2015. Conflict, Exclusion and Dissent in the Linguistic Landscape. Basingstoke: Palgrave Macmillan. Schlick, Maria 2002. The English of shop signs in Europe. English Today 18: 3–7.

Scollon, Ron and Suzie Wong Scollon 2003. Discourses in Place: Language in the Material World. London: Routledge. Shohamy, Elana and Durk Gorter (eds.) 2009. Linguistic Landscapes: Expanding the Scenery. New York: Routledge. Shohamy, Elana, Eliezer Ben‐Rafael, and Monica Barni (eds.) 2010. Linguistic Landscape in the City. Clevedon: Multilingual Matters. Sin, King‐Kui and Derek Roebuck 1996. Language engineering for legal transplantation: conceptual problems in creating common law Chinese. Language and Communication 16: 235–54. Spolsky, Bernard and Robert L. Cooper 1991. The Languages of Jerusalem. Oxford: Clarendon Press. Stroud, Christopher, Amiena Peck, and Quentin Williams 2018. Introduction. In Amiena Peck, Christopher Stroud, and Quentin Williams (eds.), Making Sense of People and Place in Linguistic Landscapes. London: Bloomsbury Academic, pp. 1–7. Vertovec, Steven 2007. Super‐diversity and its implications. Ethnic and Racial Studies 30.6: 1024–54. Wong, Alicia S. H. and Susan S. S. Chan 2018. From “the world of Suzie Wong” to “Asia’s world city”: tracing the development of bilingualism in Hong Kong’s linguistic landscape (1957–2014). International Journal of Bilingualism 15.4: 435–54. Yu, Yvonne Y. 2019. Protest on paper. The Times Literary Supplement. Zhang, Bennan and Robin Yang 2004. Putonghua education and language policy in postcolonial Hong Kong. In Minglang Zhou and Hongkai Sun (eds.), Language Policy in the People’s Republic of China: Theory and Practice since 1949. Boston: Kluwer Academic, pp. 143–62.

Part II – Case Studies of Contact

15 Contact and Early Indo‐European in Europe BRIDGET DRINKA 1 Introduction As amply documented in this volume, language contact plays a more essential role in ­motivating change than it is usually given credit for. In previous work (Drinka 2013, 2017), I  have attempted to demonstrate how pervasive the influence of contact has been in the ­history of Europe. In the present chapter, I will concentrate on a smaller set of data, with a focus on contact in the pre‐historic and ancient Indo‐European (IE) languages of Europe. While this examination is designed as a series of three interrelated case studies, it will allow us, at the same time, to examine some larger theoretical issues. Among the most important research questions which will emerge are the following: • Is contact reconstructable for prehistory? Can we actually identify points of contact among ancient varieties with confidence? How would these considerations alter our view of relatedness among the Indo‐European languages? • How can we determine with certainty whether contact or genetic relatedness is responsible for an innovation? What role do social prestige and socio‐economic influence play in the outcome of change introduced by contact? • To what extent are formal vs socio‐historical factors responsible for motivating change? This chapter addresses each of these questions by examining successive chronological stages in the pre‐history and early history of Europe, in an attempt to discover what role linguistic and cultural contact has played in the introduction of innovation. Section 2 presents evidence for considerable contact in pre‐historic and early historical times among the major Western European families of Indo‐European – Germanic, Celtic, and Italic  –  through an examination both of lexical resemblances and of structural similarities which presumably arose due to contact. Evidence for a “Northwestern European area” is examined in some detail. Section  3 examines contact among the varieties of the Italian peninsula, both Indo‐ European and non‐Indo‐European, seeking to sort out the role of genetic inheritance vs areal influences in shaping the linguistic outcomes in the Italic varieties. The superstratal influence of Rome on other Italic languages demonstrates the variable effects of prestige and power on varieties in its realm of influence. Section  4 explores the role of contact between Greek and Latin: while Latin played a ­dominating role in Italy and across Western Europe, it, too, was influenced by contact with a

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

304  Bridget Drinka prestige variety, Greek. During the time of the Roman Empire, however, Latin also exerted influence on Greek. This “mutual influence” demonstrates that influence can shift when social power and prestige shift, and that sociolinguistic factors play a crucial role as languages come in contact with each other. Section 5 presents some larger conclusions, and suggests what can be learned through an analysis of the ancient IE languages of Europe concerning language contact in general and the role of contact in linguistic change.

2 Language Contact Among the Ancient Indo‐European Languages of Europe 2.1  Foundations of Areal Approaches in Indo‐European Studies Language contact and the spread of areal features across the map of Eurasia has been central to the study of Indo‐European linguistics almost since the beginning of the discipline. Alongside the majority view of a static characterization of Proto‐Indo European (PIE), there has long been a recognition of a need to view the proto‐language as a variable, developing entity with inherent geographical diversity, displaying clear signs of language contact both among Indo‐European varieties and with contiguous varieties outside the IE family. Grassmann (1863: 119) was one of the first Indo‐Europeanists to document the special bond between Sanskrit and Greek, implying special contact between them, as reflected in language, literature, and myth. Schmidt (1872) famously proposed the “Wave Model” (the Wellentheorie) as an alternative to Schleicher’s (1861–2) representation of Indo‐European linguistic relationship as a family tree (the Stammbaumtheorie); Schmidt viewed innovations as radiating from a central point, like waves emanating from a site of inception. Communities nearest to the place of origination would be expected to adopt the change most fully; those at a more distant location might adopt the innovation later and only partially, or only in particular contexts. This early conception of areal diffusion gave language contact a central role as an explanatory process for Indo‐European linguistics. Meillet (1908), in his depiction of PIE as dialectally diverse, found substantial evidence for an eastern vs western areal distribution of many morphological features, implying late contact, after the time of PIE unity. Pisani (1933, 1940, 1949) pointed to the need for both temporal and spatial factors to be incorporated into the reconstruction process, presenting clear evidence for close contacts between Latin and Greek. Porzig (1954), building upon the foundation established by these scholars, surveyed the phonology, morphology and lexicon of the major IE languages with a view to discovering contact phenomena, and found extensive evidence of the east/ west split noted by Meillet. Porzig’s comprehensive collection of the evidence for early ­contacts among the predecessors of the Italic, Celtic, and Germanic varieties will serve as a useful starting point for our examination of contact among the IE languages of Europe.1

2.2 Evidence for Early Contacts Among the Western Indo‐European Languages In his classic work on the linguistic areas of Indo‐European, Porzig presents substantial evidence for an ancient northwestern European linguistic community:2 Germanic, Celtic, and Italic languages share a number of lexemes pointing to social, economic, religious, and political connections. Words for ‘soothsayer, prophet, poet’, deriving from *ṷāti‐ ‘divine inspiration’, for example, illustrate this apparent early linguistic and cultural relationship (1):

Contact and Early Indo‐European in Europe  305 (1) Lat OIr OE ON

vātēs ‘soothsayer, seer, poet’ fāith ‘poet’ wōþbora ‘poet, orator, prophet’ ōðr ‘inspired mental activity’; ōðinn, ‘inspired god, prince of poets’ (Porzig 1954: 127; Polomé 1972: 59)

Frequently, however, evidence for linguistic and cultural contact exists only between subgroups, between Italic and Celtic, Italic and Germanic, and Celtic and Germanic. A sampling of the shared features of these subgroups will point to the type of contact that existed between them.

2.2.1  Italic and Celtic (Porzig 1954: 98–105) • assimilation of *p…kṷ > kṷ…kṷ PIE *penkṷe ‘five’ > Lat. quinque ‘five’, OIr. cóic ‘five’ • formation of superlatives in *semo Osc. nessimas ‘nearest’, Welsh nesaf ‘next’ • terms for special weaving techniques, connected with knotting, plaiting, and coiling PIE *ṷeg‐ ‘weave’ > Lat. vēlum ‘a sail’, vexillum ‘flag’, OIr figim ‘weave’ Meillet (1908) was a strong proponent of the “Italo‐Celtic Hypothesis,” but Calvert Watkins (1966: 33) called some of these similarities into question, identifying both the above‐ mentioned assimilations and superlative formations of Latin and Celtic as independent innovations. Porzig’s view was that these similarities do not represent genetic relationship, but, rather, reflect language contact. Ternes (1998: 268) summarizes the two key perspectives as focusing on ob die offensichtlichen sprachlichen Übereinstimmungen auf einer vorhistorischen stammbaummäßigen Einheit (d.h. einer relativ späteren Aufspaltung von Italisch und Keltisch) beruhen oder auf einer (ebenfalls vorhistorischen) Kontaktsituation nach der Aufspaltung.3

In earlier scholarly work on the issue, Ternes notes, the former claim was preferred; in recent times, the latter. However, Kortlandt (2007), Weiss (2012), Ringe et  al. (2002), and others ­continue to support the validity of the Italo‐Celtic designation.

2.2.2  Italic and Germanic (Porzig 1954: 106–17) • formation of a collective number word in *‐no‐ PIE *dṷis ‘two’ → * dṷis‐no ‘twice’ Lat. bı̄nı̄ ‘two’ OE getwinne ‘twins’ • ‘indicate’ > technical legal term for ‘speak’ (i.e. ‘pronounce solemnly’) PIE *dei̭k̑‐ ‘point’ Lat. dı̄cō ‘say’ Osc. deíkum ‘to say’, Goth gateihan ‘declare’ developed legal interpretations, as in Lat. causam dı̄cere ‘discuss a case’, sententiam dı̄cere ‘give an opinion’; OHG zı̄han ‘indict’ • The verb ‘have’ Lat. habēre: habeō habēs habet habēmus habētis habent (< PIE *gheHb‐) OHG habēn: habēm habēs habēt habēmēs habēt habent (< PIE *keh2p‐) The Lat and OHG verbs for ‘have’ are remarkably similar, both formally and semantically, yet must derive from different IE roots (*gheHb‐ and *keh2p‐, respectively). Contact is clearly responsible for this similarity, with influence flowing from Latin to Germanic. Also significant is the fact that Italic and Germanic share early words for ‘metal’ (Lat. aes, aeris ‘copper, copper

306  Bridget Drinka ore, money’, OE ār, ǣr ‘ore’), forms which are not shared by Celtic (Roberge, this volume). Polomé (1972, 49, 60–2) notes that Latin and Germanic, but not Celtic, also share a considerable number of technical terms, for hunting and fishing, but especially for agriculture and law. A shared development based on agriculture, for example, is the word for ‘year’: Lat. annus (< *atnos), Osco‐Umbrian akno (‐kn‐ < ‐tn‐), Goth. aþn. With regard to the development of legal applications, such as with dı̄cere, Italic and Germanic developed a judicial meaning for the present participle of ‘be’ ‘being, true’ to mean ‘guilty’ (Lat. sōns, ON sannr); in some varieties, nominal derivatives went on to develop the religious meaning of ‘sin’: OS sundia, OHG suntea, Eng sin.

2.2.3  Celtic and Germanic (Porzig 1954: 118–23) • ‘a plastered, previously stamped‐down, floor in a house or stall’ PIE *plāro‐ > OIr lār, N. Mid.Welsh llawr, ON florr, OE flōr • name for ‘oak god’ PIE *perkṷunos > Continental Celtic ̕Aρκύνια ὄρoς (Aristotle Meteor. 1.13) ‘Hercynian forest’, Goth. fairguni ‘mountain’ OE firgen‐ ‘forested mountain’4 • name for ‘wagon’, derived from agent noun of PIE *ṷeg̑h‐ ‘travel’ PIE *ṷeg̑hno‐ OIr fēn, OE wægn ‘wagon’ Porzig (1954: 214–16) claims that the shared features of Celtic and Germanic are far fewer than those shared by Italic and Celtic or by Italic and Germanic, making it likely that the predecessors of the Italic speakers were somehow more centrally located between Celtic and Germanic. In addition, all the exclusive correspondences between Celtic and Germanic are on the lexical level, according to Porzig, rather than on the phonological or morphological level, suggesting that contact between these two varieties occurred later (Polomé 1972: 64; Evans 1981: 252–3; Roberge, this volume). Nevertheless, some 50 or 60 words, perhaps more, were borrowed from Celtic into Germanic, possibly connected with more advanced metallurgy practices among the Celts (‘lead’, ‘iron’,5 ‘wire’), and a number of terms shared exclusively by Celtic and Germanic concern either technological terms (‘spear’, ‘sword’, ‘fork’, ‘harrow’, ’wagon’ (see example above)) or social terms (‘servant’, ‘bond’, ‘free’, ‘village’, ‘hostage’) (Salmons 2003: 96–7, 111). Evidence for contact between Baltic and Slavic and the Western European languages points to early contacts, especially with the Germanic peoples, though there are fewer features shared exclusively by Germanic, Baltic, and Slavic than there are among the western varieties of Celtic, Italic, and Germanic (Polomé 1972: 54).

2.2.4  Celtic, Italic, and Germanic (Porzig 1954: 123–7) • similar phonological development of word for ‘long’ < PIE *dlaght‐ Lat. longus, Goth laggs, OE, OHG lang, OBr Λóγγoϛ (river name) • special word for ‘man’: *ṷiros Lat. uir, OIr fer, Welsh gwr, Goth. wair, OE, OS, OHG wer ‘man’ • innovative word for ‘chin’ from PIE *men‐ ‘project’ (as ‘projecting part of the face’) Lat. mentum ‘chin’, Welsh mant ‘jaw, mouth’, Goth munþs OE mū ð ‘mouth’ OHG mund ‘mouth’ < *mn̥to‐ Porzig documents few examples of phonological or morphological features that all three families share, but finds a number of lexical similarities, such as shared cultural developments of the terms for ‘furrow’(*pr̥kā), and ‘captive, slave’, as well as for ‘sieve’, and ‘fat, butter’ (Lat. unguen, Umbr. umen ‘fat, ointment’, OIr imb ‘butter’ OHG ancho ‘butter’ (see Polomé 1972: 59–60 for further discussion).

Contact and Early Indo‐European in Europe  307

2.3 Additional Evidence for an early Northwestern European Area: Stress Shift In addition to Porzig’s and Polomé’s convincing evidence for the existence of a “northwestern” dialect area among the Indo‐European languages, the important contribution of  Joe Salmons (2003) adds additional support for the claim. Besides adding a number of shared lexical items (*mak‐ ‘leather bag, stomach’, *gr̥‐no‐m ‘grain’ *bhei‐ ‘bee’ *mori‐ ‘sea’, *bhardhā‐ ‘beard’, etc.), Salmons’ most important contribution to the establishment of this contact zone is his study of the similar patterns of stress shift: building upon the findings of numerous scholars (Streitberg 1896; Devoto 1940 [1968]; Sommerfelt 1962; Kuryɬowicz 1968, and others), he presents clear evidence for similar patterns of stress shift to the first syllable in Germanic and Celtic, with the possible participation of Italic.6 Several key arguments merit special attention. In Germanic, one of the most convincing pieces of evidence for an innovative stress shift to the first syllable is the operation of Verner’s Law (the voicing of voiceless obstruents when the accent followed in PIE) (Verner 1978 [1877]). A comparison of Germanic paradigms affected by Verner’s Law7 to those of Vedic Sanskrit, where the PIE accentual system is assumed to have been maintained, reveals that Germanic inherited the variable accentual system of Indo‐European, but subsequently shifted the stress uniformly to the first syllable, as illustrated in (2): (2) Verner’s Law in Germanic PIE *dei̭k̑‐ ‘show’ PERF.1SG *de‐dóik̑a PERF.1PL *de‐dik̑mé

Vedic didḗša didiśimá

Old English tāh tigon

Old High German zēh zigum

(examples from Szemerényi 1996: 77; Salmons 2003: 137) Other evidence is found in ablaut (vowel gradation) patterns and in the pervasive use of alliteration in poetry, as illustrated in its obligatory use in Beowulf (3): (3) Beowulf (lines 1–3) Hwæt. We Gardena Lo, we Spear‐Dane.GEN.PL.M

in geardagum, in yore‐day.DAT.PL.M

þeodcyninga, people‐king.GEN.PL.M

þrym gefrunon, glory.ACC.SG.M hear.PST.PL

hu ða æþelingas How the nobles.NOM.PL.M

ellen fremedon. valor.ACC.PLN perform.PST.PL

‘Lo, we have heard of the glory of the Spear‐Danes in days of yore and of the kings who ruled them, how the nobles performed valorous deeds.’ The caesura creates two half‐lines, with the key alliterating element appearing in the third stressed syllable of the line. Although the /g/ in geardagum has palatalized to [j], it still establishes alliteration as /g/ in the first half‐line, in Gardena. The third line illustrates the fact that all vowels alliterate with each other. Both the trochaic tetrameter of the verse and the alliteration pattern coincide well with the presence of word‐initial stress in the language. In Celtic poetry, we find similar evidence for this shift: the early poetic line in Old Irish resembles the Germanic half‐line; both Old Irish and Germanic rely on stressed pairs, with a caesura following the second stress; both make extensive and similar use of alliteration: in

308  Bridget Drinka both systems, all vowels alliterate with each other, and consonant clusters follow similar patterns (Campanile 1979: 197–200; Salmons 2003: 164). Both of these families appear to have inherited particular poetic traditions and tendencies from PIE (Watkins 1982), but the use of alliteration is obligatory in Germanic poetry, in contrast to its optional and more “ornamental” status in Old Irish (Suzuki 1988). The presence of alliteration in the oldest layers of Celtic is ensured by such evidence as (4), a formulaic equation: (4) Gaulish‐Irish alliteration bnanom brictom ~ brichtu ban / brechtaib ban ‘magic by women’ (example from Stifter 2016: 83) Poetic alliteration in Welsh likewise suggests accentual shift, as in (5): (5) Welsh alliteration gwyrtheu goleu gwelhattor ‘bright wonders will be seen’ Hen Gerddi Drefyddol, xxxvi, 3 (example from Salmons 2003: 158) Stifter (2016: 85) strongly endorses the recognition of alliteration as connected to and engendered by initial stress, and views this development as the result of areal spread in Western Europe: From a purely European point of view there is a strong correlation between stress‐initial languages and systematic, not just occasional, alliteration. […] Rather than thinking of it as an Indo‐European inheritance, it is more likely that the systematic employment of alliteration had diffused at some unspecifiable time from an unspecifiable source across an area that had a shared prosodic feature, namely initial stress.

Turning to the possible participation of Italic in the stress shift alongside Germanic and Celtic, we note that the claim is an old and enduring one (cf. Streitberg 1896; Devoto (1940 [1968]); Kuryɬowicz (1968) and others). Arguments for this areal connection usually focus on the weakening of vowels in medial syllables, which are assumed to have resulted from such a shift (6). (6) faciō ‘I do, make’ : reficiō ‘I rebuild, restore’ hostipotis > hospots > hospes ‘guest, stranger, host’ (example from Salmons 2003: 146) The difficulty with including Italic as a participant, however, is that other languages on the Italian peninsula also underwent a similar shift. Clackson and Horrocks (2011: 47) support the claim that Early Latin did, indeed, have stress on the first syllable, but add that syncope and vowel weakening seem not to have been in place yet in the earliest Latin texts. In the Lapis Satricanus, ca. 500 bce, for example, forms such as Mamartei appear in unweakened form, as do other forms from seventh‐ and sixth‐century Latin inscriptions. Furthermore, Sabellian, the most important member of which was Osco‐Umbrian, and Etruscan also developed initial stress accent, leading to syncope of final syllables. Significantly, a sufficient number of Etruscan texts exist to allow for the specification of a date when the syncope of short vowels in medial syllables must have occurred to the first half of the fifth century (7). (7) pre‐fifth century rendition of the name Aulus: later rendition:

Avile or Avele Avle

Contact and Early Indo‐European in Europe  309 Similar evidence exists from Sabellian (8): (8) Sabellian inscription of a name from Capua ca. 500 bce: Peracis later Oscan text from the same locale: perkium8 (examples from Clackson and Horrocks 2011: 48) While it appears, then, that the Italic stress shift is a later phenomenon, and that these varieties did not participate in the earlier Western European areal development, the evidence for a shared innovation between Germanic and Celtic remains convincing.

3  Languages in Contact on the Italian Peninsula While the stress shift in languages of the Italian peninsula described in the previous section does not appear to be tied to a similar trend in Germanic and Celtic, it is still a valuable example for our discussion of ancient IE languages in contact, since these Italian varieties all underwent this innovation at about the same time and in close proximity to one another. The Italian peninsula, in fact, provides considerable evidence for many types of contact across time, extending as far back as the eighth century bce. Contact between Greek and the languages of the Italian peninsula is documented as early as 770 bce, and contact among other Indo‐European and non‐Indo‐European languages on the peninsula is also documented soon after.

3.1  Contact with Etruscan Etruscan, a non‐Indo‐European language which has yet to be fully interpreted, was widely diffused across Etruria, with most inscriptions coming from Chiusi and Perugia (Devoto 1940: 199). While only four or five Latin inscriptions exist before 600 bce, more than 150 Etruscan inscriptions have been found from that early period. Even in 100 bce, the Etruscan inscriptions greatly outnumber the Latin ones, at a ratio of 9,000 to 3,000 (Clackson and Horrocks 2011: 38). The influence of Greek culture and language on Etruscan is well‐documented. The Etruscans appear to have been responsible for adopting the alphabet from Greek colonists, and the writing system spread from there to other languages of the peninsula. Thus, Etruscan mediated the adoption and spread of writing in Italy, and this spread implies contact. Of the approximately 10,000 total Etruscan inscriptions, surprisingly few Latin‐Etruscan bilingual documents have been found: of the 28 bilingual and 6 digraphic inscriptions, all are funerary, consisting exclusively of personal names, all date to the first century bce,9 and all were produced in northern Etruria. The monuments list two names referring to the same person, an Etruscan and a Latin name, the latter usually following the typical pattern used for Roman citizens, as illustrated in (9): (9) pup velimna au cahatial P. Volumnius. A. f. Violens ǀ Cafatia. natus. This fact suggests not only linguistic contact, but also cultural and political assimilation (Rix 1956, 1991; Benelli 1994; Langslow 2002: 31–2). Rix (1994) points out, however, that no morphological borrowing, other than derivational endings, occurred between Etruscan and the other languages of the Italian peninsula.

3.2  Contact between Latin and Sabellian Besides Etruscan and several other non‐Indo European languages which existed in Italy in the first millennium bce (e.g. Raetic and North Picene), a number of Indo‐European

310  Bridget Drinka languages are also attested, some rather robustly and some in vestigial form. Among the most important are Faliscan, the closest relative to Latin, and Sabellian, which included, especially, Osco‐Umbrian, and which was broadly diffused across central and southern Italy. One of the key issues that emerges with regard to languages in contact on the Italian peninsula is the nature of the relationship of these other IE languages with Latin: do they all derive from a common genetic ancestor, Italic, or should they be regarded as separate IE languages which developed their similarities through more recent contact? While most  scholars accept that the similarities which exist between Latino‐Faliscan and the Sabellian varieties can be accounted for as representing shared genetic features, others (most notably Devoto 1940) insist that these similarities are due to language contact in  more recent times. Clackson and Horrocks (2011: 65–74) weigh the value of both of these approaches, as well as the possibility of independent development, and suggest careful analysis on a case‐by‐case basis is called for, since some features are best explained  as due to contact, such as the late development of stress shift and resultant medial vowel weakening (see 2.3), while others point to an older affinity, such as the ­similar formation of the gerundive (Lat. faciendam, Oscan úpsannam ‘to be made’) or the  reformulation of the paradigm of copular ‘be’ (especially 1sg *esom, in contrast to *esmi in other IE languages). In what follows, we will briefly examine some of the e­ vidence from Umbrian and Oscan, and will then be in a position to assess the nature of their ­relationship with Latin.

3.2.1 Umbrian Most of our evidence for Umbrian comes from the Iguvine Tables, seven bronze tables dating from the third to the first centuries bce, found in Gubbio (ancient Iguvium). The tables describe the acts and rituals of the Atiedian brothers, and the script itself illustrates at least two levels of contact: the earlier portions are written in the Umbrian alphabet, based on the Etruscan alphabet, and are written from right to left. In the middle of Table V, the Umbrian lettering ceases, and the writing resumes in the Latin alphabet, left to right, reflecting, visibly, the fact that Umbria had been conquered by Rome in 266 bce. This indication of Roman hegemony is reinforced by the text itself: on the same table, Table V.a.23, a quaestor, a Roman civic officer, is listed as presiding over the banquet of the Brotherhood, alongside the Brother Superior (10). (10) Umbrian (Iguvine Table Va 22–4) et : ape : frater : çersnatur : furent : /ehvelklu : feia : fratreks : / and when brothers dine.PPP.NOM.PL.M be.FUT3PL vote.ACC SG. do.PRES.SUBJ.3s brother‐superior : ute : kvestur sve : rehte : kuratu :  si : sve mestru : karu: or quaestor if right provide.PPP.NOM.SG.N be  if greater part ‘and when the brothers have dined the brother‐superior or the quaestor shall call for a vote whether (the dinner) has been provided in a satisfactory manner.’ Umbrian undergoes more phonological change than Oscan. For example, rhotacism (s > z > r) appears intervocalically in Umbrian as in Latin10 (e.g. Umb. kuratu: Lat. curato), whereas Oscan only shows in similar environments, in segments written in Latin alphabet. In the later Umbrian tables, beginning with Table Va, Umbrian goes beyond Latin in rhotacizing s even in final position: tuder, totar, popler, ocrer, sir (Poultney 1959: 74). The late productivity of rhotacism in Latin and the more robust adoption of the change in later Umbrian, during the time of Roman dominance, suggest that this was a shared innovation.

Contact and Early Indo‐European in Europe  311 In Sabellian, in general, the reduction of vowels as a result of the shift of stress to the first syllable, referred to on p. 308, does not so much appear as vowel weakening, as in Latin, but much more often as syncope (Poultney 1959: 45) (11a, b): (11a) Umb. (11b) Umb.

fratreks < *fratrikos ostendu ustentu vs Lat. ostendito < *obs‐tendetōd ‘show.FUT.IMPER’

The description of the Umbrian rituals also points to shared cultural norms, since they resemble old ritual practices of Rome, including the invocation of similar gods (Jupiter, Mars, Ceres), the triple circumambulation around the city with fire and victims, the veneration of three ritual gates, similar augury practices and sacrificial animals, and cognate names for sacrificial breads and the altar. Several noteworthy borrowings from Osco‐Umbrian into Latin demonstrate that the rustics of the Italian peninsula had some influence on the urban population, as well: Latin bos (bov‐) ‘ox, cow’, bufalus ‘antelope’, and lupus ‘wolf’ are all assumed to have been brought in from Osco‐Umbrian (Solodow 2010: 25).

3.2.2 Oscan Oscan, the most widely attested of the Sabellian languages, spread across southern Italy in the fourth century bce. Oscan shares several morphosyntactic innovations with Latin: the development of the past‐marked subjunctive, for example, and, possibly, the creation of a new imperfect tense. The past‐marked subjunctive is illustrated in (12): (12) Oscan (Excerpt from Cm 1 A 1‐18) Oscan ekss kúmbened   […] puz […] fusíd thus agree.PERF.INDIC CONJ be.IMPERF.SUBJ.3S Latin translation ita conuenit ut esset ‘it was agreed…that…it should be’ (example from Clackson and Horrocks 2011: 59) This creation of the new imperfect and pluperfect subjunctives allowed for the “sequence of tenses” to be established (using a past‐marked subjunctive with a past tense matrix verb), not unlike the distinction observed in Greek. Clackson and Horrocks (2011: 71) note that this text is highly bureaucratic, and that the creation of this innovative past‐marked subjunctive in *‐sē‐ in both Latin and Oscan “accords well with the shared development of a legalistic, bureaucratic and religious idiom which took place in the cultural contact of the first ­millennium b.c.” The formation of the new imperfect in Latin, using an infectum stem and the suffix ‐bā‐, may also have been shared with Oscan, as exemplified in (13): (13) Oscan (Excerpt from Cm 1 A 1‐18) pús […]   lígat[ús]   fufans who       legates were (example from Clackson and Horrocks 2011: 61) The form fufans could represent the creation of an imperfect built with a stem fu‐ ‘be’ (as seen in example (12) above) plus a suffix in ‐fa (like Latin ‐ bā‐). Because this is the only potential example of an innovative imperfect in Sabellian, however, its association with the Latin imperfect is not assured, and, as Clackson and Horrocks (2011: 61–2) point out, alternative explanations are possible.

312  Bridget Drinka Besides the past‐marked subjunctive and, potentially, the imperfect, the similar formation in Latin and Sabellian of the gerundive (the “quasi‐participial” use of the future passive participle), presented on p. 310 as evidence in support of the “Italic theory,” also provides convincing evidence for later contact between Latin and Oscan in its use in official documents at the end of the Roman Republic (second to first centuries bce) (14) in clear imitation of a frequently used official Latin formula (15). The Latin translations are provided by Clackson and Horrocks (2011: 62) to demonstrate the structural parallelism. (14) Oscan (Po3)

kvaísstur. púmp/aiians.     trííbúm.     ekak.     kúmben /nieís. tanginud. quaestor.NOM.SG Pompeian.NOM.SG house.ACC.SG this.ABL.SG senate.GEN.SG decision.ABL.SG domum       hanc      (de) conuentus  sententia quaestor Pompeianus      úpsannam./ make.FUT.PASS.PART.ACC.SG faciendam

deded. give.PERF.3SG dedit

ísídum.    prúfatted same.nom.sg  approve.PERF.3SG idem     probauit

‘[Vibius Vinicius,] the Pompeian quaestor, arranged for this house to be built by the decision of the senate and the same man approved it.’ (example from Clackson and Horrocks 2011: 62) (15) Latin (CIL I2 passim) faciendum curauit eidemque probauit make.FUT.PASS.PART.ACC.SG arrange.PERF.3SG same_CONJ approve.PERF.3SG ‘he arranged for [something] to be made and the same man approved it.’ While the forms of the cognate gerundives (Oscan –nn‐, Latin –nd‐) probably reveal a more ancient relationship, since it is unlikely that the form itself was borrowed from Latin into Sabellian, the use of these forms in official documents in very parallel fashion signals the establishment of Roman dominance at the end of the Republican period, like that noted for Umbrian. Lexical evidence supports this conclusion: the term for quaestor has been borrowed, as in Umbrian, and the semantic reanalysis of the noun tanginud from ‘thought’ to ‘decision’ may have followed the model of Latin sententia, which ­underwent the same development. The phrase kúmbennieís tanginud closely replicates the Latin formula de  senatus sententia ‘by the decision of the senate’ (Clackson and Horrocks 2011: 63).

3.2.3 Sabellian and Latin Perfects: Early Shared Innovation or Later Development? Finally, it is noteworthy that Latin and Sabellian have reformed their verbal systems according to similar, innovative patterns, utilizing the infectum stems for the formation of the present, future, and imperfect tenses, and the perfectum stem for perfects, pluperfects, and future perfects. This fact may provide evidence for the relatedness of the Italic dialects. However, it is also important to note that individual verbs do not necessarily form their perfects according to the same patterns in Latin and Sabellian. The s‐perfect, derived from the s‐aorist of IE (cf. Lat. dı̄xı̄ ‘said’, mānsı̄ ‘remained’), is not well attested in Sabellian,11 and the –ui perfect (cf. Lat. amāuı̄ ‘loved’ audı̄uı̄ ‘heard’) is not attested at all (Coseriu 2008: 14). What Sabellian relied upon instead was the archaic reduplicated perfects (Osc. deded ‘gave’ = Lat. dedı̄t) and long‐vowel perfects (Osc. kúmbened ‘agreed’ = Lat. conuēnı̄t), both of which Latin also retained to some extent, and the tt‐perfect (cf. Osc. prúfatted ‘approved’), not found in Latin (cf. the Latin perfect for the cognate verb, probauit, using the ‐ui perfect.) The disparity

Contact and Early Indo‐European in Europe  313 of perfect formation between Latin and Sabellian is made especially clear in a table assembled by Meiser (2003: 73), where only three correspondence types of a possible 11 are actually attested. The reorganization of the verb system according to similar principles, with an amalgamation of the IE perfect and aorist categories and the rise of new future and future perfect categories, then, probably indicates early shared innovation, but the lack of unity in perfect formatives across Latin and Sabellian suggests later, separate development of some parts of the perfect morphology.12

4  Greek and Latin in Contact 4.1  Early Contacts Contact between Greek and the languages of the Italian peninsula is documented as early as 770 bce in the first inscription to be found anywhere in Italy: a Greek inscription on a pot at a woman’s burial in Osteria dell’Osa near Rome, with the single word eulin carved on it, presumably eúlinos ‘good at spinning’ (Gleba 2008; Clackson and Horrocks 2011: 39). Other early Greek inscriptions have also been found in Rome and throughout Latium and southern Etruria. The earliest influence of Greek upon Latin came through the intermediary of Etruscan, since the Etruscans had considerable political power and had established strong relations with the Greeks. It is possible, in fact, to sort out earlier Greek borrowings from later ones precisely because of the traces of Etruscan influence which may remain in the earlier borrowings. For example, in the Carmen Arvale, the ancient chant of the Arval brothers, the vocative triumpe is repeated five times. This form comes ultimately from Greek θρίαμβος (thríambos) ‘hymn to Dionysus, triumph’ but by way of Etruscan, which devoiced the Greek b to p. Likewise Lat. amurca ‘dregs of olive oil’ was borrowed from Greek ἀμóργη (amórgē), but, again, by way of Etruscan, with its devoiced k (< g). After Etruscan rule ended in the sixth century bce, the Romans came to have more direct contact with Greek, and borrowings reflected this fact: while Etruscan translated the Greek name Πoλυδεύκης (Poludeúkēs) as Pultuke, Latin translated it as Pollux, dialectal Polouces, Poloces, that is, no longer with the devoicing of the d > t, but with its weakening (Devoto 1940: 89–90).

4.2  Bilingualism and the Prestige of Greek in the Republican Era As Roman power extended across Italy, Latin came in contact with Greek in the cities and territories of Magna Graecia in southern Italy; conquests abroad brought Greek‐speaking slaves to Italy. Remarkably, then, Greek came to be used by both the lower stratum of society and the upper classes, as knowledge of Greek came to be a sign of education in the second century bce. During the Republican era (509–27 bce), bilingualism grew, with Greek holding a position of higher prestige. Highly‐educated Romans like Cicero could code‐switch into Greek with ease, especially to converse about literature or medicine, but also to express emotion or to discuss familial matters, as exemplified in a dialog reported by Cicero to his friend Pomponius Atticus, (so named because of his affinity for and long residence in Athens) (16): (16) Ad Atticum 13.42.1 Uenit ille ad me ‘καὶ μάλα κατηφής’ et ego ‘σὺ δὲ δὴ τί σύννους’; ‘rogas?’ inquit. He came to see me, ‘right down in the mouth’. I greeted him with ‘You there, why so pensive?’ ‘Need you ask?’ was the answer.

314  Bridget Drinka This reported dialog between Cicero and his nephew began with several switches to Greek, perhaps to convey familiarity, and then continued in Latin (Swain 2002: 149–50). According to Cicero (Att.i.19.10), well‐educated Romans were sometimes so skilled in speaking and writing in Greek that they would intersperse “barbarisms” purposely into their Greek, so as not to sound overly‐Greek (Biville 2002: 87). Cicero himself was proud of his ability to use Greek successfully, and states that he would never engage in such behavior. For example, he wrote the history of his consulship entirely in Greek, and noted that he had “dumbfounded the whole Greek community.” (i.21.2) (Swain 2002: 163). As Swain (2002) thoroughly demonstrates, Cicero’s Roman identity was intertwined with and dependent on displaying his complete command of Greek.

4.3  Greek and Latin in Contact Outside of Rome An extremely informative example of contact and the complex construction of identity in ancient bilingual societies is provided by Adams (2002) in his description of the bilingual inscriptions among the Italici and the Graeci on the island of Delos from the second to the first century bce. Many of the 221 Italici were identified as negotiatores, apparently existing in a kind of loosely‐bound network and maintaining Latin, alongside Greek, as an important part of their identity. The widespread use of particular formulas, with their own spelling and phraseological conventions, point not only to the existence of this network but also to the continued importance of their maintaining an Italic identity within the Greek‐speaking world (Adams 2002: 105–13). The use of Latin may have also been associated with social status: at three of the collegia, the professional associations established by the negotiatores at Delos, dedications to the gods were written in bilingual Latin and Greek texts; at a fourth, the Competaliastai, only Greek was used. Adams (2002: 114–15) concludes that, while the first three collegia had primarily freemen or freedmen as members, the fourth collegium consisted especially of slaves and freedmen, whether Italian or Greek. The use of Latin, then, was connected not only with place of origin but also with social distinction and prestige. The complexity of the social relationship of Greeks and Italians at Delos is also captured in their use of naming conventions: Greek inscriptions sometimes incorporate υἱóς (huiós) ‘son’, following Latin conventions, but evidence also exists of Italici abandoning this convention, as on a dedicatory inscription on a statue to a local deity, where the Greek tradition is followed (Adams 2002: 124). Accommodation and influence clearly existed in both directions. Greek and Latin also interacted in telling ways in the Greek city of Naples from the time of its founding by Cumean settlers in the fifth century bce well into the Christian era: even as late as the third century ce, Greek was still used in official decrees, agonistic lists of winners of athletic events, inscriptions by phratria (guilds with religious or social function), and epigrams. As these institutions and genres gradually disappeared or were replaced by Roman ones, the use of Greek diminished, but did not cease to exist: in the first to third centuries ce, dedicational, votive, or funerary inscriptions could appear in Latin but frequently still appeared in Greek. Imperial texts, on the other hand, were usually written in Latin. By the fourth century ce, Greek is still found in some Christian catacombs, but Latin had become the prevalent language in most contexts (Leiwo 1994: 170).

4.4  The Growing Prestige of Greek in the Imperial Period At the beginning of the imperial period in the first century bce, the linguistic and cultural situation slowly began to change, with the prestige of Greek becoming even more evident. The upper classes became completely bilingual in Latin and Greek. Quintilian, in the first century ce, said instruction for young Romans should begin with Greek, because they would

Contact and Early Indo‐European in Europe  315 Table 15.1  Latin calques based on Greek models (after Clackson and Horrocks 2011: 198). Greek model prólēpsis poiótēs posótēs di(h)aíresis epítheton hypokeímenon

Latin calque lit.‘pre‐apprend‐ing’ lit. ‘whatkindof‐ness’ lit. ‘howmuch‐ness’ lit. ‘part-take-ing’ lit. ‘upon‐put‐pass.adj’ i.e. ‘applied’ lit. ‘under‐place‐ed’, i.e. ‘supposed’

anti‐cip‐atio ‘preconception’ quali‐tas ‘quality’ quanti‐tas ‘quantity’ dis‐tribu‐tio; di‐ui‐sio ‘partition’ ad‐iect‐iuus ‘adjective’ sub‐iect‐iuus ‘of the subject’

soak up Latin naturally. Young upper‐class men went to Athens to receive their higher education (Marrou 1977: 553). It was expected that an educated man should know both of “our languages”: utraque lingua; uterque sermo (Kramer 2011: 61). The linguistic effect of this intimate contact was evident especially in higher registers of Latin. For example, in order to facilitate the enormous influx of lexemes, Latin speakers paired nouns with light verbs (facere ‘make/do’, habere ‘have’, etc.) : mentionem facere ‘make mention’ iter habere ‘have (take) a journey’. This periphrastic device resembles similar constructions found frequently in the Koiné (cf. Greek poieîsthai ‘make/do’) (Clackson and Horrocks 2011: 185–6). While the vernacular shows signs of more direct borrowing (Gk. kólaphos ‘blow’ => Lat. colpus, replacing ictus; Gk. pétrā ‘stone’ => Lat. petra, replacing lapis, etc.), the upper classes tended to calque, instead (see Table 15.1). The use of calquing is not limited to the lexicon; numerous examples of syntactic calques also exist. For example, Old Latin had a very meager participial system, consisting only of present active participles, past passive participles, and future participles, with the present participle being used almost exclusively as an adjective. Greek, on the other hand, had an abundant assortment of participles  –  present, aorist, and future, active and medio‐ passive – and used them much more frequently. Under the influence of Greek, speakers and writers of Classical Latin reconfigured the function of the present participle to match the Greek model, using these structures to replace adverb clauses or to form absolutive constructions. The present participles used in these constructions also grew in complexity, showing greater use of modification or complementation. As Clackson and Horrocks (2011: 192) observe, the new functions of the present and past participles represent further steps towards replicating the full array of Greek options with the limited resources available to Latin, thereby introducing greater flexibility and symmetry into what was originally a highly defective ‘system’ with only two members of restricted functional range and only partial functional correspondence.

4.5  The Role of Religion Christianity was also a conduit of Greek: the language of the early Christians was Greek, and the first teachings were in that language. The use of Greek by Christians is instructive: in the third century ce, 41.2% of the Christian inscriptions in Rome were in Greek; in the fourth century, the amount of Greek had diminished to 29.8%; in the fifth through seventh centuries, it dropped to about 1%. In contrast, Greek was more prevalent in contemporary Jewish inscriptions: of the almost 600 Jewish inscriptions found in Rome from the third and fourth centuries ce, 70% were in Greek (Leiwo 2002: 179).13 The influx of Christian concepts from Greek into Latin took place both through direct borrowing and calquing. The complexity of the process is well illustrated by the Greek lexeme

316  Bridget Drinka charisma, referring originally to the concrete manifestation of the gift of divine grace in the community. When this word was adopted into Latin, it was construed on a more abstract level, and was translated especially by Latin gratia ‘grace, good will, favor’. The Greek borrowing itself ceased to be used, especially because it came to be connected with the heresy of Montanism, with its adoption of the charisma prophetiae as a central tenet (Mohrmann 1965: 141). The related term eucharistéō ‘bestow a favor on, be thankful’ was the normal term for ‘to thank’ in the Septuagint and New Testament; from the nominal form eucháreia ‘grace, charm’ was formed eucharistía ‘eucharist, the celebration of thanksgiving’. Following this pattern, gratia became the basis for expressing thanks in several Romance languages, Italian grazie, Spanish gracias.

4.6  Mutual Influence of Latin and Greek Although the influence of Greek on Latin was pervasive and long‐lasting, it is also significant that the influence of Latin on Greek became important during the time of the Roman Empire. In the eastern areas of the Roman Empire, Latin was used in the military and for legal matters, even while Greek continued to fill many other roles (Adams 2003: 608). The presence of many Greek‐Latin glossaries demonstrates that some knowledge of Latin was important, though not necessarily widespread (Kramer 2011: 62). Evidence for this influence is also found in the preference for certain Latin structural patterns by Greek writers like Strabo (ca. 64 bce–10 ce) and Plutarch (ca. 46–120 ce), both of whom showed a strong affinity to Rome. For example, Plutarch precisely imitates Livy’s use of a past passive participle plus have auxiliary to construct a periphrastic perfect (17a) by using a medio‐passive participle plus have auxiliary (17b): (17a)  Livy (History of Rome 39, 51, 5) ut  iter semper aliquod praeparatum fugae haberet, so_as path.N.SG always some prepare.PPP.N.SG flight have.3SG.IMPERF.SUBJ. septem exitus e domo fecerat seven exits out of house made.3SG.PLUPERF.ACT ‘in order to have always some path prepared for flight, he had made seven passages from his house’ (17b)  Plutarch (Lives Titus Flamininus 20,4) tḕn oikían éti próteron exódois heptà katageíois the house.F.ACC.SG even before exits seven underground suntetrēménēn ek  tēs͂   heautoũ diaítēs  eı̃chen connect.PRF.MEDIOPASS.PTCP.F.ACC.SG from his  own  dwelling have.3SG.IMPRF.ACT ‘even before this time he had connected his house with seven underground exits leading from his own chambe.’ Plutarch could easily have chosen an active participle from the array of participles available in Greek; instead, he uses a medio‐passive participle, modeling it precisely on the Latin past passive participle.14 What this represents, then, is the operation of “mutual influence” between Greek and Latin. In earlier times, Greek exerted a strong influence on the usage and style of elite Romans. During the time of the Roman Empire, on the other hand, the reverse occurs, with some Greek writers adopting decidedly Latin‐style structures. Greek’s exposure to Latin was more limited, since it was used especially for administrative purposes.15 Bureaucratic Greek underwent more influence than other registers, since many large

Contact and Early Indo‐European in Europe  317 administrative documents composed originally in Latin needed to be translated into Greek. What should be noted is that this “mutual influence” which Greek and Latin exerted on each other was not completely simultaneous and was not evenly distributed across the two languages: the influence of Greek on Latin stretches back to the beginning of the written tradition, while the influence of Latin on Greek is confined especially to the period of the Roman Empire, in the first centuries ce. Eventually, such influences ceased, but even as the two halves of the Roman Empire became mostly monolingual by the mid sixth century, tenacious remnants of Roman rule in the east persisted: Latin was still to be found on coins until the eighth century; military commands were also given in Latin until the eighth century; furthermore, some Latin was used in the Byzantine Court Ceremony until the tenth century (Kramer 2011: 79).

5 Conclusions Several specific conclusions can be drawn from this three‐part analysis: • Substantial evidence exists for early contact among the Germanic, Celtic, and Italic languages, the Indo‐European languages of Western Europe. While contact is most visible in the shared lexicon of these languages, it is also detectable in the systematic layers of the grammar, including the shared accentual patterns found especially in Germanic and Celtic. Since most isoglosses do not encompass all of these language families but subsets of them, one should not envision a “Northwestern European area” as a holistic entity, but rather as a geographical space where numerous sub‐areal contacts occurred. The evidence presented here demonstrates that contact can, indeed, be reconstructed for pre‐ historic societies, and that areal factors need to be taken into consideration as models of relationship are envisioned. • The Italian peninsula provides extensive evidence of contact among the Indo‐European and non‐Indo‐European varieties, already revealed in the earliest inscription in Italy, dating to 770 bce and written in Greek: cultural and linguistic contact existed between the Etruscans and the Greeks at an early time. The structural similarities of shared innovations in Latin, Oscan, and Umbrian point to extensive contact, as does the demise of the latter two varieties in the wake of Roman ascendancy. Some similarities, such as the use of reduplication to form perfects, were undeniably inherited from ancient times, but others, such as the formation of past‐marked subjunctives, were clearly recent, shared innovations which spread across the Italic varieties. Above all, the superstratal influence of Rome demonstrates the variable effects of prestige and power on varieties in its realm of influence. • While Latin played a dominating role in Italy and across Western Europe, it, too, was influenced by contact with a prestige variety, Greek. This influence was pervasive and profound, introducing structural and stylistic patterns and an array of lexical forms and calques, especially into the speech of upper‐class Romans. During the time of the Roman Empire, however, Latin also exerted an influence on Greek. This “mutual influence” demonstrates that the direction of influence can shift when social power and prestige are altered, and that sociolinguistic factors play an essential role as languages come in contact. Several larger conclusions also emerge from this analysis: • The traditional view of a unified proto‐language, depicted as a family tree, is not adequate as a representation of the complex histories of the Indo‐European languages. Proto‐Indo‐European was not a unified, single entity in space or time, and the languages

318  Bridget Drinka which developed from it have been in contact, developing shared features throughout their existence. • Areal diffusion is a more essential path of innovation than is often thought, and socio‐ historical factors are largely responsible for motivating the movement of these innovations across geographical space and across populations. • In all of the examples gathered here  –  from the shared connection of ‘seer’ to ‘poet’ among the Western IE languages to the extension of Oscan ‘thought’ to ‘decision’ based on Latin sententia, from the calquing of Greek poiótēs ‘whatkindof‐ness’ as Latin qualitas to the straightforward borrowing of Latin familia as Greek phamilia, from the increased use of Latin light verbs as in the Greek Koiné to the Greek co‐opting of passive participles to form perfects as in Latin – in all of these cases and in innumerable others, we recognize the clear and essential role of contact as a trigger for linguistic change.

NOTES 1 See Drinka (2013) for an examination of the role of contact across all IE languages, and a discussion of the suitability of phylogenetic vs areal models to represent relatedness. 2 Porzig’s analysis of the northwestern European area resembles yet differs in significant ways from the almost contemporaneous visualization of a “Nordwestblock” proposed by Kuhn (1959, 1961). In Kuhn’s view, the presence of aberrant forms like Low German pū e and pū ne ‘a boil’ (cf. Lat. pū s ‘pus’) and OE prætt ‘cunning’ (> pretty), which did not undergo Grimm’s Law (PIE *p > Gmc. f) as expected, points to the existence of another language in the area, exerting substratal influence upon both Germanic and Celtic. In a critique of Kuhn’s claims, Meid (1986: 203) suggests that such exceptions represented, not an additional variety but a small number of archaic pockets within Germanic; Lühr (2018: 188) likewise voices skepticism, demonstrating that many of these apparently exceptional forms derived from *s + consonant, and so were not anomalous at all: PGmc. *splutta‐ (> OE splott ‘patch, land parcel’) alongside *plutta‐ (> OE plott ‘patch of land’). While Porzig (1954: 106), like Kuhn, assumes that a substratal language served as an intermediary among Italic, Celtic, and Germanic – an outdated view by today’s standards – he takes a more forward‐looking approach than Kuhn in his comprehensive analysis of shared features of the northwestern area, investigating lexical, phonological, and morphosyntactic similarities, and assessing possible genetic vs areal explanations in each case. He thus moves far beyond the facile assumption that language contact entails substratal influence. 3 “[W]hether the obvious linguistic correspondences are based on a prehistoric genetic unity (entailing a relatively later splitting up of Italic and Celtic) or on a (likewise prehistoric) contact situation after the split.” 4 These last two Celtic examples, lār and óros, taken together, illustrate the Celtic loss of *p in initial position; the second example demonstrates that the earlier rule of of *p…kṷ > kṷ…kṷ was no longer in effect (Porzig 1954: 120). The second example also provides evidence that, if the source for Goth. fairguni truly was Celtic *perkṷúni‐, as has been claimed, then this borrowing occurred before the operation of Verner’s Law, and therefore before the shift of stress to the first syllable (see 2.3) (Lehmann 1977: 288). 5 It is surely significant that the words for ‘ore, copper, bronze’, noted on p. 305–6, are only found in Germanic and Italic, not Celtic, implying contact during the early bronze age (as early as the second millennium bce), while the shared terms for ‘iron’, developing from *ı̄sarno‐, are only found in Germanic and Celtic, not Italic. Evidence like this reinforces the claim made by Polomé 1972 and others that the Italic languages, in contact with Germanic at an earlier time, must have then moved southward, out of the realm of contact, and that the Germanic‐Celtic contacts are, in general, later. The importance of mapping layers of contact through space and time emerges, once again, as crucial, whether these contacts occurred in prehistory or in the light of history (Drinka 2017). 6 Salmons (2003: 180) makes the thought‐provoking claim that contact with Finno‐Ugric may have played a role in initiating this trend in the northwestern Indo‐European languages.

Contact and Early Indo‐European in Europe  319 7 Gothic is largely excluded from the operation of Verner’s Law. See Bennett (1968), Kortlandt (1988), Voyles (1988) for discussion of this exclusion. 8 Standard conventions are followed here of using bold typeface for texts written in native script and italics for those written in Latin script. 9 Although the last Etruscan inscription is found in Perugia on the Volumnii tomb, and dates to around 10 bce, the language clearly continued to be spoken after this time (Polomé 1983: 520). 10 The “Cippo del Foro,” 575–550 bce, an ancient cornerstone in the Roman Forum, still had esed (> later erit). Rhotacism was officially recognized in Latin in 312 bce by Appius Claudius Caecus, who instituted orthographic reform (Devoto 1940: 99). 11 While Coseriu (2008: 14) claims that Osco‐Umbrian does not provide evidence of the s‐perfect, Meiser (2003: 106; 127–8) lists two examples: *sess‐ ‘took a seat’ Umbrian VI a 5.7 (ander‐) sesust, (future) ‘will have taken the place (in between)’ and *ōps‐ ‘made’ Oscan Po 8, Ve 16, etc. upsens Umbrian Um 6, Ve 234, etc. opse(n)t ‘produced’. 12 See Drinka (1995: 54–75) for a chronologization of the development of the Latin perfect, with the reduplicated perfect representing the oldest layer, along with the amalgamation of the perfect with the s‐aorist. Euler (1993: 102–3) views the similar development of the IE perfect > preterite in both Latin and Sabellian as evidence for their shared history. 13 Rutgers (1995: 182–94) points to the gradualness of the transition from Greek to Latin in Jewish inscriptions in Rome, and the patterning of later Jewish inscriptions written in Latin on those written in Greek. For example, epitaphs written in Greek tended to list the dedicator first, followed by the name of the deceased; Jewish epitaphs written in Latin tended to follow this Greek pattern rather than the usual Latin one. Vernacular tendencies exist in both the Greek and Latin Jewish inscriptions, suggesting close contact with the non‐Jewish community. Fascinating contrasts do exist, however, such as the preference for Greek glukútatos ‘sweetest’ as an extremely popular epithet in pagan and Christian inscriptions, in contrast to Greek hósios ‘holy, devout’ in Jewish ones – a descriptor never found in pagan or Christian burial inscriptions. 14 Bruno (2012: 370) likewise regards features like the invariability of the participle, its usual agreement with the direct object, and its inert diathesis as potentially indicative of Latin influence. 15 Löfstedt (1959: 108–10) presents an array of Greek administrative terms borrowed from Latin: military terms, e.g. kentēnárion (centenarium), words for money and weights and measures, e.g. skála (scala); legal terminology, e.g. fourth‐century Greek papyri brébion (breve), etc., as well as some general terms, e.g. phamília (familia), etc. To this list Löfstedt (1959: 119) adds the name that Greeks call themselves even to this day, Rhomiós, which harkens back to the division of the Roman Empire into East and West by Constantine.

REFERENCES Adams, J. N. 2002. Bilingualism at Delos. In J. N. Adams, Mark Hanes, and Simon Swain (eds.), Bilingualism in Ancient Society: Language Contact and the Written Text. Oxford: Oxford University Press, pp. 103–27. Adams, J. N. 2003. Bilingualism and the Latin Language. Cambridge: Cambridge University Press. Benelli, Enrico 1994. Le iscrizioni bilingui etrusco‐ latine (Biblioteca di Studi Etruschi 27). Florence: Leo S. Olschki. Bennett, William H. 1968. The operation and relative chronology of Verner’s Law. Language 44.2: 219–23. Biville, Frédérique 2002. The Graeco‐Romans and Graeco‐Latin: a terminological framework for

cases of bilingualism. In J. N. Adams, Mark Hanes, and Simon Swain (eds.) Bilingualism in Ancient Society: Language Contact and the Written Text. Oxford: Oxford University Press, pp. 77–102. Bruno, Carla 2012. On a Latin‐Greek diachronic convergence: the perfects with Latin habeo/ Greek échô and a participle. In Claudine Chamoreau and Isabelle Léglise (eds.), Dynamics of Contact‐Induced Language Change. Berlin: de Gruyter Mouton, pp. 359–75. Campanile, Enrico 1979. Indogermanische Metrik und altirische Metrik. Zeitschrift für celtische Philologie 37: 174–202. Clackson, James and Geoffrey Horrocks 2011. The Blackwell History of the Latin Language. Oxford: Wiley‐Blackwell.

320  Bridget Drinka Coseriu, Eugenio 2008. Lateinisch – Romanisch: Vorlesungen und Abhandlungen zum sogenannten Vulgärlatein und zur Entstehung der romanischen Sprachen. Tübingen: Günter Narr. Devoto, Giacomo 1940/1968. Storia della lingua di Roma. Bologna: Accademia. (Translated in 1968 by I. Opelt as Geschichte der Sprache Roms. Heidelberg: Winter.). Drinka, Bridget 1995. The sigmatic aorist in Indo‐European: evidence for the Space‐Time Hypothesis. Journal of Indo‐European Studies Monograph 13. Drinka, Bridget 2013. Phylogenetic and areal models of Indo‐European relatedness: the role of contact in reconstruction. Journal of Language Contact 6: 379–410. Drinka, Bridget 2017. Language Contact in Europe: The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Euler, Wolfram 1993. Oskisch‐Umbrisch, Venetisch un Lateinisch – grammatische Kategorien zur inneritalischen Sprachverwandtschaft. In Helmut Rix (ed.), Oskische‐Umbrisch Texte und Grammatik. Wiesbaden: Reichert, pp. 96–105. Evans, D. Ellis 1981. Celts and Germans. The Bulletin of the Board of Celtic Studies, University of Wales 29: 230–55. Gleba, Margerita 2008. Textile Production in Pre‐ Roman Italy. Oxford: Oxbow. Grassmann, Hermann 1863. Über die Aspiranten und ihr gleichzeitiges Vorhandensein im An‐ und Auslaut der Wurzel. Zeitschrift für vergleichende Sprachforschung. 12: 81–138. Kortlandt, Frederik H. H. 1988. Proto‐Germanic obstruents. Amsterdamer Beiträge zur Älteren Germanistik 27: 3–10. Kortlandt, Frederik H. H. 2007. Italo‐Celtic Origins and Prehistoric Development of the Irish Language. Amsterdam: Rodopi. Kramer, Johannes 2011. Von der Papyrologie zur Romanistik. Berlin: Walter de Gruyter. Kuhn, Hans 1959. Vor‐ und frühgermanische Ortsnamen in Nord‐Deutschland und in den Niederlanden, Westfälische Forschungen 12: 5–44. Kuhn, Hans 1961. Anlautend p im Germanischen. Zeitschrift für Mundartforschung 28: 1–31. Kury3owicz, Jerzy 1968. Indogermanische Grammatik. II. Akzent‐Ablaut. Heidelberg: Winter. Langslow, D. R. 2002. Approaching bilingualism in corpus languages. In J. N. Adams, Mark Hanes, and Simon Swain (eds.), Bilingualism in Ancient Society: Language Contact and the

Written Text. Oxford: Oxford University Press, pp. 23–51. Lehmann, Winfred P. 1977. Language contacts and interference in the Germanic period. In Herbert Kolb and Hartmut Lauffer (eds.), Sprachliche Interferenz. Festschrift für Werner Betz. Tübingen: Niemeyer, pp. 278–91. Leiwo, Martti 1994. Neapolitana: A Study of Population and Language in Graeco‐Roman Naples. Helsinki: Societas Scientiarum Fennica. Leiwo, Martti 2002. From contact to mixture: bilingual inscriptions from Italy. In J. N. Adams, Mark Hanes, and Simon Swain (eds.), Bilingualism in Ancient Society: Language Contact and the Written Text. Oxford: Oxford University Press, pp. 168–94. Löfstedt, Einar 1959. Late Latin. Oslo: Instituttet for Sammenlignende Kulturforskning. Lühr, Rosemarie 2018. The Language of the Nordwestblock. In Gus Kroonen, James P. Mallory, and Bernard Comrie (eds.). Talking Neolithic. Journal of Indo‐European Studies Monograph Series 65. Washington DC: Institute for the Study of Man, pp. 174–95. Marrou, H. I. 1977. A History of Education in Antiquity. New York: New American Library. Meid, Wolfgang 1986. Hans Kuhns “Nordwestblock” Hypothese: Zur Problematik der Völker zwischen Germanen und Kelten. In Heinreich Beck (ed.), Germanenprobleme in heutiger Sicht, 2nd edition. Berlin: de Gruyter, pp. 183–212. Meillet, Antoine 1908. Les dialects indo‐européens. Paris: Honoré Champion. Meiser, Gerhard 2003. Veni vidi vici. Die Vorgeschichte des lateinischen Perfektsystems. Munich: C. H. Beck. Mohrmann, Christine 1958–77. Études sur le latin des chrétiens. I–IV. Rome: Edizioni di Storia e Letteratura (four volumes, published in 1958, 1961, 1965, 1977). Pisani, Vittore 1933. Studi sulla preistoria delle lingue indeuropee. Atti della Reale Academia Nazionale dei Lincei. Memorie. Ser. VI. Classe di scienze morali Vol. IV. Fasc. VI. Pisani, Vittore 1940. Geolinguistica e indeuropeo. Atti della Reale Academia Nazionale dei Lincei. Memorie. Ser. VII. Classe di scienze morali. IX, fasc. II. Pisani, Vittore 1949. La question de l’Indo‐Hittite, et le concept de parenté linguistique. Archiv Orientálni 17.251–64. (Reprinted in Saggi di linguistica storica 1959: 43–60). Polomé, Edgar C. 1972. Germanic and the other Indo‐European languages. In Frans van

Contact and Early Indo‐European in Europe  321 Coetsem and Hebert L. Kufner (eds.), Toward a Grammar of Proto‐Germanic. Tübingen: Max Niemeyer, pp. 43–69. Polomé, Edgar C. 1983. The linguistic situation in the western provinces of the Roman Empire. In Hildegard Temporini and Wolfgang Haase (eds.), Aufstieg und Niedergang der römischen Welt. II. Berlin: de Gruyter, pp. 509–53. Porzig, Walter 1954. Die Gliederung des indogermanischen Sprachgebiets. Heidelberg: Winter. Poultney, James 1959. The Bronze Tables of Iguvium. Baltimore: The American Philological Association. Ringe, Don, Tandy Warnow, and Ann Taylor 2002. Indo‐European and computational cladistics. Transactions of the Philological Society 100: 59–129. Rix, Helmut 1956. Die Personennamen auf den etruskisch‐lateinischen Bilinguen. Beiträge zur Namenforschung 7: 147–72. Rix, Helmut 1991. Etruskische Texte: Editio minor (2 vols.) Tübingen: Günter Narr. Rix, Helmut 1994. Latein und Sabellisch: Stammbaum und/oder Sprachbund. Incontri Linguistici 7: 13–29. Rutgers, Leonard 1995. The Jews in Late Ancient Rome: Evidence of Cultural Interaction in the Roman Diaspora. Leiden: Brill. Salmons, Joseph 2003. Accentual Change and Language Contact: Comparative Survey and a Case Study of Early Northern Europe. London: Routledge. Schleicher, August 1861–2. Compendium der vergleichenden Grammatik der indogermanischen sprachen. Weimar: Böhlau. Schmidt, Johannes 1872. Die Verwandtschaftverhältnisse der indogermanischen Sprachen. Weimar: Böhlau. Solodow, Joseph 2010. Latin Alive: The Survival of Latin in English and the Romance Languages. Cambridge: Cambridge University Press. Sommerfelt, Alf 1962. Diachronic and Synchronic Aspects of Language. The Hague: de Gruyter Mouton. Stifter, David 2016. Metrical systems of Celtic traditions. North‐Western European Language Evolution 69.1: 38–94. Streitberg, Wilhelm 1896. Urgermanische Grammatik. Einführung in das vergleichende

Studium der altgermanischen Dialekte. Heidelberg: Winter. Suzuki, Seiichi 1988. The Indo‐European basis of Germanic alliterative verse. Lingua 75: 1–24. Swain, Simon 2002. Bilingualism in Cicero? The evidence of code‐switching. In J. N. Adams, Mark Hanes, and Simon Swain (eds.), Bilingualism in Ancient Society: Language Contact and the Written Text. Oxford: Oxford University Press, pp. 128–67. Szemerényi, Oswald 1996. Introduction to Indo‐ European Linguistics. (Translated from Einführung in die vergleichende Sprachwissenschaft, 4th edition, 1990, with additional notes and references). Oxford: Oxford University Press. Ternes, Elmar 1998. Keltisch und Romanisch. In Günter Holtus, Michael Metzeltin, and Christian Schmitt (eds), Lexicon der Romanistischen Linguistik VII. Kontakt, Migration, und Kunstsprachen. Kontrastivität, Klassifikation und Typologie, pp. 266–91. Verner, Karl 1877 [1978]. An Exception to Grimm’s Law. In Philip Baldi and Ronald Werth (eds.), Readings in Historical Phonology: Chapters in the Theory of Sound change. University Park, PA: Pennsylvania State University Press, pp. 32–63. (Translation of Eine ausnahme der ersten lautverschiebung. 1877. Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der Indogermanischen Sprachen 23.2: 97–130.) Voyles, James B. 1988. Early Germanic changes in unstressed word‐final syllables: problems and ramifications. Lingua 76: 63–90. Watkins, Calvert 1966. Italo‐Celtic revisited. In Henrik Birnbaum and Jaan Puhvel (eds.), Ancient Indo‐European Dialects. Berkeley: University of California Press, pp. 29–50. Watkins, Calvert 1982. Aspects of Indo‐European poetics. In Edgar Polomé (ed.), The Indo‐ Europeans in the Fourth and Third Millennia. Ann Arbor, MI: Karoma, pp. 104–20. Weiss, Michael 2012. Italo‐Celtica: linguistic and cultural points of contact between Italic and Celtic. In Stephanie W. Jamison, H. Craig Melchert, and Brent Vine (eds.), Proceedings of the 23rd Annual UCLA Indo‐European Conference. Bremen: Hempen, pp. 151–73.

16 Contact and the History of Germanic Languages PAUL ROBERGE 1  The Germanic Languages The Germanic languages are, as the phrase suggests, a group of languages that trace their origin to a common ancestor and constitute a branch of the Indo‐European language family. Prior to the beginning of the present era, as per the traditional view, Proto‐Germanic (in German Urgermanisch), the hypothetical parent language, is supposed to have been “a fairly homogenous linguistic and cultural unit” (Prokosch 1939: 26) existing at a given point in time. Lehmann (1977: 287) asserts that “we … have good evidence to conclude that Germanic was little influenced in structure by external contacts leading to interference until approximately the middle of the first millennium b.c.” Yet, none of this is as straightforward as it seems. Over the course of approximately a millennium and a half, the ancestral language fragmented into dialects, which ultimately gave rise to the universally recognized independent languages of the European metropole, in all their varieties: English, German, Dutch, and Frisian comprise the West Germanic group; Danish, Norwegian, and Swedish  –  together with the insular Scandinavian languages of Faroese and Icelandic  –  make up the North Germanic group. A third group, East Germanic (of which Gothic is the principal representative), has been extinct since at least the sixteenth century. The mechanisms of linguistic diversification are conventionally understood in terms of speciation following the separation of ­peoples from the core group. Accordingly, the structural distance between the daughter dialects is the aggregate of a series of (mostly) incremental system‐internal mutations that have taken place during the uninterrupted intergenerational transmission of grammar. Yet, none of these languages has developed without contact with others.

2  Delimiting Prehistoric Germanic in Time and Space The question of Indo‐European origins and the related problems of proto‐culture and diffusion have generated an enormous body of scholarship. Yet, we still cannot say with certainty where and when Proto‐Indo‐European was spoken. Ringe (2006: 4) asserts that a reasonable guess “would be the river valleys of Ukraine in the centuries around 4000 bc, though one can’t absolutely exclude a somewhat earlier date, nor a place somewhat further east.” The subsequent period of some 2,000 years saw the advance of Indo‐European groups

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

324  Paul Roberge into Anatolia, into Persia and the Indian subcontinent, the Balkan peninsula, and across Europe all the way to the Atlantic. Lehmann estimates a final date of ca. 2500 bce for Proto‐Indo‐European (1961: 73). According to Anthony (2007), the pastoralist Yamnaya culture emerged from previous cultures on the steppe and spread over a vast region, their mobility having been enabled by the wheel and the horse. The shared vocabulary for cart technology across the Indo‐European family – with the notable exception of Anatolian – suggests that its languages descend from an ancient population that made use of wagons. Studies of ancient DNA show that the Yamnaya carried “exactly the type of ancestry that needed to be added to early European farmers and hunter‐gatherers to produce populations with the mixture of ancestries in Europe today” (Reich 2018: 108). Genetic data have also established a link between the Yamnaya and Corded Ware cultures, which show striking parallels, such as the construction of large burial mounds, economic activity grounded in horse domestication and pastoralism, and weaponry (maces buried in some graves). In genetic terms, at least, the creators of the Corded Ware culture were a westward extension of the Yamnaya (Reich 2018: 110). The genetic findings support the idea that a mass migration of people from the steppe north of the Black Sea and Caspian Sea into Europe occurred after 5,000 years ago through which the Yamnaya and/or groups closely related to them introduced varieties that would develop into the Indo‐European languages (Reich, 2018: 119). With these considerations in mind, we may revisit Polomé’s belief (1987: 234–6; 1992a: 48) that the population movement that brought “Indo‐Europeans” to northern Europe started towards the end of the fourth millennium bce or in the third millennium bce, being tied to the Corded Ware culture that flourished ca. 2500 bce and is a parent of “Germanic” and other groups. However, the relevant groups did not arrive en masse directly from the Indo‐European homeland but rather separated from the core speech community at staggered points in time. At this early stage, dialectal variation in the vernacular spoken by Indo‐European migrants would have been minor. We can speak of a regionally slightly differentiated linguistic continuum across northwestern Europe until the second millennium bce, from which time the major linguistic changes that would lead to thoroughly differentiated dialect clusters were taking place. The period between Lehmann’s terminus ad quem for Proto‐Indo‐European and the middle of the first millennium bce is often (and vaguely) characterized as “pre‐Germanic.” The beginnings of Germanic are supposed to have overlapped with early stages of other northwestern Indo‐European dialects. The founder population of the Germanic habitus is commonly assumed to have established itself along the northern flank of the Indo‐European spread in the late Bronze Age (second millennium bce), extending over what is today southern Sweden and Denmark – with their neighboring islands – the coastal fringe of Norway, and Schleswig‐Holstein, bounded by the lower Elbe River and perhaps running along the Baltic Sea coast to Rügen. From there, some groups advanced toward the Oder River while others followed the coast of the North Sea, toward the Weser. If we follow Musset (1975: 8), Germanic peoples had expanded into the northern German lowlands between the Ems and Pomerania by 1000 bce; by 800 bce they had reached Westphalia in the west and the Vistula River region in the East; and by 500 bce they had settled in the lower Rhine, Thuringia, and lower Silesia. In this area we find the Jastorf culture, from roughly 600–300 bce, which developed directly out of the northern Bronze Age culture (Mallory 1989: 87). By 300 bce Germanic peoples had migrated as far as the Netherlands and northern Belgium. The Harpstedt culture of northwest Germany and eastern Holland is thought to be related to the Jastorf culture, and both “have provided Germanicists with a generally agreed‐upon Germanic homeland” (Mallory 1989: 87). Chronologically, the middle of the last millennium bce has been proposed for the formation of Proto‐Germanic (e.g. Polomé 1992a: 47; Ringe 2006: 213). Late Proto‐Germanic (known also as Common Germanic, in German Gemeingermanisch) extends through the

Contact and the History of Germanic Languages  325 second or first century bce (Penzl 1972: 23), possibly reaching into the present era (see Polomé 1987: 219, with references). It is conventionally assumed that Proto‐Germanic formed in southern Scandinavia and in the northern German lowlands, in the regions above the Hercynius saltus (Hercynian Forest) to which Julius Caesar (Commentarii de Bello Gallico, chs. 6, 24–5), Tacitus (Germania, chs. 28, 30), and other classical writers make reference (cf. Prokosch 1939: 26). For comparative purposes, distinguishing Proto‐Germanic geographically has been less important than establishing a relative chronology (Lehmann 1961: 68). In principle, an assessment of prehistoric language contact would require that temporal and spatial delimitations of cultures, population movements, and hypothesized speech communities be brought into some kind of accord (Evans, 1981: 233; Polomé 1985: 49–50). In practice, of course, neither archeology nor genetics can reveal what languages people spoke. What these disciplines can do is establish that migrations occurred. If people moved, it means that language contact occurred, too. Prehistoric contact in northwestern Europe begins with linguistic and cultural encounters between immigrants and autochthonous (substratal) populations. There followed extensive contacts between northwestern Indo‐ European dialects and Finno‐Ugric languages, and a somewhat later period of contacts amongst the northwestern Indo‐European languages themselves, cf. Baltic, Slavic, Celtic, Italic, and Germanic.

2.1  Non‐Indo‐European Substratal Influences There is no doubt that however northern Europe was Indo‐Europeanized, the in‐migrating vanguard “initially constituted a mere adstratum or superstratum to a long‐established set of peoples. When and why the language shift took place remains a widely open question, but one thing is certain: it did not take place without leaving clear traces of the prior language(s) in the lexicon” (Polomé 1990: 337). The most compelling linguistic argument for substrate influence rests on a substantial number of terms in Germanic that do not admit satisfactory Indo‐European etymologies. Such lexis is quite plausibly ascribed to earlier populations in northern Europe that were Indo‐Europeanized or displaced following a series of in‐migrations from the Eurasian steppe (see Polomé 1986, 1989, 1992b; Vennemann 2003: 1–20). The Germanic term for ‘apple’ and its congeners in Celtic, Baltic, and Slavic may obtain from a non‐Indo‐European form (Markey 1989b: 599–600; Vennemann 2003: 466–8). The Germanic term for ‘silver’ likely reflects substratal (possibly Vasconic) influence (cf. Polomé 1987: 229; Vennemann 2003: 346– 7). Some Germanic words appear not even to have plausible cognates in the other Indo‐ European languages, as, for example, the Germanic term for ‘eel’ (cf. Polomé 1992a: 51–2; 1992b: 78–9). Salmons (2004) cautions that even when lexical items appear to satisfy the criteria for substratum status (set out by Polomé 1989: 54–5), they could still reflect internal neologisms. At the same time, a word originating in a substrate language could have become so completely integrated as to behave like an original Indo‐European root, as appears to be the case with *bhar‐ ‘spinous, prickly’, *bhares‐ ‘barley’ in northwestern Europe (Markey 1989b: 589–93; Polomé 1992b: 70–6; Seebold 2015: 13–16). Our expectation is that words belonging to an archaic vocabulary would refer to the natural environment, to certain human feelings and activities, and to objects “pertaining to the level of culture in the prehistoric period of symbiosis between the newcomers and the people already established in the area” (Polomé 1986: 661). On the assumption that the subsratum comprised a settled farming population, retention of local agricultural terms would be a further expectation (cf. Markey 1989b; Kroonen 2012). Germanic maritime and nautical terminology has drawn special attention as a potential repository of non‐Indo‐European material. PGmc. *saiwa‐ (or *saiwi‐) ‘sea’, with reflexes in all of the daughter dialects, has no accepted Indo‐European etymology despite numerous proposals (summarized in Lehmann

326  Paul Roberge 1986: 292). Sausverde (1996) argues that pre‐Germanic peoples and their Indo‐European ancestors were not acquainted with the sea. In her view the Germanic word for ‘sea’ is a borrowing from the language of autochthonous pre‐European inhabitants of the Baltic and North Sea coasts. A number of other terms in this lexical field represent new meanings assigned to existing forms or are neologisms constructed out of existing morphemes. Yet, linguists armed with the same set of data have drawn quite different conclusions. Witczak (1996: 174–5) dismisses claims about the substratal origin of ‘eel’ and of the semantic group associated with the sea and seafaring. Our expectation is that the oldest linguistic material in a geographic area would be toponymic, with names for rivers and lakes forming a particularly conservative layer. With respect to place names in Western and Central Europe, Nielsen (1989: 22) suspects that Krahe (1964) may have “underrated the role played by the pre‐Indo‐European nomenclature,” while Vennemann (2003) relates many of Krahe’s “Old European” (i.e. West Indo‐European) hydronyms to substratal Vasconic. Some linguists have posited prehistorical contacts between Indo‐European and non‐ Indo‐European population strata of sufficient intensity as to leave an imprint on the structure of a pre‐Germanic recipient language. The Germanic consonant shift has sometimes been considered a phonological transfer from a substratum (e.g. Witczak 1996: 167–9) or a superstratal imposition of a phonological pattern on the developing system of an indigenous pre‐Germanic population in the course of language shift (Feist 1928, 1932). Mainstream comparative Germanic linguistics has consistently resisted such proposals (e.g. Polomé 1992b: 77–8, 1997; Mees 2003: 19–21). A longstanding postulate holds that Germanic lost a number of original Indo‐European verbal categories – such as the imperfect, the aorist, the forms of the subjunctive, and the medio‐passive (save for a vestigial presence in Gothic) – in addition to a reduction of the nominal case system. Neither the social situation (to the extent that we can reconstruct it) nor the linguistic facts indicate creolization by any rational definition of the term (cf. Vennemann 2012: 422–45). It could well be that the Indo‐European component of Germanic does, in fact, reflect the oldest stage of Indo‐European spoken by the first wave of emigrants to leave the homeland. In this scenario, the early ancestors of the Germanic peoples would have departed before the Indo‐European verbal system had developed its three‐stem pattern (present, aorist, perfect) that is imputed to the protolanguage on the basis of Greek and Sanskrit, the preterital categories of imperfect and aorist, and the modal categories of subjunctive and future (but after the formation of the optative); cf. Polomé (1987: 234–6). In numerous publications (collected 2003, 2012), Vennemann tackles the fundamental questions that arise in connection with pre‐Indo European strata, as part of a research program: who were these people? Where did they come from? Which languages did they speak? In Vennemann’s theory, languages of three filiations were spoken in prehistoric Europe north of the Alps: Old European (a branch of Vasconic, of which Modern Basque is the sole survivor), Atlantic (a branch of “Semitidic,” i.e. Mediterranean Hamito‐Semitic), and West Indo‐European.

2.2  Germanic Contacts with Finno‐Ugric In addition to encounters between dialects within its own genetic spread, northwest Indo‐ European came into a sustained contact with Finno‐Ugric at the beginning of the first millennium bce or perhaps even as early as the latter half of the second millennium bce (cf. Koivulehto 2002: 583–5). This contact continued “apparently without noticeable interruptions” through the pre‐, Proto‐, and Common Germanic stages and would extend through the historical periods until the present time (Koivulehto 2002: 590). Germanic influences manifest themselves most visibly in the form of loanwords in (mainly) Finnic and/or Sámi.

Contact and the History of Germanic Languages  327 These loans pattern themselves in chronological layers, which can be adduced from the phonological features that the words exhibit in the recipient languages. The lowering of PGmc. *ē1 to *ā took place in Northwest Germanic after the breakup of Proto‐Germanic and is assignable perhaps to the first century bce (see Koivulehto 1981). A cross‐cultural term for ‘sword’ that is ultimately related to Greek μáχαιρα ‘knife, dagger’ entered Germanic as a secondarily suffixed *meg‐(i)yo‐ (from a primary root *meg‐) with the lengthened grade (Go. mēki ‘short sword’ beside ON mækir, OE mǣ ce, OS māki ‘sword’). From Germanic, *mēk(i)ja‐ ‘sword’ spread to neighboring Slavic and Finnic dialects (Markey 1999: 146–53). We may infer from the vocalism that Finnish miekka ‘sword’ was part of an early layer of borrowings from Germanic, having been adopted in the mid first millennium bce but obviously prior to the lowering of PGmc. *ē1 to *ā. Owing to the conservatism of some Finnish dialects over the past two millennia, the oldest Germanic loanwords constitute important evidence in the reconstruction of Proto‐ Germanic, even if their value has been overestimated at times (see Juntune 1973 for discussion). Finnish rengas, ‘ring’, for example, is supposed to confirm *hrengaz as the correct reconstruction of this lexical item in Proto‐Germanic, whereas all attested Germanic languages show i in the tonic syllable before a nasal plus a consonant (ON hringr, OHG/OE/OS hring). Finnish rengas, kuningas ‘king’ (< PGmc. *kuningaz; cf. ON konungr, OHG cuning, OS kuning, OE cyning) appear to preserve the Proto‐Germanic masculine nominative singular ending *‐az (< PIE *‐os), which, following the fixation of accent on the first or root syllable, became reduced or lost altogether in the daughter dialects. Koivulehto has proposed numerous etymologies (2002: 585–90) that would show the Germanic contribution to Finnic and Sámi lexicons to be sizable; see further Kallio (2012, 2015). The deep lexical impact of early Germanic on Finnic might lead one to question whether the standard view of adstratal influence adequately captures the contact situation in prehistoric northwestern Europe (Juntune 1994). It has in fact been suggested that Germanic also influenced Finnic phonological development owing to the existence of a Germanic‐ speaking superstratum in regions populated by Finnic groups. Koivulehto (2002: 590–1) seeks to explain the Finnic consonant gradation (the lenition of p, t, k, s in certain environments) as the culmination of internal phonetic tendencies that were enhanced by external influence, namely, from a Germanic Verner’s Law that antedates the Germanic consonant shift proper (cf. Koivulehto and Vennemann 1996), which of course is in itself a controversial proposal. The unidirectionality of early Germanic lexical transfer suggests that this putative Germanic superstratum did not maintain its position for a long period and was gradually assimilated into the native population. Taking a somewhat different tack, Salmons (1992: 180) has written that prehistoric northwestern Europe constituted “in some ways a (weak) … linguistic area, sharing features evolved and borrowed during the last millennium before our era or early in the first millennium of this era. Much key cultural and technical vocabulary from the Iron Age (as well as some from slightly earlier and later) is shared across a wide area that included the Germanic, Celtic, Finnish, Lappish, Baltic, Slavic, and Italic areas.” Within this area, Germanic, Celtic, and Finno‐Ugric converged on a prosodic trait, namely, fixed initial stress. As concerns chronology, Salmons’s hypothesis squares with the period of Celtic‐Germanic contact, from the mid first millennium bce the source of this accentual pattern cannot be determined with certainty, but Salmons concludes that “a Finno‐Ugric language (or languages) may seem perhaps more plausible than the alternatives, given the reconstructed Proto‐Finno‐Ugric accent and early and profound contacts by trade and as long‐time neighbors” (1992: 173). Schrijver (2014: ch.5) takes matters a step further by proposing that Germanic, as a separate branch of Indo‐European, arose as a result of contact with Balto‐Finnic, “probably as a result of Indo‐European–Finno‐Sámic bilingualism and a subsequent switch to Indo‐ European” (2014: 197). Accordingly, the early fragmentation of Proto‐Germanic probably

328  Paul Roberge occurred when speakers of a “lost northern European language or languages … came into contact with Germanic, to which they ultimately switched” (2014: 197). A lost language with similar phonological properties was probably involved in turning Finno‐Sámic into Proto‐ Sámi. The absorption of this group of languages accompanied the spread of Finno‐Sámic in Finland and in central and northern Sweden and Norway, and the spread of Germanic in southern Scandinavia and presumably into coastal Germany.

2.3 Language Contact Within the Northwest Indo‐European Spread Linguistically, Indo‐European “constitutes a complicated blend in which the proportions of the common elements vary greatly from branch to branch” (Prokosch 1939: 23). Some Indo‐ European language groups show greater affinities to one another than to other groups within the family. Comparative linguistics has long been concerned with (inter alia) the earliest relationships between the ancestors of the various branches of Indo‐European. Germanic shares significant commonalities with Baltic, Slavic, Celtic, and Italic, along with some possible links to Illyrian and Venetic; see Polomé (1972) and Nielsen (1989: 18–28) for concise surveys. These correspondences are preponderantly lexical. Because we cannot establish with precision what position should be accorded to Germanic in the Indo‐European complex, it will be difficult in many cases to know for sure whether a given correspondence came about due to common heritage or borrowing between adjacent or co‐territorial sister dialects. Polomé (1972: 54) cautions that “the acceptable lexical evidence exclusively shared by the Gmc., Baltic, and Slavic tribes is hardly sufficient to draw any definite conclusions as to their close relationship or even the level of civilization that they had reached at the time of their contact.” There seem to be just 10 to 20 old Germanic loans in Baltic, by Koivulehto’s count (2002: 591), belonging to different chronological layers; and several of these are thought to have passed through Slavic, in which there are considerably more Germanic forms. Slavic borrowed the Germanic word for ‘bread’ (PGmc. *xlaibaz), as attested by Old Church Slavonic chlěbь ‘bread’. Lithuanian klı̃epas ‘loaf of bread’ is conventionally seen as a younger loan via Belarusian (cf. Lehmann 1986: 186), though Koivulehto (2002: 592) raises the possibility of an older borrowing directly from Germanic, which, a fortiori, would be compelling if Latvian klàips ‘bread’ is drawn from early Germanic rather than from Gothic hlaifs, gen. hlaibis. Germanic, Celtic, and Italic have preserved correspondences that are suggestive of early contact between speakers of all three branches. Relations between Italic and Germanic peoples during the Bronze Age are suggested by the fact that Italic and Germanic (but not Celtic) share a common word for a metal: Latin aes, aeris ‘copper, copper ore, money’, Go. aiz ‘money, metal coin’, ON eir ‘copper’, OHG/OS ēr, OE ār, ǣ r‘ore’. After the southward migration of Italic groups from the last third of the second millennium bce, one might suppose that Celtic and Germanic would have developed together for a longer period as closely related, mutually intelligible dialects in a situation of contact (Pokorny 1936: 508). One problematic aspect of this hypothesis is that there are no incontestible common innovations between them in phonology and grammar, to the exclusion of the other Indo‐European branches; Germano‐Celtic correspondences are confined to the lexicon (Polomé, 1972: 64; Evans, 1981: 242–3). For Evans (1981: 252–3), this is in itself a strong indication that the two groups were not in close association until a fairly late period. From 400 bce to 100 bce, the advance of Germanic peoples southward brought them into contact with the Celts, who were moving north. It was once believed that Celtic hegemony over the Germanic world accounted for a long list of lexical correspondences (see Mees, 2003: 18–19, 25–6 for references and discussion of the older literature). Some contemporary scholars have averred that Germanic must have occupied a culturally subordinate position

Contact and the History of Germanic Languages  329 with regard to Celtic (Polomé 1972: 67; Lehmann 1977: 288–9, 1987: 80). Evans (1981: 253–4) insists that there is no historical justification at all for assuming the domination of the one group by the other, but cultural preeminence does not imply political control (cf. Polomé 1985: 52). Salmons (1992: 95–6) observes that “subordinate” lends itself to different interpretations ranging from “subjugated” to “less prestigious.” Minimal “subordination” in the form of proximity to a somewhat more advanced civilization in terms of technology and social organization – combined with utility and/or novelty – would be sufficient motivation for lexical borrowing. In later work Polomé (1987: 221) adopts a more conservative approach to interpreting Germano‐Celtic agreements in vocabulary: “The type of relationship that existed between the Germanic people and the Celts is … hard to define by linguistic means: when W. P. Lehmann [1987] argues from legal and institutional terminology that the Celts were the givers and the less developed Germanic people the receivers in the linguistic interchange, he may be overstating his case.” Estimates of the Celtic Lehngut in early Germanic range from 50 or 60 items (Salmons 1992: 96) to a mere 10 or fewer (Lane 1933: 263–4; Evans 1981: 248). In keeping with Polomé’s conservatism, shared terms are best considered cognates belonging to a larger institutional framework of similar, possibly inherited, legal, social, and political structures or reflecting common regional innovations and localized archaisms  –  unless there is linguistic and/or cultural evidence for their being borrowed (Polomé 1987: 221). The Germanic term *rı̄k‐ ‘ruler’ (Gothic reiks ‘ruler’ ON ríkr ‘mighty’, OHG masc. rı̄hhi ‘ruler, king’, neut. rı̄hhi ‘kingdom’, etc.) is arguably an early borrowing from Celtic. Other Germanic words that are widely accepted as Celtic loans include PGmc. *ambahtaz ‘servant’ and PGmc. *ı̄sarna‐ ‘iron’ (Polomé 1985: 57). Lexical transfer in the opposite direction, from Germanic into Celtic, amounts to perhaps two or three examples (Lane 1933: 264). So‐called Wanderwörter represent a special category of loanwords that spread across languages, usually in connection with trade or the adoption of external technological, economic, or cultural practices. Germanic absorbed lexis (and associated items of material culture) that traveled through contiguous Indo‐European (and non‐Indo‐European) ­language groups well after their individualization. The cultivation of hemp is possibly Scythian in origin, and its introduction to Germanic areas (or at least knowledge of the plant), perhaps by way of a neighboring language, antedates the Germanic consonant shift: PGmc. *xanap‐ < *kan(n)ab‐ (Greek κáνναβις, Latin cannabis, cannabum) yielding OHG  hanaf, OS hanap, OE hænep, ON hampr ‘hemp’. A non‐Indo‐European etymon for *kannab‐ is not unthinkable, but source languages for such cross‐cultural words are often obscure; see further in Seebold (2015: 2–3).

3 Language Contact Following the Breakup of Proto‐Germanic Lehmann (1977: 285) claims that one of the most notable characteristics of the Germanic branch is its lack of dialect differences at the time of Proto‐Germanic: “The unity of Germanic is striking when compared for example with the diversity of Greek.” This uniformity is best accounted for by assuming “a stable cohesive community, presumably that located around the Baltic for a millennium or more.” Major innovations – such as the fixation of accent, the Germanic consonant shift, and dual (strong/weak) adjectival inflection based on definiteness rather than stem class – diffused across the entire speech community; dialect differences arose in Germanic at a late period. Yet, imputation of a single, unified speech community largely free of dialect variation to Germanic prehistory (Lehmann 1977: 285–6) is doubtless an oversimplification (cf. Polomé 1972: 45).

330  Paul Roberge The actual situation may have been more akin to a chain of interrelated dialects with no clear internal boundaries. The traditional visualization of such a unity incrementally fracturing into three discrete subgroups (North, West, and East Germanic) without contact or mutual influence has long been abandoned (Kufner 1972: 94–5; Scardigli 2002). Although there are isoglosses that connect North and East Germanic, North and West Germanic exhibit a numerically preponderant set of linguistic affinities that are not shared by East Germanic and are more heavily weighted in their probative value. East and West Germanic have virtually no isoglosses in common with each other, save for some pronominal forms in Gothic and Old High German (Go. 3 sg. masc. nom. sg. is, acc. sg. ina, OHG ir/er, in(an) beside OS hē, ina(n), OE hē, hine; ON nom./acc. hann). A period of common development in North and West Germanic languages after the migration of the Goths from eastern Scandinavia is widely assumed, at least as a working hypothesis (e.g. Kuhn 1955; Kufner 1972: 94; Haugen 1976: 111–12; Ringe 2006: 213). For its part, Gothic developed more or less separately along the shores of the Baltic Sea, east of the Vistula, from the second century bce to the second century ce. Intermediate relationships in Germanic, which have long been the subject of scholarly debate (see especially Nielsen 1989: ch. 4 for a critical review of the literature), do not directly concern us here. Suffice it to say that Germanic peoples “lived at various times after the PGmc. stage in differing constellations of proximity, [and] retained close contacts with each another even after migrations, so that linguistic borrowing [between dialects] is likely to have taken place for long periods of time” (Kufner 1972: 74). With fragmentation into dialects and increasing temporal remove from the ancestral language, the concept Germanic comes to refer not to a historical object but rather a genetic affiliation. The history of contact in Germanic languages becomes the diachronic record of discrete speech communities, the outcomes of which depend on the nature, duration, and intensity of the linguistic encounters. Salmons (1992: 90) makes essentially this point when he writes that by the early Roman period, Celtic‐Germanic contacts cover much less territory than earlier contacts might have, indicating greater complexity. The evidence from the historical sources makes it clear that in Roman times contact continued between Celts and Germanic peoples living along the common frontier between them. Caesar described Gaul at the time of his conquests (58–51 bce) as being divided into three parts inhabited by the Aquitani, the Gauls (who in their own language are the Celts), and the Belgae, each differing from the other in language, customs, and laws (Commentarii de Bello Gallico: ch. 1, 1). His sources informed him that one Belgic tribe, the Remi, are descended from Germanic peoples whose forebears had crossed the Rhine to settle in a more fertile region, driving out the Gaulish natives; this tribe is said to have resisted the Cimbri and Teutoni, who would go on to harry Roman territories from ca. 113–101 bce. Other Belgic tribes – the Condrusi, Eburones, Caeroesi, Paemani – are also said to be of Germanic descent (Commentarii de Bello Gallico: ch. 4, ch. 6: 32). Caesar’s report invites speculation over whether the Belgae were an ethnically mixed Gaulish‐Germanic group. Tacitus retreated somewhat from his (in)famous supposition that the Germani were an ethnically homogenous, indigenous people (Germania: chs. 2 and 4) when he alludes to migration, mixing, and language shift on the part of certain groups. During the first centuries of the common era, Latin was the medium of communication in the “vertical” domains of administration, cultured discourse, and the military in territories west of the Rhine and south of the Danube. One is tempted to conjecture that there must have been a Latin‐based lingua franca or trade jargon – mixed with Germanic and Gaulish forms – along the long frontier between the Roman Empire and unoccupied Germania and along the North Sea coast. Unfortunately, that question is largely moot. Roman authors took little interest in linguistic matters; “the amount of information we have about the complex language situation in the Empire is appallingly poor” (Polomé 1980: 193). Multilingualism was probably the norm in the Roman towns and trading centers along the Rhine, Moselle,

Contact and the History of Germanic Languages  331 Meuse, and Danube, and among Germanic tribes resettled on Roman territory, like the Ubii (cf. Polomé 1980: 194). The domination of the Romans and the preeminence of their culture assured that the influence of Latin would be substantial (see Green 1998: chs. 10 and 11). Northern Gaul and Roman Germania were constitutive of an economic and cultural sphere within the Roman Empire, out of which a plethora of Latin loanwords entered West Germanic dialects in the domains of administration (OHG keisur ‘emperor’ < Latin Caesar), the military (OHG kampf ‘battle’ < Latin campus ‘plain, field, battlefield’), the domestic sphere (OE pytt, OHG pfuzzi ‘well, pit’ < put‐tius < putjus < Latin puteus), commerce (OS munita, OHG munizza ‘coin’ < Latin moneta ‘coined metal, money’), agriculture (OHG/OS fruht ‘fruit’ < Latin fructus ), and some days of the week (e.g. Saturni dies, English Saturday, Dutch Zaterdag). The early adoption and nativization of these words are indicated by the effects of the West Germanic consonant gemination and the High German consonant shift. Beyond the lexicon, Schrijver (2014: ch.4) discerns a connection between the High German consonant shift and the agentivity of fifth‐ to sixth‐century Latin speakers who had acquired German. The first direct testimonials to any kind of Germanic language are inscriptions in the runic orthography beginning in the first century ce, most of which were found in present‐day Denmark, Norway, and Sweden, but with scattered finds in Germany, France, and Eastern Europe, along the route of East Germanic migrations. The prevailing theories on the origin of the writing system tend to fall into three categories: (a) It is an adaptation of – or autonomous creation inspired by – the Latin alphabet, on the impetus of increased cultural contact at the turn of the present era between Germanic peoples and Romans living along the Rhine; (b) it is based on North Etruscan epigraphic practice that reached northern populations along the amber trade route between Italy and the Baltic coast; (c) it is derivable from a Mediterranean script (Latin or Greek) that was imported to the North by sea rather than passing through the European continent by land (cf. Antonsen 2002: 116). Whichever position one favors, one must recognize that individuals in the borrowing oral culture had to have learned the language(s) of the literate culture and did not casually imitate a foreign writing system.

4  The Migration Period (ca. 200–600 ce) The emergence of the Germanic peoples upon the scene of recorded history “as a progressively intrusive migratory movement was one of the most fundamental ethnographic events in early European history. Quite obviously, the cultural face of Europe was to be forever altered by this intrusion” (Markey 1986: 248–9). The origins and causes of Völkerwanderung are variously identified as overpopulation and competition for resources in the homeland, coastal inundation due to climate change, the whiff of wealth from the south, and a power vacuum created by a Roman Empire in retreat and inexorable decline. Early in the present era a number of Germanic groups settled in the eastern Baltic and along the Elbe migrated southward. At about 200 ce, the Goths left their homeland along the lower Vistula and settled in the plains north of the Black Sea, eventually occupying a swath of territory between the mouth of the Danube and the Don River. The Visigoths pressed into Dacia (275), took refuge from the Huns in Lower Moesia, and marched through Greece and Italy, sacking Rome in 410. From there they wandered into southern Gaul and the Iberian peninsula. The Ostrogoths, having recovered their independence after the repulse of the Huns in 451, penetrated southeastern Europe. In 493, they seized Ravenna, which had come under control of the East Germanic chieftain Odoacer. In 535 the Byzantine emperor Justinian I commissioned a war of restoration in the former Western Empire. With the final defeats of the Ostrogoths in the 550s and the end of Visigothic domination after the Moorish conquest

332  Paul Roberge of Spain (711), the Goths disappear from history. In the Crimea a form of Gothic was still spoken as late as the sixteenth century, but in the west the language died much earlier. In the last quarter of the third century the Vandals, a collective East Germanic group, migrated south into Dacia and Pannonia. Under pressure from the Huns, they entered Gaul (406) and crossed the Pyrenees into Hispania in 409–10. The incursion of the Visigoths forced the Vandals into northern Africa, from which they went on to raid the coastal areas of the Eastern and Western Empires, sacking Rome in 455. The restoration of imperial control over northern Africa in 534 meant the expulsion and dispersal of the Vandals. By the mid fifth century the Burgundians, yet another East Germanic group, had settled in what is today the French‐speaking part of Switzerland and in the south of the French Jura around Geneva. From there they spread out over the Rhone region towards the south. The Langobards, a Herminonic group, departed from the Elbe and established themselves in northern Italy in 568, in the aftermath of the ruinous Ostrogothic wars. Generally, Germanic Mediterraneans were transitory, minority superstrata that were absorbed into local populations. The enclavement that promotes language maintenance on foreign soil was never a significant long‐term factor. Native texts are extant only for Gothic, which is represented by the fourth‐century New Testament translation – preserved in manuscripts from the late fifth and sixth centuries  –  and minor fragments. For the other Mediterranean Germanic dialects, we have only the isolated words and names that are preserved in historical and geographic writings from Late Antiquity and in co‐territorial Romance dialects (cf. Söhrman 2015; Falluomini 2015). Biblical Gothic evinces Greek influence in the design of the writing system, in lexis (cf. Snædal, 2015), and in syntax, especially word order; occasional Latinisms are also discernible. It is reasonable to suppose that Germanic vernaculars (as opposed to the high variety of religious texts) exported south of the Alps, to northern Africa, and to the Balkans were already “well mixed and stirred before or immediately after confrontation with the non‐ Germanic languages of the Mediterranean basin” (Markey 1989a: 62). Nielsen (2015) claims that the vocalic system of Gothic (and that of Old High German) did not develop from early Germanic in a manner that we would expect in consequence of the language contact to which speakers were exposed in new settlement environments. Meanwhile, two Germanic groups of composite origin appear on the scene in the third century. The Alamanni (‘all men’) were an amalgamation of Germanic tribes between the upper Danube and the middle Rhine who entered into prolonged conflict with the Romans. Between 300 and 450, they pressured and eventually succeeded in occupying Roman territories in southwestern Germany, present‐day Switzerland, and Alsace. Another tribal confederation, the Franks, moved into Germania Inferior and northern Gaul during the fifth century. By 537 virtually the whole of Gaul was under the authority of the Franks, by which time the Gaulish language was on the cusp of extinction. Romance and Germanic vernaculars mingled with one another in the plurilingual areas of Gaul during Merovingian times. Assimilation of the Frankish superstratum proceeded from the Latinization of vertical communication (i.e. between supra‐regional institutions and local units) starting in the fifth century to the Romanization of horizontal communication (between and within local affiliations) by the eighth century (Banniard 1996; Van Durme 2002: 12). The Baiovarii  –  first mentioned even later  –  were an Elbe‐Germanic group (descended perhaps from the Marcomanni) that entered history after they crossed the limes in the mid sixth century. Their name appears to be derived from that of a Celtic tribe, the Boii, which would suggest Bohemia as their initial expansion area. A handful of lexical agreements raise the possibility of ancillary contact with Gothic with the introduction of Christianity. The North Sea coast served as the staging area for yet another great migration that gave rise to the first permanent extraterritorial Germanic language of the common era. This population movement began with the Saxons (yet another merged group that first appears

Contact and the History of Germanic Languages  333 in the mid second century), who raided northern Gaul and southeastern England from the latter part of the third century. The emigration of small parties of North Sea Germanic peoples across the English Channel from the early to mid fifth century intensified into a large‐ scale colonization of Britain. The linguistic consequences for the Romano‐Celtic population were rather more pernicious than was the case in speech communities once dominated by Germanic Mediterraneans. As early as the fifth century, Latin faded away along with the Roman superstratum that introduced its language. For Britons who were not killed or displaced during the invasions, the Anglo‐Saxon conquest meant asymmetrical bilingualism, gradual language shift, and eventual assimilation. The areas occupied by the southwestern Britons were reduced by the expansion of Wessex in the centuries following the Battle of Deorham in 577. At the same time, some scholars argue that Brythonic, the language spoken in England by the Celts at the time of the Germanic invasions, had a significant effect on the development of English phonology (umlaut and breaking, as per Schrijver 2014: ch.2, though see Nielsen 2015) and syntax (e.g. McWhorter 2011: ch. 9). Vennemann (2012) writes of a “contact transitivity” that links some syntactic features of English back to Insular Celtic and Semitidic. For a cogent assessment of the so‐called “Celtic hypothesis”, see Hickey (2012).

5  The Projection of Norse Power (ca. 750–1050) Norse expeditions of the so‐called Viking Age were an exercise in wealth accumulation by multiple means – trade, piracy, colonization – that were enabled by advances in maritime technology. Of linguistic significance is the exportation of North Germanic varieties of the period, known collectively as the dǫnsk tunga lit. ‘Danish’ but more precisely ‘Norse tongue’ (Bibire 2001: 89), to territories outside of mainland Scandinavia. Norse was introduced to the eastern shores of the Baltic, Finland, and deep into Russia, where Swedish traders known as Rús (ostensibly derivable from Finnish Ruotsi ‘Sweden’, ON róðr ‘rowing’) and Scandinavian mercenaries constituted a superstratum that melted into the Slavic population after the mid tenth century. It is unclear to what extent the Swedish spoken natively by nearly 300,000 people in Finland and once spoken in Estonia (until World War II) derives from language spread during the Viking Age vis‐à‐vis secondary migration in the Middle Ages (Barnes 2005: 183). The ninth century saw an influx of (mainly) Norwegians in Shetland, Orkney, the Hebrides, Isle of Man, Scotland, and Ireland. In areas where settlement was dense and the immigrants formed cohesive polities, Norse may have continued in use for several generations. In the first two offshore island groups and on the north coast of mainland Scotland, a descendant of the immigrant language called Norn was maintained until the eighteenth century. But as a rule, Norse disappeared, though not without leaving its mark on the local languages to which it succumbed. Only in the hitherto sparsely populated Faroe Islands and Iceland did Norse take permanent root and develop into new languages. Icelanders and Norwegians established toeholds in Greenland (ca. 986) and from there explored parts of eastern Canada. From archeological evidence, it appears that there was robust trade between the Greenlandic Norse and native people, which hints at an erstwhile jargon. Today, the sporadic Old Norse loanwords in Inuktitut (catalogued by van der Voort 1996) are the monument to the Norse speech community in Greenland, which perished in the fifteenth century. Danish and Norwegian invaders settled in the north and east of England between 865 and 955. Indications of an intense contact situation in the Danelaw reside in place names, heavy borrowing (content words, pronouns, closed‐class functional items), and possibly deflection. For historical and linguistic details see McColl‐Millar (this volume).

334  Paul Roberge

6 Superposition The introduction of Christianity placed Latin in a privileged relationship with respect to the metropolitan Germanic languages, if only in specific domains – as hagiolect, the language of literacy, and as a quasi‐official language of state. The period of Latin superposition commenced with the Christianization of Anglo‐Saxon England, from the beginning of the seventh century, and the establishment of monastic culture. It is bracketed by the northern Renaissance of the fifteenth and sixteenth centuries, during which Latin served as a language of learned disquisition. Latin contributed religious and specialized terminology to the local Germanic languages, and it also exerted influences on vernacular written style. With the vernacularization of literacy, the Bible, and official documents, which took place at different times and circumstances in Germanic‐speaking regions, the position of Latin gradually attenuated. The Roman Catholic church, however, continued to use Latin for liturgical purposes and for official purposes. French‐speaking Normans with Breton and Flemish allies conquered England between 1066 and 1070 and proceeded to Ireland a century later. As the language of the court, the higher nobility, of the legal and ecclesiastical superstructure, and of European culture, French enjoyed enormous prestige. This situation lasted for some 300 years. Gradually, the resident body of Anglo‐Norman French speakers in the British Isles became assimilated, although the language hung on in some domains until around 1430 (Ingham 2012). English came to acquire an immense number of French borrowings (as well as Latin and Greek forms through the medium of French); see McColl‐Millar (this volume). The emergence of Irish English and the heavy influence of Irish are treated in detailed studies by Hickey (1997, 2010). The Hanseatic League united Lübeck, Hamburg, Bremen, and other northern German cities in an alliance for the purpose of mutual trade and protection. The cartel came to dominate mercantile activity in the North Sea and Baltic between 1250 and 1450. The importance of Middle Low German as a medium of communication in the domains of commerce, law, and administration secured its long‐term prestige and influence in Visby, Stockholm, Kalmar, Bergen, and other emerging cities in Scandinavia. Through widespread bilingualism within the merchant class and civil bureaucracy and significant German presence in urban areas, the Middle Low German imprint on Danish, Norwegian, and Swedish is every bit as deep as that of French on English (Thomason and Kaufman 1988: 315; Braunmüller 2000). As the Protestant Reformation got under way in Germany in 1517, High German eclipsed Low German as an important source of loanwords in the north. The Kalmar Union of 1397 united the kingdoms of Denmark, Sweden, and Norway under a single sovereign. Although the union broke up in 1523, Denmark retained control over Norway until 1814, and Danish served as the H variety in what was effectively a diglossic relationship with spoken Norwegian. One of the two standard varieties of Norwegian known as bokmål ‘book language’ (also riksmål ‘official language’) took shape after 1814 as a Dano‐Norwegian koine and is sometimes referred to by that name. The position of Icelandic was subordinated when Iceland entered political unions with Norway (1263–1385) and Denmark (1385–1944). According to Haugen (1976: 32), the extraordinary (if, at times, exaggerated) conservatism of Icelandic with regard to the importation of foreign words was nourished in the nineteenth century as a symbol of resistance to Danish domination. Danish was the sole official language of Greenland and the Faroe Islands until home rule in (respectively) 1948 and 1979. Spoken Faroese has been heavily influenced by Danish, especially in its lexicon. Bandle et  al. (2005: §21) devote a section to inter‐Scandinavian language contacts, alongside contact with minority and immigrant languages in situ, and with non‐Scandinavian languages in Europe and the British Isles.

Contact and the History of Germanic Languages  335

7  Germanic Languages Beyond the European Metropole From the late fifteenth century maritime European nation‐states – most importantly, Portugal, Spain, France, England, and the Netherlands and private interests operating therefrom – commissioned exploratory and trading expeditions that would in time lead to large‐ scale movements of people and the exportation of economic and cultural practices across five continents and the Pacific. Conquest, colonization, and mercantilism would transform indigenous linguistic ecologies. From the seventeenth through the nineteenth centuries, England spread its language more widely ex patria than any other European imperial power (cf. Hickey 2004, 2014). Colonial varieties of English in the United States, Canada, Australia, and New Zealand developed among settlers from the British Isles, with limited influence from indigenous languages (and in North America adstratal European languages) that English displaced in the course of the eighteenth, nineteenth, and twentieth centuries. See Salmons and Purnell (this volume) on the imprint of immigrants’ non‐native English on emergent regional varieties of American English and on the formation of ethnic varieties shaped by bilingualism, e.g. African American English. So‐called “new” (indigenized) Englishes are second‐language varieties that developed under British and American colonialism, specifically in multilingual territories in sub‐ Saharan Africa (see Mesthrie, this volume) and Asia (India, Singapore, the Philippines) where English was introduced in the official sphere – rather than by a large settler population for whom it was a first language – and advanced through formal education to serve as a lingua franca in high functions. New Englishes also arose alongside pidgins and creoles in territories and regions where English is an official language, such as in the Caribbean, West Africa, the southwest Pacific, and Australia (Schneider 2003: 236). Owing to emigration, German is spoken in North America, South America (Brazil, Argentina, Paraguay), and southern Africa. German‐speaking Anabaptists arrived in North America in large numbers from 1683, a year after the establishment of Philadelphia. Lutheran and Reformed Germans followed from 1727. These immigrants fanned out across Pennsylvania and then southward into parts of Appalachia, comprising the largest non‐ English‐speaking European group by the American Revolution (1775–83). Only the most conservative Anabaptist sects have maintained German into our time. From the 1840s, massive secondary German immigration created German‐speaking communities in the American Midlands (especially Ohio, Wisconsin, and Iowa), the Plains, and Texas. Heritage German was robust in many communities until the early twentieth century, but changes in community structures accelerated the transition from German to English in public domains. Daily use of German had largely ceased by the end of World War II, and the language has not been passed on to younger generations. Secondary immigration also brought Dutch to the American Midwest (Michigan, Wisconsin, Iowa) in the nineteenth and early twentieth centuries, but the language gradually fell out of regular use and has largely vanished (see Smits and van Marle 2013). More or less contemporaneous Norwegian and Swedish immigrant communities in the United States Upper Midwest and Canada have shown similar patterns of shift to English since the 1920s. Although the Dutch erected a far‐flung overseas mercantile operation that reached its apotheosis in the mid seventeenth century, its linguistic legacy worldwide is modest. In New Jersey and New York vestiges of New Netherland (1614–64) Dutch survived into the early twentieth century, including a subdialect that was spoken by descendants of slaves. Today, Dutch retains official status in the island territories that comprise the former Netherlands Antilles  –  cf. the special municipalities (Bonaire, Sint Eustatius, Saba) and the constituent countries (Aruba, Curaçao, Sint Maarten) of the Kingdom of the Netherlands – but its reach

336  Paul Roberge does not extend beyond administration and law. Dutch has left its mark on the lexicon of Papiamentu, the Iberian (i.e. Portuguese and Spanish) based creole spoken on the ABC‐ islands north of Venezuela, and contributed to its syntax a passive construction used mainly in writing (Holm 1988: 84). Dutch is the sole official language of postcolonial Suriname and is used alongside of some 20 ethnic home languages. Suriname Dutch spans a continuum, the poles of which are Standard Dutch and a locally colored vernacular form. Such varieties have been indigenized through the adoption of lexis, phonological patterns, and grammatical features from local languages, along with some restructuring as the product of L2‐acquisition processes (see De Kleine 2013). Over three centuries of commercial activity and colonial rule, Dutch was employed in the East Indies only to a limited extent as compared with English in India. Settler and locally born Eurasian communities were confined to relatively few localities of which Batavia (today Djakarta) was the most significant. Creole Portuguese and Bazaar Malay served as the principal lingue franche. When Indonesian independence came in 1949, Dutch lost its official status and was replaced by Bahasa Indonesia. Nevertheless, Dutch has had a strong impact on the Indonesian lexicon and a lesser one on its grammar (Tadmor 2007: 303).

8  Germanic Contact Languages and Varieties In the present century Germany, the Low Countries, the United Kingdom, Denmark, Sweden, and Norway have a vast array of languages spoken within their borders. Of particular interest to linguists are the verbal repertoires of immigrants. A particularly well‐ studied case is that of Foreign Worker German (formerly labeled Gastarbeiterdeutsch), which was spoken by laborers who migrated from southern Europe and Turkey during the 1950s and 1960s. Gilbert and Pavlou (1994) have characterized Foreign Worker German as an industrial pidgin on the basis of its structural properties and obstacles to targeted L2‐ acquisition due to the partial ghettoization of its speakers. However, the current consensus is that the German of these immigrants has been constitutive of a continuum of jargons and interlanguage varieties, which is being superseded by ambilingualism in generations born in the host country. Multi‐ethnolects are linguistic styles that are part of the repertoires of speakers of different ethno‐linguistic backgrounds. These varieties have their base in the dominant language of the society but draw from the other languages in contact. Exempla are new varieties that have emerged among second and third generations of immigrants in the Germanic metropole, usually in urban environments (see Nortier and Dorleijn 2013). A pidgin is the linguistic creation of a contact community that has need for a common means of communication but does not share a preexisting language that fulfills this function. According to Holm (1988: 6), a creole has a pidgin in its ancestry and “is spoken natively by an entire speech community, often one whose ancestors were displaced geographically,” typically in European overseas colonies that put in place economies based on plantation agriculture and utilizing nonindigenous slave or indentured labor. The idea that creoles derive from antecedent pidgins is warmly disputed. Whether creoles are truly new languages that have no genetic affiliation (Thomason and Kaufman 1988) or are to be situated along the phylogenetic branches of their lexifiers is likewise a contentious issue. English‐lexified pidgins and creoles  –  along with other English contact varieties  –  are grouped by region and assigned separate chapters in this handbook (Schneider and Hickey, Siegel); they are also surveyed in Holm (1989) and Michaelis et al. (2013). Russenorsk was a seasonal pidgin used by Norwegians and Russians involved in the fish trade in northern Norway from the late eighteenth century up until the Russian revolution of 1917. Its lexicon is drawn in roughly equal measure from the two stock languages, though

Contact and the History of Germanic Languages  337 there are also lexical items from Sámi, Low German/Dutch, French, Swedish, and English (Broch and Jahr 1984; Jahr 1996: 110). Reports of other erstwhile Arctic trade jargons with a Germanic determinant are in Broch (1996), Hancock (1996), and Kusmenko (2017). In Europe German gradually spread eastward from the ninth century until its abrupt retreat after World War II. One consequence was the creation of German‐speaking Sprachinseln in Russia, the Czech and Slovak republics, Slovenia, Hungary, and Romania (e.g. the Siebenbürger Sachsen). But another consequence was the genesis of transitory restructured forms of German used for interethnic communication. Halbdeutsch refers to interlanguage versions of German spoken by Estonians and Latvians from the late Middle Ages through the nineteenth century (Stammler 1922; Mitzka 1923; Lehiste 1965). Mitrović (1972) describes a German‐based sabir balkanique that crystalized in Bosnia during the late nineteenth century among workers from all parts of the Austro‐Hungarian Empire. Germany’s bid for an overseas colonial empire was late and short‐lived (1871–1918). Its linguistic legacy consists in the first instance of a settler dialect of German still spoken as a first language in the former colony of South West Africa (now Namibia), alongside a fading “Kiche Duits” (Küchendeutsch) that emerged among Africans in the workforce around 1900; see Deumert (2009). Incipient German‐lexified pidgins are attested in the Kiautschou (Jiaozhou) Bay Pachtgebiet in China and in German New Guinea from the late nineteenth century up to the outbreak of World War I in 1914; see Mühlhäusler (1984, 2012), who marshals the few examples that have been preserved for posterity. Mühlhäusler (1980: 176–84) lists more than 150 words of German origin in the Tok Pisin source material, of which a few have survived. Unserdeutsch is the name for what is an untypical German‐lexified creole that developed from the end of the nineteenth century at a mission school and orphanage near Rabaul, in German New Guinea (see Maitz and Volker 2017; Siegel, this volume). As of this writing, Unserdeutsch is critically endangered, with perhaps 100 speakers (not all fluent), most of whom live in Australia and are over the age of 70. There are but three known Dutch‐lexified creoles in the Caribbean (cf. Bakker 2014). Negerhollands was spoken on St. Thomas and St. John in the Danish Antilles (the US Virgin Islands since 1917), which were originally settled by Dutch planters and their slaves (in 1672 and 1717, respectively). Its last fluent speaker died in 1987. In Guiana Dutch settlements along the Essequibo (1618) and Berbice (1627) rivers saw the development of wholly discrete creole vernaculars during the seventeenth century (see Kouwenberg 2015). The death of scantily documented Skepi Dutch Creole can be dated circa 1998. Berbice Dutch Creole became extinct with the passing of its last speaker in 2005. So‐called mixed languages (Bakker, this volume) have emerged in situations of community bilingualism and evince a split ancestry (e.g. the core lexicon from one source language and grammatical elements from another). They serve as salient markers of new or retained group identity. The best known exempla involving Germanic languages are spoken in the British Isles and other Anglophone countries, cf. Angloromani of the traveling Romanis (Romani lexis and an English grammatical frame, cf. Matras 2010) and Shelta (Irish lexis, English structure) of the Irish Travelers, which died out in the twentieth century. After Indonesian independence in 1949, many Dutch colonials and Eurasians moved to the Netherlands, bringing with them varieties ranging from Standard Dutch to so‐called “Indies Dutch” to mixed Dutch‐Javanese (Javindo) and Dutch‐Malay (Petjo) vernaculars (Salverda 2013: 801, 809). Afrikaans poses a chimera‐like tableau – a West Germanic language that has a creole history. Some scholars (e.g. Holm 1989) consider Afrikaans a semicreole, albeit on the basis of static comparison of synchronic states rather than formative processes. In 1652 the Dutch East India Company established a station at the Cape of Good Hope, which saw the development of an inchoate fort colony, with complex economic and social relationships between Europeans and the indigenous Khoekhoen. But the disrupted linguistic ecology took a

338  Paul Roberge different trajectory with the formal introduction of African and Asian slave labor in 1658. Mutual accommodation at the interface between an expanded pidgin and a co‐territorial settler variety of Dutch produced a Cape Dutch Vernacular, which itself showed a high degree of variation. By the mid nineteenth century, the sectarian Muslim community in Cape Town had developed a tradition of composing religious texts in the Cape Dutch Vernacular, using Arabic orthography. Starting in the 1870s, Afrikaner nationalists cultivated their own dialect – which they called Afrikaans – as a written medium and promoted its use in domains hitherto reserved for English or Dutch. This movement culminated in standardization and official recognition of Afrikaans in 1925 which was subsequently codified in reference works (dictionaries and grammars) and for which there is a Language Commission (Taalkommissie) in South Africa. Weinreich (2008) fixed the origin of Yiddish in the Rhineland (Loter). Roughly 1,000 Jews speaking a Judeo‐Romance language (Loez) originating in France, with Hebrew‐Aramaic as their mediating language, settled in areas along the Rhine where the co‐territorial non‐Jewish population spoke different varieties of German. These languages shaped the formation of an ethnolect in a process that Weinreich called fusion, with German as the main determinant. The linguistic system of Yiddish was never that of any specific variety of German, for Jews migrated in distinct groups and were in contact with various dialects at different times. As Ashkenazy Jews moved into Eastern Europe, starting in the mid thirteenth century, a fourth component, Slavic, was added. Weinreich’s model of Yiddish formation has not gone uncontested. Prince (2001: 286n.) maintains that – perhaps for sociopolitical reasons – Weinreich “lacked notions of genetic classification.” More recent views posit a separate development for the westernmost Judeo‐ German dialects in the Rhineland and fix the origins of Yiddish in eastern German territory, based on the presence of Bavarian and to a lesser extent East Central German features; see Spolsky (2014: ch. 10 and pp. 178–86). Wexler’s (1991, 2002) hypothesis that Eastern Yiddish arose as a kind of mixed language via the relexification of Judeo‐Slavic by German‐speaking Ashkenazy Jews is radical and controversial. Regardless of the position one takes with respect to its geographic “home,” Yiddish is the result of fully proficient Germanic speakers borrowing items from other languages they knew, “all the while maintaining and fitting these items into the Germanic system” (Prince 2001: 264). Pennsylvania German is the home language of roughly 330,000 bilingual speakers in Old Order Amish and Old Order Mennonite communities in the United States (Pennsylvania, New York, Ohio, Indiana, among other states) and Canada (mainly southern Ontario) (Louden 2016: 50). Pennsylvania German derives from German dialects spoken by immigrants from southwestern Germany and Switzerland during the late seventeenth and eighteenth centuries, with dialects of the southeastern Palatinate exerting a strong founder effect. Koineization, independent development, and adlexification from American English have produced what Louden (2016: 12) regards as a language rather than an extraterritorial variety on grounds of its autonomy from metropolitan German. English influence is detectable in nominal case syncretism, the tense‐mood‐aspect system, and nonfinite complementation. But despite universal bilingualism, this variety has otherwise remained largely resistant to contact‐induced change in the core areas of morphosyntax (cf. Louden 2016: 28–49).

9 Conclusion Contact with co‐territorial languages has been a key element in the development of Germanic in its diffusion across northwestern Europe and the British Isles. Originally confined to these regions, the Germanic languages – English in particular – have spread secondarily overseas to the Americas, Africa, Asia, and the Pacific due to colonization and migration. English has

Contact and the History of Germanic Languages  339 become the world’s most important international language. Metropolitan urban environments have brought together people from different language backgrounds, and these contact situations provide settings that give rise to new ways of speaking (Wiese, this volume). Interlanguage varieties spoken by adult immigrants and language use by subsequent generations in urban centers and suburbs will enjoy increasing attention in multidisciplinary research.

REFERENCES Anthony, David W. 2007. The Horse, the Wheel, and Language: How Bronze‐Age Riders from the Eurasian Steppes Shaped the Modern World. Princeton: Princeton University Press. Antonsen, Elmer H. 2002. Runes and Germanic Languages. Berlin: de Gruyter Mouton. Bakker, Peter 2014. Three Dutch creoles in comparison. Journal of Germanic Linguistics 26: 191–222. Bandle, Oskar et al. (eds.) 2002–2005. The Nordic Languages: An International Handbook of the History of the North Germanic Languages, 2 vols. Berlin: de Gruyter. Banniard, Michel 1996. Die Franken zwischen Spätlatein und Altfranzösich. In Die Franken: Wegbereiter Europas. Vor 1500 Jahren, König Chlodwig und seine Erben. Mainz: Philipp von Zabern, pp. 574–78. Barnes, Michael 2005. Language. In Rory McTurk (ed.), A Companion to Old Norse‐Icelandic Literature and Culture. Malden, MA: Blackwell, pp. 173–89. Bibire, Paul 2001. North Sea language contacts in the early Middle Ages: English and Norse. In Thomas R. Liszka and Lorna E. M. Walker (eds.), The North Sea World in the Middle Ages: Studies in the Cultural History of North‐Western Europe. Dublin: Four Courts Press, pp. 88–107. Braunmüller, Kurt 2000. Was ist germanisch heute? Sprachwissenschaft 25: 271–95. Broch, Ingvild 1996. Solombala‐English in Archangel. In Ernst Håkon Jahr and Ingvild Broch (eds.) Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: de Gruyter Mouton, pp. 93–8. Broch, Ingvild and Ernst Håkon Jahr 1984. Russenorsk: a new look at the Russo‐ Norwegian pidgin in northern Norway. In P. Sture Ureland and Iain Clarkson (eds.), Scandinavian Language Contacts. Cambridge: Cambridge University Press, pp. 21–65. de Kleine, Christa 2013. Dutch in Suriname. In Frans Hinskens and Johan Taeldeman (eds.),

Language and Space: An International Handbook of Linguistic Variation. Volume 3. Dutch. Berlin: de Gruyter Mouton, 841–58. Deumert, Ana 2009. Namibian Kiche Duits: The making (and decline) of a Neo‐African language. Journal of Germanic Linguistics 21: 349–417. Evans, D. Ellis 1981. Celts and Germans. The Bulletin of the Board of Celtic Studies, University of Wales 29: 230–55. Falluomini, Carla 2015. Die langobardischen Sprachreste in Italien und ihr Beitrag zur Kenntnis des langobardischen Konsonantismus. In John Ole Askedal and Hans Frede Nielsen (eds.), Early Germanic Languages in Contact. Amsterdam: Benjamins, pp. 109–23. Feist, Sigmund 1928. Die Ausbreitung des indogermanischen Sprachstammes über Nordeuropa in vorgeschichtlicher Zeit. Wörter und Sachen 11: 29–53. Feist, Sigmund 1932. The origin of the Germanic languages and the Indo‐Europeanising of north Europe. Language 8: 245–54. Gilbert, Glenn G. and Pavlos Pavlou 1994. Gastarbeiterdeutsch “Foreign Workers’ German”: an industrial pidgin, . In Carol Aisha Blackshire‐Belay (ed.), The Germanic Mosaic: Cultural and Linguistic Diversity in Society. Westport, CT: Greenwood Press, pp. 147–54. Green, D. H. 1998. Language and History in the Early Germanic World. Cambridge: Cambridge University Press. Hancock, Ian 1996. The special case of Arctic pidgins. In Ernst Håkon Jahr, and Ingvild Broch (eds.), Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: de Gruyter Mouton, pp. 15–29. Haugen, Einar 1976. The Scandinavian Languages. London: Faber and Faber. Hickey, Raymond 1997. Arguments for creolization in Irish English. In Raymond Hickey and Stanisław Puppel (eds.), Language

340  Paul Roberge History and Linguistic Modelling: A Festschrift for Jacek Fisiak on His 60th Birthday. Berlin: de Gruyter Mouton, pp. 969–1038. Hickey, Raymond (ed.) 2004. Legacies of Colonial English. Cambridge: Cambridge University Press. Hickey, Raymond 2010. Irish English: History and Present‐Day Forms. Cambridge: Cambridge University Press. Hickey, Raymond 2012. Early English and the Celtic hypothesis. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 497–507. Hickey, Raymond 2014. A Dictionary of Varieties of English. Malden, MA: Wiley‐Blackwell. Holm, John 1988–9. Pidgins and Creoles. 2 vols. Cambridge: Cambridge University Press. Ingham, Richard P. 2012. The Transmission of Anglo‐Norman: Language History and Language Acquisition. Amsterdam: Benjamins. Jahr, Ernst Håkon 1996. On the pidgin status of Russenorsk. In Ernst Håkon Jahr, and Ingvild Broch (eds.) Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: de Gruyter Mouton, pp 107–22. Juntune, Thomas W. 1973. The informational value of Germanic loanwords into Finnish. Amsterdamer Beiträge zur älteren Germanistik 4: 1–48. Juntune, Thomas W. 1994. Languages and peoples in contact: Early Germanic and Early Finnic. In Carol Aisha Blackshire‐Belay (ed.), The Germanic Mosaic: Cultural and Linguistic Diversity in Society. Westport, CT: Greenwood Press, pp. 85–93. Kallio, Petri 2012. The prehistoric Germanic loanword strata in Finnic. In Riho Grünthal and Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: Société Finno‐Ougrienne, pp. 225–38. Kallio, Petri 2015. The stratigraphy of the Germanic loanwords in Finnic. In John Ole Askedal and Hans Frede Nielsen (eds.), Early Germanic Languages in Contact. Amsterdam: Benjamins, pp. 23–38. Koivulehto, Jorma 1981. Reflexe des germ. /ē1/ im Finnischen und die Datierung der germanisch‐finnischen Lehbeziehungen. Beiträge zur Geschichte der deutschen Sprache und Literatur 103: 167–203, 333–76. Koivulehto, Jorma 2002. Contact with non‐ Germanic languages II: Relations to the East. In Oskar Bandle et al. (eds.), The Nordic Languages: An International Handbook of the

History of the North Germanic Languages, 2 vols. Berlin: de Gruyter, pp. 583–94. Koivulehto, Jorma and Theo Vennemann 1996. Der finnische Stufenwechsel und das Vernersche Gesetz. Beiträge zur Geschichte der deutschen Sprache und Literatur 118: 163–82. Kouwenberg, Silvia 2015. Dutch Guiana: Demographics and living conditions and the emergence of Dutch creoles during the first one hundred years, 1580–1675. Journal of Language Contact 8: 70–90. Kroonen, Guus 2012. Non‐Indo‐European root nouns in Germanic: evidence in support of the agricultural substrate hypothesis. In Riho Grünthal and Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: Société Finno‐Ougrienne, pp. 239–60. Krahe, Hans 1964. Unsere ältesten Flußnamen. Wiesbaden: Otto Harrassowitz. Kufner, Herbert L. 1972. The grouping and separation of the Germanic languages. In Frans van Coetsem and Herbert L. Kufner (eds.), Toward a Grammar of Proto‐Germanic. Tübingen: Max Niemeyer, pp. 71–97. Kuhn, Hans 1955. Zur Gliederung der germanischen Sprachen. Zeitschrift für deutsches Altertum und deutsche Literatur 86, 1–47. Kusmenko, Jurij 2017. Borgarmålet: a Sámi‐ Swedish pidgin from the beginning of the 18th century. NOWELE 70: 39–56. Lane, George S. 1933. The Germano‐Celtic vocabulary. Language 9: 244–64. Lehiste, Ilse 1965. A poem in Halbdeutsch and some questions concerning substratum. Word 21: 55–69. Lehmann, Winfred P. 1961. A definition of Proto‐ Germanic: a study in the chronological delimitation of languages. Language 37: 67–74. Lehmann, Winfred P. 1977. Language contacts and interference in the Germanic period. In Herbert Kolb, Hartmut Lauffer, et al. (eds.), Sprachliche Interferenz: Festschrift für Werner Betz zum 65. Geburtstag, 278–91. Tübingen: Max Niemeyer. Lehmann, Winfred P. 1986. A Gothic Etymological Dictionary. Leiden: E. J. Brill. Lehmann, Winfred P. 1987. Linguistic and archeological data for handbooks of proto‐ languages. In Susan Nacev Skomal, and Edgar C. Polomé (eds.), Proto‐Indo‐European: The Archeology of a Linguistic Problem. Studies in Honor of Marija Gimbutas. Washington, DC: Institute for the Study of Man, pp. 72–87.

Contact and the History of Germanic Languages  341 Louden, Mark L. 2016. Pennsylvania Dutch: The Story of an American Language. Baltimore: Johns Hopkins University Press. Maitz, Péter and Craig Alan Volker 2017. Documenting Unserdeutsch: reversing colonial amnesia. Journal of Pidgin and Creole Languages 32: 365–97. Mallory, J. P. 1989. In Search of the Indo‐Europeans: Language, Archaeology and Myth. London: Thames and Hudson. Markey, Thomas L. 1986. Social spheres and national groups in Germania. In Heinrich Beck (ed.), Germanenprobleme in heutiger Sicht. Berlin: de Gruyter, pp. 248–66. Markey, Thomas L. 1989a. Germanic in the Mediterranean: Lombards, Vandals, and Visigoths. In F. M. Clover and R. S. Humphreys (eds.), Tradition and Innovation in Late Antiquity. Madison: University of Wisconsin Press, pp. 51–71. Markey, Thomas L. 1989b: The spread of agriculture in western Europe: Indo‐European and (non‐) pre‐Indo‐European linguistic evidence. In David R. Harris and Gordon C. Hillman (eds.), Foraging and Farming: The Evolution of Plant Exploitation. London: Unwin Hyman, 585–606. Markey, Thomas L. 1999. Studies in runic origins 2. From gods to men. American Journal of Germanic Linguistics and Literatures 11: 131–203. Matras, Yaron 2010. Romani in Britain: The Afterlife of a Language. Edinburgh: Edinburgh University Press. McWhorter, John H. 2011. Linguistic Simplicity and Complexity: Why Do Languages Undress? Berlin: de Gruyter Mouton. Mees, Bernard 2003. Stratum and shadow: a genealogy of stratigraphy theories from the Indo‐European west. In Henning Andersen (ed.), Language Contacts in Prehistory: Studies in Stratigraphy. Amsterdam: Benjamins, pp. 11–44. Michaelis, Susanne Maria, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) 2013. The Survey of Pidgin and Creole Languages, vol. 1. English‐Based and Dutch‐Based Languages. Oxford: Oxford University Press. Mitrović, Paul 1972. Deux sabirs balkaniques. Linguistique 8: 137–40. Mitzka, Walther 1923. Studien zum baltischen Deutsch. Marburg: Elwert. Mühlhäusler, Peter 1980. German as a contact language in the Pacific (with special reference to the influence of German on Tok Pisin.). Michigan Germanic Studies 6: 163–218.

Mühlhäusler, Peter 1984. Tracing the roots of Pidgin German. Language and Communication 4: 27–57. Mühlhäusler, Peter 2012. Sprachliche Kontakte in den Missionen auf Deutsch‐Neuguinea und die Entstehung eines Pidgin‐Deutsch. In Stefan Engelberg and Doris Stolberg (eds.), Sprachwissenschaft und kolonialzeitlicher Sprachkontakt: Sprachliche Begegnungen und Auseinandersetzungen. Berlin: Akademie‐ Verlag, pp. 71–100. Musset, Lucien 1975. The Germanic Invasions: The Making of Europe ad 400–600. (Edward James and Columba James, trans.) University Park: Pennsylvania State University Press. Nielsen, Hans Frede 1989. The Germanic Languages: Origins and Early Dialectal Interrelations. Tuscaloosa: University of Alabama Press. Nielsen, Hans Frede 2015. The vowel systems of Old English, Old Norse and Old High German compared. In John Ole Askedal and Hans Frede Nielsen (eds.), Early Germanic Languages in Contact. Amsterdam: Benjamins, pp. 261–76. Nortier, Jacomine and Margreet Dorleijn 2013. Multi‐ethnolects: Kebabnorsk, Perkerdansk, Verlan, Kanakensprache, Straattaal, etc. In Peter Bakker and Yaron Matras (eds.), Contact Languages: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 229–71. Penzl, Herbert 1972. Methods of comparative Germanic linguistics. In Frans van Coetsem and Herbert L. Kufner (eds.), Toward a Grammar of Proto‐Germanic. Tübingen: Max Niemeyer, pp. 1–42. Pokorny, Julius 1936. Zur Urgeschichte der Kelten und Illyrier. Zeitschrift für celtische Philologie 20: 315–52, 489–522. Polomé, Edgar C. 1972. Germanic and the other Indo‐European languages. In Frans van Coetsem and Herbert L. Kufner (eds.), Toward a Grammar of Proto‐Germanic. Tübingen: Max Niemeyer, pp. 43–69. Polomé, Edgar C. 1980. Creolization processes and diachronic linguistics. In Albert Valdman and Arnold Highfield (eds.), Theoretical Orientations in Creole Studies. New York: Academic Press, pp. 185–202. Polomé, Edgar C. 1985. Methodological approaches to the ethno‐ and glottogenesis of the Germanic people. In P. Sture Ureland (ed.), Entstehung von Sprachen und Völkern: Glotto‐ und ethnogenetische Aspekte europäischer Sprachen. Tübingen: Max Niemeyer, pp. 45–70.

342  Paul Roberge Polomé, Edgar C. 1986. The non‐Indo‐European component of the Germanic lexicon. In Annemarie Etter (ed.), o‐o‐pe‐ro‐si: Festschrift für Ernst Risch zum 75. Geburtstag. Berlin: de Gruyter, pp. 661–72. Polomé, Edgar C. 1987. Who are the Germanic people? In Susan Nacev Skomal and Edgar C. Polomé (eds.), Proto‐Indo‐European: The Archaeology of a Linguistic Problem. Studies in Honor of Marija Gimbutas. Washington, DC: Institute for the Study of Man, 216–44. Polomé, Edgar C. 1989. Substrate lexicon in Germanic. NOWELE 14: 53–73. Polomé, Edgar C. 1990. The Indo‐ Europeanization of northern Europe: the linguistic evidence. Journal of Indo‐European Studies 18: 331–38. Polomé, Edgar C. 1992a. Germanic, Norwest‐ Indo‐European and pre‐Indo‐European substrates. In Rosina Lippi‐Green (ed.), Recent Developments in Germanic Linguistics. Amsterdam: Benjamins, pp. 47–55. Polomé, Edgar C. 1992b: Indo‐European and substrate languages in the west. Archivio Glottologico Italiano 67: 66–85. Polomé, Edgar C. 1997. How Indo‐European is Germanic? In Irmengard Rauch and Gerald F. Carr (eds.), Insights in Germanic Linguistics II: Classic and Contemporary. Berlin: de Gruyter Mouton, pp. 197–206. Prince, Ellen 2001. Yiddish as a contact language. In Norval Smith and Tonjes Veenstra (eds.), Creolization and Contact. Amsterdam: Benjamins, pp. 263–89. Prokosch, Eduard 1939. A Comparative Germanic Grammar. Philadelphia: Linguistic Society of America. Reich, David 2018. Who We Are and How We Got Here: Ancient DNA and the New Science of the Human Past. New York: Pantheon. Ringe, Don 2006. From Proto‐Indo‐European to Proto‐Germanic. Oxford: Oxford University Press. Salmons, Joseph C. 1992. Accentual Change and Language Contact: Comparative Survey and a Case Study of Early Northern Europe. Stanford: Stanford University Press. Salmons, Joseph C. 2004. How (non‐)Indo‐ European is the Germanic lexicon? … And what does that mean? In Irma Hyvärinen, Petri Kallio, and Jarmo Korhonen (eds.), Etymologie, Entlehnungen und Entwicklungen: Festschrift für Jorma Koivulehto zum 70. Geburtstag. Helsinki: Société Néophilologique, pp. 311–21. Salverda, Reinier 2013. Between Dutch and Indonesian: Colonial Dutch in time and space.

In Frans Hinskens and Johan Taeldeman (eds.), Language and Space: An International Handbook of Linguistic Variation, vol. 3: Dutch. Berlin: de Gruyter Mouton, pp. 800–21. Sausverde, Ērika 1996. Seewörter and substratum in Germanic, Baltic and Baltic Finno‐Ugric languages. In Karlene Jones‐Bley and Martin E. Huld (eds.), The Indo‐Europeanization of Northern Europe. Washington, DC: Institute for the Study of Man, pp. 133–47. Scardigli, Piergiuseppe 2002. Nordic‐Gothic linguistic relations. In Oskar Bandle, et al. (eds.), 2002–2005. The Nordic Languages: An International Handbook of the History of the North Germanic Languages, 2 vols. Berlin: de Gruyter, pp. 553–58. Schneider, Edgar W. 2003. The dynamics of new Englishes: from identity construction to dialect birth. Language 79, 233–81. Schrijver, Peter 2014. Language Contact and the Origins of the Germanic Languages. New York: Routledge. Seebold, Elmar 2015. Entlehnung und Urverwandtschaft im vorliterarischen germanischen Wortschatz. In John Ole Askedal and Hans Frede Nielsen (eds.), Early Germanic Languages in Contact. Amsterdam: Benjamins, pp. 1–22. Smits, Caroline and Jaap van Marle 2013. The Dutch language in the USA. In Frans Hinskens and Johan Taeldeman (eds.), Language and Space: An International Handbook of Linguistic Variation, vol. 3: Dutch. Berlin: de Gruyter Mouton, pp. 821–41. Snædal, Magnús 2015. Gothic contact with Greek: loan translations and a translation problem. In John Ole Askedal and Hans Frede Nielsen (eds.), Early Germanic Languages in Contact. Amsterdam: Benjamins, pp. 75–90. Söhrman, Ingmar 2015. What is Visigothic and what is Frankish in medieval and later Spanish? In John Ole Askedal and Hans Frede Nielsen (eds.), Early Germanic Languages in Contact. Amsterdam: Benjamins, pp. 125–35. Spolsky, Bernard 2014. The Languages of the Jews: A Sociolinguistic History. Cambridge: Cambridge University Press. Stammler, Wolfgang 1922. Das “Halbdeutsch” der Esten. Zeitschrift für deutsche Mundarten 17: 160–72. Tadmor, Uri 2007. Grammatical borrowing in Indonesian. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross‐Linguistic Perspective. Berlin: de Gruyter Mouton, pp. 301–28. Thomason, Sarah Grey and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic

Contact and the History of Germanic Languages  343 Linguistics. Berkeley and Los Angeles: University of California Press. van der Voort, Hein 1996. Eskimo pidgin in West Greenland. In Ernst Håkon Jahr and Ingvild Broch (eds.) 1996. Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: de Gruyter Mouton, pp. 157–258. van Durme, Luc 2002. Genesis and evolution of the Romance‐Germanic language border in Europe. In Jeanine Treffers‐Daller and Roland Willemyns (eds.), Language Contact at the Romance‐Germanic Language Border. Clevedon: Multilingual Matters, pp. 9–21. Vennemann, Theo 2003. Europa Vasconica – Europa Semitica. Berlin: de Gruyter Mouton. Vennemann, Theo 2012. Germania Semitica. Berlin: de Gruyter Mouton.

Weinreich, Max 2008. History of the Yiddish Language. 2 vols. (Paul Glasser, ed., Shlomo Noble and Joshua A. Fishman, trans.). New Haven: Yale University Press. Wexler, Paul 1991. Yiddish – the fifteenth Slavic language: a study of partial language shift from Judeo‐Sorbian to German. International Journal of the Sociology of Language 91: 9–150. Wexler, Paul 2002. Two‐Tiered Relexification in Yiddish: Jews, Sorbs, Khazars and the Kiev‐ Polessian Dialect. Berlin: de Gruyter Mouton. Witczak, Krzysztof Tomasz 1996. The pre‐ Germanic substrata and Germanic maritime vocabulary. In Karlene Jones‐Bley and Martin E. Huld (eds.), The Indo‐Europeanization of Northern Europe. Washington, DC: Institute for the Study of Man, pp. 166–80.

17 Contact in the History of English ROBERT McCOLL MILLAR 1 Introduction Given the mixed origins of a considerable part of English lexis, it is unsurprising that the primary cause of this multiplicity – language contact – has been discussed since the early days of scholarly coverage of change across history in that language (these treatments include, in recent times, Miller 2012, Hickey 2012a, and Durkin 2014). English seems both prolific and eclectic in its borrowing practices. While certain languages, such as French, predominate as sources for borrowings into English, there are very few languages with a large number of speakers (and a considerable number of languages which have fewer speakers) which have not provided words and phrases to the language. Other aspects of contact‐ induced change affect English, including changes in morphosyntax and phonology, are less well covered in the literature or perceived by interested laypeople, perhaps because the probability of language contact being a catalyst for change in these circumstances is far more difficult to prove and assess. Moreover, it can take only a few minutes to borrow a word or phrase; the influence of contact upon structures in the influenced language may take generations before it is fully apparent. This chapter will attempt to cover all of these features, discussing different types of lexical borrowing (the origins of many examples are derived from the Oxford English Dictionary and the Dictionary of the Scots Language) while also providing a deeper analysis of a set of morphosyntactic changes through which English passed in the period from around 800 to 1100 ce, suggesting a contact‐derived catalyst underlying them. It will also consider some of the effects different varieties of the same language can have upon each other when they come into contact. This chapter will not, and is not intended to, document all contacts through which English has derived material or been altered (Durkin 2014 is a particularly good example of this type of treatment); instead, it is intended to illustrate some of the patterns of contact visible in the language.

2  Lexical Borrowing There are at least three routes by which external vocabulary (occasionally pronunciation and structure) have come into the English language. The first of these is immigration or at the very least contact between speakers of the “home” English language and “foreign” languages in the “home” environment. The second type comes about through migration “out” from

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

346  Robert McColl Millar the “home” territories, contact with speakers of “foreign” languages in the “foreign” environment with a form of feedback mechanism relaying this newly acquired material into the “home” environment. Finally, there are occasions when items originating in cultural expression, material culture, food preparation, and so on, have been borrowed into English without many speakers of these languages migrating into the “home” environment or large‐scale and long‐term transfer of “home” population into the source community. On a number of occasions, as this chapter demonstrates, borrowing may include a combination of two or more of these tendencies.

2.1  Lexical Transfer Through Contact with Immigrant Languages The English language has been affected considerably by the languages of immigrants. Some of these have proved of great consequence in relation to its later development. A number of these will be discussed in greater detail later in this chapter, such as the influence felt by English from Norse and French. Most influences are rather less marked, however, while still having some effect on the language. Dutch (and Low German), for instance, has had a longstanding but rather shallow influence upon English and Scots (for a brief treatment, see McArthur, Lam‐McArthur, and Fontaine 2018, s.v. Dutch; for the Early Modern period, see Joby 2005). This influence can differ between dialects, in particular with older borrowings. Thus Scots (see Murison 1971) has borrowed a small number of words (such as crag ‘neck’ or bonspiel ‘curling match’) which Standard English has not; within Scots, Eastern dialects possess vocabulary items, such as loon ‘boy, young man’, which are not found in western dialects (the geographical position of the various dialects compared to that of the Low Countries explains this). Given the particular nature of the contact and the occupations of many Dutch speakers, it is unsurprising that words connected to sailing or sea‐based trades (such as yacht or, in many varieties, mattie or madge ‘(immature) herring’) or art (landscape and easel are particularly appropriate examples) should have been borrowed. The latter group represents high status borrowings, which we will discuss in relation to Italian. Other immigrant groups have had less influence upon English. Beyond a few words related to food (such as chow mein or, indirectly and externally, tea) or referring to specifically Chinese cultural concepts (such as feng shui or, viewed from an external position and through an earlier imperialist viewpoint, kowtow), the various Chinese varieties spoken throughout the English‐speaking world for a hundred years and more have had very limited influence on the English of native speakers (naturally varieties such as Singlish (Deterding 2007), the vernacular English of Singapore and parts of Malaysia, have considerably more Chinese input, particularly in terms of phonology, perhaps). This lack may be due to the two ­languages’ typological differences (although too much can be made of this), but a combination of racism (in particular, but not exclusively, in the United States in the late nineteenth and first half of the twentieth centuries) in the host community and the inward‐looking nature of elements within the Chinese community is likely to have contributed to this lack of contact transference (for a series of treatments, see Wang and Wang 1998). Equally striking is the lack of influence on the language by the considerable number of French Protestant refugees (‘Huguenots’) into the English‐speaking world in the course of the seventeenth century (Gwynn 2001), particularly when compared to the immense effects French had on English in the High Middle Ages and also the considerable prominence of people of Huguenot extraction in British life (including Peter Mark Roget, of thesaurus fame, and Laurence Olivier, renowned twentieth‐century actor). It is likely, in fact, that it was this very prominence which led to the lack of influence. In the end, the difference between White middle‐class English‐speaking Protestants and White middle‐ class French‐speaking Protestants was not great in the late seventeenth and early

Contact in the History of English  347 eighteenth centuries; particularly when the latter group rapidly learned the former group’s language. Thus both social inclusion and exclusion may lead to lower levels of borrowing in these contexts.

2.2 Lexical Borrowing through Emigration from the Original “Home” Environment and Spread of Speakers Given the spread of English, generally associated with the seizure of economic and political power, across a large part of the globe from the seventeenth century on, it is unsurprising how many languages have influenced the English language, from Inuit kayak ‘seagoing canoe’ to gnu (originally from a Khoisan language, via Afrikaans), referring to the cow‐like antelope of southern Africa, otherwise known by the Afrikaans borrowing wildebeest. Largely and, again, inevitably, this influence is demonstrated by lexical borrowing. The process can be traced across time and space. In the first instance, the speakers of other languages used in the Atlantic Archipelago came increasingly under the power of a centralizing London‐based state, a process which inevitably involved the spread in use of Standard English, along with its speakers, but also some borrowing into that variety from the indigenous languages. Words from the Celtic languages, such as shebeen ‘unlicensed drinking establishment’, Tory (originally an Irish word for a ‘bandit’ or ‘renegade’, later applied in a contemptuous way to supporters of the legitimate Stuart succession; nowadays used in a number of countries to refer to the local conservative party) and slogan are diverse in meaning and may have been borrowed more than once in different places at different times (both in Scotland and Ireland). Even Scots, a close relative of English, transferred a number of words into Standard English during this period, including raid (originally the Scots equivalent of road), Whig (originally an armed supporter of the (hardline) Presbyterian cause in Scotland but eventually used to refer to those who supported a degree of political and social change) and glamour. More recent borrowings include mingin ‘offensively smelling, generally unpleasant’ and numptie ‘idiot, fool’. Expansion of English speakers into North America led to a considerable number of lexical borrowings, concerned with local flora and fauna, such as racoon (from Virginia Algonquian) and moose (from Eastern Abenaki), topography, such as bayou (an oxbow lake on the lower Mississippi; from Choctaw), and cultural items, such as wampum (beads often used as money historically by native Americans; from northerly dialects of Algonquian). Occasionally turns of phrase were apparently borrowed in a calqued form, such as, possibly, in this neck of the woods. But while there have been, and remain, varieties of English associated with speakers of Native American languages (given the linguistic diversity of pre‐Columbian North America, this topic is unlikely ever to be discussed fully. An illuminating recent discussion of some of these features is Newmark, Walker, and Stanford 2015), influence of these varieties upon mainstream varieties of North American English has been, at most, slight and ephemeral. Similar patterns can be found practically anywhere where English‐language colonization and exploitation developed. Māori words have been borrowed into local (and occasionally global) varieties of English, as can be seen with words such as kiwi itself, while more recent enlightened times have led to cultural borrowings, such as kōhanga reo, literally ‘language nest’, referring to Māori immersion nursery school provision. But despite the fact that New Zealanders of European origin refer to themselves as Pakehas, New Zealand English is a mainstream Southern Hemisphere variety. Indeed, it could be argued that, beyond a patina of “exotic” lexis, New Zealand English is, like most “colonial” varieties, essentially a direct descendant of colloquial Southeast England English varieties similar to, and in contact with, Standard English (see Hay, Maclagan, and Gordon 2008; for Māori English, see Holmes 1997).

348  Robert McColl Millar Material from other European languages was also borrowed into English as a result of imperial expansion due to conquest, or at least takeover, of territories previously ruled by speakers of other languages. The word boss, for instance, so commonplace in the English‐ speaking world, was probably borrowed twice from Dutch (or Afrikaans) in both New York state and South Africa. Other New York borrowing have essentially been nativized, including Santa Claus and coleslaw; some Afrikaans borrowings have followed the same trajectory, in particular in relation to southern African flora, fauna, and physical features, including veld (referring to the high plains of the region) or springbok, an antelope renowned for the height of its leaps and something of an identity symbol for (White) South Africans. Other words, such as laager, referring to a defensive position formed with wagons (a highly evocative cultural concept for Afrikaners), are less well known outside South Africa. On this occasion, Afrikaans has remained a major language in the region and appears to have influenced South African English, including its phonology (for a discussion, see Bowerman 2008). Afrikaans has also acted as a conduit for African words into English. German spoken in North America, in particular historically in Pennsylvania, has had a similar effect on English, as words such as Katzenjammer ‘hangover’ demonstrate. Some German effects on local pronunciation and syntax are reported, although it is difficult to know how far these changes have passed beyond the original ethnic group (for a brief discussion, see Lance 1993).

2.3  Cultural Influence as an Encouragement to Borrowing Cultural prestige has also encouraged borrowing and other phenomena into English. In English, as with other western and central European languages, borrowing from Greek and in particular Latin has (and to a degree remains) commonplace, certainly during the imperial period and because of the spread of (Latin) Christianity, but in particular in the Early Modern period. While Latin borrowings are not as numerous (or commonly employed) as their French equivalents, they are omnipresent, particularly in elevated and official spoken and written contexts. Latin structures may also have affected their English equivalents, in particular in relation both to multiple negative constructions and the prescriptive preference for subject pronouns in constructions such as this is she (the native this is her being dispreferred, particularly in American schoolrooms), even if evidence now exists that the former change may only have been influenced by Latin patterns after native forces began to be applied (for a discussion in context, see Mazzon 2004). Other languages from which material was taken by English speakers because of prestige include French, both in the Middle Ages and in the Modern era, in the fields of art, cuisine, government, and economics. Other sources could be put forward, however; strikingly, this could include Arabic words such as alcohol and zenith, borrowed into Iberian Romance languages (and to a lesser extent southern Italian and Sicilian Romance varieties) and passed from there into French and other languages, representing the advanced nature of Muslim society at the time in comparison with its Christian equivalents. Other Arabic loanwords into English come from different, often imperialist (on the part of English speakers), sources. These include mufti (in Arabic, a scholar of Islamic Law; in English, someone not dressed in uniform > someone dressed in plain or everyday dress) and mullah (in Arabic, a colloquial and possibly demeaning term for an Islamic scholar; in English any Muslim ‘cleric’). Many of these are likely to have been borrowed from Muslim speakers of South Asian languages, although Britain exerted at the very least authority in Egypt and many southwest Asian territories in the late nineteenth and twentieth centuries. Later borrowings, such as jihad (often taken to refer to holy war, although the orthodox Muslim interpretation is considerably more nuanced) or fatwah (a judgment given by a Muslim jurist; analyzed by English speakers, in the wake of the Rushdie affair in the late 1980s, as a decree from such an authority demanding the execution of a person or

Contact in the History of English  349 persons considered as having blasphemed against the teachings of Islam) are, for most English speakers, I imagine, conceptual borrowings of the last half century. Running parallel to these tendencies are occasions where internationalisms (such as oxygen or telescope, often largely Ancient Greek in origin (although the words themselves were generally not used in Greek) and coined to name new technologies and scientific discoveries, have spread to many languages, with English often acting as the vehicle for that spread.

2.4  Lexical Borrowing Complexity Naturally the analyses presented represent a schematization. There are occasions when contact with a particular language or set of languages can be affected by all of these tendencies and potentially more. When we consider South Asian languages as a whole (while respecting the complex realities of culture and relationship between different speakers and different languages) as source languages for English lexis. India (taken in its broadest sense) was at the very least influenced by British interests for over 250 years. Many English speakers lived permanently or for considerable periods in the Indian Empire. While learning local languages was not as prevalent as it might have been, there was considerable interchange of words and phrases between languages, to the extent that a complex and sizable “Anglo‐ Indian” vocabulary development became a reality (as illustrated by the great Hobson Jobson dictionary of 1903). Most of these borrowings stayed in India (or lived on in the minds of those who had gone ‘home’); others became widespread, including bungalow or Blighty ‘home, England’). Some was belittling and aggressive, such as the regular use of jildy ‘get a move on’; others, such as fakir, ‘religious aesthetic, often fraudulent’, demonstrated some interest, from an imperialist and rather condescending viewpoint, in Indian society (even if this knowledge was partial and biased). South Asian lexis has also spread through migration of the languages’ speakers, who are also found practically everywhere in the English‐speaking world. Languages like Punjabi, Urdu/Hindi, Bengali and Tamil have millions of speakers; many other South Asian languages are regularly heard across the world (for discussions, see Jayaram 2004). Because the spread of South Asian migrants was spearheaded (and is often associated) with “Indian” cuisine (even when most South Asian migrants and their descendants have little or no connection with that trade), words connected to cooking – naan (a type of bread), tandoor(i) (a form of oven used in northern India and Pakistan), daal (a simple everyday lentil dish), and so on – are widespread in host‐variety English (others, such as roti (a kind of bread), are only really widespread in the Caribbean, where South Asian and Afro‐Caribbean communities shared similar peasant (if not coolie) status). This level of influence is new, even if British imperial expansion led to the borrowing and mainstreaming of words, such as kedgeree, referring to smoked fish on a bed of curry flavored rice, and mulligatawny, a spicy soup, in South Asia itself. Conversely, words associated with cultural exchange which may not involve migration in either direction are also present in borrowings from South Asian languages. Some of these refer to music, such as sitar (a stringed instrument similar to a lute) and raga (an Indian classical musical form; or, in a more popular vein, bhangra, a form of dance music), but it is philosophy and religion which have provided the most prominent of these borrowings. Some, such as yoga (a discipline conceived to combine and develop the connection between body and spirit) and perhaps guru (a religious teacher, now referring, in English, to any charismatic expert or proponent) are widely known; others, such as mantra (a repeated set of words, word or syllable intended to let the user transcend immediate concerns and experience), may expect greater knowledge. Some, such as dharma (a difficult word to translate, essentially implying ‘correct action’), demand considerable understanding of Hindu or

350  Robert McColl Millar Buddhist belief systems. They are regularly heard in some circles, however. Many of these words have been spread largely through writing or the movement of a small number of prestigious individuals.

2.5 A Comparison: French and Italian Lexical Influence on English Any history of the English language, no matter how brief and summary (for instance, Leith 1997), or methodologically and analytically conservative (for instance, Baugh and Cable 2013), recognizes the importance of French influence on the nature of English as it stands today. French has had only limited influence upon either the phonology (where the phonemic split between voiced and voiceless fricatives is likely to be due to the importation of this distinction along with minimal pair borrowings) or structure (where the impersonal pronoun one may be derived from what is now French on – although the connection between the two is not without problems). The variable stress which English now exhibits is likely to be due to French influence through lexical borrowing. On the lexis of English, however, French has had an immense impact; since a language is likely to be seen, by non‐specialists, as primarily a collection of words, the importance of French influence upon English is self‐ explanatory and self‐evident; for a recent discussion of these issues, see English Language and Linguistics (2018: 22:3). If we analyse the routes by which French words came into English, we need to accept the precept that, with the exception of Scots (due to close ties between Scotland and France in the later medieval and Early Modern period) and a number of varieties spoken in places like Louisiana, where English and French are still, or recently have been, in contact in interaction between locals, most French borrowings are shared by all English dialects. This implies the percolation of elements of the language of an overclass relatively evenly spread across a wide territory. Second, the period between 1300 and 1450 is when the highest proportion of French words entered English (Coleman 1995) represents, the time when the English form(s) of French were ceasing to be spoken as a first language (as Chaucer’s prioress demonstrates, by the end of fourteenth century French in England remained only as a prestigious code of little essential use, in a sense similar to that of Latin in nineteenth‐ and twentieth‐century schools). This process represents an unusual example supporting Sasse’s views (contained in his paper of 1992) on the influence speakers of an abandoned language have on the now dominant language to which they have switched. We often associate this process with the “death” of low‐status languages and with substratal linguistic influence; it is enlightening to see essentially the same processes at work in connection with a superstratal relationship, there has never been a time between 1000 (in other words considerably before the Norman Conquest of England in 1066–7) and now when French words, associated with prestigious occupations, pastimes, and ideas, were not being borrowed and used by English speakers. The effects of the language of the ruling elite in the High Middle Ages already mentioned have regularly been outweighed by the purveyors and promoters of fashion and style. Italian, another Romance language, has had influence upon English in essentially two ways, similar to those discussed earlier in this section. In the first instance, this influence is connected to mass immigration into urban areas in the English‐speaking world by Italian speakers; the other is connected with both the migration of a small number of Italian‐ speaking experts (often in music, art and cuisine) from the seventeenth century on, along with the prominence and prestige given to Italian culture due to this migration and indirect communication transference of innovation through the arts. Large‐scale migration of Italian speakers into the English‐speaking world was particularly prevalent in the late nineteenth and early twentieth centuries; it has continued, albeit at

Contact in the History of English  351 lower levels, since then (see Gabaccia 2013). While, as was suggested earlier, some of these migrants were associated with prestige occupations, most came from relatively lowly rural and urban backgrounds, in particular originating in those areas associated with systemic instability and poverty. Illiteracy in their own language was by no means uncommon. Simplifying somewhat, when and where there was a sufficient number of immigrants speaking the same language and associated with similar geographical and cultural backgrounds, as in the great cities of the eastern and central parts of the United States in the late eighteenth and early nineteenth centuries, neighborhoods developed in the poorer parts of these cities, inhabited by these groups, often in direct contact and sometimes confrontation with nearby groups from different origins. While affluence necessarily attenuated these ties over time through the move away from the original ethnic neighborhood, and particularly in relation to generational transfer, similar communities in leafier locales – in New Jersey and Brooklyn, for instance – often represent the continuation of these ties. Italians were, perhaps, particularly given to maintaining cultural connections within defined neighborhoods because they represented the first large‐scale migration from southern Europe, associated by outsiders with Catholicism and organized criminal activity. They were obvious targets for what has been termed, in an American context, the “Party of Fear,” associated by the host culture with a series of prejudicial perceptions of the “other,” often caused by a panic‐based fear of profound changes in the ethnic, cultural, and social nature of the United States (see Bennett 1988). Inevitably Italian migrants imported features of their native varieties into their English. The most marked of these in phonological terms, perhaps, was the use of /t/ and /d/ instead of /θ/ and /ð/. This particular pronunciation, combined with similar pronunciations by Irish and Slavonic immigrants, among others, has sometimes also passed into the language of “native” urban working‐class speakers in New York City, Boston, and Chicago at least, possibly as a marker of class and place identity. It is severely criticized by linguistic and social (particularly educational) mainstream linguistic arbiters (for a discussion of these tendencies, see Lippi‐Green 1997). Some elements of Italian lexis from this immigrant source have been borrowed into mainstream English. Many of these, such as capeesh?, ‘Do you understand?’, may have come into the language through the influence of film, television, and other portrayals of Italian American life and culture, no matter how inaccurate. Inevitably many Italian words for street food have also entered the language through contact. Italian also influences English through several kinds of interaction in relatively elevated contexts. This is particularly the case with music, perhaps. While the terminology of the Classical musical tradition is inundated with Italian terminology in English, whether that be compositional forms, such as sonata or concerto or composer’s instructions on modes of playing, such as andante or presto, opera in many ways dominates, employing both these instructional conventions while also providing material for stage directions, in particular in relation to the use of the voice (most strikingly with terms like recitavo, the half sung, half spoken “conversations” between characters carrying the story between arias). While it would be possible for English speakers to produce native lexis for any or all of these terms (as is often true for German), sociolinguistic concepts of appropriateness mean that this has been, and is, unlikely to be successful. Terminology related to other forms of high art in English, including painting and sculpting, is also often Italian, as are words connected to particular types of architecture and construction, such as stucco (although this does not imply that borrowings of this type are inevitably high status. In my dialect, for instance, a plaster cast is often termed a stookie). Cuisine of various sorts has also been profoundly influenced by Italian. As we have seen, many of these terms may well have come via Italian immigrant groups in English‐speaking territories. In my experience, many English speakers attempt to achieve a ‘native’ pronunciation when ordering, even though very few English speakers have ever studied,

352  Robert McColl Millar never mind achieved any fluency in, Italian; strikingly, however, this tendency towards “authenticity” is not always replicated among ethnic Italians in the English‐speaking world. I have regularly heard pizza, for instance, pronounced by Italian Americans as /ˈpiza/, rather than /ˈpitsa/. It is possible that the first pronunciation is an attempt on the part of the immigrant population to acculturate towards the host culture and language, even as members of the host culture take on the original pronunciation as a marker of “authenticity.” In many ways the effects of the influence of French and Italian upon English lexis are similar, if we disregard the considerable discrepancy in number of borrowings. Because Italian has never been the minority but dominant language where the majority speak English, however, prestige borrowings tend to be carried across by the predominance of Italians in particular fields, rather than group prestige, no matter the subject or context.

3  Dialect Contact A feature of practically all languages with more than a few thousand speakers is that different varieties of the same language regularly come into contact with each other. Often this occurs on the boundaries between dialects and may evince features of long standing. For instance, the Scots dialects of where I live  –  southern Kincardineshire  –  exhibit features shared with dialects to the north, such as the pronunciation of mainstream Scots /ʍ/ as /f/ (although only with the relative and interrogative pronouns, not, as is the case in Aberdeenshire, with other words, such as white or whitterick ‘ferret, weasel, stoat’). On the other hand, the equivalent of English moon (Scots mune) is /men/, the norm in the north East Central dialects, rather than the /min/ of the mainstream Northern dialects. The negativizer for many speakers of South Northern dialects is, moreover, no rather than the peculiarly northeastern nae. This brings them into line with all other dialects of Scots. Moreover, a sociolinguistic overlay contact is also present, derived from (Scottish) Standard English. People in southern Kincardineshire often say who instead of faa, ferret instead of whiterrick, /mun/ instead of /men/ and not instead of no. On a grander scale, populations of geographically distinct (often distant) territories where the same language is spoken may come into comment with each other. Until recently this had to take place face‐to‐face through travel. Thus Glasgow Scots doll, referring affectionately to a young woman or to someone’s beloved of whatever age, regularly used in vocative constructions such as Whaur er ye, Mary doll, appears to have been carried to the West of Scotland through direct contact between sailors and travelers from North America and local people. This borrowing is also likely, however, to represent an early example of influence from the United States over the rest of the English‐speaking world, through the mass media, film, and popular music being so dominated by United States productions by the 1930s. Both sources may have acted upon and reinforced each other. This kind of contact is often quite banal. In contemporary British English, for instance, mad is increasingly being used by younger speakers to mean ‘angry’, the mainstream North American meaning. Even more markedly, mean has come to imply ‘unpleasant in terms of behavior’. More profound results can be observed when new varieties of the colonial language are formed, often in a new territory, distant from the metropolitan center. Of course other languages can have some influence upon the new contact variety, both native (such as the Māori words in New Zealand English already mentioned) and from other immigrant languages (as with the German and Yiddish words and phrases found in North American English but not, historically at least, in other varieties). Primarily, however, in most English‐language dominant varieties the most striking influence in the creation and development of local English stems from contact between different varieties of English.

Contact in the History of English  353 Let us consider New Zealand English. This is among the “newest” varieties of the l­ anguage, since English settlers only began to come to the islands in any numbers in the first decades of the nineteenth century (for sometimes contentious discussion on the history of this variety, see Gordon et al. 2004 and Trudgill 2004). Beyond one area in the southern part of the South Island, where settler origins and settlement patterns were rather different from elsewhere, having a particularly large Scottish component, New Zealand English is remarkably homogenous, with most variation being social rather than geographical in origin. When other native speakers of English are exposed to the variety, most immediately perceive its close affinity to Australian English (and, to a lesser extent, to the South African equivalent), which a great many would recognize its probable origins in nonstandard working‐class varieties of London in particular (“Cockney”) and the rest of the southeast of England. These similarities are palpable – all, for instance, have a diphthong in words such as say which is pronounced considerably lower in the mouth than is the case for highly prestigious England English varieties, such as Received Pronunciation (RP). But the “colonial varieties” do not share all features with “Cockney.” No Australian or New Zealand varieties represent mainstream /θ/ as /f/; apart from a small number of Australian speakers, moreover, “h‐dropping” in stressed positions in words like hat does not occur (although there is some evidence that this feature may at some point have been more common). The latter feature has been a mainstream feature of vernacular forms of the English of southeastern England English since well before the British takeovers in Australia and New Zealand; there is a small amount of evidence that this was also the case with the former. We have to assume, therefore, that the dominant forms of English brought to New Zealand were molded to some extent by the presence of other varieties (on this occasion varieties which had maintained historical use of /θ/ and /h/; many of these would have come from Scotland and Ireland; some would have come from parts of England where these features were retained). All of these usages were added to a “feature pool” which produced the new variety (the actual process for the choice of which features is contentious and will not be entered into here). Elements of this contact can also be found in lexis. While most New Zealand lexis is mainstream (not surprising, perhaps, when mass education was the norm from an earlier period in the Colony’s history and, moreover, many – perhaps most – of the original settlers spoke varieties derived from the same regions as the Standard developed in), certain features, such as the use of Scots byre for ‘cow house’, demonstrate alternative preferred sources, even outside the primary Scottish settlement centers in the southernmost parts of the country. The product of these “negotiations” might be seen as a koine, a variety which aids mutual comprehension by foregrounding common features from different varieties in the new variety while downplaying features which might cause difficulties in understanding.

4  The Typological “Transition Period,” ca. 850–1300 We have not as yet given much attention to morphosyntactic change caused by contact. As an in‐depth study, let us consider, therefore, the morphosyntactic changes through which the English language passed from the tenth to the thirteenth centuries. These are, perhaps, most marked within the noun phrase: grammatical gender was lost entirely, while grammatical case marking became, if it survived at all, vestigial (as in the partial marking of function with some personal, relative, and interrogative pronouns and the retention of an ancestral genitive inflection as a possessive marker); the distinction between “strong” and “weak” adjective paradigms was lost (although some remnants of this distinction were maintained in the mid fourteenth century London English employed by Chaucer (Samuels 1989a)), as was the ability to mark plurality on the adjective; a split occurred between one part of the inherited

354  Robert McColl Millar “simple demonstrative” paradigm  –  what became a discrete definite article  –  and another – what became the “pure demonstrative” that (Millar 2000). These changes are, perhaps, less marked with the inherited verb system, although a number of changes need to be recognized, not least the apparent breakdown (or at least “simplification”) of the connection between person and number marking on the pronoun and noun and equivalent inflectional realization on the verb, marked in many dialects (albeit in different ways) for the present tense (and be in the past tense) and also categorical (with the exception of be) in the past tense. Typologically, English experienced a striking change in direction from a rather synthetic morphosyntactic nature (where functional relationships are expressed through inflection) to one which exhibited strong analytic tendencies (where functional relationships are expressed through element order rather than through inflection). Interestingly, these changes took place in different parts of England at different times, with the North being affected much earlier than the South and the Midlands being affected in the period between these extremes. Even more interestingly, where there is sufficient evidence, the changes involved, while apparently spreading gradually across the language as a whole, appear to have passed through individual dialects at a (for morphosyntax) rather rapid pace, often taking no more than three generations and quite possibly fewer (for a discussion, see Millar 2016). None of these changes necessarily demands a contact‐based explanation. Languages by their nature change typologically. The processes involved could well be system internal; at one level, searching for external explanations for change is unnecessary and may not be helpful. The singularity of the changes involved on this occasion, along with their speed of change within the various dialects still calls for some kind of explanation beyond “these things happen,” however. Before these typological changes occurred in English (and in the period when they were first taking effect), three major external influences affected the language: Celtic, Norse, and French (Latin could also be included, but this contact was primarily, at least in mainstream analyses, of a learned sort associated with lexical percolation from Church usage into everyday practice after the period of English settlement on Britain). These contacts were of different types (one  –  that between Old English and Viking Norse being essentially adstratal, although see Lutz 2012), the other involving domination by speakers of one language over speakers of another (in one case, the apparent domination of speakers of Celtic dialects by speakers of Old English and on the other by speakers of French on speakers of English). Can any of these contacts be seen as the catalyst for the major typological changes already described? In many ways French appears the most obvious catalyst for these changes. The lexical influence of French upon English is, as we have seen, overwhelming and much commented upon since the first scientific attempts to describe the historical development of the language. During the 1970s, indeed, French was on occasion seen as central to the creation of what was analyzed as the “creolization” of English (for instance, Bailey and Maroldt 1977 and Dominigue 1977). There is a serious – indeed practically incontestable – problem with this interpretation, however: the changes described began centuries before the results of contact between English and French became tangible (or, indeed, possible). While obviously of considerable importance to our understanding of contact interpretations of the history of the English language, therefore, French must be abandoned as anything other than peripheral to the changes under discussion here, although its superstratal nature after 1066–7 may have in the later stages of the changes involved, in particular in relation to the final breakdown of grammatical gender, as discussed in Millar (2002). In terms of scholarly historiography, Norse had until the last 20 years been considered the most viable catalyst for the changes presently focused upon (see, for instance, Hines 1991). From an historical point of view there are a number of advantages to this supposition. Large‐ scale Scandinavian settlement and its concomitant societal disruption not only took place at

Contact in the History of English  355 the time when we assume the transition began – around the end of the eighth century; its focus on the North and northern Midlands of England also suits the linguistic evidence (a connection made particularly clearly by Samuels (1989b) in his discussion of the “Great Scandinavian Belt”). We will return shortly to what contact between speakers of two closely related languages living in relative equality and often in close, if not intimate proximity might mean. If we consider, as many scholars do, that borrowing – the wholesale transfer of structural materials from one language to another – is central to profound language contact, then Norse contact demonstrates few such phenomena, beyond disputed claims to influence over change in element order and the development of a discrete definite article (made as a minor point by Millar 2000; recently questioned by Sommerer (2018)). If we think in terms of interference, however, the possibility of Norse contact influence becomes greater. Such an interpretation was championed by, among others, Trudgill in the 1980s (1983 and 1986; although, as we will see, he has now abandoned this view). While Thomason and Kaufman (1988) played down this contact’s importance by employing the term Norsified English, they certainly did not suggest that the interaction between speakers of English and Norse had negligible results for the future trajectory of English (for a review of their views, see Millar (2000: Chapter 2)). Because of the meager number of early borrowings (those recorded from before around 1250) from the Celtic languages into English, along with a variety of other explanations (some of which, such as the view that the Celtic inhabitants of Britannia were at best forced off their lands and forced west, or even that genocide took place, are not now supported by all but a small number of scholars), scholarly fashion did not, until the late 1990s, generally favor Celtic influence as the impetus for typological change in English. This has now changed for many scholars, with a growing conviction that Celtic influence is at the heart of these major typological changes. Interestingly, direct structural borrowing is not commonly stressed with this explanation either. Instead, interference caused by the “simplifications” created by non‐native speakers is central to a process whereby majority features are favored over less common ones (so that, for instance, the relatively complex means by which Old English represented noun plurality was replaced apparently rapidly in many dialects by a system which, at most, possesses three means of marking that feature), with “mistakes” in interpretation of structure and morphology being commonplace. This apparent insight has led to a plethora of publications over the last two decades. While many of these treatments are of a high standard, a number demonstrating a rigorous awareness of the theoretical problems involved (see Lutz 2009; Schrijver 2007; Tristram 2007; Hickey 2012b), some scholarship of this type could be accused of being either piecemeal – attempting to explain a relatively small‐scale feature by a large‐scale explanatory construct which has not been fully thought through – or developing such a construct without establishing its viability in relation to what we know happened to English, as well as the norms of language contact phenomena in the theoretical literature (see, for instance, Woolf 2007). While space does not allow me to consider the whole argument behind this issue, the following should provide some sense of the issues involved in accepting examples of this type wholesale. No one would now claim that contact between Celtic and the dialects of the Germanic settlers of southeastern Britannia in the post‐imperial period did not have considerable strength and its results being felt for some duration. At times there may well have been more speakers of English as a second than as a first language, thus provoking examples of TA phenomena, the transfer of phenomena from the language previously associated with the group as they moved to the new prestigious language associated initially with the ethnic majority (Sasse 1992). The problem lies in perceiving the products of this lengthy period of bilingualism as central to the transition phenomena already described. There is an issue with time in relation to large parts of England: Celtic‐English contact must have been most marked

356  Robert McColl Millar in the fifth, sixth, and seventh centuries. There is also an issue with geography: the most intense early contact between Celtic and Germanic would have taken place in southern and central England (these regions would also have been the most populous), while the transition at least appears to have begun in the northern parts of that country. Supporters of the Celtic hypothesis attempt to get round the first issue by proposing (in the words of some) an apartheid which led to the Celtic‐influenced English being kept inside its ghetto (despite the matter of numbers of speakers and proportions in relation to the proposed numbers of speakers of “pure” Germanic Old English) for a considerable period; following this line of thought, the northern inception for the changes involved is only apparent because in that region the conservative native‐speaker rulers were neutralized as a result of the Norse settlement. It needs to be stated, however, that evidence for this “ghetto” is not only (to be generous) thin on the ground; it is also not supported by either the relative fluidity of “Englishness” during the period or the fact that Britons, while certainly of less power within the commonwealth, were not all by any means invisible near‐slaves throughout the period. From a linguistic viewpoint, moreover, the fact that the changes involved were less advanced in southern than northern England holds against the idea of the universal spread at an early period of the changes in question. The most extreme claims made by proponents of this viewpoint are worryingly close to conspiracy theory. Probably the most coherent arguments in favor of the Celtic influence hypothesis are found in Filppula, Klemola, and Paulasto (2008), a central publication following on Filppula, Markku, Klemola, and Pitkänen (eds, 2002). In addition, the “Celtic hypothesis” is treated in Hickey (2012b) as well as partially in Trudgill (2010) and, in particular, Trudgill (2011) in the context of his views on simpliciation vs complexification. Trudgill maintains the idea that simplification of linguistic structure is primarily found in societies where many adults are learning a new language; complexification, on the other hand, is generally associated, he suggests, with inward‐looking societies where the normal means by which the community language is acquired through inter‐generational transfer. If a language is learned through contact, children acquiring will learn it in a similar process to native language acquisition, developing towards greater inflectional morphological complexity. Following this argument, the contact between Celtic and English in early Anglo‐Saxon England would have caused simplification. This suggestion is tempting, but it does ignore what happens when the “adult” contact between Old Norse and Old English later in the first millennium was analogous (although not identical) to what happened in contact between closely related dialects: koineization (for discussion, see Siegel 1985, 2001 and Kerswill 2002). There is an ongoing debate over the extent to which Norse and English were mutually intelligible a thousand years ago and more (see Moulton 1988; Eyþórsson 2002, and Townend 2002). The chances that intelligibility was straightforward are low, at least following an analysis of the evidence (bearing in mind that the earliest attestations we have for Old Norse are somewhat later than what we would classify as archetypal Old English and are likely to evidence greater divergence from the ancestor it shares with Old English than Viking Norse did). That there was considerable mutual recognition of relationship is much more likely: untutored observers do exactly this when faced with English and German today, even though the differences between the two varieties are much greater than with Norse and English in the period 750–1050. When attempts were made – as must regularly have been the case, particularly where speakers of either language cohabited and had children together – to bridge the linguistic divide between residents of the same place, recourse must regularly have been made to koineization in relation to phonology and, perhaps in particular, structure. Similarities would be accentuated while dissimilarities might even be jettisoned and would definitely be downplayed significantly. These activities would have been to a degree improvised, but also predictable. Within families where both languages were present, the process would be repeated across and within

Contact in the History of English  357 generations, the differences being accentuated and encouraged. Within two generations in a largely non‐literate community framework, a variety structurally (and also typologically) different from the inputs would have been produced. Given that it was produced in England, this variety would have been a linear, if distinct(ive), descendant of Old English, not of Old Norse, the latter being rather extravagantly claimed in recent years, largely due to a rather limited understanding of the history of the dialects of English during the period (Emonds 2011 and Emonds and Faarlund 2014). This koine, entering into further contacts with other dialects of English, formed new koines which were less radical in form and usage, essentially because the native English input became increasingly dominant. In this way a somewhat diluted version of the changes found in the original contact variety spread across England incrementally. Thus language contact‐based explanations for structural change through interference rather than borrowing are problematical, in particular in terms of being possible to provide an exact trajectory for change. That such contacts act as catalysts or “quickeners” for change which might already be potential or inherent in a language is difficult to dismiss, however. The typological transition in the way described is just such an event (or series of events). Change of this type, I would argue, would be particularly prevalent when the varieties in contact are closely related, because of the typological fit between them and of the possibility of processes similar to koineization.

5  Towards a Conclusion That language contact is a central impetus to change in at least most languages is incontestable. Because of its lengthy history and evident documentation with speakers of other languages around the world and “at home,” English represents a useful “test bed” for our understanding of the pathways of linguistic change, with both borrowing and interference features being present at all levels across its history. As we have also seen, interference in particular can be analyzed as having a profound effect on the language’s typological development under the influence of a series of disorienting social changes. While an understanding of the development of English is aided by its long‐written heritage and the presence of its speakers around the world, in regular contact with speakers of many languages, it would be very surprising if similar processes were not present in the histories of other languages across the globe.

REFERENCES Bailey, Charles J. and Karl Maroldt 1977. The French lineage of English. In Jürgen Meisel (ed.), Langues en contact – pidgins – creoles [Languages in contact – pidgins – creoles]. Tübingen: Narr, pp. 21–53. Baugh, Albert C. and Thomas Cable 2013 [1935]. A History of the English Language. Abingdon: Routledge. Bennett, Davis Harry 1988. The Party of Fear: from Nativist Movements to the New Right in American History. Chapel Hill: University of North Carolina Press.

Bowerman, Sean 2008. White South African English: phonology. In Rajend Mesthrie (ed.), Varieties of English, vol. 4: Africa, South and Southeast Asia. Berlin: de Gruyter Mouton, pp. 164–76. Coleman, Julie 1995. The chronology of French and Latin loan words in English. Transactions of the Philological Society 9: 95–124. Deterding, David 2007. Singapore English. Edinburgh: Edinburgh University Press. Dominigue, Nicole Z. 1977. Middle English: another creole? Journal of Creole Studies 1: 89–100.

358  Robert McColl Millar Durkin, Philip 2014. Borrowed Words. A History of Loanwords in English. Oxford: Oxford University Press. Emonds, Joseph Embley 2011. English as a North Germanic Language: from the Norman Conquest to the Present. In Roman Trušník, Katarína Nemčoková and Gregory Jason Bell (eds.), Theories and Practice. Proceedings of the Second International Conference on English and American Studies, September 7–8, 2010, Tomas Bata University in Zlín, Czech Republic. Zlín: Universita Tomáše Bati ve Zlíně, pp.13–26. Emonds, Joseph Embley and Jan Terje Faarlund 2014. English: the Language of the Vikings. Olomouc: Palacký University. Available at http://anglistika.upol.cz/vikings2014 Eyþórsson, Þórhallur 2002. Hvaða mál talaði Egill Skalla‐Grímsson á Englandi? [What language did Egill Skalla‐Grímsson speak in England?], Málfríður 18: 21–6. Filppula, Markku, Juhani Klemola, and Heli Paulasto 2008. English and Celtic in Contact. London: Routledge. Filppula, Markku, Juhani Klemola, and Heli Pitkänen (eds) 2002. The Celtic Roots of English. Studies in Language 37. University of Joensuu: Faculty of Humanities. Gabaccia, Donna R. 2013. India’s Many Diasporas. London: Routledge. Gordon, Elizabeth et al. 2004. New Zealand English: Its Origins and Evolution. Cambridge: Cambridge University Press. Gwynn, Robin D. 2001 [1985]. Huguenot Heritage: The History and Contribution of the Huguenots in Britain. Brighton: Sussex Academic Press. Hay, Jennifer, Margaret Maclagan and Elizabeth Gordon 2008. New Zealand English. Edinburgh: Edinburgh University Press. Hickey, Raymond 2012a. Assessing the role of contact in the history of English. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 485–96. Hickey, Raymond 2012b. Early English and the Celtic hypothesis. In Terttu Nevalainen and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 497–507. Hines, John 1991. Scandinavian English: a creole in context. In P. Sture Ureland and George Broderick (eds.), Language Contact in the British Isles. Proceedings of the Eighth International Symposium on Language Contact in Europe,

Douglas, Isle of Man, 1988. Tübingen: Max Niemeyer Verlag, pp. 403–27. Holmes, Janet 1997. Maori and Pakeha English: some New Zealand social dialect data. Language in Society 26: 65–101. Jayaram, N. (ed.) 2004. The Indian Diaspora: Dynamics of Migration. New Delhi and London: Sage. Joby, Christopher 2005. The Dutch Language in Britain (1550–1702): A Social History of the Use of Dutch in Early Modern Britain. Leiden: Brill. Kerswill, Paul 2002. Koineization and accommodation. In J. K. Chambers, Peter Trudgill, and Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change. Oxford: Blackwell, pp. 669–702. Lance, Donald M. 1993. Some dialect features in the speech of Missouri Germans. In A. Wayne Glowka and Donald M. Lance (eds.), Language Variation in North American English: Research and Teaching. New York: Modern Language Association of America, pp. 187–97. Leith, Dick 1997 [1983]. A Social History of English. London: Routledge. Lippi‐Green, Rosina 1997. English with an Accent: Language, Ideology, and Discrimination in the United States. London: Routledge. Lutz, Angelika 2009. Celtic influence on Old English and West Germanic. English Language and Linguistics 13: 227–49. Lutz, Angelika 2012. Norse influence on English in the light of general contact linguistics. In Irén Hegedű s, and Alexandra Fodor (eds.), English Historical Linguistics 2010: Selected Papers from the Sixteenth International Conference on English Historical Linguistics. Amsterdam: Benjamins, pp. 15–42. McArthur, Tom, Jacqueline Lam‐McArthur, and Lisa Fontaine 2018 [1996]. The Oxford Companion to the English Language. Oxford: Oxford University Press. Mazzon, Gabriella 2004. A History of English Negation. London: Pearson Education. Millar, Robert McColl 2000. System Collapse, System Rebirth: The Demonstrative Systems of English 900–1350 and the Birth of the Definite Article. Bern: Peter Lang. Millar, Robert McColl 2002. After Jones: some thoughts on the final collapse of the grammatical gender system in English. In Jacek Fisiak (ed.), Studies in English Historical Linguistics and Philology. A Festschrift for Akio Oizumi. Bern: Peter Lang, pp. 293–306. Millar, Robert McColl 2016. Contact: The Interaction of Closely Related Varieties and the

Contact in the History of English  359 History of English. Edinburgh: Edinburgh University Press. Miller, D. Gary. 2012. External Influences on English. Oxford: Oxford University Press. Moulton, William G. 1988. Mutual intelligibility among speakers of early Germanic dialects. In Daniel G. Calder and T. Craig Christy (eds.), Germania. Comparative Studies in the Old Germanic Languages and Literatures. Wolfeboro, NH: D. S. Brewer, pp. 9–28. Murison, David. 1971. The Dutch element in the vocabulary of Scots. In A. J. Aitken, Angus McIntosh, and Hermann Pálsson (eds.), Edinburgh Studies in English and Scots. London: Longman, pp. 159–76. Newmark, Katrina, Nicole Walker, and James Stanford. 2015. “The rez accent knows no borders.” Native American ethnic identity expressed through English prosody. Language in Society 35: 633–45. Samuels, Michael L. 1989a. Chaucerian final ‐e. In Jeremy J. Smith (ed.), The English of Chaucer and his Contemporaries. Essays by M. L. Samuels and J. J. Smith. Aberdeen: Aberdeen University Press, pp. 7–12. Samuels, Michael L. 1989b. The great Scandinavian belt. In Margaret Laing (ed.), Middle English Dialectology: Essays on Some Principles and Problems by Angus McIntosh, M. L. Samuels and Margaret Laing. Aberdeen: Aberdeen University Press, pp. 106–15. Sasse, Hans‐Jürgen. 1992. Theory of language death. In Matthias Brenzinger (ed.), Language Death. Factual and Theoretical Explorations with Special Reference to East Africa. Berlin: de Gruyter Mouton, pp. 7–30. Schrijver, Peter 2007. What Britons spoke around 400 ad. In Nicholas Higham (ed.), Britons in Anglo‐Saxon England. Woodbridge: Boydell and Brewer, pp. 165–71.

Siegel, Jeff 1985. Koinés and koinéisation. Language in Society 14: 357–78. Siegel, Jeff 2001. Koiné formation and creole genesis. In Norval Smith (ed.), Creolization and Contact. Amsterdam: Benjamins. Sommerer, Lotte 2018. Article Emergence in Old English. A Constructionalist Perspective. Berlin: de Gruyter. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Townend, Matthew 2002. Language and History in Viking Age England. Linguistic Relations between Speakers of Old Norse and Old English. Turnhout: Brepols Publishers. Tristram, Hildegard 2007. Why don’t the English speak Welsh? In Nicholas Higham (ed.), Britons in Anglo‐Saxon England. Woodbridge: Boydell and Brewer, pp. 192–214. Trudgill, Peter 1983. On Dialect. Social and Geographical Perspectives. Oxford: Blackwell. Trudgill, Peter 1986. Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 2004. New‐Dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press. Trudgill, Peter 2010. Investigations in Sociohistorical Linguistics. Stories of Colonisation and Contact. Cambridge: Cambridge University Press. Trudgill, Peter 2011. Sociolinguistic Typology. Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. Wang, Ling‐Chi and Wang Gungwu 1998. The Chinese Diaspora: Selected Essays. Singapore: Times Academic Press. Woolf, Alex 2007. Apartheid and economics in Anglo‐Saxon England. In Nicholas Higham (ed.), Britons in Anglo‐Saxon England. Woodbridge: Boydell and Brewer, pp. 115–29.

18 Contact and the Development of American English JOSEPH C. SALMONS AND THOMAS PURNELL

Writing in American Speech in 1929, E. C. Hills argued vociferously against Meillet’s “theory of a linguistic ‘substratum’ that underlies and modifies or has modified certain languages” for English spoken in the United States, asserting that aside from a few situations like French in Louisiana or Spanish in the Southwest: … English has no substratum in the United States. It is true that in certain large cities there are recently arrived colonies of people of non‐English speech, such, for instance, as the Italians of San Francisco and the Poles of Chicago. But in these colonies those of the second generation speak English with little or no foreign accent and those of the third generation generally lose the foreign speech completely. It could not be otherwise with the extreme mobility of our population and the great economic pressure that is put on our immigrants to learn English. (Hills 1929: 432)

This is broadly so with regard to influence from Indigenous languages because American English is what Denis and D’Arcy (2018) refer to as a settler colonial variety, one where original languages of the area were marginalized or even erased (though see below for complexity on this point). At the same time, immigrant languages have had effects on American English, even if work on the history of these varieties long tended to bracket out contact as an issue, though this has changed in recent years, as we show here. The present analysis of language contact in the development of American English highlights especially recent work. We focus on the local level, often meaning the presence of both a limited geographical area and some identifiable social group. Below we review research showing that demarcative differences occur on this level, especially in the realm of structural language‐ contact effects beyond the lexicon. Even at the regional level, we will argue here, the effects of contact on many varieties of American English may be modest, even subtle, on first glance, but they exist in ways that are interesting and important for understanding American language and society today, in addition to advancing our understanding of language contact generally. We discuss possible structural impacts on the regional English of monolingual Americans that are likely to have originated in other languages and, to a lesser extent, in other English dialects. We first lay out some conceptual and theoretical preliminaries (Section 1), and touch on structural effects of lexical borrowing (Section 2). With that, we turn to ethnolects in the US that reflect the linguistic heritage of particular groups (Section 3). That sets up a more

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

362  Joseph C. Salmons and Thomas Purnell extended case study of the Upper Midwest (Section 4), where we see features originally associated with the most widely spoken immigrant languages establishing themselves as regional markers. Finally, we conclude with remarks on the bigger picture (Section 5). From its beginnings, we argue, English spoken in the present‐day United States has been forged by language contact.

1 Background One major challenge for our topic is determining what is and is not “contact‐induced” change. On the one hand, traditional historical linguists sometimes preferred “internal” accounts of change unless there is compelling evidence that something is contact‐induced. On the other, some now see essentially all language and dialect change as driven by contact, usually in the sense that the spread of change involves diffusion. The former view rested on a dichotomy between “internal” and “external” motivations, where structural and contact‐ related changes are rigidly separated. We avoid “the weakness of simplistic dichotomous thinking” (Dorian 1993: 152) to trace how “internal” and “external” factors interact in change (as have others, including Rickford 1986 and Mufwene and Gilman 1987). The nature of the interplay differs depending on the type of contact and varieties in contact. The latter view  –  where all change arises from contact  –  risks trivializing the notion of “language contact,” rendering it vacuous (see Labov 2001: 20). We can avoid this problem by placing analytical importance on both the ultimate sources of features and the transmission of features through communities. Both play central roles (see below). Turning to terminology, we take Hills’ term substrate as a cover term, referring to the residues of language shift, where an adult’s first language (L1) influences L2 acquisition (e.g. Mufwene 2001). Some argue that substrate effects are motivated not by social identity but by structural accommodation in the L2 whereby language learners engage in relexifying portions of their native lexicon through transfer of frequent and perceptually salient or congruent patterns (Trudgill 1986, 2008; Mesthrie 2001; see criticism in Hickey 2003). Speaker age naturally plays a role in the transmission of features, whether under contact or not (Labov 2001: 415–45). Under the apparent‐time hypothesis (cf. Labov 1966), linguistic differences between speakers of various ages collected at one point in time reflect different eras of acquisition, making change in progress readily observable. Longitudinal studies over speakers’ lifespans show real‐time changes even among adults, particularly in terms of patterns of variation, for example the Canadian [ʍ ~ w] merger (Chambers 2013: 309) and Montreal French shift of apical > dorsal variants of /r/ (Sankoff and Blondeau 2007). We assume that early acquisition follows from input of caretakers and peers, with reshaping along emerging lines of social identity later with only limited flexibility. Substrate influence, thus, is most likely to occur with adult acquisition of the new language or dialect, and a cohesive community involving enough adults to help shape the grammars of a generation of L1 learners. We draw on three aspects of contact theory to help us understand how source features have been transmitted into American English: imposition (ultimately, source features showing up in the developing variety, but see Section 1.1 for a fuller description), the process of koineization (variety formation via leveling and reallocation), and timing (a post‐immigration delay allowing for leveling stabilization).

1.1 Imposition A more nuanced alternative to “substrate” is “imposition” (van Coetsem 1988, 2000; Howell 1993; Winford 2005), arising in situations where a group brings L1 features into their L2 repertoire, features which are, in turn, adopted by L1 speakers of the earlier L2 as part of the

Contact and the Development of American English  363 broader speech pattern. Much structural interference can be characterized as “imposition” resulting from this imperfect L2 acquisition.1 As van Coetsem puts it, “the source language speaker is the agent, as in the case of a French speaker using his French articulatory habits while speaking English” (1988: 3, also in Winford 2005: 376). This focus on “agentivity” – where borrowing is “recipient language agentivity” and imposition “source language agentivity” – helps constrain notions of which kind of linguistic item is more or less likely to be borrowed or to be imposed during contact. Howell (1993: 189) represents the inverse relation between borrowing and imposition like this: (1) Stability: Borrowing versus imposition More open to borrowing → Less open to borrowing Less affected by imposition ← More affected by imposition Less stable domains: More stable domains: lexical items, derivational phonology, inflectional morphology, semantic morphology system, syntax

1.2 Koineization While van Coetsem’s model predicts structural types of contact effects, recent work on koineization in closely related dialects helps illuminate developments under contact. Koineization, or new‐dialect formation, proceeds from dialect contact through leveling and simplification (Trudgill 1986 2004; Britain and Trudgill 1999; Kerswill 2013; Kerswill and Trudgill 2005; Hickey 2003; Schreier 2015), while Trudgill (2011) argues for specific settings that allow for complexification. A Founder Principle applies to mixing situations where dialect features persisting generally come from the first speech communities contributing to the koine (Mufwene 1996). For example, Kwa‐speaking enslaved people, who were brought to the west early, shaped plantation African American English (AAE) much more than more numerous Bantu speakers who arrived later (Mufwene 2001). Generally, dialect features with psychological significance (stereotypical or stigmatized, Dillard 1972; Hickey 2000; Kerswill and Trudgill 2005) tend to be smoothed over and leveled. This phase thus militates against substrate effects. Consequently, research on koineization claims that the only substrate effects persisting into a new koine serving as an incipient standard (in contrast to a vernacular) will be those without psychological baggage. If it is true that this process works only when the koine serves as a supraregional variety (Hickey, p.c.), then we are unsurprised at substrate effects being incorporated into AAE or vernacular AAE features (e.g. from sources such as hip‐hop) into general American English. This approach allows for substrate effects; once the koine is stabilized via the smoothing out of cross‐dialect variation, some variables emerge with new purpose (“focusing” in Kerswill and Trudgill 2005, or “reallocation” for some others). Thus an effect emerges over the course of two or more post‐contact generations. Kerswill and Trudgill (2005: 200; schematized in (2) below, from Trudgill 1999, elsewhere) argue that it is important for migration to have stabilized by that point as well. This waiting period allowing for stabilization of the koine often occurs with colonialization, hence Trudgill’s term Colonial Lag. In Section 4.2, we apply this to another apparent imposition, final laryngeal neutralization in the Upper Midwest. (2)   The path of koine formation Stage I II III

Speakers involved adult migrants (first generation) first native‐born speakers (second generation) subsequent generations

Linguistic characteristics rudimentary leveling extreme variability and further leveling focusing, leveling, and reallocation

364  Joseph C. Salmons and Thomas Purnell As new varieties form, whether through contact among a variety of very different languages or a few closely related dialects, the social significance we assign to particular features changes, sometimes dramatically. Consider an example of dialect contact in American English, rhoticity.2 Rhoticity is a cover term for the consonantal articulation of coda /ɹ/ (r‐ fullness) or non‐articulation of /ɹ/ in codas (r‐lessness). Although coda‐final /ɹ/ was likely in recession (see Hickey 2010), early colonial settlement on the Atlantic coast brought speakers of both r‐ful and r‐less dialects  –  typically r‐less varieties in areas settled from southern England and typically r‐ful in those settled by northern English and Scots‐Irish, for instance. R‐lessness was regarded as “prestigious” (Kurath and McDavid 1961), and spread inland in particular directions from urban centers. For example, r‐lessness spread northward from Boston into Maine but not westward across Massachusetts or into upstate New York (Bloch 1971), and also spread westward from the Tidewater region in South Carolina from Charleston through the plantation region toward the Blue Ridge Mountains (McDavid 1948). Even assuming the original distribution reflected settler dialects, explanation of subsequent changes – like the diffusion of r‐lessness into once r‐ful geography from eastern Massachusetts north to Maine and inland from Charleston, South Carolina  –  requires focusing on social factors like identity and affiliation to the source location, not migration alone. We then see patterns of diffusion defined along such parameters. McDavid (1948) noted that younger speakers, female speakers, and urban speakers were r‐less, characteristics now associated with carriers of innovations.

1.3  Time Lag We can still observe the changing status of r‐fullness and r‐lessness. In the South, the switch of r‐less prestige to r‐ful “prestige” has been well documented, and social variation differs by time and place. Compare McDavid’s (1948) picture of South Carolina where female speakers are more r‐less to Schönweitz’s (2001) survey of data from The Linguistic Atlas of the Gulf States (LAGS 1986–92) where female speakers are more r‐ful. In New York City, Labov (1966, 1972) found that innovating r‐lessness co‐varies with social class and argued for a top‐down innovation suggestive of hypercorrection. Finally, variation and change in r‐lessness among African‐Americans confirms the notion that analysis needs to take place locally (Nguyen 2006). Wolfram’s (1969) Detroit study shows that r‐lessness is sensitive to social class and gender as well as ethnicity (see also Levine and Crockett 1966; Anshen 1969; Schönweitz 2001). Wolfram and Thomas (2002) found older AAE speakers in North Carolina exhibit a wide range of variability from those who are almost exclusively r‐less to those who are almost entirely r‐ful. Additionally, subtle differences have been found recently in numerous locales. In coastal North Carolina (Wolfram and Thomas 2002) younger AAE speakers have increased in r‐lessness while Whites are moving towards r‐fullness. AAE contact with White vernaculars can lead to r‐fullness among AAE speakers, a pattern found in an Appalachian enclave community (Childs and Mallinson 2004) and Northern urban areas (Nguyen 2006; Purnell 2009). Moreover, accommodation (when r‐less speakers use r‐ful forms) entails consideration of a number of situational factors in a conversation. For example, Baugh (1988) found that r‐fullness increased when an African American English speaker was talking to an unfamiliar person of any race or a non‐Black speaker. Downes (1998: 175) rightly argues that “postvocalic r is a different sociolinguistic variable” across communities, with its own history in each case. In that spirit, our window of analysis must be local. We show below that systematic indeterminacies follow similar patterns of social and structural change under contact at the local level. By shifting the focus from lexical items to imposition in language structure and to processes and timing of koineization, we will find the crisscrossing of social and geographic strata by variables yielding distinct subtypes of American English.

Contact and the Development of American English  365

2  A Note on Loanwords The one effect of contact readily apparent and universally accepted as such is lexical. Some of the best‐known discussions of contact in American English focus almost solely on that, for example, Mencken’s “Loan‐words and Non‐English influences” (1937: 150–63). Aside from scattered remarks on structure, like the intonation of Pennsylvania German English, he restricts discussion largely to the lexicon, as does Romaine (2001).3 In and of themselves, loanwords have had marginal impact on American English – they have not changed our stress patterns the way Norman French did for English and their morphological integration is seldom distinctive, at least for American English (but see Cannon 1984 on zero plural marking on nouns borrowed from Japanese). Loanwords have contributed few and minor new phonotactic patterns, and differential phonological integration can be found. Perhaps limitations on loanword upward influence is due to a general restriction by domain, that is, to place names, terms about nature or cultural terms (Schneider 2007; Denis and D’Arcy 2018), and not to the lexicon as a whole or to grammatical forms. Consider first differences depending on whether a given borrowing came in by written or oral transmission: among German borrowings, for instance, danke schön has generally appeared in American speech with the last vowel /e:/. This reflects the unrounding of German umlauts found in most dialects imported to this country, where Standard German schön [ ʃø:n] is pronounced [ ʃe:n]. This is certainly the case in the Pennsylvania German area, but also a natural way of importing front rounded vowels into English. In contrast, German über ‘over’ (often spelled uber in English), long known in philosophical usage and as a part of loan compounds (cf. Übermensch), has become a productive prefix. It is pronounced with the back vowel [u] rather than front [i], reflecting transmission through writing rather than speaking. Japanese is the source language for American skosh [sko:ʃ] ‘[a] little bit’, and both the pronunciation and spelling reflect the loss of the first and last voiceless high vowels of Japanese rather than the transcription of the Japanese form, sukoshi. Likewise, our name for the rapidly growing Japanese vine is kudzu, approximating the native pronunciation rather than the Romanization kuzu. Second, numerous foreign words and names have been taken into American English (and other varieties) with “hyperforeign” pronunciations, as described in Janda, Joseph, and Jacobs (1994). They note that some forms retain relatively native pronunciations (or have had them reintroduced historically) such as Bach with the velar fricative [x] or milieu as [mɪlˈjø]. In many other cases, Americans and other English speakers have extended generalizations about particular languages to over‐adapt loans. For instance, the knowledge that French often does not pronounce written final consonants (English ballet [bæˈle]) and its phonological counterpart that French prefers open syllables prompts most Americans to produce coup de grâce without its final [s], turning a ‘stroke of mercy’ into a French ‘stroke of grease’. While the voiced alveolo‐palatal fricative [ʒ] is found in some very well‐ established English words, like measure and pleasure, its foreignness is apparently still evident, by its regular extension to Bei[ʒ]ing rather than (the more Chinese‐like) Bei[dʒ]ing, and humorously (together with stress shift) in garbage [ˈgaɹbɨdʒ] > [gaɹˈbaʒ], or the chain store Target as [thaɹˈʒe]. Studies have shown that loanword use and pronunciation are influenced by non‐linguistic factors. Weinreich (1964 [1953]: 27) suggests that not only might education and prestige influence a pronunciation closer to the source language pronunciation, but that the status of the source language itself may influence the pronunciation. Prestige as motivator for pronunciation was echoed by Boberg (1999) who examined borrowings involving the grapheme pronounced variably as [æ, a], as well as Janda, Joseph, and Jacobs (1994) mentioned above. Hall‐Lew, Coppocki, and Starr (2010) found that loanword pronunciation (specifically of the vowel in Iraq) reflected a speaker’s political affiliation.

366  Joseph C. Salmons and Thomas Purnell Still, the widespread scholarly assumption of a lack of influence beyond the lexicon does not match some folk perceptions about American English in some parts of the country. In the Upper Midwest, as developed in Section 4, it is unremarkable for members of communities with strong ethnic/immigrant identities to assume that their personal speech reflects their heritage, even if they are monolingual English speakers. If asked about some distinctive‐sounding pronunciation, such as “stopping” of interdental fricatives or final laryngeal neutralization in a word like beer[s], speakers may matter‐of‐factly say “oh, that’s just the Polish/German/ Norwegian coming out in me.” We turn now to such imposition‐sourced structural patterns.

3  American Varieties Shaped by Bilingualism As noted, much attention to language contact in the history of American English has been addressed to what are sometimes called “ethnic dialects,” such as French influences on the English of southern Louisiana, Pennsylvania German influence on the English of southeastern Pennsylvania, and Spanish influences on the English of broad areas of the Southwest. This literature has often focused on L2 English. The traditionally discussed English of “Cajuns” or “Pennsylvania Dutchmen” was non‐native English, reflecting direct imposition from the relevant L1(s), not the residue of earlier language shift. Well into the twentieth century, such communities were often not merely bilingual, but often heavily non‐English speaking. In the Upper Midwest, communities much smaller and less isolated than the Cajuns or Pennsylvania Dutch remained monolingual more than 70 years after immigration. For instance, up to a quarter of Wisconsinites in some German immigrant communities are reported in the 1910 US Census as non‐English speakers, many of them second and even third generation in the United States (Wilkerson and Salmons 2008; Frey 2013; Wilkerson et al. 2014).4 The nonstandardness of English spoken under such circumstances reflects L1 interference; it does not reflect stable patterns of American English as a native dialect, but rather typically transitional phenomena, so that almost all such features have been thought to recede and disappear as communities become proficient in English and then shift to English (see Brown, forthcoming for a model and case studies of shift in such settings). As such, these patterns are of interest largely because they provide a seed‐bed from which later features may emerge via the processes of koineization laid out in (2) above. For example, in relatively homogeneous immigrant communities where large numbers of adult speakers learned English as L2 during a single generation, this may provide enough impetus to plant features in a new generation. As discussed above, established features in a given community may later become ethnically unmarked regional features or speakers may become bidialectal (e.g. Anderson 2014). The patterns of distinctiveness and non‐distinctiveness of American English as spoken by English monolinguals in a whole array of broad and internally diverse communities – for example, “Native American English” and “Jewish English”5  –  provide insights into roles played by contact in the history of American English. We will briefly treat both sets of patterns, followed by a word about dialect contact in “Latino Englishes” and current AAE. We focus on phenomena found among monolingual English‐speaking members of communities, which typically begin as ethnic or social features, due to a shared linguistic heritage. Crucially, such features can generalize to become local or regional markers beyond a particular ethnic community.

3.1  “Native American English” In the midst of widespread and vigorous revitalization efforts today, almost all Native American communities in the United States are today in advanced stages of shift to English, and a majority in almost every community is made up of native speakers of English. In a

Contact and the Development of American English  367 classic work on what he calls American Indian English, Leap (1993: 281–2) begins his summary with these points: 1. American Indian English is an aggregate of English varieties, which differ, as a group and individually, from standard English (as expressed through the language of the metropolis) and from the varieties of English spoken by non‐Indians in American society. 2. The distinctive characteristics of these codes derive, in large part, from their close association with their speakers’ ancestral language traditions. In many cases, rules of grammar and discourse from that tradition provide the basis for grammar and discourse in these English codes – even in instances where the speakers are not fluent in their ancestral language. 3. Other components of Indian English grammar and discourse resemble features of nonstandard English; usually, however, these features express meanings not attested in other nonstandard codes. The similarities in form should not overshadow the significance that these features hold in each case. Native Americans speak or spoke hundreds of languages (see Mithun, this volume), are in contact with a range of American English dialects, have acquired English under widely different circumstances, and so on.6 Given this situation, a central question is whether shared features span tribes and regions. An important historical event bears on the answer to this question. In the late nineteenth century, the federal government forced many Native children into “boarding schools,” where students were made to use English and punished for using their native tongues. Leap concludes from a survey of structural features found in boarding school student letters that most reflected L1 interference. The situation led to the rise of what Leap calls “codes‐under‐construction,” as students learned English using all available input (1993: 162), but not to the rise of any broader unified dialect. More recently, Newmark et al. (2016) note a likely role for broader patterns of intertribal contact past and present. They present clear evidence for the ongoing emergence of distinct prosodic patterns, which they interpret as a way of constructing an ethnic identity, where a distinct variety of English functions “as a form of resilience and resistance for Native people” (2016: 655). Three phonological case studies illustrate paths that English is taking in Indigenous communities in the United States. First, Rowicka (2005) presents evidence that members of the Quinault Indian Nation in Washington State have begun to use [ʔ] for /t/ in ways shared by other Native American varieties of English. Similar “glottaling” is familiar from an array of other varieties, including White American vernaculars, so this pattern may fit under Leap’s point (3), above, perhaps with koineization. Second, Anderson (1999) compares the diphthongs /ai/ and /oi/ in Cherokee English (westernmost North Carolina) with neighboring non‐Cherokee dialects. Monophthongization of /ai/ to [a:] is one of the most characteristic features of Southern US English generally, but the two communities show somewhat different patterns between [ai] and [a:] (along with different patterns of monophthongization of /oi/): Cherokee English has [ai] in hiatus or utterance finally, while Anglos monophthongize across all environments. Anderson argues that this reflects a combination of both Cherokee influence and accommodation to local norms. Third, Newmark et al. (2016: 654–5) find commonalities in the prosodic systems of Native American and Canadian First Nations peoples, notably “contour pitch accent, high‐falling, high‐rising and mid terminals, lengthened intonation‐unit‐final syllables, and syllable timing.” Quinault English lacks demonstrable imposition on the emerging variety, while the Cherokee example may reflect some transfer, but mediated and mitigated by patterns found in local varieties. The broader prosodic features may be tied to shared tonality among a number (but not all) of the Native American and Canadian First Nations languages in question (Newmark et al. 2016). What these three have in common, then, is that the current

368  Joseph C. Salmons and Thomas Purnell situation shows relatively subtle linguistic differentiation from surrounding Anglo vernacular usage. To our knowledge, nowhere have uniquely Native American features of English yet spread into systematic usage among non‐native neighbors or contact populations. Newmark et al. (2016) combined their analysis of US and Canadian Englishes. Genee and Stigter (2010) examined Canadian Blackfoot English, one variety of what they refer to as Indigenous Englishes. By comparing the English of Blackfoot speakers to the grammar of Blackfoot, Genee and Stigter observe the following differences from standard and nonstandard forms of Canadian English: uninflected verbs, substitution of present participles for past participles, omission of infinitive marker to and copular/auxiliary/existential to be. Moreover, they report changes in marking for gender, possession, definiteness and plurality, tracing each back to a potential source in Blackfoot.

3.2  “Jewish English” A clear contrast to the Native American English situation is found in “Jewish English.” These varieties  –  perceived with coherent features since at least the 1800s  –  are often associated with Yiddish influence, but the designation covers a vast range of groups of very varied cultural and linguistic heritage (e.g. Benor 2015, 2018). Many discussions correlate religious observance roughly with linguistic distinctiveness, from Hasidic communities where Yiddish is learned as L1 to those who consider themselves “culturally Jewish” but are not observant and whose speech may differ from other local varieties only by a few words. For part of this spectrum, Benor (2004, chs. 4, 5; see also 2018) gives a concise picture of English spoken in Orthodox communities, covering lexical and structural features, as well as discourse and pragmatic patterns. Benor’s (2018) explanation of the development of Jewish English follows the processes and paths noted elsewhere in this chapter. Specifically, there were two periods of immigration followed by a quiet integrational period (1931–80) leading to the “Distinguishing period” (1981 to present). In this latter period, English is repurposed “especially in in‐group communication” through the employment of “Yiddish, textual Hebrew/Aramaic and Modern Hebrew” (2018: 416). A number of features strongly associated with Jewish speech have become unremarkable (if recognizable) parts of English for non‐Jewish Americans. Consider these examples from syntax and phonotactics, domains susceptible to imposition. First, a stereotype of Jewish speech is the topicalization of indefinites, as in these examples (all examples from Feinstein 1980: 15): (3) Indefinite topicalization Some milk you want? A hotel she lives in. Feinstein’s questionnaire results indicate that New Yorkers found such sentences more acceptable than non‐New Yorkers. While Jewish New Yorkers reported the highest use, he still considers this a general New York feature. Though associated with Yiddish, this is no simple transfer. First, other languages spoken in New York City allowed similar topicalization, including Germanic ones, and Irish‐derived communities would have shown a preference for a similar topicalization, via clefting. Second, other varieties of American English allow a closely related pattern of topicalization, namely of definites: (4) Definite topicalization The book by Bellow I already read. Him I like.

Contact and the Development of American English  369 Feinstein argues that “Yiddish helped to extend the domain of an analogous existing rule in English” (1980: 22), rather than introducing a basic change. Moreover, the extension remains incomplete: Yiddish allows topicalization of any element, including nonfinite verbs, while the New York pattern allows only more constrained patterns. Benor (2018: 420) observes that topicalization is not present in contemporary Orthodox Jewish speech. Turning to phonotactics, English has historically had word‐initial /s/ + consonant clusters but not /ʃ/ + consonant clusters. Numerous Yiddish loanwords have become ubiquitous without assimilating to the native sC pattern  –  shlep ‘to drag’ and shmooze ‘to chat, especially currying favor’ with [ʃ]. The Yiddish pattern matches patterns common to other immigrant languages, such as German, where most varieties allow only /ʃ/ + consonant clusters, as found in common pronunciation of proper names like Schlitz and Schmidts (both family names and brands of beer). Intriguingly, Durian (2007) reports a change in progress of /s/ to [ ʃ] before consonants, especially sCr‐ clusters like strong but any possible connection between the change in progress and language contact is too tenuous to entertain. Thomas (1932) compared speech of Jewish and non‐Jewish students attending Cornell in terms of pronunciation. In addition to a number of New York features, Thomas observed dentalization of alveolars, overaspiration of [t] after alveolar sonorants [n] and [l] after a stressed vowel, a range of [s] variants, and stopping after the velar nasal. All these features have generalized beyond formerly Yiddish‐speaking communities, but the history is far richer than direct transfer or imposition from Yiddish (or other Jewish languages) onto L1 English. In the first example, an existing English pattern was extended. In both, changes were supported by parallels in other languages present in the communities. Other features have spread, including shm‐ reduplication (fancy shmancy), or features supported by multiple linguistic sources, like coda [ŋg] for expected [ŋ] in singer or long (also a Slavic pattern). More recent work by Benor (2016) describes the intersectionality of Black Jews. In addition to standard wide‐ranging AAE grammatical features for such speakers, Benor (2016: 175) notes the expected employment of Hebrew and Yiddish loanwords, features from Yiddish that are shared by German substrate (e.g. in Wisconsin, below) such as the preposition by to mean at or with, final laryngeal neutralization and stop affrication, as well as perhaps uniquely Yiddish prosody.

3.3  “Latino Englishes” Up to this point, our examples have shown connections between source languages and American English where the connection has weakened for such reasons as the synchronic diminishment of substrate languages and diachronic distance. In contrast, Latino English (Fought 2006; Thomas 2019) holds a markedly different relation precisely because of the proximity to and robust ongoing relations with Spanish‐speaking geographies of Mexico, of Central and South American countries, and of the Caribbean Sea islands. Specifically, Chicano English (ChE; e.g. González 1988; Wald 1984; Fought 2006; Thomas 2019 among others) reflects connection to Mexican Spanish whereas Cuban English (e.g. Otheguy, García, and Roca 2000) and Puerto Rican English (e.g. Urciuoli 1996) would reflect connections of English to other respective forms of Spanish. Nevertheless, given the predominance of Chicano English in terms of numbers of speakers and their relationship with Mexico, as well as the lack of study explicitly on other varieties and their interaction with each other, ChE is sometimes (mistakenly; see Rosa 2019: 149–50) used as a cover term for the range of Spanish‐ influenced Englishes. Another cover term, Spanglish, permits avoidance of a distinction between Spanish source varieties and avoidance of strict adherence to Spanish grammar and practices (Rosa 2019: 148–9). As many researchers have noted, Latino Englishes explicitly refer to dialects of American English by virtue of its speakers being English first language and English-dominant speakers

370  Joseph C. Salmons and Thomas Purnell influenced by the koneization of a distinct variety of US English and not second language learners of English (Santa Ana 1993; Santa Ana and Bayley 2004: 417–18). In terms of origins, ChE reflects a language‐contact scenario in regions of the English-speaking United States with a historical connection to speakers of Spanish. This would specifically refer to the crescent of states from northern or central California to eastern Texas where connections to Mexico have existed since the centuries‐old colonialization of the Western Hemisphere. Researchers refer to this region as being the locus of ChE (Fought 2006: 79). The time lag from the onset of contact leads us to also expect forms of Latino Englishes to occur in Florida (primary contact with varieties of Caribbean Spanish) and New York City (also primary contact with varieties of Caribbean Spanish). Chicago may also display some effects having had Spanish‐speaking communities for some time. What effect contact Spanish will have on the English spoken in areas of heavy new migration of Spanish speakers from Mexico such as North Carolina remains to be seen (see Thomas 2019). Given the contact time‐depth and intermingling settlement patterns of English and Spanish speakers in specific regions, it is no surprise that we see both borrowing and imposition in Latino Englishes communities: borrowing from ongoing relations between English and Spanish (e.g. the prevalence of culinary words such as taco, burrito, etc., González 2000, and pronunciation of such words, Baird et  al. 2018) and imposition from longstanding contact (e.g. absence of tense‐lax and long‐short distinctions in vowels, Santa Ana and Bayley 2004; consonant pronunciation omitting the [dʒ] affricate, Godínez 1984; devoicing of [z], Bayley and Holland 2014). While readers are encouraged to consult the growing body of work on Latino Englishes, we highlight here a small sample of substrate effects on English from Spanish that reflect imposition as “source language agentivity” noted above. First, vowel reduction and centralization  –  rampant in unstressed spoken Standard American English  –  is much less common in ChE (Santa Ana and Bayley 2004). Spanish speakers do not reduce unstressed vowels to schwa. Thus, the infrequency of schwa reduction can be attributed to Spanish substrate. Santa Ana (1991) “hypothesized that the extent to which [ChE] speakers accommodated to the general US schwa‐mean centralization pattern corresponded to the amount of social contact and personal identification that an individual had with Euro‐American dialect speakers” (p. 177, cited in Santa Ana and Bayley 2004: 421). Given considerably diminished reduction, it is also no surprise that the “light” diphthongal tense mid vowels would lack the offgliding typical of American English so that SAE [ej] is produced as [eː] (Fought 2006: 81). Fought (2003) notes a few areas where Spanish may have had a substrate influence. First, drawing on research by Wald (1996), Fought notes a parallelism with the modal would. If an if‐clause has a stative verb in the present tense, then the modal and verbs are parallel in both clauses. The oft‐cited example from Wald (1996: 520; cited in Fought 2003: 99): (5) If Thurman Thomas wouldn’t’ve dropped those fumbles, then the Bills woulda won. is linked to Spanish, as evidenced in hubiera (would‐have‐1P‐SG) and hubieran (would‐have‐ 2P‐PL) below. (5’) Si Thurman Thomas no hubiera perdido la pelota, los Bills hubieran ganado. NEG would‐have would‐have Second, Fought notes that some patterns of preposition usage in ChE derives from Spanish and not another source. To highlight this connection we mention only the use of for in the example below (Fought 2003: 100):

Contact and the Development of American English  371 (6) For my mom can understand. For she won’t feel guilty. Fought points out that for is interpreted as ‘so that’ deriving directly from the Spanish para que (‘so that’). Third, Fought (2003: 105) mentions the use of tell instead of ask that both she and Wald (=(14), 1987: 63) observed and propose deriving from decir, meaning ‘say, tell, ask’ in Spanish. (7) a. Les decían que por que no iban. ‘They asked themi (that) why theyi weren’t going.’ b. I told (=asked) heri that – that why does shei always make me do it? c. So she goes to the doctor n the doctor tells (=asks) heri why is shei nervous?

3.4  “African American English” The above examples show clear, direct links between bilingualism and contemporary speech. AAE had equally clear roots in language contact, as speakers of many African languages became speakers of English (Wolfram and Thomas 2002: 12–31; Lanehart 2015). The development of AAE is also consistent with imposition, koineization, and timing. Mufwene (2000, 2015) argues that AAE results more from the convergence of American English and founding African varieties at various times and locations before the Reconstruction of the South than directly from an American or Caribbean creole. The resultant reallocated forms are neither wholly American English (e.g. gon(na), Poplack and Tagliamonte 1996) nor wholly African (e.g. the associative plural, Boretzky 1993). Today, AAE varieties are more closely tied to local linguistic ecology than to a perceived, supranational or translocal AAE variety (Mufwene 2015). AAE is in contact with essentially all other varieties of American English. Recent literature suggests that AAE speakers are both accommodating in part to local White vernaculars while distinguishing themselves by nonparticipation in changes sweeping through the rest of the local population (Bailey 1997; Childs and Mallinson 2004; see also Part II of Lanehart 2015). For example, AAE speakers who live in the “Northern Cities” area are not fully participating in the vowel restructuring characteristic of urban, White speech (Gordon 2000; Thomas 2001).7 Recent research in the Upper Midwest has begun to explore the role of contact (e.g. Purnell 2009) and work to date suggests that these groups are both participating in limited ways, and retaining either unifying translocal AAE features or adopting local White features. Since the late 1980s, controversy has flared about the relationship of AAE to other varieties of American English, with some arguing that AAE was/is diverging from White vernaculars and others arguing for convergence. This controversy is today perhaps best understood at the local level because the psychological value of reallocated features is as local as the contact between speakers within a single community (Bailey and Maynor 1989; Rickford 1999; Mufwene 2015: 76). We conclude that the shifting awareness and evaluation of features, along with the leveling of features that appear to have lost significance under contact, reinforce linguistic differences across communities. The generalization across all the three settings mentioned above is that speakers are navigating a complex cluster of structural and social options in the ways they speak. Much of what we see is imposition, save for its absence in Quinault English. Borrowing is also present in these communities, as English L1 speakers acquire elements of their ancestral Indigenous languages in Native American communities, for instance when borrowings, which presumably began as emblematic code‐switching with greetings and leave‐takings, take place. Another instance would be where some Yiddish or Hebrew is acquired in the case

372  Joseph C. Salmons and Thomas Purnell of “Jewish English” (see Benor 2004 and 2015 on sociolinguistic complexities of this process). These, then, help shape the discourse and pragmatic systems of the variety.

4 Case Study: From Immigrant Languages to Regional Varieties We now shift focus to an area where features that a couple of generations ago were directly identified with immigrants’ non‐native English have now become regionally or socially aligned. These features are also widely used beyond members of the particular ethnic communities that introduced the features into the mix. The potential role of influence from immigrant languages has been noted in the recent literature, notably by Mufwene (2006: 178, emphasis ours): … nationalistic parochialism among European colonists, which lasted until the early twentieth century, must have reduced the extent of influence that continental Europeans could have exerted on structures of North American Englishes. Although the overall demographic proportion of these immigrants exceeded that of the Britons in America by the nineteenth century, the incremental pattern of their growth, including the later immigration of some groups and their gradual absorption into the prevailing Anglo socio‐economic structure, weeded out much of the substrate influence they could have exerted on their present English vernacular.

This section sketches some patterns that have resisted this weeding out process and the reasons why they might have. We draw here on the Upper Midwest,8 where the demographic history of immigration is relatively well understood and the emergence of salient regionalisms is relatively recent and well documented (Schuld et al. 2016; Braun 2020). (See Litty 2017 for the most detailed account to date of the full history of language and language contact in the region.) Southern and eastern Wisconsin received repeated surges of immigrants. The largest immigrant group in the region, Germans in eastern Wisconsin, provide a handy focus for this discussion. Consider census reports of German nativity in the largest city, Milwaukee, between 1870 and 1950, in Figure 18.1, where nativity increases from 1870 to 1910, then falls. While each community’s history is unique, the chronology is typically similar across the region, as Radzilowski (2007: 217) confirms: “Between 1870 and 1900 the settler immigration that helped fill up the rural areas of the Midwest gradually tapered off.” Let us correlate this timing to the source of features and timing of the emergence of reallocated features of the contemporary dialect of the region. As laid out in Wilkerson et  al. (2014), the learning of English came long after settlement for many communities and the eventual shift to English was abrupt, so that many children were acquiring English from adults with clear L2 patterns, an ideal situation for carrying features over into an emerging monolingual population. We sketch four patterns associated with the Upper Midwest, each with a distinct relationship to contact: the “stopping” of interdental fricatives (Section 4.1), final laryngeal neutralization (Section 4.2), the use of with as a verbal particle with verbs of motion (Section 4.3), and innovations in the use of other adverbs (Section 4.4).

4.1  Interdental Fricative Stopping When asked what makes for a regional accent in various parts of the Upper Midwest – including Michigan’s Upper Peninsula, Wisconsin, and Minnesota – one of the most common responses from laypeople is “dem,” “dere,” “dose,” the use of [d] or [dð] for

Contact and the Development of American English  373

8

×104

7

Population

6 5 4 3 2 1 0

1870 1880 1890 1910 1920 1930 1940 1950 Census year

Figure 18.1  German nativity in Milwaukee as reported on the national decennial census, 1870–1950 (Immigration numbers were extracted from US Census data, accessed through the University of Virginia Library, http://fisher.lib.virginia.edu/collections/stats/histcensus).

interdental fricative /ð/, along with [t] or [tθ] for /θ/, as in upnort ‘up north’, a stereotypical reference to Wisconsin’s Northwoods. Among older speakers in one heavily German‐ ethnic community in eastern Wisconsin, Rose (2006) sees this feature, which she calls “stopping,” as key to “performing Germanness” among her speakers, who included German–English bilinguals and English monolinguals.9 Rose finds that speakers of English or Irish heritage are equally likely to use stops for fricatives. In particular the use among Anglo‐Americans indicates that the feature no longer only conveys ethnicity, but has been partially reinterpreted as social variation. Among other things, the feature varies by gender – more widely used by men than women – and by social context – with high rates among those playing skat (a German card game popular in Wisconsin) and bingo, but low among bridge players, for example. While Rose studied a community with heavy German (and Dutch) roots, stopping is likewise a central part of performing other ethnicities. It was common in many languages brought to Wisconsin and some English dialects, including Irish English (Hickey 2004). In koine formation, features present in multiple input varieties, thus not readily identified with one particular group, tend to survive the leveling process into the new variety (Kerswill and Trudgill 2005). The multiple origins of stopping promoted its propagation in the state among even those of English heritage. From a traditional historical linguistic perspective, it would be easy to deny that the ­present‐day stopping in the Upper Midwest was directly triggered by immigrant‐language interference, given that varieties of English spoken around the globe have lost interdental fricatives, sometimes where appeal to language contact is implausible at best, like southeastern England. Moreover, language‐internal accounts of this change abound, notably perceptual and articulatory accounts. All these factors favor stopping. At the same time, the feature is stigmatized in some settings, and may be receding in some areas, though it is certainly used among young speakers. As Blevins (2006: 20) concludes generally about these segments in English:

374  Joseph C. Salmons and Thomas Purnell In an imaginary natural history of English, untainted by literacy, prescriptive norms, social conventions, and language contact, the loss of /ð/, /θ/ … would likely be complete in all varieties of Modern English.… However, external factors have intervened – among others, the infiltration of American Broadcasting English to ever more remote corners of the Earth. As a consequence, these phonemes and the contrasts they take part in are hanging onto life….

In the Upper Midwest, speakers apparently transformed stopping, once a sign of L1 interference, into a widespread but socially stratified feature found throughout a broader region among speakers of varying ethnic backgrounds. Again, we do not see direct transfer of an immigrant‐language feature onto the regional dialects. Instead, as in Section 3, a socially and structurally more complex story is required: as noted, features found in the speech of many different groups – so not identified with any single group – preferentially find their way into the koine. As with all features, psychological significance changes during reallocation. Stopping has become disconnected to some extent, as Rose shows, from ethnicity, being found now also among Anglos, but has taken on new social meaning.

4.2  Final Laryngeal Neutralization Another “reallocated” feature integral to the koineization of Upper Midwestern English is final laryngeal neutralization of the “voicing” contrast in word‐final position.10 American English contrasts words like his and hiss or bat and bad, for example, while in German Bad and bat are pronounced [ba:t], Rad and Rat as [ʁa:t]. While devoicing in word‐final position has been attributed directly to German influence, fortition or devoicing can also be found in Polish, Dutch, partially in Scandinavian, and in some dialects of Yiddish. The following are stereotypical pronunciations. (5) Stereotypical final laryngeal neutralization in the Upper Midwest Da Bear[s]! (Saturday Night Live television skit, Chicago) I’m going to wash my hair[s]. (Milwaukee) The fact that this feature was used as a stereotype of regional speech on national television suggests a perception of devoicing as being connected to the Upper Midwest. Those skits traded, as well, in stereotypes of specifically Polish‐American speech, while the Milwaukee example portrays German‐American speech.11 Phonetically, “voicing” is not only evident in vocal fold vibration during the coda obstruent, but also in the length of vowels preceding the consonant (e.g. Heffner 1937; Parker 1974). These two features – pulsing and vowel duration – can be considered as overall measures of the change in voicing over successive generations of speakers from southeastern Wisconsin (see Figure 18.2 below, from Purnell et al. 2005b; see also Purnell et al. 2005a). This figure depicts changes across four sets of speakers: English speakers with birthdates between 1866 and 1892, 1899 and 1918, 1920 and 1939, and 1966 and 1986. The top point represents the average location for /d/ for that group and the bottom point is the average location for /t/. What is important here is how the speakers differ from the canonical standard relation between /t/ and /d/. We expect that the /t/ points would be in the lower left corner and the /d/ in the opposite corner, the upper right corner. Those speakers born before the peak in German nativity (1866 to 1892, Figure 18.2) use glottal pulsing for voicing and largely ignore vowel duration. The middle chronological group is most like the generally assumed English pattern. They were born after German immigration had waned. The youngest group trades off between vowel duration and pulsing, but it is clear that their /d/ is more /t/-like. These speakers are two generations removed from the middle group born in the German nativity decline. Once again, a pattern that survives into

Contact and the Development of American English  375 1.0 0.9

1899–1918 1920–1939

Glottal pulsing ratio

0.8 0.7 A

0.6 0.5 0.4

B

0.3

1866–1892

0.2 0.1

1966–1986

0 0.05

0.1

0.15

0.2

Vowel duration (sec)

Figure 18.2  The historical development of final laryngeal neutralization in Wisconsin (from Purnell et al. 2005b: 331).

contemporary vernacular represents a feature present in a variety of immigrant source languages. We interpret this as the reallocation of final laryngeal neutralization.

4.3  Verbal Particle with Consider now a grammatical structure, the use of adverbial/particle with together with verbs of motion, as in these sentences:12 (6) Verbal particle with We’re going now. Are you coming with? Are you taking your phone with? Most English speakers would use along in this context, or just the verb without an adverb at all. This construction is widespread not only in the Upper Midwest but also in the Lower Midwest, for example in Illinois and Indiana and in historically Germanic areas of Pennsylvania (Wolfram and Schillings 2016).13 This feature is also popularly associated with German heritage, presumably based on familiarity with the parallel construction in German, which has so‐called “separable prefix” verbs with mit‐ ‘with’: mitkommen ‘to come along’, mitnehmen ‘to take along’, and so on. Like stopping, it has multiple immigrant sources, namely across the Germanic family: (7) Germanic models Er kommt mit. German: Hij komt mee. Dutch: Han kommer med. Danish:

‘He comes along/with.’

This feature has close parallels in English verb particles, such as to come to ‘to regain consciousness’ so that this constitutes a lexico‐semantic innovation for a set of verbs, not a more broadly syntactic one. Most Germanic languages have large, productive sets of such “separable prefixes,” yet no American English dialect has acquired any beyond with that we know of. The

376  Joseph C. Salmons and Thomas Purnell feature differs socially from stopping in not being stigmatized, perhaps in part (as Raymond Hickey suggests to us) because it is less frequent and thus less salient than stopping. Perhaps because this feature is not stigmatized, it appears to be increasingly widespread throughout the Midwest and beyond, though it remains foreign, to our knowledge, in the South. Many speakers in the Upper Midwest use it across a wide range of styles and registers, and some are surprised to learn that it is a regional feature. In this case, a seed sown in the region’s immigrant past has flourished and spread as predicted by the koineization process.

4.4  Other Adverbial Changes and Discourse Patterns We have explored examples where more than one source language contributed to the success of a feature in regional vernaculars. Having a source in multiple input languages is not necessary for change via contact. Especially in eastern Wisconsin (but stretching westward), we find direct imposition of features from German (with further support from Dutch), where “modal particles” like mal are used to soften requests and convey other information about speaker attitudes and intentions. The English translation of that word, once, is today used in the same function: (8) Modal‐particle‐like adverbs come (over) here once (German: komm mal her) ‘just come over here; come over here, won’t you?’ Similar pragmatically oriented transfers are found throughout the region, including tag questions, like Wisconsin ainna? < from ‘ain’t it?’, calqued from German nicht wahr?: “Hot today, ainna?” (from Cassidy and Hall 1985: s.v.). The same appears to be true with exclamations, like (American‐) Norwegian uff da! ‘oh darn!’ Still, such single‐source features appear to be less common and less central grammatically.

4.5  Emergence of the Regional Pattern We have argued that the Upper Midwest’s immigrant past helped guide the region’s path to increasing linguistic distinctiveness. That path is indirect, involving koine formation set in motion at the end, rather than beginning, of major immigration to the region. At the outset of this chapter, we noted that three characteristics of language contact are emphasized by the data we presented in this chapter. These characteristics include the following. 1. Imposition on particular types of structural domains 2. The process of leveling and reallocation/hypercorrection from potentially multiple sources 3. The timing of the lag after immigration for leveling stabilization The examples laid out in Section  4 exemplify these patterns. First, imposition is found in examples from phonology and syntax (i.e. from the right side in van Coetsem’s scheme in (1)). The Upper Midwest also shows borrowings, but they are less instrumental to regional dialect formation (cf. Cassidy and Hall 1985 – on kaffeeklatsch, lutefisk, etc.). Second, the process of koineization is exemplified by multiple immigrant languages contributing to the leveling process by sharing final laryngeal neutralization, stopping, and the come with pattern. Third is the timing of the development of the new variety. The historical patterns of immigration appear to match the “lag” associated with koine formation, where new varieties arise only generations after immigration ceases. At the same time, it warrants mention that such changes do not necessarily arise in one place and spread from there. Purnell, Raimy,

Contact and the Development of American English  377 and Salmons (2017) provide evidence of this in the Upper Midwest for a set of vowel changes, including data suggesting that the monophthongal GOAT vowel often associated with immigrant influence may not correlate with the relevant immigrant settlement patterns.

5  Summary and Conclusions We opened with Hills’ arguments against substrates in American English. We close by noting Meillet’s response to Hills, who had sent Meillet a draft of the paper (in Hills’ translation): In order that there may be a linguistic substratum, it is not necessary that there should be  only one. A complex substratum, especially, brings modifications which it becomes impossible to evaluate by reason of their complexity and their variety. The fact that French is spoken at the present time in Paris by a majority of provincials and descendants of ­provincials and of foreigners and descendants of foreigners is, I believe, of great importance. At first sight, the effects are not appreciable, but the fundamental result is that “Parisian” is disappearing, drowned in a sort of Koiné (common speech), just as Attic formerly disappeared drowned in the Greek Koiné. The idiomatic character of “Parisian” is being progressively effaced. I can scarcely believe that the great mixing of population that is taking place in the United States will not have a similar effect. A banal Koiné is being produced. (Hills 1929: 431)

These comments are similar to the views of many in that era (e.g. Mencken 1937: 356). One thrust of modern variationist sociolinguistics has been demonstrating the increasing overall diversity of American English. We now understand the diversification of American English today in no small part as the slow‐motion resolution of the contacts encoded in our history. The geographical differentiation may spring in part from the local working out of koine formation and language shift in these communities. Non‐English features enter the pool of variants available for incorporation into local and regional speech. Over post‐immigration generations, these variables have often been reallocated, taking on social and regional import rather than ethnic/heritage‐language meanings. American English has not fully crystalized into one coherent whole, nor should it be expected to. Rather it is developing a set of parochial or, in contemporary terms, regional koines. Once in the system, the data suggest, any given feature can spread, recede, or be reallocated along any available line. But this is only visible with the kind of rich empirical record over a long period of time, in other words, the kinds of evidence that are only now becoming available. These complex historical paths do not arise ex nihilo from contact but often involve extensions of patterns historically present in English, where acquirers or adults, seeking to accommodate to new patterns of speech, could tweak the inherited pattern to produce the new structures. We see this syntactically in both New York indefinite topicalization (drawing on topicalization of definites) and Upper Midwestern verbal particle constructions of the come with type (drawing on verbal particles of the come to type.) Neither is really a fundamental syntactic change such as word order changes, and neither has developed the full range found in the immigrant language (topicalization of any constituent in Yiddish, much larger sets of verbal particles in German). The negotiation of structural change from contact in American English gives every impression of expanding possibilities and generalizing on restricted processes. Also temporally, these effects show clear signs that they are not mundane transfers from other first languages categorically imposed on English. In fact, the features were never actually absent from the local varieties, but they may appear to be submerged for a generation

378  Joseph C. Salmons and Thomas Purnell or two, only to re‐emerge later. Similar time gaps between contact and the rise of such ­features once counted as a reason to discount “substrate” accounts, but they fit nicely with koineization theory. The features treated follow van Coetsem’s stability gradient generally, but add some nuance. Borrowing of lexical items is widespread, we see derivational morphology ­borrowed (prefixal über), and we find imposition in phonetics/phonology (final laryngeal neutralization), for example. But borrowing has created new phonotactic patterns (like /ʃ/C‐) and imposition has been gentler and more indirect than we might expect, whether in syntactic extensions or the time lag in final laryngeal neutralization. Time and again, we see the interplay between “internal,” or structural, and “external,” or social factors, in the origins and transmission of change.14

Acknowledgments We owe Raymond Hickey our gratitude for an invitation to revise our original contribution for this second edition. We thank the following for a wide variety of comments, discussion and inspiration on this topic and this chapter: Angela Bagwell, Matt Bauer, Erica Benson, Jennifer Delahanty, Bernd Heine, Rob Howell, Neil Jacobs, Brian José, Monica Macaulay, Salikoko Mufwene, Mike Olson, Eric Raimy, Kate Remlinger, Becky Roeder, Mary Rose, Luanne von Schneidemesser. The usual disclaimers apply.

NOTES 1 Note that the onset of what is arguably the most widespread vowel change in American English – the low back merger of vowels in word pairs such as caught versus cot – has been argued to be an instance of substrate influence by Slavic immigrants on the local variety of western Pennsylvania (Herold 1990, cited in Labov 1994: 318). Moreover, the role of repertoire and the practice of mingling repertoires (e.g. Benor 2016) reinforces L2 features into an eventual substrate imposition. 2 See Downes (1998: 150–75) for details on rhoticity as a variable in American English. 3 Much remains to be done even on loanwords. The key resource for tracking lexical results of contact is the Dictionary of American Regional English (Cassidy and Hall 1985–), and it becomes especially valuable when used in conjunction with the volume indices like DARE (1993) and von Schneidemesser (1999). 4 Note that such situations challenge the widespread assumption that time of immigration reflects the beginning of bilingualism. 5 “Scare quotes” are used as a reminder that each of these terms is a cover for myriad very different dialects and sociolects. 6 The Lumbee (Wolfram 1996, among others) provide an example of Native Americans who are monolingual English speakers, maintaining unique speech patterns while being influenced by contact with outsiders (Schilling‐Estes 2002). 7 Briefly, the Northern Cities Chain Shift is a restructuring of the vowel space where the most salient shift, /æ/ raising, is accompanied by /ɪ/ > /ε/ > /ʌ/ > /ɔ/ > /ɒ/ > /æ/. 8 We define “Upper Midwest” here, following the Center for the Study of Upper Midwestern Cultures, to wit as: “Although the exact contours of the Upper Midwest are open to debate, most arbiters apply the term to Minnesota, Wisconsin, and the Upper Peninsula of Michigan (with overlap into lower Michigan, Ontario, Manitoba, the Dakotas, Iowa, Illinois …).” 9 See also McCarthy (2007). 10 We follow Iverson and Salmons (1995, 2007) and Salmons (2019), among many others, in taking laryngeal distinctions in English and German to be better captured by considering aspiration

Contact and the Development of American English  379

11 12 13 14

(“spread glottis,” “fortis,” etc.) to be the distinctive feature rather than “voice,” as found in Dutch, Yiddish, Polish, and so on. This reflects a nonstandard use of hair as a count rather than a mass noun, at least in part attributable to German influence, and represents part of a broader pattern, cf. a scissors and various similar forms. The syntactic analysis of this construction – whether the with is better understood as a “verb particle” construction or adverb – is irrelevant for present purposes. It is also reported as a feature of South African English, presumably by a similar historical path from Afrikaans. In one case, that of Quinault English, it appears that we have language shift without lingering L1 transfer into the new variety, but at the same time, the emergence of a new linguistic trait (glottalization) that draws on a familiar pattern in English and shows affinity to certain groups, namely other Native communities. Such an outcome appears to be the exception, not the rule.

REFERENCES Anderson, Bridget L. 1999. Source‐language transfer and vowel accommodation in the patterning of Cherokee English /ai/ and /oi/. American Speech 74: 33968. Anderson, Vicki M. 2014. Bidialectalism: An Unexpected Development in the Obsolescence of Pennsylvania Dutchified English. Durham, NC: Duke University Press. Anshen, Frank 1969. Speech Variation among Negroes in a Small Southern Community. Ann Arbor, MI: University Microfilms. Bailey, Guy 1997. When did Southern American English begin? In Edgar W. Schneider (ed.), Englishes around the World: Studies in Honor of Manfred Görlach, vol. 1: General Studies, British Isles, North America. Amsterdam: Benjamins, pp. 255–375. Bailey, Guy and Natalie Maynor 1989. The divergence controversy. American Speech 64: 12–39. Baird, Brandon, Marcos Rohena‐Madrazo and Caroline Cating 2018. Perceptions of lexically specific phonology switches on Spanish‐origin loanwords in American English. American Speech 93.1: 79–107. Baugh, John 1988. Black Street Speech: Its History, Structure and Survival. Austin, TX: University of Texas Press. Bayley, Robert and Cory Holland 2014. Variation in Chicano English: the case of final (z) devoicing. American Speech 89.4: 385–407. Benor, Sarah Bunin 2004. Second style acquisition: the linguistic socialization of newly Orthodox Jews (Stanford University Ph.D. dissertation). Benor, Sarah Bunin 2015. How synagogues became shuls. In Janne Bondi Johannessen and

Joseph Salmons (eds.), Germanic Heritage Languages in North America: Acquisition, Attrition and Change. Amsterdam: John Benjamins, pp. 217–333. Benor, Sarah Bunin 2016. Jews of color: performing Black Jewishness through the creative use of two ethnolinguistic repertoires. In H. Samy Alim, John R. Rickford, and Arnetha F. Ball (eds.), Raciolinguistics: How Language Shapes Our Ideas About Race. Oxford: Oxford University Press, pp. 171–84. Benor, Sarah Bunin 2018. Jewish English in the United States. In Benjamin Hary and Sarah Bunin Benor (eds.), Languages in Jewish Communities, Past and Present. Berlin: de Gruyter, pp. 414–30. Blevins, Juliette 2006. New perspectives on English sound patterns: “natural” and “unnatural” in evolutionary phonology. Journal of English Linguistics 34: 6–25. Bloch, Bernard 1971. Postvocalic r in New England speech, a study in American dialect geography. In Harold Byron Allen and Gary N. Underwood (eds.), Readings in American Dialectology. New York: Appleton‐Century‐ Crofts, pp. 196–215. Boberg, Charles 1999. The attitudinal component of variation in American English foreign (a) nativization. Journal of Language and Social Psychology 18: 49–61. Boretzky, Norbert 1993. The concept of rule, rule borrowing, and substrate influence in creole languages. In Salikoko S. Mufwene (ed.), Africanisms in Afro‐American Language Varieties. Athens, GA: University of Georgia Press, pp. 74–92.

380  Joseph C. Salmons and Thomas Purnell Britain, David and Peter Trudgill 1999. Migration, new‐dialect formation and sociolinguistic refunctionalisation: reallocation as an outcome of dialect contact. Transactions of the Philological Society 97: 245–56. Braun, Sarah 2020. Mapping perceptions of language variation in Wisconsin: On “Goin’ Fishin’ at my Cabin Don’t Cha Know” and “Normal Like Me”. American Speech 95: 82–102. Brown, Joshua R. (ed.) forthcoming. Verticalization: A Model for Language Shift. Oxford: Oxford University Press. Cannon, Garland 1984. Zero plurals among the Japanese loanwords in English. American Speech 59: 149–58. Cassidy, Frederic G. and Joan Houston Hall (eds.) 1985–. The Dictionary of American Regional English. Cambridge, MA: Harvard University Press. Chambers, J. K. 2013. Patterns of variation including change. In J. K. Chambers and Natalie Schilling‐Estes (eds.), The Handbook of Language Variation and Change, 2nd edition. Oxford: Blackwell, pp. 297–323. Childs, Becky and Christine Mallinson 2004. African American English in Appalachia. English World‐Wide 25: 27–50. DARE = Dictionary of American Regional English 1993. An index by region, usage, and etymology to the Dictionary of American Regional English, Volumes I and II. Durham, NC: Duke University Press. (= Publications of the American Dialect Society 77.) Denis, Derek and Alexandra D’Arcy 2018. Settler colonial Englishes are distinct from postcolonial Englishes. American Speech 93: 3–31. Dillard, J. L. 1972. Black English: Its History and Usage in the United States. New York: Vintage House. Dorian, Nancy 1993. Internally and externally motivated change in language contact settings: doubts about the dichotomy. In Charles Jones (ed.), Historical Linguistics: Problems and Perspectives. London: Longman, pp. 131–55. Downes, William 1998. Language and Society. London: Fontana. Durian, David 2007. Getting [s]tronger every day? Urbanization and the socio‐geographic diffusion of (str) in Columbus, OH. University of Pennsylvania Working Papers in Linguistics 13: 65–79. Feinstein, Mark 1980. Ethnicity and topicalization in New York City English. International Journal of the Sociology of Language 26: 15–24. Fought, Carmen 2003. Chicano English in context. New York: Palgrave Macmillan.

Fought, Carmen 2006. Language and Ethnicity. Cambridge: Cambridge University Press. Frey, Benjamin E. 2013. Towards a general theory of language shift: a case study in Wisconsin German and North Carolina Cherokee (University of Wisconsin–Madison Ph.D. dissertation). Genee, Inge and Shelley Stigter 2010. Not just “Broken English”: Some grammatical characteristics of Blackfoot English. Canadian Journal of Native Education 32: 62–82. Godínez, Manuel, Jr. 1984. Chicano English phonology: Norms vs. interference. In Jacob Ornstein‐Galicia and Allan Metcalf (eds.), Form and Function in Chicano English. Rowley, MA: Newbury House, pp. 42–8. González, Félix 2000. Spanish contribution to American English word stock: an overview. American Speech 75.3: 292–5. González, Gustavo 1988. Chicano English. In Dennis Bixler‐Márquez and Jacob Ornstein‐Galicia (eds.), Chicano Speech in the Bilingual Classroom. New York: Peter Lang, pp. 71–82. Gordon, Matthew J. 2000. Phonological correlates of ethnic identity: evidence of divergence? American Speech 75: 115–36. Hall‐Lew, Lauren, Elizabeth Coppock, and Rebecca L. Starr 2010. Variation in the Iraq vowels. American Speech 85: 91–102. Herold, Ruth 1990. Mechanisms of merger: the implementation and distribution of the low back merger in Eastern Pennsylvania, (University of Pennsylvania Ph.D. dissertation). Heffner, Roe‐Merrill Secrist 1937. Notes on the length of vowels. American Speech 12: 128–34. Hickey, Raymond 2000. Salience, stigma and standard. In Laura Wright (ed.), The Development of Standard English 1300–1800: Theories, Descriptions, Conflicts. Cambridge: Cambridge University Press, pp. 57–72. Hickey, Raymond 2003. How do dialects get the features they have? On the process of new dialect formation. In Raymond Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press, pp. 213–39. Hickey, Raymond 2004. Development and diffusion of Irish English. In Raymond Hickey (ed.), Legacies of Colonial English. Cambridge: Cambridge University Press, pp. 82–117. Hickey, Raymond 2010. Attitudes and concerns in eighteenth‐century English. In Raymond Hickey (ed.), Eighteenth Century English: Ideology and Change. Cambridge: Cambridge University Press, pp. 1–20.

Contact and the Development of American English  381 Hills, E. C. 1929. Linguistic substrata in American English. American Speech 4: 431–3. Howell, Robert B. 1993. German immigration and the development of regional variants of American English: using contact theory to discover our roots. In Joseph Salmons (ed.), The German Language in America. Madison, WI: Max Kade Institute, pp. 190–212. Iverson, Gregory K. and Joseph C. Salmons. 1995. Aspiration and laryngeal representation in Germanic. Phonology 12: 369–96. Iverson, Gregory K. and Joseph C. Salmons 2007. Domains and directionality in the evolution of German final fortition. Phonology 24: 1–25. Janda, Richard D., Brian D. Joseph, and Neil G. Jacobs 1994. Systematic hyperforeignisms as maximally external evidence for linguistic rules. In Susan Lima, Roberta Corrigan, and Gregory K. Iverson (eds.), The Reality of Linguistic Rules. Amsterdam: John Benjamins, pp. 67–92. Kerswill, Paul 2013. Koinéization and accommodation. In J. K. Chambers and Natalie Schilling (eds), The Handbook of Language Variation and Change, 2nd edition. Oxford: Blackwell, pp. 519–36. Kerswill, Paul and Peter Trudgill 2005. The birth of new dialects. In Peter Auer, Frans Hinskens, and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press, pp. 196–220. Kurath, Hans and Raven McDavid, Jr. 1961. The Pronunciation of English in the Atlantic States. Ann Arbor: University of Michigan Press. Labov, William 1966. The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics. Labov, William 1972. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William 1994. Principles of Linguistic Change, vol. 1: Internal Factors. Malden, MA: Blackwell. Labov, William 2001. Principles of Linguistic Change, vol. 2: Social Factors. Oxford: Blackwell. Lanehart, Sonja (ed.) 2015. The Oxford Handbook of African American Language. Oxford: Oxford University Press. Leap, William L. 1993. American Indian English. Salt Lake City: University of Utah Press. Levine, Lewis and Harry Crockett 1966. Speech variation in a Piedmont community: postvocalic /r/. In Stanley Lieberson (ed.), Explorations in Sociolinguistics. The Hague: de Gruyter Mouton, pp. 76–98.

Linguistic Atlas of the Gulf States 1986–92, ed. Lee A. Pederson, 7 vols. Athens: University of Georgia Press. Litty, Samantha M. 2017. We talk German now yet: the sociolinguistic development of voice onset time and final obstruent neutralization in Wisconsin German and English varieties, 1863–2013, (University of Wisconsin–Madison Ph.D. dissertation). McCarthy, Corrine 2007. Social correlates of vowel shifting in Chicago: a pilot study. McGill Working Papers in Linguistics 21: 21–34. McDavid, Raven I., Jr. 1948. Postvocalic /‐r/ in South Carolina: a social analysis. American Speech 23: 194–203. Mencken, H. L. 1937. The American Language: An Inquiry into the Development of English in the United States. New York: Knopf. Mesthrie, Rajend 2001. Koinés. In Rajend Mesthrie (ed.), Concise Encyclopedia of Sociolinguistics. Amsterdam: Elsevier, pp. 485–49. Mufwene, Salikoko S. 1996. The Founder Principle in creole genesis. Diachronica 13: 83–134. Mufwene, Salikoko S. 2000. Some sociohistorical inferences about the development of African American English. In Shana Poplack (ed.), The English History of African American English. Oxford: Blackwell, pp. 233–63. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Mufwene, Salikoko S. 2006. The comparability of new‐dialect formation and creole development. World Englishes 25: 177–86. Mufwene, Salikoko S. 2015. The emergence of African American English: Monogenetic or polygenetic? With or without “decreolization”? Under how much substrate influence? In Sonja Lanehart (ed.), The Oxford Handbook of African American Language. Oxford: Oxford University Press, pp. 57–84. Mufwene, Salikoko S. and Charles Gilman 1987. How African is Gullah, and why? American Speech 62: 120–39. Newmark, Kalina, Nacole Walker, and James Stanford 2016. “The rez accent knows no borders”: Native American ethnic identity expressed through English prosody. Language in Society 45: 633–64. Nguyen, Jennifer 2006. The changing social and linguistic orientation of the African American middle class (University of Michigan Ph.D. dissertation). Otheguy, Ricardo, Ofelia García, and Ana Roca 2000. Speaking in Cuban: the language of

382  Joseph C. Salmons and Thomas Purnell Cuban Americans. In Sandra McKay and Sau‐ ling Wong (eds.), New Immigrants in the United States. Cambridge: Cambridge University Press, pp. 165–88. Parker, Frank 1974. The coarticulation of vowels and stop consonants. Journal of Phonetics 2: 211–21. Poplack, Shana and Sali Tagliamonte 1996. Nothing in context: variation, grammaticalization and past time marking in Nigerian Pidgin English. In Philip Baker (ed.), Changing Meanings, Changing Functions: Papers Relating to Grammaticalization in Contact Languages. London: University of Westminster Press, pp. 71–94. Purnell, Thomas C. 2009. Convergence and contact in Milwaukee: evidence for select African American and white vowel space features. Journal of Language and Social Psychology 28: 408–27. Purnell, Thomas, Eric Raimy and Joseph Salmons 2017. Upper Midwestern English. In Raymond Hickey (ed.), Listening to the Past: Audio Records of Accents of English. Cambridge: Cambridge University Press, pp. 298–324. Purnell, Thomas C., Joseph C. Salmons, and Dilara Tepeli 2005a. German substrate effects in Wisconsin English: evidence for final fortition. American Speech 80: 135–64. Purnell, Thomas C., Joseph C. Salmons, Dilara Tepeli, and Jennifer Mercer 2005b. Structured heterogeneity and change in laryngeal phonetics: Upper Midwestern final obstruents. Journal of English Linguistics 33: 307–38. Radzilowski, John 2007. Late nineteenth century Europeans, Mexicans, and Asians. In Richard Sisson, Christian Zacher, and Andrew Clayton (eds.), The American Midwest: An Interpretative Encyclopedia. Bloomington: Indiana University Press, pp. 217–20. Rickford, John 1986. Social contact and linguistic diffusion. Language 62: 245–90. Rickford, John 1999. African American Vernacular English: Features, Evolution and Educational Implications. Oxford: Blackwell. Romaine, Suzanne 2001. Contact with other languages. In John Algeo (ed.), The Cambridge History of the English Language, vol. 6: English in North America. Cambridge: Cambridge University Press, pp. 154–83. Rosa, Jonathan 2019. Looking Like a Language, Sounding Like a Race: Raciolinguistic Ideologies and the Learning of Latinidad. Oxford: Oxford University Press.

Rose, Mary Aleene 2006. Language, place and identity in later life (Stanford University Ph.D. dissertation). Rowicka, Grażyna J. 2005. American Indian English: the Quinault case. English World‐Wide 26: 301–24. Salmons, Joseph 2019. Laryngeal phonetics, phonology, assimilation and final neutralization. In B. Richard Page and Michael T. Putnam (eds.), The Cambridge Handbook of Germanic Linguistics. Cambridge: Cambridge University Press, pp. 119–42. Sankoff, Gillian and Hélène Blondeau 2007. Language change across the lifespan: /r/ in Montreal French. Language 83: 560–88. Santa Ana, Otto 1991. Phonetic simplification processes in the English of the barrio: a cross‐ generational sociolinguistic study of the Chicanos of Los Angeles (University of Pennsylvania Ph.D. dissertation). Santa Ana, Otto 1993. Chicano English and the nature of the Chicano language setting. Hispanic Journal of Behavioral Sciences 15.1: 3–35. Santa Ana, Otto and Robert Bayley 2004. Chicano English phonology. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar Schneider and Clive Upton (eds.), A Handbook of Varieties of English Vol. 1 Phonology. Berlin: de Gruyter Mouton, pp. 417–34. Schilling‐Estes, Natalie 2002. On the nature of isolated and post‐isolated dialects: innovation, variation, and differentiation. In Lesley Milroy (ed.), Investigating Change and Variation through Dialect Contact. Special issue of Journal of Sociolinguistics 6.1: 64–85. Schneider, Edgar 2007. Postcolonial English: Varieties Around the World. Cambridge: Cambridge University Press. Schönweitz, Thomas 2001. Gender and postvocalic /r/ in the American south: a detailed socioregional analysis. American Speech 76: 259–85. Schreier, Daniel 2015. Historical phonology and koineization. In Patrick Honeybone and Joseph Salmons (eds.), The Oxford Handbook of Historical Phonology. Oxford: Oxford University Press, pp. 619–36. Schuld, Danielle, Joseph Salmons, Thomas Purnell, and Eric Raimy 2016. “Subliminal accent”: Reactions to the rise of Wisconsin English. Journal of Linguistic Geography 4: 15–30. Thomas, C. K. 1932. Jewish dialect and New York dialect. American Speech 7: 321–6.

Contact and the Development of American English  383 Thomas, Erik R. 2001. An Acoustic Analysis of Vowel Variation in New World English. Durham, NC: Duke University Press. Thomas, Erik R. (ed.) 2019. Mexican American English: Substrate Influence and the Birth of an Ethnolect. Cambridge: Cambridge University Press. Trudgill, Peter 1986. Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 1999. A window on the past: “colonial lag” and New Zealand evidence for phonology of nineteenth‐century English. American Speech 74: 227–39. Trudgill, Peter 2004. New‐Dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press. Trudgill, Peter 2008. Colonial dialect contact in the history of European languages: on the irrelevance of identity to new‐dialect formation. Language in Society 37: 241–80. Trudgill, Peter 2011. Sociolinguistic Typology: Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. Urciuoli, Bonnie 1996. Exposing Prejudice: Puerto Rican Experiences of Language, Race, and Class. Boulder, CO: Westview Press. van Coetsem, Frans 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. van Coetsem, Frans 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Winter. von Schneidemesser, Luanne 1999. An Index by Region, Usage, and Etymology to the Dictionary of American Regional English, Volume III (Publications of the American Dialect Society 82). Durham, NC: Duke University Press. Wald, Benji 1984. The status of Chicano English as a dialect of American English. In Jacob Ornstein‐Galicia and Allan Metcalf (eds.),

Form and Function in Chicano English. Rowley, MA: Newbury House, pp. 14–31. Wald, Benji 1987. Spanish‐English grammatical contact in Los Angeles: The grammar of reported speech in the East Los Angeles English contact vernacular. Linguistics 25: 53–80. Wald, Benji 1996. Substratal effects on the evolution of modals in East LA English. In Jennifer Arnold, Renée Blake, Brad Davidson, Scott Schwenter, and Julie Solomon (eds.), Sociolingistic Variation: Data, Theory, and Analysis: Selected papers from NWAV 23 at Stanford. Stanford, CA: Center for the Study of Language and Information, pp. 515–30. Weinreich, Uriel 1964 [1953]. Languages in Contact: Findings and Problems. The Hague: de Gruyter Mouton. Wilkerson, Miranda, Mark Livengood and Joseph Salmons 2014. The socio‐historical context of imposition in substrate effects. Journal of English Linguistics 42: 1–23. Wilkerson, Miranda and Joseph C. Salmons 2008. “Good old immigrants of yesteryear” who didn’t learn English: Germans in Wisconsin. American Speech 83.3: 259–83. Winford, Donald 2005. Contact‐induced changes: classification and processes. Diachronica 22: 373–427. Wolfram, Walt 1969. A Sociolinguistic Description of Detroit Negro Speech. Washington, DC: Center for Applied Linguistics. Wolfram, Walt 1996. Delineation and description in dialectology: the case of perfective I’m in Lumbee English. American Speech 71: 5–26. Wolfram, Walt and Natalie Schilling 2016. American English. Third edition. Oxford: Blackwell. Wolfram, Walt and Erik R. Thomas 2002. The Development of African American English. Oxford: Blackwell.

19 Contact and African Englishes RAJEND MESTHRIE 1  Background to Anglo‐African Contact Although the topic of African Englishes is a potentially vast one, several factors make it a manageable and coherent one. One is that the main influence of English has been in Sub‐ Saharan Africa, rather than the more northern countries in which French and Arabic predominate as cultural and contact languages. Formal British colonialism touched mainly West and Central Africa and almost all of East and Southern Africa in the establishment of colonies and protectorates. This chapter assesses the role of contact in the development of English in these places, in relation to other significant factors. According to Spencer (1971: 8) English was probably first taken to Africa in the 1530s when William Hawkins the Elder passed there on his way to Brazil. This would have been a form of Elizabethan English. A regular trade in spices, ivory, and slaves began in the mid 1500s when British ships sailed along the Guinea coast (Schmied 1991: 6). European forts were subsequently established along the West African coast. Pidgin Portuguese was the earliest form of a European language used there. As British supremacy in trade gradually grew, English began to gain a foothold. During this time West Africans were taken in small numbers to Europe to be trained as interpreters. An account in Hakluyt (1598–1600, vol. 6) cited by Spencer (1971: 8) suggests that by 1555 five West Africans had been taken to England for over a year for this purpose. Within Africa the earliest contacts between English speakers and the locals were informal and sporadic. There was no expectation of a permanent settlement or of colonization (and therefore formal education) until centuries later. In this first phase pidgins and “broken English” (i.e. early‐fossilized interlanguages) were the outcomes of contact. These would not necessarily be ephemeral: West African Pidgin English whose roots lie in the seventeenth and eighteenth centuries is today more widespread (in Cameroon, Ghana, and Nigeria) than is English as a second language. Pidgin English was not the only code used, as the African interpreters returning from training in England would probably have used English as a second language (ESL) rather than pidgin. Two later influential varieties in West Africa (from about 1787 onwards) were the forms of creole English spoken by manumitted slaves who were repatriated from Britain, North America, and the Caribbean. Krio was the English creole of slaves freed from Britain who were returned to Sierra Leone, where they were joined by slaves released from Nova Scotia and Jamaica. Liberia was established in 1821 as an African homeland for freed slaves from the US, becoming independent in 1847. The creole English that the returnees brought with them was most likely related to African American Vernacular English (Hancock and Kobbah 1975: 248, cited by Todd 1982: 284; Singler 2004). Today, American rather than British forms of English continue to dominate in Liberia (see Singler 2004). Todd (1982) describes four types of

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

386  Rajend Mesthrie English in West Africa: pidgin; ESL; Standard West African English (mostly oriented to the UK, with the exception of places like Liberia), and Francophone West African English. In Southern and East Africa, by contrast, no pidgin English emerged, as other important lingua francas came into being at different times – e.g. Swahili in East Africa and Fanakalo, a Southern African pidgin in which Zulu is the main lexifier. In addition, urban forms of certain languages like Town Bemba in Zambia and youthful mixed forms like Sheng in Kenya play a role in cutting across ethnicities. Four traditional language phyla are identified within Africa: Afro‐Asiatic, Niger‐Congo, Nilo‐Saharan, and Khoe‐San. Of these phyla the second is of key relevance to English studies, as it accounts for most cases of “Afro‐Saxon bilingualism” (my term for the co‐existence of an African language with English in societies). Niger‐Congo itself is made up of many ­subfamilies, the best‐known and most influential one being that of the Bantu languages. These are the most widespread languages in key territories like South Africa, Zimbabwe, Lesotho, Botswana, Zambia, Malawi (in Southern Africa), Nigeria, Ghana, and Liberia (West Africa), Cameroon and the Democratic Republic of the Congo (Central Africa), and Kenya, Uganda, and Tanzania (East Africa). It would appear that, for demographic reasons, it is the  Bantu substratum that has had the most influence on English in Africa. However, a ­comparative study of the influence of non‐Bantu languages, especially of West Africa would be desirable before one could be conclusive about this. This chapter assumes – as do most scholars (see e.g. Bokamba 2011; Simo Bobda 2000a; Anchimbe 2006) – that it is possible to speak of Sub‐Saharan English as a typological entity. Henceforth I will use the abbreviation SSE for this loose grouping of English as a second language in this territory.

2  Contact vs Non‐Contact Effects This chapter concurs with most previous writing that contact is a major ingredient in the formation of SSE (e.g. Sey 1973; Schmied 1991; Bokamba 1992). However, it is important not to prejudge the influence of contact, and to examine other possible factors that may have shaped SSE significantly. One consideration is that English was often introduced in the classroom, rather than via the sizable presence of its L1 (first language) speakers. The learning of English was therefore a relatively controlled process, in which the influence of the primary context (education rather than contact with speakers) could be expected to play some role. Janina Brutt‐Griffler (2000) coined the term “macroacquisition” for this kind of controlled group learning of a superstrate. To my knowledge no close empirical study has been done of the unfolding of classroom English of this kind. This would require a longitudinal study of a group of students undergoing “macroacquisition” to describe and understand its characteristic processes, in relation to stages of acquisition, nature of the input, etc. in situations in Africa and Asia where native speakers have largely departed. In the area of phonology there is a common assumption in the literature on “World Englishes” that the introduction of English via reading and writing has resulted in “spelling pronunciations,” see e.g. (Schmied 1991: 53–4; Gough 1996: 59, Van Rooy and Van Huyssteen 2000: 23). If this hypothesis is correct then we would indeed be faced with a new kind of influence that goes beyond that of inter‐speaker contact (see Section 3.2 for a detailed analysis of this possibility). There are other positions to consider in weighing up the relative influence of contact and its opposites in SSE. One concerns the influence of variation from within the superstrate itself. Most World English studies set on determining the features of a variety operate within a contrastive paradigm in which the local variety of L2 (second language) English is compared to the standard form of the superstrate. But given that contact in Africa goes back to the sixteenth century, we should be alert to features that are historical retentions that have since been lost in the standard varieties of English. Some such lexical examples include trinket ‘item of jewelry’ (not necessarily a cheap one); station ‘place of abode’, and not so?

Contact and African Englishes  387 (Davy 2000; Mesthrie 2003). As far as syntax is concerned the form can be able which is widespread in parts of Africa (as an apparent equivalent of the auxiliary can) could possibly go back to Elizabethan English. Visser (1963–73, III: 1738) gives examples of such usage from Thomas More, Shakespeare, Congreve, and Dryden, etc. The example from Dryden’s Satires (1693) provided by Visser is: That favour is sufficient to bind any grateful man, …to all the future service, which one of my mean condition can ever be able to perform. As far as phonology goes, Simo Bobda (2003) provides a long list of possible phonetic links between SSE and Anglo‐ Saxon varieties that are masked if one were to use the modern standards (or prestige varieties – RP or “General American”) as the yardstick. Some of his examples are convincing: the monophthongal realizations of the FACE and GOAT diphthongs may well be linked to the older forms of these diphthongs ([e] and [o]) in Northern English, the Celtic Englishes, and the Caribbean. Likewise, Harris (1996) describes the possible dialect input to the realization of the STRUT vowel as [ɔ] in West Africa. Other examples of parallels that Simo Bobda (2003) cites (e.g. l‐vocalization, fronted variants of NURSE) are less compelling, though his overall position that these parallels should not be entirely ignored is a valid one. The final position is a comparative one that considers SSE in relation to L2 varieties of English across the globe. Platt, Weber, and Ho (1984) identified recurrent similarities in the grammatical and other features of the following countries/regions: India, Singapore, and Malaysia, Sri Lanka, and East and West Africa. Some of these features (statives as progressives, resumptive pronouns, occasional gender neutralization of pronouns, and a high frequency of left dislocation and topicalization) are discussed and exemplified in Section 4.1. Although Platt, Weber, and Ho (1984) excluded South Africa on the grounds that it had a considerable body of L1 speakers of English, the L2 English of many Southern African countries, including South Africa do exhibit the features they documented. Such an approach that stresses similarities in L2 Englishes world‐wide diminishes the importance of the substrates, since it would be difficult to maintain that a large number of substrates belonging to different families and types of languages all produced the same effects. Williams (1987) drew these threads together in a psycholinguistic approach that looked at processing properties like maximizing salience, avoiding redundancy within clauses, making cross‐clausal relations more explicit by double marking, etc. Today the terms “Angloversals” (Mair 2003; Sand 2004), “New Englishisms” (Simo Bobda 2000b) and “Universals of New Englishes” (Szmrecsanyi and Kortmann, 2009) can be found for the persistence of such properties in L2 Englishes, despite their infrequency or absence in more mainstream standard forms. Finally, although some scholars have provided accounts that attempt to minimize the differences between pidgin‐creoles and L2 varieties (Mufwene 2001; DeGraff 2003), the data from Africa and Asia suggests that there is a typological disjunction between the two types (Mesthrie 2004a; Szmrecsanyi and Kortmann, 2009). Within Africa the difference between pidgin/creole varieties like West African Pidgin English on the one hand and SSE on the other is vast. For reasons of space these pidgin/creole varieties will not be discussed in this chapter.

3  Effects of Contact in Sub‐Saharan English Phonology Substrate influence is clearly discernible in SSE phonetics and phonology. Among the relevant salient phonetic/phonological tendencies in Bantu languages are: 1. Five‐ or seven‐vowel systems in which length is not distinctive and diphthongs are absent (Clements 2000: 135) 2. Rarity of vowel reduction to schwa 3. Rich tonal rather than stress systems; a tendency towards syllable‐timing 4. A tendency to CV syllable structure, or NCV where N is a nasal1 In this section I will concentrate on characteristic (1).

388  Rajend Mesthrie

3.1  Vowel Systems in SSE Most of the varieties of SSE, especially those spoken by people with lower levels of formal education and fewer contacts with native speakers, display a five‐vowel system, with few or no diphthongs. Whilst there is internal differentiation according to the trichotomy East Africa, West Africa, and Southern Africa, the very existence of this broad typology argues for substrate influence. African languages tend not to differentiate vowels according to length, tend to have few instances of vowel reduction to schwa, and tend to avoid diphthongs. There are two subtypes of the five‐vowel system, depending on particular mergers. Figure  19.1 shows the system in terms of the subset of short vowels in Wells’ (1982) lexical sets. Type 1, with merger of TRAP and STRUT is found in East and Southern Africa (with one difference that I will discuss presently). Simo Bobda (2001) surveys varieties from the following countries: Zimbabwe, Zambia, and Malawi (Southern Africa); and Tanzania, Uganda, and Kenya (East Africa) showing this basic pattern. Type 2 with merger of LOT and STRUT is found in West Africa (Southern Nigeria and Ghana) and Cameroon, the neighboring country from Central Africa. The vowel systems in northern Nigerian English are different (Gut 2004), relying more upon the vowel systems of languages like Hausa, an Afro‐Asiatic language. Liberian Settler English, spoken largely by descendants of repatriated slaves of the nineteenth century (Singler 2004), is an offshoot of African American Vernacular English and is therefore different from other varieties of English in Africa. Figure 19.1 shows only the short‐vowel reflexes of Wells’ lexical sets. There is inevitably a certain amount of idealization in the sketches. Allophones of the mid‐vowels do occur: e.g. [ε] and [e] and [ɔ] and [o] in South Africa. These might be influenced by raising rules of [e] and [o] before high vowels in local languages like Zulu and Xhosa. In Southern Africa the TRAP vowel tends to have realizations [e] or [ε] in addition to [a]. The variant [e] occurs frequently in words like trap, matter, cat, and is usually given as the basic value of the set by most authors on the subject. Hundleby (1963: 68) notes a range of articulation points for this vowel amongst Xhosa speakers of English, from a lowered [e] to a point back of [a], and suggests that the former is a “weak” form and the latter the “strong” form. Van Rooy and Van Huyssteen (2000) give the basic value of TRAP as [e] with [ε] and [a] as lesser alternatives. Van Rooy (2004) gives the basic value as [e] but cites [æ] as a significant alternant. However, it seems to me that [e] or [ε] is not really as widespread as the [a] equivalent. I propose that the basic form is [a], with some fluctuation according to words (often monosyllables) and speakers. I also suggest that [æ] is a form more characteristic of the acrolect. Here is a list of words taken from educated Black South African English (henceforth BSAE) speakers speaking on national radio as anchor person or expert commentators:2

KIT

FOOT

KIT

FOOT

DRESS

LOT

DRESS

LOT/STRUT

TRAP/STRUT Type 1

Figure 19.1  Two idealized vowel templates in SSE.

TRAP Type 2

Contact and African Englishes  389 (1) a. b. c. d. e. f. g. h. i. j. k. l.

salaries balance analyst affidavit adventure Andrew Natal Maritzburg matter that’s taxi standing

[a] [a] [a] [a] [a] [a] [a] [a] [a] ~ [e] [æ] ~ [e] [æ] [æ]

[salari:z] [bala:ns] [anali:st] [afidavi:t] [advɛ:ntʃa] [andru:] [nata:l] [maritsbe:g] [ma:ta] ~ [me:ta] [ðæts] ~ [ðets] [tæksi:] [stændi:ŋ]

In addition, a mesolectal speaker I reported on in Mesthrie (2005), showed several variants in his reallocation of the TRAP vowel to either [e], [ε], or [a]. The following examples occur with [ɛ̢]: cannot, have, understand. The following had [a]: grammar, actually, emancipation. Attestations of [e] were rare: happy, trap. Attestations of [æ] were even rarer, occurring only in the word family. Such patterns of reallocation show a strategy of “mapping” as speakers try and relate new target‐language vowels to their L1 system. These two short preliminary case studies suggest that it is reasonable to consider [a] the basic value, with some alternation (possibly lexically governed, with monosyllabic bases) with [ɛ] or [e], and some speakers starting to go beyond the five‐vowel system to produce incipient [æ]. However, there is no serious flaw in proposing the alternative that [e] is basic in the TRAP set, with [a] a variant. A third alternative – that TRAP is redistributed over DRESS and TRAP – is also possible. However, it is a less likely solution, since there does appear to be some phonological rationale (rather than lexical arbitrariness) in choosing between [e] (largely with monosyllabic bases) and [a] in most other instances. Perhaps we should give the phonological casting vote to the first vowel in the words Africa and African. Here [a] wins hands down for all BSAE speakers. As far as the other vowels of mesolectal BSAE are concerned the following trends emerge. Length is neutralized so that KIT and FLEECE have /i/, FOOT and GOOSE have /u/, and LOT and THOUGHT have /ɔ/, with some allophonic variation in the last set. BATH tends to merge with /a/ (which is TRAP/STRUT in type 1 or just TRAP in type 2). Again there are exceptions and complications. However, [a] is recorded as the main variant in East Africa, Southern Africa, and much of West Africa (southern Nigeria, Ghana) as well as the Cameroons (Central Africa). In South Africa heart and hut can sound the same, taking [a]. Finally, central vowels are eschewed in SSE. NURSE has the reflexes [a] in East Africa, [e] or [ɛ] in Southern and much of West Africa (southern Nigeria and Ghana) and [ɛ] or [ɔ] in the Cameroons (Simo Bobda 2004: 888). Finally, schwa is rare in SSE, usually being replaced by [a] or a close variant like [ɐ] in all territories. In some instances it takes on the full value of [i] (e.g. in certain), [u] (e.g. in people), or [e] (e.g. in fountain), with predictable lengthening according to syllable structure (described on p. 390). See also the transcriptions of the underlined vowels in words of the TRAP set in example (1) above. At this stage, taking into account vowel‐length neutralization, we have the monophthongal systems shown in Figure 19.2. Type 1A is Southern African;3 type 1B is East African, and type 2 is West African. In the Cameroons we have a further type (Figure 19.2d), distinguished by the treatment of NURSE, which is usually [ɔ], though some words in this set take [ɛ] (Simo Bobda 2004). The L1 English diphthongs are not as easily generalizable in SSE. A noticeable tendency is to monophthongize FACE and GOAT. These are given as [e] and [o] respectively in East and

390  Rajend Mesthrie (a)

(b)

KIT/FLEECE

FOOT/GOOSE

DRESS/NURSE

LOT

KIT/FLEECE

DRESS

TRAP/STRUT/BATH Type 1A

(c)

FOOT/GOOSE

LOT

TRAP/STRUT/BATH/NURSE Type 1B

(d)

KIT/FLEECE

FOOT/GOOSE

DRESS/NURSE

LOT/STRUT

KIT/FLEECE

TRAP/BATH Type 2

FOOT/GOOSE

DRESS

LOT/STRUT/NURSE

TRAP/BATH Type 3

Figure 19.2  Four main types of five‐vowel systems in SSE.

West Africa, the Cameroons, and South Africa. It would be interesting to see whether these are phonemically distinct as /e/ and /o/ from the other sets which might be conceived of as /ɛ/ and /ɔ/. This appears to be the case in Southern Nigeria (Gut 2004: 819), and for South Africa (van Rooy 2004: 947). Finally, although vowel length is not a distinctive feature in SSE, there is a regular phonological rule that lengthens vowels in heavy final syllables (CVC or VCC) or else in the penultimate syllable (see Van Rooy 2004: 950). It would be difficult to find a reasonable explanation for the broad similarities in the vowel systems of SSE other than the effects of language contact, with subsequent regularizations. However, substrate influence is mediated by other factors. Take the vowel‐length rule of BSAE as an example. The broad impetus for penultimate lengthening comes from languages like Zulu and Xhosa (where this is a regular feature, secondary to tonal effects on the syllable in question). Nonetheless, the rule has to be modified in the emerging L2 dialect to take into account the presence of heavy final syllables, lacking in the substrates (see further Van Rooy 2004: 950). Finally, I agree with Harris (1996) and Simo Bobda (2003) that not all influences should automatically be attributed to substrate influence. Some of the phonetic variation in realizations of individual lexical sets might have been due to differential input. Harris (1996: 7–8) points out that even where interlingual phonemic identification between substrate and L2 English appears to indicate substrate influence “lock, stock and barrel,” the very fact that the L2 lexis comes almost entirely from the superstrate means that “on the basis of shared lexical stock it, it is still possible to establish regular phonological correspondences between the contact variety and its lexical donor.”4 He proposes that the realization of TRAP and STRUT as [a] and [ɔ] respectively in West Africa and [e] or [ɛ] and [a] in Southern Africa is due to such differences in superstratal values in the seventeenth century for West Africa and the nineteenth century for Southern Africa. This is a plausible argument, except that the Southern African realizations of TRAP are (as pointed out on p. 388–389) more complex than an examination of just the (monosyllabic) token trap indicates. A large number of tokens in the set have [a] which is more likely to favor the substrate pattern, rather than the shape of the superstrate in nineteenth‐ century South Africa (showing raising of the TRAP vowel).

Contact and African Englishes  391

3.2  Possible Roles for Spelling Pronunciations The final topic in my exploration of language contact phonology and its opposites is a consideration of what I term the “spelling form hypothesis.” A number of writers have tacitly assumed and sometimes openly asserted that spelling pronunciations are rampant in the variety of SSE that they describe. The position is that since these varieties were learned via the education system, they show the influence of the written medium more than other instances of L2 acquisition involving interactional contexts with at least some mother tongue speakers. This is an important position to consider since it poses a major challenge to contact and historical linguistics: that substrate and superstrate are subject to a “third force” of reading in L2 formation under macroacquisition. I tackle this issue in Mesthrie (2005), where I agree that some spelling forms are inevitable for words that one sees in print but seldom ever hears pronounced: rare names of people, places, and things, for example. But there are many unwarranted assumptions in generalizing this influence to all or even a large section of the vocabulary of an individual L2 variety. The unwarranted assumptions are as follows: 1. Literacy is widespread. 2. L2 speakers learn the spelling of words before they formulate (mental) phonological classes of words. 3. Spelling is easy; pronunciation is difficult. 4. English letters “have” sounds, about which there is consensus amongst all L2 speakers. 5. People access orthographic forms as they speak (in early stages of acquisition). 6. L2 diachrony is equated with L2 synchronic processing. 7. Spelling pronunciations are random. In fairness to authors cited above on SSE, their observations are often made in passing, when attempting to give a descriptive picture of stress and vowel realizations, rather than seriously proposing a theory of acquisition of individual sociolects. One might, more appropriately, distinguish between a strong form of the hypothesis, holding that spellings are of considerable significance in L2 phonology, and a weaker form, specifying that under certain conditions, some words show a spelling pronunciation in the sense that their orthographic representation influences their phonological form in a particular L2. Assumption 1 is not tenable for Sub‐Saharan Africa. South Africa, for example, had a functional illiteracy rate amongst adults of 29% in 2001. In discounting assumption 2, it should be stressed that the “first learners” of BSAE did not learn the target language (mainly) via the written mode: there had to be teachers using spoken English (including Teacher Talk). For BSAE the first generation of teachers were mainly missionaries from a variety of European backgrounds (Magura 1984; Mesthrie 1996). Their pupils would have tried to imitate the pronunciations of their teachers as best they could, using their own L1 phonology as a rough guide to the system they were learning. If, for example, their native phonology had no schwa then they would find its nearest equivalent, using a slowly developing “mapping procedure.” The nearest equivalent settled upon appears to be [ɐ], a close variant of /a/ as a default, pending other principles (discussed on p. 389). Assumption 3 is counteracted by anecdotes of experiences of students I interviewed, some of whom claimed that they had few problems learning the pronunciation of words, but had to work especially hard to learn the spelling forms of an erratic target language. An interview with one speaker even suggested that in his case, rather than L2 spelling generally driving the phonology of the L2, it may well be that L1 phonology can be used to help learn L2 spelling. Examples from university students, even at postgraduate level, occasionally reflect the lack of vowel‐length distinctions: beeches for “bitches” (in an essay on slang); whizzing for “wheezing” (in a dissertation on health); feast for “fist” (in an essay on sport) etc. These “pronunciation spellings” as I term them call into question the assumed direction of influence in 3.

392  Rajend Mesthrie Assumption 4 is contradicted by the fact that different varieties have come up with different realizations of the same spelling forms (as in the account of the vowels on p. 388–390). The question one would want to explore is why are there (alleged) spelling pronunciations for, say, unstressed vowels in BSAE, but no spelling pronunciations for, say, instances of postvocalic /r/. Clearly the influence of the L1 is a significant factor in the nonrhotic nature of the L2 variety being developed. Assumptions 5 and 6 fail to draw a distinction between language description and psycholinguistic processing. That is, for L2 studies, the phonological forms have to be analyzed as part of an interlocking system, irrespective of whether they originate from imitation of target‐language speakers, L1 transfer, overgeneralization of rules, analogical formation, or the occasional spelling pronunciation. Such properties are of immense historical interest, but should not be assumed to be “active” in the synchronic phonological processing of the L2 variety being studied. Assumption 7 is related to assumption 4, for which I propose that spelling pronunciations in second language acquisition are not unconstrained. Spellings might suggest certain pronunciations, but these pronunciations are realized in terms of permissible patterns of the L2 being developed and the L1 of learners. Thus if L2 learners of English come across a new word in print, they will not give it a spelling pronunciation if the forms suggested by the spelling are phonologically unacceptable in their L1 and in their L2. They have no option but to rephonologize it as best they can. In Mesthrie (2005), on the basis of analyses of three speakers of Black South African English, I propose a developmental profile that shows the initial importance of contact, and the gradual restructuring away from a basic contact variety. Speakers initially have a five‐vowel system with some diphthongs. In the basilect itself some minor variations occur with schwa, [æ] and [ʌ] occurring incipiently. In the mesolect these three develop further (i.e. are more frequently used), whilst still remaining variants within the five‐ vowel system. Finally, the acrolect fleshes out the peripheral vowels, with [ʌ] and [æ] occurring frequently, though not invariantly. Furthermore, /æ/ occurs regularly in the acrolect as [æ] medially and [a] initially. Even in the acrolect [ɜ:] is absent; and target‐language schwa corresponds to [ɐ] in exposed position, and to the unstressed variants /i/, /u/, /e/ depending on the surrounding segments, or by convention with certain suffixes. This developmental study gives only minor support for the notion of spelling pronunciation being a significant part of the dialect.

4  Contact and Sub‐Saharan English Syntax I argue that substrate influence is less clear in syntax than in phonology, though some such influence doubtlessly exists. A useful starting point for discussions of the syntax of English in Africa are the collection of articles in the Handbook of Varieties of English (Kortmann et al. 2004), and the Mouton World Atlas of Variation in English (Kortmann and Lunkenheimer 2012). These two volumes contain descriptions of English in a wide range of Sub‐Saharan countries, and detailed synoptic interviews by Mesthrie (2004a) and Huber (2012) respectively.

4.1  Four Syntactic Features of Sub‐Saharan English In this section I describe four features that are found in all five territories covered in Kortmann et al. (2004), with exemplification and offer an assessment of the extent to which they can be said to be outcomes of contact. (This does not exhaust the store of recurrent syntactic similarities, for which see Mesthrie 2004a.)

Contact and African Englishes  393

4.1.1  Use of Resumptive Pronouns As with many varieties world‐wide (including informal L1 English) resumptive pronouns surface in relative clauses, where Standard English uses a “gap”: (2) a. There is our glue which we are getting them near. (East Africa – Schmied 2004: 934) b. The other teacher that we were teaching English with her went away. (Cameroon – Mbangwana 2004: 906) c. The old woman who I gave her the money […] (Ghana – Huber and Dako 2004: 858) Resumptive pronouns have not been reported in subject position in any variety of SSE. Sentences like (3) below are unattested in a wide variety of sources consulted: (3)

? This is a man that he loves dogs.

The occurrence of resumptive pronouns in SSE illustrates the difficulties in deciding between contact and other factors in L2 dialect emergence. As resumptive pronouns are abundant in Bantu languages there is indeed a case for substrate influence. However, general L2 processing might well account for the phenomenon, since the use of a pronoun is a more transparent form of syntax over a gap. In Zulu (an example of an influential substrate for BSAE), relative clauses do not allow a “gap.”5 They have to be filled by a preposition plus resumptive pronoun like na‐ye ‘with him/her’, wa‐khe ‘of him/her’, and nga‐lo ‘by means of this’ (in non‐subject and direct object functions). The direct object and subject NPs in the relative clause take the usual subject or object concords. Although these concord markers are less salient than the free pronouns exemplified on p. 396, they do exemplify the absence of a gap (and can be considered the equivalents of bound pronouns). Subject concords occur in all clauses and cannot be dropped (e.g. umama ucula iculo ‘mother sings a song’ – the first u is the noun class marker, the second u is its subject concord). Object concords occur mainly for focusing effects, with umama uyicula iculo translating into something like ‘mother sings it, the song’). The object concord yi in bold is dropped in sentences unmarked for focus. It is not possible to use a free pronoun with subject relative NPs (mother who‐SC sings but not *mother who‐SC she sings’ (where SC is the subject concord u; and ‘she’ the free pronoun yena). The same consideration holds for object case relative NPs. Substrate influence thus explains why resumptive pronouns are common in BSAE in oblique cases other than direct object. It also suggests why they do not occur in subject position. However, they do not cover the direct object case, since BSAE allows a resumptive pronoun here (though infrequently), while the substrate has no free pronoun in this function. It might be countered that the object concord marker is a sort of equivalent of a bound pronoun: however its occurrence is for special focusing, rather than in unmarked sentences. So in coming to a full understanding of substrate influences, pragmatic functions in both substrate and L2 variety have to factored in. The psycholinguistic processing model appears to be relevant here, since the hierarchy posited by Keenan and Comrie (1977) suggests that it is a typological universal that resumptive pronouns appear with the oblique cases (indirect object > genitive > direct object) rather than the easiest position of subject. (Sentences 2b and 2c above exemplify resumptives with indirect object and sentence 2a with direct objects.) The avoidance of resumptives might also be related to another internal factor in SSE: the predilection for left dislocation. It is this that we turn to next.

4.1.2  Left Dislocation This construction isolates a “topic” and follows it up by a “comment,” which in the grammatical sense is a full sentence, containing a copy or appositional pronoun relating to

394  Rajend Mesthrie the topic NP. In sentence (4a) below the topic and comment are separated by a hyphen, and the copy pronoun highlighted: (4) a. The students – they are demonstrating again. (Nigeria – Alo and Mesthrie 2004: 823) b. That woman – she cheated me. (Ghana – Huber and Dako 2004: 862) c. The people – they got nothing to eat. (South Africa – Mesthrie 1997: 127) Such copy pronouns typically occur in subject positions in the comment, but other positions on the noun phrase hierarchy are also possible, as in (5): (5)

Q: Where did you learn Tswana? A: Tswana, I learnt it in Pretoria. (South Africa – Mesthrie 1997: 127)

As with resumptive pronouns, several difficulties surround any one explanation for the occurrence of left dislocation. A superstratist might aver that this is simply a continuation of a construction common in informal English that carries pragmatic effects of contrast, itemizing NPs from a list or reintroducing given information after a stretch of discourse (see Prince 1981; Finegan and Besnier 1989). However, my experience of listening to decades of English suggest that some varieties use the strategy far more frequently than others: notably Black South African English and Indian South African English (Mesthrie 1992, 1997). In some instances BSAE comes close to grammaticalizing the use of copy pronouns, especially with noun phrases like the people or with complex noun phrases (or both as in (6) below). (6) The people who are essentially born in Soweto – they can speak Tsotsi. (South Africa – Mesthrie 1997: 132). The fact that complex noun phrases induce copy pronouns might suggest a processing explanation: copy pronouns help keep track of the salient topic noun phrase better than an S–V–O syntax might, especially if the subject is a complex NP. A third explanation does involve contact. When a Bantu language like Zulu needs to express special effects like focusing or contrast, a free pronoun form comes into play. An unmarked declarative sentence like ‘the cattle died’ is Izi‐nkomo za‐fa, with subject concord marker zi ­(assimilated with the past tense vowel to form za) and no absolute (or free) pronoun. In a (marked) contrastive clause like the following, such a pronoun does surface (Taljaard and Bosch 1988: 78): (7)

Izi‐mvu za‐sinda kodwa zi‐nkomo Cl10-sheep SC-survive but cl10-cattle ‘The sheep survived, but the cattle died.’

zona SC

za‐fa PAST-die

The BSAE equivalent sentences are thus likely to be substrate‐induced, since the form in Zulu is indeed a pronoun and not – say – a focusing particle. However, the pronoun form is at a pinch a possibility in equivalent sentences in the superstrate too. Although one does not generally say “The sheep survived but the cattle – they died,” it is not wildly ungrammatical to do so. With special intonation and pause structure within a certain discourse context, it is a possibility.

Contact and African Englishes  395

4.1.3  Stative and Habitual Forms of Be + ‐ing In most New Englishes the distinction between present progressive forms and stative/ habitual forms is more fluid than in Standard English, as in sentences (8a)–(8c): (8)

a. … it produces a lot of smoke … heavy smoke and it is smelling … (East Africa – Schmied: 2004: 930) b. The rural areas are not having access to higher education. (Ghana – Huber and Dako 2004: 855) c. People who are having time for their children … (South Africa – Mesthrie 2004b: 963)

In Standard English there is a somewhat rigid difference between stative and nonstative verbs, especially in their acceptance of the progressive: know, understand, love, like, have (it) but not (generally) *am knowing, understanding, loving, liking, having (it). Leketi Makalela (2004) notes that habitual and progressive are not usually differentiated in Sepedi (aka Pedi or North Sotho). This would explain why the distinction between habitual and progressive senses of the same verb (I work vs. I am working) is overridden in early interlanguage English of its speakers. Translations of the form Ndi‐ya‐sebenza in the Bantu language Xhosa (aka isiXhosa) cover both possibilities. And conversely, Ndi‐ya‐kha ‘I am building’ and Ndi‐ya‐zi ‘I  know’ (respectively dynamic versus stative) have the same morphosyntax (‘I  –  long ­present  –  verb’). If one wishes to emphasize the ongoing nature of an activity then the ­“persistative” (or “persistive”6 – see Nurse 2008) prefix sa would be used (‘am still working’, ‘am working right now’). Bertus Van Rooy has devoted the most attention to this phenomenon in BSAE from a synchronic and diachronic perspective. Van Rooy (2014) proposes a general “extended timeframe semantics” missed in earlier accounts. While acknowledging possible conventionally progressive meanings and the extension to statives in individual instances, he constructs a more general semantic characterization for the data. The BSAE data suggest a different prototype meaning from the standard, which has ‘extended duration’ at its core. This meaning is related to the persistative aspect in Bantu languages as follows (Van Rooy 2014: 160). The persistive, sometimes called progressive in the literature on the Zone S languages, is formed by the verbal prefix ‐sa‐ in the Sotho and Nguni groups of languages in South Africa. Louwrens (1996) argues that the ‐sa‐ prefix in any case does not encode an ordinary progressive, because the meaning of an on‐going activity is not salient. Rather, the possibility that an event or state may end at some future point, and is not unlimited, is profiled. However, the period prior to the endpoint need not be short, as is typically the case in Inner Circle varieties of English.

In a subsequent diachronic study Van Rooy and Piotrowska (2015) show use of such extended timeframe usage in the earliest phase of BSAE in the nineteenth century. Van Rooy and Kruger (2016) reinforce the persistative reading by showing that while certain features of BSAE disappear with increased proficiency, this is not true of the extended timeframe semantics of be + ‐ing. Although there is thus substantial support for substrate influence, some superstratal factors also play a small role. One factor is that the superstratal differentiation of the categories, rather than being transparent to new learners, is variable and subject to change. Hear for example is a stative verb, that may allow be + ‐ing in certain contexts: What I’m hearing is that people in the room are unhappy is a possible sentence of Standard English (whose own norms regarding be + ‐ing are slightly in flux). Another example concerns subtle variants like

396  Rajend Mesthrie I love this song versus I’m loving every minute of this song. Likewise, the Macdonald’s slogan of not so long ago, I’m loving it was slightly opaque to my students of L1 and L2 English background, who could not see the subtle grammatical play (between stative and non‐ stative). Thus different explanations hold for ‐ing variation: language contact (as the main force), complemented by L2 processing, and variable (if not misleading) signals from the superstrate.

4.1.4  Pronoun Gender Conflation The occasional conflation of he and she is reported for all the varieties studied here, except Cameroon. (9) He is called Mary. (Ghana – Huber and Dako 2004: 859) Huber and Dako mention a trend towards gender selection of the pronoun according to the head noun: (10) a. He was looking for her aunt. (Ghana – Huber and Dako 2004: 859; std Eng: … his aunt) b. She thought his husband had traveled. (Ghana – Huber and Dako 2004: 859; std Eng: … her husband) There is a strong case for substrate influence here; Bantu languages do not differentiate sex gender with pronouns at all. They do differentiate a variety of pronoun forms in concord with the class of the noun they refer to. But this does not apply to humans, which prototypically occur within the same noun class (1 and 2). Thus the Xhosa pronoun equivalent to the inanimate it of English comes from a large set of forms determined by class affiliation of the equivalent noun: wona (class 3), lona (class 5), sona (class 7), yona (class 9), and bona (class 14). On the other hand the equivalents of third person human forms he and she (and him and her) are yena alone. So the case for substrate influence is strong. One case study (Mesthrie 1999) enables us to evaluate how this influence operates from a developmental perspective using the elementary English of Xhosa working‐class speakers in Cape Town. In early interlanguage data he is the basic form for humans (male or female); it is always used for nonhumans. This accords closely with the substrate. She emerges in subsequent acquisition for female humans, with some hypercorrection (she for males but not for nonhumans). The full pattern is eventually acquired, although some “backsliding” occurs. Up till recently even university professors in my observation may sometimes under duress or extreme relaxation occasionally use he for she or vice versa. All of this shows a significant influence of the substrate Xhosa grammar. The pattern identified for early English interlanguage of Xhosa speakers does not appear to be the same as for children’s L1 acquisition of English, where no gender variation has been reported (see Huxley 1970: 151). That substrate influence even in early interlanguage is not total can be seen in realizations of number. Speakers in Mesthrie’s (1999) database occasionally conflated singular and plural (he for they), in sharp contrast to Xhosa grammar which categorically differentiates pronouns by number. To end this subsection, the general point can be made that SSE syntax shows the mutual influences of substrate, superstrate, and processing universals.

4.2 Anti‐deletions Given the search in modern linguistics for general principles (and parameters) of syntactic organization, the effects of contact should ideally be meaningfully integrated within an

Contact and African Englishes  397 unfolding grammar. In this section I propose an approach that integrates a large proportion of the syntactic features of BSAE with a general syntactic property that I term “anti‐deletion” (based on Mesthrie 2006). In this formulation, the syntax of mesolectal BSAE accords largely with that of L1 English in respect of underlying patterns and structures. However, it differs on the surface in disfavoring zero elements in phrases and clauses. The first property that I  identify is an “undeletion”  –  a tendency to restore elements that by almost all syntactic accounts involve a gap or deletion in Standard English. Here are some brief examples with discussion: further details can be found in Mesthrie (2006). The examples were selected from mesolectal speakers of BSAE. (11)

a. Come what may come. (‘Come what may Ø.’) b. He made me to do it. (‘He made me Ø do it.’) c. The fact has made me to conclude that my idea is sound. (‘… made me Ø conclude …’) d. So she was warning us that, “You’d better learn this language because, like, you’re going to Cape Town.” (that for Ø) e. They’ll just tell you that, “We have been using Fanakalo.” (that for Ø) f. As you know that they are from the Ciskei. (that for Ø) g. As I made it clear before, I am going to talk about solutions, not problems. (‘As I made Ø clear …’) h. As it is the case elsewhere in Africa, much can still be done for children. (‘As Ø is the case …’)

These examples show that in common with other SSE varieties BSAE speakers disfavor the deletion of elements like infinitive to (11b), dummy it after verbs in constructions like make clear (11g), dummy it before the verb be (11h), complementizer that before direct speech (11d) and (11f), and adjunct that after as you know, as I said etc. (11f). Resumptive pronouns in relative clauses also qualify as undeletions but are not exemplified here, as they have been discussed in Section 4.1.1. Because of these properties I suggested the following principle: Principle 1: If a grammatical feature can be deleted in Standard English, it can be undeleted in BSAE mesolect.

Reference to Standard English in the formulation acknowledges its role as the form of the superstrate aimed at (but not always achieved) in “macro‐acquisition.” A second category concerns nondeletions in the standard dialect: i.e. places where the grammar of English dialects seems to readily permit deletion, but not the standard language. This category includes well‐known features like pro‐drop, copula deletion, ellipsis, gapping, and lack of do‐support. As none of these tendencies exists in BSAE, they would not ordinarily be thought of as features of the variety. Yet they fit the anti‐deletions tendency very well, and suggest a second principle: Principle 2: If a grammatical feature can’t be deleted in Standard English, it almost always can’t be deleted in BSAE mesolect.

The last property I describe is insertion (the opposite of deletion). Looking at the remaining features described by linguists working on BSAE (Gough 1996; De Klerk 2005; Makalela 2004), I found very few instances of deletion or permutation of elements at the syntactic and morphological level. By contrast, a large majority of features involved the insertion of

398  Rajend Mesthrie a morpheme, where Standard English has none. This pertains to constructions like the following: 1. the double marking of clauses, well known in the New Englishes of Asia and Africa (sentence 12a); 2. the use of can be able for ‘can’ (sentence 12b); 3. the frequent use of that one for anaphoric that (sentence 12c); 4. the existence of occasional double conjunctions like supposing if ‘if’; unless if ‘unless’; because why ‘because’; double comparatives like more better; and occasional double negatives; 5. the presence of “underlying” prepositions with verbs like mention (about), discuss (about), voice (out). (12) a. Although I’m not that shy, but it’s hard for me to make friends. b. … how am I going to construct a good sentence so as this person can be able to hear me clearly … c. Q.: Do you know more Xhosa or Zulu? A.: That one I can’t tell you. These insertions lead to a last principle: Principle 3: If X is a grammatical feature of BSAE mesolect that is not covered by Principles 1 and 2, then X almost always involves the presence of a morpheme not found in Standard English.

I show further in Mesthrie (2006) that none of these principles characterize the other major English varieties of South Africa, and that of all the constructions mentioned in this section only two occur in another variety. These are left dislocation and occasional resumptive pronoun use in South African Indian English, but not to any great extent in any other variety of the country. I proposed that The “anti‐deletion” concept and the more specific “undeletion” will, I believe, help characterise other New Englishes too, especially other varieties in Africa … Mesolectal BSAE shows a pendulum swing between L1 tendencies that favour undeleting and the influence of settings required by the standard form of the TL. (Mesthrie 2006: 142)

In terms of the investigation of language contact, anti‐deletions can be attributed to the nature of the substrata, in which deletions are rare (see van der Spuy 1997 for Zulu; du Plessis and Visser 1992 for Xhosa). However, a closer comparison has still to be undertaken. Andrew van der Spuy, a specialist in Zulu syntax, is of the opinion (personal communication, ca. 2009) that many – but not all – of the properties would be supported by a contrastive analysis with Zulu.

5 Conclusion African Englishes are clearly very important varieties in characterizing language contact. They have stabilized in situations where native speakers have become increasingly rare (except in South Africa), and remain the most important media in pan‐African communication (notwithstanding the importance of Swahili, Arabic, French, and Portuguese). In this chapter I have tried to characterize the recurrent similarities in Sub‐Saharan English, and to examine the relation of substrates, superstrates, and processing universals to each other. As

Contact and African Englishes  399 Mufwene (1986) argued convincingly for another set of varieties (pidgins and creoles), all three spheres of linguistic influence play a complementary role in molding a subset of New Englishes of the world.

Acknowledgment I am grateful to the NRF (SARCHI Chair grant no. 64805) for funds that enabled this research and the subsequent writing of this chapter.

NOTES 1 Clements (2000: 140) goes further in defining the structure of prefixes as (C)V; roots as CV(N)C, derivational elements as ‐VC‐ and the obligatory final vowel suffix ‐V. 2 The data was collected over a three‐day period in June 2008 from the early morning news program. 3 Subject to the discussion on p. 388 that monophthongal tokens (mainly) of the TRAP set often take [e] or [ɛ]… 4 And probably higher for speakers of English as an L2… 5 In this section I use examples from either Zulu or Xhosa – two closely related and mutually intelligible Bantu languages of the Nguni subgroup, which are the two most numerous languages of the country in terms of native speakers. 6 The usual term in Bantu linguistics is persistive – I prefer the additional morpheme in line with adjectives like resultative and consultative. There is one other spelling in the literature – persistitive.

REFERENCES Alo, M. A. and Rajend Mesthrie 2004. Nigerian English: morphology and syntax. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2. Berlin: de Gruyter Mouton, pp. 813–27. Anchimbe, Eric A. 2006. Socio‐pragmatic constraints to native or indigenous language education in Cameroon. In Olaoba F. Arasanyin and Michael A. Pemberton (eds.), Selected Proceedings of the 36th Annual Conference on African Linguistics]. Somerville, MA: Cascadilla Proceedings Project, pp. 133–41. Bokamba, Eyamba G. 1992. The Africanization of English. In Braj B. Kachru (ed.), The Other Tongue: English across Cultures. Urbana: University of Illinois Press, pp. 125–47. Bokamba, Eyamba G. 2011. Ukolonia in African language policies and practices. In Eyamba G. Bokamba et al. Selected Proceedings of the 40th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project, pp. 146–67.

Brutt‐Griffler, Janina 2000. World English: A Study of its Development. Clevedon: Multilingual Matters. Clements, George N. 2000. Phonology. In Bernd Heine and Derek Nurse (eds.), African Languages: An Introduction. Cambridge: Cambridge University Press, pp. 123–60. Davy, Jim 2000. A conservative view of the New Englishes. Paper presented at the First International Conference on Linguistics in Southern Africa, 12–14 January. University of Cape Town. DeGraff, Michele 2003. Against Creole exceptionalism. Language 79.2: 391– 410. De Klerk, Vivian 2005. Procedural meanings of well in a corpus of Xhosa English. Journal of Pragmatics 37.8: 1183–205. Dryden, John 1693. The Art of Satire. London. Du Plessis, Jan Adriaan and Marianna Visser 1992. Xhosa Syntax. Pretoria: Via Afrika. Finegan, Edward and Niko Besnier 1989. Language: Its Structure and Use. Orlando: Harcourt, Brace, Jovanovich.

400  Rajend Mesthrie Gough, David 1996. Black English in South Africa. In Vivian de Klerk (ed.), Focus on South Africa. Amsterdam: Benjamins, pp. 53–77. Gut, Ulrike 2004. Nigerian English: phonology. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 1. Berlin: de Gruyter Mouton, pp. 813–30. Hakluyt, Richard 1598–1600. The Principal Navigations, Voyages, Traffiques and Discoveries of the English Nation. London: Dent and Sons (1927 reprint). Hancock, Ian and Piayon E. Kobbah 1975. Liberian English of Cape Palmas. In Joey L. Dillard (ed.), Perspectives on Black English. The Hague: de Gruyter Mouton, pp. 248–71. Harris, John 1996. On the trail of short u. English World‐Wide 17: 1–40. Huber, Magnus. 2012. Regional profile: Africa. In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Mouton World Atlas of Variation in English. Berlin: de Gruyter Mouton, pp. 806–23. Huber, Magnus and Kari Dako 2004. Ghanaian English: morphology and syntax. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2. Berlin: de Gruyter Mouton, pp. 854–65. Hundleby, C. E. 1963. Xhosa‐English pronunciation in the South‐East Cape (Rhodes University Ph.D. dissertation). Huxley, Renira 1970. The development of the correct use of subject personal pronouns in two children. In Giovanni B. Flores d’Arcais and Willem J. M. Levelt (eds.), Advances in Psycholinguistics. Amsterdam: North Holland, pp. 141–65. Keenan, Edward and Bernard Comrie 1977. Noun phrase accessibility and Universal Grammar. Linguistic Inquiry 8: 63–99. Kortmann, Bernd and Kerstin Lunkenheimer (eds.) 2012. The Mouton World Atlas of Variation in English. Berlin: de Gruyter Mouton. Kortmann, Bernd, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.) 2004. A Handbook of Varieties of English, 2 vols. Berlin: de Gruyter Mouton de Gruyter. Magura, Benjamin 1984. Style and meaning in African English: a sociolinguistic analysis of South African and Zimbabwean English (University of Illinois (Urbana‐Champaign) Ph.D. dissertation). Mair, Christian 2003. Kreolismen und verbales Identitätsmanagement in geschreibenen

jamaikanischen Englisch. In Elisabeth Vogel, Antonia Napp, and Wolfram Lutterer (eds.), Zwischen Ausgrenzung und Hybridisierung. Würzburg: Ergon, pp. 79–96. Makalela, Leketi 2004. Making sense of BSAE for linguistic democracy in South Africa. World Englishes 23.3: 355–66. Mbangwana, Paul 2004. Cameroon English: morphology and syntax. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2. Berlin: de Gruyter Mouton, pp. 898–908. Mesthrie, Rajend 1992. English in Language Shift: The History, Structure and Sociolinguistics of South African Indian English. Cambridge: Cambridge University Press. Mesthrie, Rajend 1996. Imagint excusations: missionary English in the nineteenth century Cape Colony, South Africa. World Englishes 15: 139–57. Mesthrie, Rajend 1997. A sociolinguistic study of topicalisation phenomena in Black South African English. In Edgar W. Schneider (ed.), Englishes around the World (Studies in Honour of Manfred Görlach), vol. 2: Caribbean, Africa, Asia, Australasia. Amsterdam: Benjamins, pp. 119–40. Mesthrie, Rajend 1999. Male workers’ English in the western Cape: interlanguage, code‐ switching, pidginisation. Special issue of Indian Journal of Applied Linguistics 25.1–2: 151–74. Mesthrie, Rajend 2003. The World Englishes paradigm and contact linguistics –  refurbishing the foundations. World Englishes 22.4: 449–62. Mesthrie, Rajend 2004a. Synopsis: morphological and syntactic variation in Africa and south and southeast Asia. In Bernd Kortmann et al. (eds.), vol. II, pp. 1132–41. Mesthrie, Rajend 2004b. Black South African English: morphology and syntax. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2. Berlin: de Gruyter Mouton, pp. 962–73. Mesthrie, Rajend 2005. Putting back the horse before the cart: the “spelling form” fallacy in second language acquisition studies, with special reference to the treatment of unstressed vowels in Black South African English. English World‐Wide 26.2: 127–52. Mesthrie, Rajend 2006. Anti‐deletions in a second language variety: a study of Black South

Contact and African Englishes  401 African English mesolect. English World‐Wide 27.2: 111–46. Mufwene, Salikoko S. 1986. The universalist and substrate hypotheses complement one another. In Pieter Muysken and Norval Smith (eds.), Substrata versus Universals in Creole Genesis. Amsterdam: Benjamins, pp. 129–62. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Nurse, Derek. 2008. Tense and Aspect in Bantu. Oxford: Oxford University Press. Platt, John, Heidi Weber, and Mian L. Ho 1984. The New Englishes. London: Routledge. Prince, Ellen F. 1981. Topicalisation, focus‐ movement and Yiddish‐movement: a pragmatic differentiation. Proceedings of the Seventh Meeting of the Berkeley Linguistics Society: 249–64. Sand, Andrea 2004. Shared morphosyntactic features of contact varieties: article use. World Englishes 23.2: 281–98. Schmied, Josef 1991. English in Africa: An Introduction. London: Longman. Schmied, Josef 2004. East African English (Kenya, Uganda, Tanzania): phonology. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 1. Berlin: Gruyter, pp. 918–30. Sey, K. A. 1973. Ghanaian English. London: Macmillan. Simo Bobda, Augustin 2000a. English pronunciation in sub‐Saharan Africa as illustrated by the NURSE vowel. English Today 16.4: 41–8. Simo Bobda, Augustin 2000b. Research on New Englishes: a critical review of some findings with a focus on Cameroon. Arbeiten aus Anglistik und Amerikanistik 25: 53–70. Simo Bobda, Augustin 2001. East and southern African English accents. World Englishes 20.3: 269–84. Simo Bobda, Augustin 2003. The formation of regional and national features in African English pronunciation: an exploration of some non‐interference factors. English World‐Wide 24.1: 17–42. Simo Bobda, Augustin 2004. Cameroon English: phonology. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 1. Berlin: de Gruyter Mouton, pp. 885–901. Singler, John 2004. Liberian Settler English: phonology. In Bernd Kortmann, Edgar W.

Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 1. Berlin: de Gruyter Mouton, pp. 874–84. Spencer, John 1971. West Africa and the English language. In John Spencer (ed.), The English Language in West Africa. London: Longman, pp. 1–34. Szmrecsanyi, Benedikt and Bernd Kortmann 2009. The morphosyntax of varieties of English world‐wide: a quantitative perspective. Lingua 119.11: 1643–63. Taljaard, P. C. and Sonja E. Bosch 1988. Handbook of Isizulu. Pretoria: J. L. van Schaik. Todd, Loreto 1982. The English language in West Africa. In Richard W. Bailey and Manfred Görlach (eds.), English as a World Language. Ann Arbor: University of Michigan Press, pp. 281–305. Van der Spuy, Andrew 1997. Syntactic traces in Zulu. In Robert K. Herbert (ed.), African Linguistics at the Crossroads: Papers from Kwaluseni. Cologne: Rüdiger Köppe Verlag, pp. 297–310. Van Rooy, Bertus 2004. Black South African English: phonology. In Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 1. Berlin: de Gruyter Mouton, pp. 943–52. Van Rooy, Bertus. 2014. Progressive aspect and stative verbs in Outer Circle varieties. World Englishes 33.2: 157–72. Van Rooy, Bertus and Haidee Kruger. 2016. The innovative progressive aspect of Black South African English. International Journal of Learner Corpus Research 2.2: 205–28. Van Rooy, Bertus and Caroline Piotrowska. 2015. The development of an extended time period meaning of the progressive in Black South African English. In Peter Collins (ed.), Grammatical Change in English World‐Wide. Amsterdam: Benjamins. Van Rooy, Bertus and Gerhard B. Van Huyssteen 2000. The vowels of Black South African English: current knowledge and future prospects. South African Journal of Linguistics, Supplement 38: 35–54. Visser, Fredericus T. 1963–73. An Historical Syntax of the English Language, 4 vols. Leiden: Brill. Wells, John C. 1982 Accents of English, 3 vols. Cambridge: Cambridge University Press. Williams, Jessica 1987. Non‐native varieties of English: a special case of language acquisition. English World‐Wide 8: 161–99.

20 Contact and Caribbean Creoles  DGAR W. SCHNEIDER AND E RAYMOND HICKEY

1 Introduction The Caribbean, defined here as the chain of Caribbean islands and the adjacent mainland of Central (cf. Holm 1983) and northern South America, has been a region marked by linguistic and cultural contact to an exceptional degree (Map 20.1). Sadly enough, the original indigenous population, native Americans of the Arawak and Carib tribes who had themselves moved to the region from South America, played practically no role in the early contact situations. They were small in numbers anyhow, and in practically all cases did not survive for long the challenge posed by European invaders and settlers, i.e. both the Europeans’ military superiority and the effect of germs against which the indigenous peoples had no resistance. Thus, contact in the Caribbean in its early phases was unusual because to all intents and ­purposes it started from a tabula rasa‐like situation. It mainly involved two broadly defined, ethnically distinct population groups between whom the distribution of power was exceptionally unequal: Europeans and Africans. European powers active in the Caribbean were basically the main seafaring peoples of the colonial period: originally the “discoverers” of the “New World,” the Spaniards, soon ­followed by the Dutch, the French, and, considerably later, the British (in the early seventeenth century). The region was attractive for its agricultural potential, especially from the mid-seventeenth century onward when the sugar industry promised immense profits. The Europeans established plantation economies on the Caribbean islands, settled by a small stratum of adventurous society leaders. These were attracted by the desire for power and profits, as was a larger number of laborers who either sought a better fortune which would have been out of their reach at home or who were shipped off involuntarily – as prisoners, indentured laborers, debtors, rebels, prostitutes, and so on (Holm 1994: 338). Politically speaking, the Caribbean has had an amazingly checkered history, with many islands changing hands repeatedly, either by force or through agreements made in any of the peace treaties of far‐away Europe throughout these centuries, with several colonial powers competing for the wealth that the region offered. This competition took many forms – from settlement initiatives and political claims of sorts via open warfare to diplomacy, and, last but not least, the illegal but common and effective tactics of buccaneering. Conversely, Africans, who soon came in much larger numbers, had no opportunities or choices. Millions of Africans from various West African coastal regions were enslaved and forcibly transported to the Caribbean to work the plantations and to satisfy the ever‐ increasing demand for cheap and exploitable labor, especially in the sugar industry. This

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

404  Edgar W. Schneider and Raymond Hickey North America

The Caribbean

The Bahamas

Turks and Caicos

Cuba Cayman Islands

British Virgin Is. U.S. Virgin Is. Anguilla Antigua and St. Kitts and Nevis Barbuda Montserrat Guadeloupe Dominica Martinique St. Lucia Barbados St. Vincent Nether lands Antilles Grenada

Dominican Republic

Haiti Jamaica Anglophone settlement after mid 17th century

Mexico Aruba

Puerto Rico

Trinidad and Tobago

First English settlers arrive in 1620s

South America

Map 20.1  Main islands of the Caribbean.

most deplorable chapter of human history generated a wide range of complex contact situations, in many different localities and involving an array of different peoples. Like the Europeans, the West African ancestors of much of today’s Caribbean population represented a range of peoples, cultural traditions, and languages. All situations, however, were marked by extremely unequal power distributions between White planters and their support staff on the one hand and Black slaves on the other. Numerically, the balance shifted in the course of time, with the Africans growing in proportion and constituting a strong majority at most localities after a certain point in time. The result of this situation is well known, and fairly well documented: creolization. Throughout the Caribbean we find creoles spoken by the majority of the populations, mixed languages with input from both the European side (usually predominantly in vocabulary; cf. Allsopp 1996; Cassidy and Le Page 1980) and from African sources (arguably most effective on the grammatical level, though, as we will see, the exact amount of this impact has been under dispute). Reflecting the positions of their speakers in the social hierarchies involved, these two components have conventionally been labeled “superstrate” and “substrate” input, respectively. While the fact that contact and mixing occurred and have fundamentally determined the language situation of the Caribbean is uncontroversial, in detail conditions varied from one locality and polity to another, and so we find a wide range of contact results, from “deep” or “radical” via intermediate or mesolectal to “light” creoles. These varieties in turn can be accounted for as products of either postcolonial approximations to the standard varieties in their vicinities, or as different degrees of restructuring in the first place. How the interrelationship between these has to be understood is not always clear and is sometimes controversial, and it definitely may be frustratingly complex. For example, in the two‐island state of Trinidad and Tobago, the most creole and “deviant” speech forms of Trinidad correspond roughly to what is only mesolectal on Tobago, where due to historical reasons an even deeper form of creole is widely available. The notion of “Caribbean English Creoles” (CEC) is a useful and widely established cover term for these varieties, though it should be conceptualized (and understood in the present context) to encompass the entire range of English‐related varieties found in the region.

Contact and Caribbean Creoles  405

2  Historical Background In the sixteenth century Spain practically held a monopoly on trade and was the predominant regional power, but the French, the Dutch, and then also England began to challenge this position. The motives for expanding into the region were in fact different: while Spain was driven by a curious mixture of desires for exploitation but also conversion of native populations, the primary interest of the Dutch and also others was trade. England’s activities soon turned to settlement, however, in the Caribbean as in North America: the New World promised land and work, while England in the seventeenth century suffered from a shortage of the former and a surplus of men (Parry and Sherlock 1981: 52). St. Kitts was the first British settlement in the Caribbean (1624), and an important early point of dispersal (Baker 1998); but others followed soon, and in rapid succession: Barbados (where Robert Powell, on his way back from Brazil, took formal possession in the name of the King in 1624 and then persuaded London merchants to launch a settlement initiative), St. Croix (1625, shared with the Dutch), Tobago (1625 saw an early but at that point unsuccessful settlement attempt, the first in a long series of varying ownership claims), Nevis (settled in 1628, and providing the earliest example of the common Caribbean pattern of internal migration, with settlers having come from St. Kitts), Antigua and Montserrat in 1632, and so on. In the early decades smallholders and indentured servants produced primarily maize and tobacco for the European market, under difficult conditions, struggling for patronage and to secure land titles in London, more often than not fighting competing investors. African slaves, the first ones imported by the Spanish a century earlier, were around, and their numbers grew, but they were still a definite minority, and so they had more exposure to, and better opportunities for, gradually adjusting to the speech forms used by Whites. However, around 1650 things changed drastically, and revolutionized the economic basis and the racial composition of the islands. The production of sugar, introduced about a decade earlier, brought immense profits – but unlike tobacco growing it requires factories to produce it, thus huge capital and large plantations were necessary in order to be successful. This meant a great demand for, and large‐scale importation of, slave labor, and thus within a relatively short period of time a complete reversal of the Black–White demographic proportions. Two locations epitomize this new state of affairs more than others. In Barbados large numbers of small proprietors and tobacco farmers gave up and sold their lands; huge estates grew, and the proportion of Blacks to Whites changed from ca. 6,000 : 40,000 in 1645 to a strong Black majority of 46,000 : 20,000 in 1685 (Parry and Sherlock 1981: 69). Jamaica, conquered by Cromwell’s troops in 1655 and soon to be Britain’s success story in the region, did not evolve as a settlement colony but rather became the prototypical sugar island with a predominantly African‐derived population, with a Black : White ratio of about 15 : 1 (Cassidy 1961: 16; Lalla and D’Costa 1990: 23). But among the Whites were deported Irish such as the 2,000 young Catholics sent by the government of Ireland shortly after the British taking of Jamaica to populate the island with British subjects (Ó Siochrú 2008). Relatively little is known about the exact details of social contacts, and especially interethnic relations, during that period, as contemporary reports tended to focus on commercial aspects and the White perspective. Slaves came from the entire West African coast between today’s Sierra Leone via the Ivory, Slave, and Gold Coasts to the Niger Delta, an area with a complex culture, considerable agricultural expertise, and also a local slavery tradition. The conditions of life for slaves on the sugar plantations were cruel, marked by rigorous discipline and brutal penalties for non‐cooperation with the colonists. The proportion of Europeans was relatively low, and relations between the races were mostly distant to hostile – fear of slave uprisings was a constant reality on the islands. In the eighteenth century there was still a small number of poor Whites around who did not fare much better and interbred with slaves (Parry and Sherlock 1981: 70). Later legislation improved their situation

406  Edgar W. Schneider and Raymond Hickey to avoid an even greater ethnic imbalance (felt to be a threat to safety). The European ­languages, the languages of those in power and as such the established lingua francas, were the common targets of the slaves’ language acquisition, but access to them was highly restricted, so on many islands the conditions for creolization were met. Throughout the eighteenth century slave importation went on continuously and in large numbers, because “slaves in those days were short‐lived” due to ill treatment and epidemics, and “replacement was more economical than rearing slave children” (Parry and Sherlock 1981: 70). As far as we can tell, political turmoil caused by American independence or volatile sugar prices did not much affect life on the plantations (and thus the language contact settings) until slave emancipation in the British Empire in 1834. Petty terror by overseers was probably more of a daily reality than constant terror, and some slaves worked on small plots of their own (and some even managed to become free), so even in terms of social relations and realities there was a cline of options and different practices, but one clearly and decidedly leaning toward the down end from the slaves’ perspective. Emancipation, long prepared by political debates in England and preceded by the official end of the slave trade in 1808, came in 1834, though another four years of “apprenticeship” followed, a period during which slaves had only limited rights, the assumption being that they would need to be prepared for freedom. For slaves freedom brought some new options and removed the constant threat to life and limb, but actually for many of them things did not change much in practice (Alleyne 1980: 186). After all, the lack of skills and/or property did not leave any real choices to most of the Black people, and working the sugar cane was not any easier afterwards. The old plantation system was dissolved only gradually and in parts, but in many places it resurfaced in new but similar forms. One thing that did emerge during this period, however, was the social pattern of small‐scale peasantry, especially in Jamaica, with former slaves who managed to secure a plot of land for themselves becoming smallholders and leading a rural life, relatively isolated from developments in the outside world (Patrick 2008: 127). To some extent this is a form of life that persists to the present day. However, in some countries the end of slavery did have severe consequences for the society’s demographic make‐up and hence the conditions of linguistic contact. The continuing demand for cheap labor led to the immigration of large numbers of indentured workers from India who then remained in their host countries and added a new factor to the contact equation. This applies to Trinidad and Guyana in particular – in these two countries roughly half of today’s population is of Indian ancestry. While these Indians set up their own communities and have upheld many of their social and religious traditions, by now many of them have given up their ancestral languages and shifted to local forms of creole English. In Guyana, the Indo‐Guyanese population is predominantly rural and speaks a basilect, while the Afro‐Guyanese are more urban and tend to speak a mesolectal variety; there have been repeated upsurges of ethnic tensions between both groups. The present‐day “Anglophone” (i.e. mostly CEC‐speaking) Caribbean consists predominantly of island and mainland states with a British colonial history – large ones like Jamaica, Barbados, Trinidad and Tobago, Guyana, the Bahamas, and Belize (former British Honduras), and smaller ones like Antigua, Barbuda, Grenada, St. Vincent, St. Lucia, St. Kitts, Nevis, the US Virgin Islands, the Turks and Caicos Islands, or the Caymans. In addition, there are a few English‐speaking pockets stranded for some reason in regions which are predominantly Spanish‐speaking: in the Bay Islands of Honduras (settled mainly from the Caymans), in Panama (where the construction of the Canal had attracted workers mainly from Jamaica), in Costa Rica (spoken by workers on banana plantations), or along Nicaragua’s Miskito Coast (Holm 1983, 1986). Again, contact conditions vary from one place to another. Ethnic composition has been a decisive factor in this  –  some degree of interbreeding has always been a part of Caribbean life, and traditionally a loose correlation between skin complexion and status has characterized the region’s social stratification.

Contact and Caribbean Creoles  407

3  Contact Parameters Language contacts in the Caribbean in the present context resulted primarily from encounters between Europeans (British people in the present case) and Africans, and secondarily also between Africans of different cultural and linguistic origins. We have very little documentation of the exact nature of such interactions in the early phase (which obviously is strongly relevant for shaping the ultimate linguistic outcome, based on the “founder principle,” Mufwene 1996), but we can make educated guesses, as it were, based on what we know about demographic data and the social structure of these communities in earlier centuries. As is well known, the basic precondition for creolization to occur is a large disproportion between a small number of target language speakers (i.e. Europeans) and a large majority of learners (i.e. African slaves) adopting the language. Certainly this was the case on many plantations, where the social setup was characterized by a preponderance of slave labor, frequently absentee ownership, and the presence of relatively few Whites with managerial functions, mainly as overseers. However, all of these factors primarily apply to sugar plantations so they fail to adequately grasp the phase prior to the sugar period and in locations with other economic bases. Therefore, we have to distinguish between an early “homestead phase” (Mufwene 2001: 34, following Chaudenson 2001), with Whites usually still constituting a majority of the population, and the later plantation phase. As was pointed out on p.  405, in Barbados the transition between these stages occurred around 1650; in Jamaica there never really was a full homestead phase, with the sugar economy dominating pretty much from the beginning of British colonization. In smaller islands the situation may have been more variable. Emphasizing differences between these phases has important consequences for the creolization scenario: the traditional creolist position of basilects having been the original forms of Caribbean creoles and approximations to English resulting from decreolization coming in only later builds upon the conception of large plantations, while a more recent understanding of variability having been there from the beginning assumes that intermediate and less creolized forms grew during the early homestead phase and have been maintained, or at least highly effective and influential, because of the founder effect. As far as the nature of the social relationship between the races is concerned, we are again faced with a stereotypical image which probably covers the most important aspects of reality but which at the same time is in need of some reflection and qualification. The basic pattern was most likely one characterized by distance, resentment and even hostility between the races  –  reflecting the slaves’ legal status as property without rights and indicated by the constant fear of Whites of slave uprisings. Clearly, this is a situation which maximally impedes communication, i.e. it results in very limited and rather superficial language contact and is thus consistent with the creolization scenario. There were exceptions, however, although in numerical terms these probably remained relatively insignificant. Not all slaves were field hands – some were house servants (numerically speaking, in Jamaica one out of eight slaves was a domestic helper, according to Roberts 1988: 8), others were trained as craftsmen, artisans, etc. (Alleyne 1980: 184); and these roles typically would have entailed a wider range of encounters with Whites and thus a higher degree of linguistic accommodation. While these hierarchical layers tended to be rigidly established and enforced, a monolithic conception of the social structure and realities of the early Anglophone Caribbean is certainly inappropriate. It is probably noteworthy that slave emancipation does not seem to have changed social relationships and communicative patterns as much as one might have expected. By that time linguistic conventions had been more or less stabilized in most regions anyhow. Demographic proportions between Blacks and Whites were not greatly affected, and it may be assumed that roughly the same applies to linguistic habits in interethnic encounters. Thus the assumption that the end of slavery would have resulted in greater exposure to anything like

408  Edgar W. Schneider and Raymond Hickey standard English seems misguided. Schooling certainly did have an effect along these lines, but it spread only slowly and not very effectively (and, in practical terms, it really only became more effective on a broader scale during the later twentieth century; Christie 2003: 13). The newly arrived Indian laborers, mostly in Trinidad and Guyana, apparently acquired the local vernaculars from Blacks whom they worked with on the plantations, so they tended to adopt the creoles.

4  Creolization and Decreolization The Caribbean is one of the world’s main regions where creole languages are spoken, so any account of language contact there overlaps to a considerable extent with what we know or think about the emergence and evolution of creoles in general. The dominant theories and concepts of creole studies have typically been developed on the basis of Caribbean languages or have been applied to them. This has been a history of controversies, so we do not stand on firm ground here. However, on zooming in more closely it can be claimed that since the beginnings of creole studies as a recognized academic discipline, roughly in the 1960s, theory formation and creolists’ debates have revolved around the same issues and questions, in various guises: how closely related are the creoles to their European lexifier languages? How much of African origin can be identified in the Caribbean creoles, especially in their grammars? And how can the structural similarities between historically unrelated creoles (with different lexical bases) and their European source languages be explained? To this last question essentially two types of response have been proposed: African substrate effects, or universal principles of human language in general. Universalist theories have tended to suggest that shared features of creoles go back to the effect of deeply rooted principles of language organization or even of human cognition in general, like tendencies toward simplicity or the avoidance of markedness which became activated in situations of natural second language acquisition. The most extreme suggestion along this line of thinking was Bickerton’s “bioprogram” hypothesis (Bickerton 1981), which proposed that preferences for certain structures are genetically encoded because, phylogenetically, they offered a competitive advantage increasing chances of survival to pre‐humans, and that it is only in plantation situations, when children are faced with insufficient linguistic input (which normally overrides the genetic code), that their genetic endowment shines through, as it were, and becomes fully effective. It seems safe to state that Bickerton’s theory, while having met with great interest and having triggered important research, has simply not been confirmed by subsequent findings and is rejected by almost all creolists today. However, the related but much wider question of whether creoles can be identified as such on purely structural grounds is still around and has found another, equally controversial revitalization in McWhorter’s (1998, 2000, 2018) suggestion that the three features of lack of inflectional affixation, of tone, and of derivational noncompositionality mark creoles as relatively young languages (with these properties only emerging with the passage of time in such languages). In any case, it seems clear that the scenario of abrupt creolization by children, on which Bickerton’s theory builds so strongly, fails to adequately describe the realities on Caribbean plantations: As was stated on p. 405, the main strategy to secure the presence of slave labor, e.g. in Jamaica, was continuous importation from Africa rather than raising slave children on location. The number of children on a sugar plantation tended to be very small, certainly insufficient for them to constitute a community of creators of a new language (Roberts 1988: 111). Obviously, the substrate thesis builds strongly upon the idea of contact between African languages and English. Whether or not contact with long‐term linguistic consequences had already taken place in Africa, i.e. whether a pidgin precursor to which Caribbean creoles can

Contact and Caribbean Creoles  409 be traced back had already arisen on African soil, is equally disputed. The strongest proponent of this thesis was Ian Hancock, who suggested that a “Guinea Coast Creole English” might have been the ancestor of these languages (e.g. Hancock 1993). More recently, McWhorter (1997) ventured to locate the major site of early contact and the genesis of a protopidgin specifically at the trading fort of Cormantin on the Gold Coast. The issues involved here are relatively speculative, as the precise historical evidence we have is even scantier than that concerning communication patterns on Caribbean plantations. In Africa, captured slaves were held and “seasoned” for a while, presumably up to a few weeks, in coastal forts until the next ship for the Middle Passage stopped by; and we know about the dreadful conditions and high mortality rates during the Atlantic crossing. All of this sounds not really conducive to even the most rudimentary language acquisition, so we are skeptical with respect to the idea of a pidgin having been imported to the Caribbean. But the question of where CECs ultimately originated seems actually not that important for weighing the African input in general, given that without doubt the African slaves brought their knowledge of their native languages with them into the contact setting. Creolization definitely can have happened on the Caribbean islands without any African‐based prior pidginization, and actually this seems much more likely than it having taken place in Africa before transportation to the Caribbean. On the other hand, the fact that substrate influences did occur and probably played an important role seems all too obvious and difficult to dispute – even if some scholars have radically denied it and the exact amount and importance of African traces is difficult to identify. Early and influential proponents of the substrate view include John Holm and, most notably in his 1980 book which adopted a historical‐comparative perspective and pointed out similarities between Caribbean creoles and African languages, Mervyn Alleyne (also, with respect to cultural traces, Alleyne 1988). Boretzky (1983) is a major scholarly study adopting the same strategy: formal and structural parallels between African and Caribbean language patterns on the levels of phonology and grammar are worked out systematically and in great detail, thus producing a strong case (by means of undogmatic assessments and carefully weighing differing interpretations) for substrate transfer effects especially of Kwa and Mande languages in the emergence of Caribbean creoles. Among the features for which he finds strong and suggestive parallels are the following: “vowel elision in Sranan, similar to what is found in Yoruba and Ewe” (Boretzky 1983: 57–8); “the conflation of progressive and habitual functions in preverbal markers, e.g. in Sranan dε/ε and in Yoruba n‐” (Boretzky 1983: 130–8); “a lexical verb meaning ‘finished’ being grammaticalized as a completive marker, especially in postverbal position, e.g. Sranan kaba and Yoruba tì/tó/tã” (Boretzky 1983: 137–8); etc. More recently, Migge (2003) and Huttar (2008: 217) convincingly argued for substrate effects on Suriname’s creoles (Ndyuka in the second case). Even more detailed cases for substrate influences have been made outside the domain of English‐based creoles (though they may be taken to be indicative of the potential of such effects): Berbice Dutch, a creole of Guyana, has been shown to be very directly related to and motivated by Eastern Ijo. Claire Lefebvre (and others) have for a long time attempted to trace Ewe‐Fon reflexes in Haitian Creole. The most recent assessments of the substrate theory are available in Migge (2007) and in Chapter 8 of Mufwene (2008); see also Singler (1988) and Mufwene (1993). The case of Berbice Dutch is so exceptional and convincing among other reasons because of the homogeneity of the African input in that region. It is known that all the slaves there brought the same African language to the contact situation, something which of course provides for a compact and tightly circumscribed substrate, with relatively few potentially interfering factors. However, in most other situations this does not apply as the majority of slaves on any plantation tended to come from widely different regional and linguistic backgrounds. In fact, a widely cited assumption has been that plantation owners deliberately attempted to mix slaves from different tribal origins and linguistic backgrounds in order to

410  Edgar W. Schneider and Raymond Hickey consciously deprive them of the option of communicating with anyone else on the plantation in their own African languages (to make secret communication and thus insurrections less likely). There is probably some truth in that (and to this extent such a strategy would have had linguistic consequences, weakening direct transfer from any single African language), but it also seems questionable whether or to what extent such a policy was, or could have been, imposed systematically. On the one hand, it is not clear to what extent planters could or would really have cared about such a policy – when they were in need of field hands they probably would have filled this need at the next available auction. Furthermore, the opposite consideration may have played a role as well: slaves from certain regions were specifically sought for special qualities and skills. For instance we know, though the evidence is scant, that some African languages did not die out in the Caribbean: at occasional festivities slaves would have practiced some African customs and used their ancestral languages (though there is no way of telling if these were not rudimentary, formulaic, or frozen patterns in the long run), in particular on Jamaica (see Cassidy 1961: 18; Lalla and D’Costa 1990: 2; Patrick 2004: 407). There, the maroons have been credited with having retained much more of their African roots, including traits of language (though, again, it is not really clear what specific African languages were used for what purposes amongst the maroons; cf. Alleyne 1988: 125–6). Clearly the contact situation must have been such that field hands on Caribbean plantations heard some but relatively little English around them (mainly from overseers, occasionally from other Whites). For new arrivals, the model variety in their language acquisition process was probably not first‐language English as spoken by Whites but rather a contact variety itself, as produced by other Africans who had been around for a longer period of time (Cassidy 1961: 19). In that process of adjusting to their environment and adopting a communicatively useful language form they picked up words and probably also patterns, generalized these through some internal process of linguistic hypothesis‐making, and compensated for whatever structural or other linguistic input was not available to them by filling in patterns, using sounds or trying to use words from their native language knowledge – thus contributing to the available “feature pool” in a given communicative setting and enriching the African component and increasing its linguistic usefulness. And some of these forms and features would have been communicatively successful (either coincidentally or because other Africans around them shared similar underlying linguistic concepts from their own native languages). They were thus strengthened, maintained, and integrated into the local linguistic repertoire of the speech community. It makes sense to assume that the relative strength of the respective input factors (English, African, or whatever), determined by demographic proportions, interaction habits, and other components of the communicative setting, was ultimately decisive in shaping or at least influencing the local linguistic repertoire. Furthermore, it seems clear that a diachronic component is also important to consider here, in line with, though perhaps not exclusively restricted to, Mufwene’s idea of a “founder effect.” During the earliest stages of the formation of a new local speech community (i.e. one big plantation – on which most field hands would have had few if any outside contacts for most of the time) no established norms existed, so the impact of speakers and their features in these situations would have been both stronger and less constrained. In the course of time, however, on the basis of the success or failure of certain linguistic forms, corroborative effects and thus habit formation would have generated a new, local linguistic norm, which in turn stabilized and was strengthened by more and more speakers tuning in to it. Those who came later already found such a norm, and for them it was more a matter of acquiring it or at least adjusting to it in the course of time (though it is clear that these people still contributed to the collection of linguistic forms surrounding them and to the development and the shaping of linguistic habits in the speech community at large).

Contact and Caribbean Creoles  411 It is also noteworthy that this process of creole formation and creole usage operated not only monodirectionally and did not involve African language acquirers only (though this is not to deny the huge disproportions here). We have reports stating that Whites picked up creole as well and used it in appropriate communicative situations (Holm 1994: 334–5). Most evidently this would have characterized the speech of poor Whites and indentured laborers who also worked the fields, especially during the early, “homestead” phase of the seventeenth‐century colonies, but not exclusively so. From the mid eighteenth century we have reports of both “ladies,” who were not sent to England for education, and of children (“for a Boy, thill the Age of Seven or Eioght, diverts himself with the Negroes, acquires ther broken way of talking,” reported from Jamaica in 1740) adopting the “gibberish” spoken by slaves (Mühleisen 2002: 63; see also Cassidy 1961: 21–3). The outcome of this can best be observed today on Barbados, where Blake (2002, 2004) studied Black–White speech relationships and usage and found that poor Whites use language forms which are definitely influenced by and similar to Black people’s creole (or dialect). Such considerations, and a range of studies engendered by similar ideas, have tended to weaken “extreme” positions in recent years and to suggest more moderate compromise hypotheses which tend to deny the “exceptionalism” of creoles (DeGraff 2003) or even the status of creoles as separate languages, implying that they are best considered varieties of their lexifiers, or at least closely related to them (Mufwene 2000, 2001), in contradistinction to an earlier, widely discussed theoretical stance by Thomason and Kaufmann (1988). In general, the recent trend has been to emphasize the contact‐induced character of creoles and thus to put them in line with other varieties and languages which are known to have been strongly shaped by language contact (Thomason 2001; Winford 2003), but to somehow sidestep the issue of their purported status as new languages, only tangentially related to their lexifiers. Winford (2000: 217) suggests that intermediate creoles should best be regarded as products of untutored second language acquisition and language shift, and as restructured forms of English. None of the major controversies in the field has really been solved, but the insight seems to be spreading that extreme positions fail to capture the whole range of observable phenomena. As to the origins question, the conference volume edited by Muysken and Smith (1986) has deepened our understanding of the issues and, to my mind, has made it clear that, in the words of the title of Mufwene’s contribution, “the universalist and substrate hypotheses complement one another” (Mufwene 1986): It would be difficult to deny that substrate forms and patterns are effective, just like superstrate input has been, while “universal” and cognitive factors, together with social factors, apparently contribute to the selection of which of these forms will ultimately be successful in the formative process. Work by Arends (1993) and others has shown that creole evolution proceeds gradually and involves some restructuring by adult speakers rather than in an abrupt fashion and exclusively by children, an observation which largely invalidates the genetic endowment predicted by the “bioprogram” hypothesis. Against the earlier, implied assumption that only deep basilects are “real” creoles it has been shown and argued that creoles come in various “degrees of restructuring”  –  which can nicely be observed when comparing different regions in the Caribbean (Schneider 1990; Neumann‐Holzschuh and Schneider 2000). Many of these observations can be convincingly accounted for in the more recent “uniformitarian” framework suggested by Mufwene (2001, 2008), which assumes that language evolution always operates under contact conditions, that speakers from different linguistic backgrounds bring their respective features to a contact situation, and from this “feature pool” some forms ultimately tend to be successful in an ongoing process of “imperfect replication” of forms in the speech community. These features are selected and reproduced while others die out. Mufwene’s framework claims to be valid for language evolution in general, even if it has been developed primarily by working on creole languages (including Caribbean

412  Edgar W. Schneider and Raymond Hickey ones). Essentially, the claim resulting from this perspective is that creoles and non‐creoles are distinguished not in principle but only by different proportions of substratal versus superstratal features, respectively. And what we find throughout the Caribbean is very clear: ­varieties with variable degrees of distance from their European donor languages (and, conversely, from their African substrates as well). This issue of the “depth” of CECs has not only a synchronic dimension but also an important diachronic side to it, namely the conventional idea of present‐day mesolectal varieties being products of “decreolization.” The traditional assumption in creole studies, still to be found in many handbooks, used to be that the extreme contact situations on newly formed plantations quite abruptly produced maximally deviant basilects, which in the course of time, through increasing exposure to the standard varieties exerting pressure from above, as it were, were “decreolized,” i.e. they adopted standard language features piece by piece, gradually replacing the erstwhile creole features. This results in “post‐creole continua,” with typical examples found in Jamaica or Guyana, where a range of variants stratified from basilectal via mesolectal to acrolectal varieties can be observed. It is important to understand that in this line of thinking all non‐creole forms are thought to be later intrusions into an erstwhile “pure” creole system, which in turn was assumed to have originated at the outset. Intermediate forms of Caribbean creoles which combine typically creole with typically acrolectal features, found very widely, have thus been interpreted as younger, mixed patterns, indicating the ultimate demise of the basilectal creoles. A characteristic example of this line of thinking is the debate on the nature of early Bajan on Barbados. Barbados is noteworthy in that its creole is obviously closer to English than that of Jamaica, for instance. Cassidy (1980) claimed that this situation is a product of decreolization, with early Bajan a full‐fledged creole. In contrast, Hancock (1980) believed that a deep creole had never been spoken on Barbados, due to the island’s special topographic and demographic conditions. The matter was considered an open question until Rickford and Handler (1994) and Fields (1995) produced evidence showing that in earlier texts from Barbados more creole features can be found than today, thus suggesting a sequence of an original creole and a later, more decreolized variety. In contrast, Winford’s (2000) historical survey suggests a much more refined and differentiated social and linguistic scenario that has produced today’s situation without resorting to extreme positions. Note that the evidence produced by Rickford, Handler, and Fields, while documenting that more basilectal forms than today were in use, does not necessarily imply that this is the whole story, i.e. it does not rule out the possibility of variability at that earlier stage as well (as postulated by Blake 2004: 502, and conceded by Rickford and Handler 1994: 228; cf. Winford 2000: 223). The decreolization scenario, as suggested by DeCamp (1971) and Bickerton (1975) and modeled in their “implicational scales” framework, has been questioned increasingly, however. Rickford (1983) doubted whether it adequately accounts for the situation. Similarly, Le Page and Tabouret‐Keller (1985) rejected the model as too simplistic, arguing essentially that the variation to be observed is multidimensional, caused by a complex range of interacting factors, rather than corresponding to the monodirectional post‐creole continuum. Patrick (1999, 2004: 410) considers the theoretical status of the Jamaican mesolect and argues that rather than viewing it as a mixture between basilectal and acrolectal features, it has to be understood as a linguistic system in its own right. Mufwene’s framework, mentioned on p. 407, also denies the validity of what he calls “debasilectalization” and implies that in all varieties there has been graded variability from the outset. Historically speaking, the evidence available clearly shows that mesolectal forms predated the basilectal forms (e.g. past ben, pl. ‐dem, possessive fi, serial verbs, or the pronoun unu, all of which appear relatively late in the record; Patrick 2008: 149–50). More generally, this ties in with the synchronic question of how the intermediate, less strongly creolized varieties, which can be found in many localities, but do not seem to be

Contact and Caribbean Creoles  413 derived from a basilectal forerunner, can be explained. Schneider (1990) showed that with respect to the English‐related varieties of the Caribbean there exists a “cline of creoleness” between “deep creole” varieties like Jamaica and Guyana (and especially the creoles of Suriname, which, however, were not the subject of that paper), via “intermediate” ones like Bajan to “light” varieties which display only a small fraction of the features typically considered characteristic of creoles, as spoken, for instance, on the Cayman Islands or the Bay Islands of Honduras. This line of thinking was continued in works such as Neumann‐ Holzschuh and Schneider (2000), arguing for the existence of “degrees of restructuring” more generally, or Holm (2004), who assumes the existence of “partially restructured” varieties.

5  Products of Contact 5.1  Characteristic Features Especially on the level of morphosyntax the CECs (and related “lighter” varieties) have a number of properties in common that set them off as a special group of language varieties, even if many of them are shared with English‐related creoles elsewhere. Most examples in this section are taken from the contributions to Schneider et al. (2004) and Kortmann et al. (2004). The most distinctive of all “creole features” are the so‐called preverbal markers, short particles placed before the verbs expressing aspectual distinctions rather than tense relationships (as English does, mostly using suffixes). The precise semantic specifications of the meanings of these categories may vary from one variety to another, but basically the forms widely found (with some examples) include anterior (JamC1 ben/men/wen, mesolectal did), progressive (JamC a/da/de; e.g. St. Vincent he mama a call im ‘his mother is calling him’), ­perfective/completive (JamC don, Sranan kaba), habitual (Gullah doz, Eastern Caribbean da/ de, mesolectal doz/is; JamC is assumed to lack a distinctive marker here; e.g. Antigua e de see e breda ‘she sees her brother regularly’), and future/irrealis (JamC go/wi, GuyC go, sa; Rickford 1987). Combinations of these forms are also possible (e.g. JamC Mi ben a ron ‘I was running’, Barbuda she a go sing ‘she will be singing’, Trinidadian bina for past imperfective), and mesolectal varieties have forms recognizably closer to English (like TriC yuuztu for the past habitual). Characteristically, Caribbean creoles have distinct copula forms depending on their functions. Actually, no copula tends to occur before adjectives (as in JamC Him lucky, Trinidad De baby sick), so that, commonly, these are also analyzed as being equivalent to static verbs (and they also take preverbal markers). However, before nouns (in “equative” structures) a special form of the copula is typically found (da/a, e.g. JamC Ebry day da fishing day ‘Every day is a fishing day’). In addition, most creoles have a distinct copula form (mostly de or similar) before locatives (e.g. im de de ‘he is there’; Barbuda we i de ‘where is he/she?’). Finally, Caribbean creoles can introduce sentences with a form of a copula followed by a constituent that is highlighted (and later repeated) in the sentence (e.g. Is tiif dem tiif it ‘It was certainly stolen’, Holm 1988/89: 179). Negation requires no do‐support but is achieved by simply placing a negator in front of the verb, typically no/na or en (probably reduced from ain’t). Object clauses after speech act verbs or mental verbs are commonly introduced by the “complementizer” say/se(h) (e.g. JamC Him swear seh him was going to tell me ‘He swore that he was going to tell me’; TriC he tell me se he na come again ‘He told me that he would not come again’) or, in Suriname, taki (from talk). In comparison with English, the inventory of pronoun forms tends to be reduced. For subject and object functions it is common to use the same pronoun form (e.g. mi ‘I, me’), and

414  Edgar W. Schneider and Raymond Hickey sometimes this includes the possessive as well. Furthermore, some creoles have reduced the gender distinction (though forms like im meaning ‘she’ are found not that widely and primarily in rather deep basilects). A distinct second person plural form unu/wunnah, presumably African‐derived, is also widespread (Hickey 2003). Creoles mostly lack inflectional morphology  –  again, most evidently in basilects. Pluralization may be achieved by forms like ‐dem (e.g. JamC di man‐dem ‘the men’); possessive relations can be expressed by a simple juxtaposing of possessor and possessed or by patterns like Jam fi‐dem (e.g. fi‐me work ‘my job’). Verbs have no third person singular suffix, and usually no suffixes for the past tense either (with time relations being expressed by anterior markers, context, or lexical means). Serial verbs, a sequence of two or more verbs in the same predication, with or without an object interspersed, are another unique kind of construction, e.g. JamC Im tek naif kot me ‘He cut me with a knife’; Sranan mi seri a oso gi en ‘I sold the house to her.’ Finally, a totally different domain of linguistic expression, rarely ever explored, deserves to be mentioned here, namely creole pragmatics: in things like greeting practices, expressions of politeness, the social meanings of seemingly aggressive statements, and some aspects of gestural and nonlinguistic expression (like “suck‐teeth,” a facial expression of strong disapproval) CECs have quite distinct properties which in some respects may ultimately be African in origin. Very little work in this interesting area of language contact effects has been done so far, with a volume edited by Mühleisen and Migge (2005) being the main exception and a solid beginning.

5.2  The Cline of Creoleness Pending the still open question of what it actually means for a language to be “a creole,” i.e. whether creoles can be identified on the basis of their structural properties or only because of their shared sociohistories related to plantation environments, it seems true that many of the Caribbean English‐related creoles are somehow “intermediate creoles” in the above sense. Actually, too much of the discussion has focused on the big and well‐documented varieties of Jamaica, Trinidad, or Guyana, and there is still room and an actual need for further fieldwork, analysis, and documentation. Initiatives like the ones by Aceto and Williams (2003), instigating investigations of “smaller varieties,” Aceto’s own work on such varieties (2003, 2004a, 2004b), or studies like those by Walker and Meyerhoff (2006) on the small island of Bequia are thus most welcome and much needed. The range of different varieties, in terms of their recognizable distance from or proximity to dialectal English, the superstrate variety, is impressive. Suriname’s creoles (cf. Migge 2003; Winford and Migge 2004; Smith and Haabo 2004) are decidedly those most distant from English, wholly corresponding to the notion of full and deep creoles. They have all the features associated with creoles, some lexical influences from African languages and also from Portuguese, and also heavily restructured phonotactics (e.g. bradi, bigi, langa ‘broad’, ‘big’, ‘long’) which renders many phrases difficult to comprehend for an outsider (e.g. A nyan kaba ‘she has eaten’). Jamaica and Guyana also have deep basilects, and both of them are characterized by the continuum situation discussed on p. 412 (to which we hesitate to apply the traditional term “post‐creole”; for an analysis of a literary representation of the Jamaican continuum see Schneider and Wagner 2006). In Guyana the positions on the continuum are ethnically stratified, with Indo‐Guyanese speakers usually being more basilectal than Afro‐Guyanese. The situation is similar in Trinidad and Tobago, where the Tobagonian Creole is more basilectal than the Trinidadian one (e.g. TobC i a teacha vs. TriC hi is teacha; James and Youssef 2002: 156). English‐based creoles can be found on many of the other, smaller islands as well, including Antigua, Barbuda, St. Lucia, St. Vincent, Grenada, the

Contact and Caribbean Creoles  415 Bahamas (Hackert 2004), and so on. As was mentioned on p. 412, Bajan, while clearly also a creole, is somewhat “lighter” in comparison, with more restricted uses of preverbal markers, serial verbs not commonly in use, pronominal distinctions retained to a considerable extent, and also some inflectional endings in use. At the other end of the cline, there are some varieties which look essentially like dialects of English but do have occasional elements typically associated with creoles in use, like preverbal been or done, uninflected nouns or verbs, the complementizer say, etc. Most notably, this applies to the variety in use on St. Eustatius, Saba, Montserrat, the Cayman Islands, and to the Bay Islands of Honduras (where there is a distinction along the cline correlating with skin color; Graham 2005). Work by Wolfram, Reaser, Torbert, and others (e.g. Reaser and Torbert 2004) has shown that on the Bahamas there is also a gradual difference between White and Black speech forms (for example, Black but not White Bahamians vocalize word‐final /‐l/ in words like steal). Finally, there are also pockets of White dialects to be found in the Caribbean, arguably exposed to and influenced by creole structures. For instance, on Anguilla Jeff Williams studied a dialect spoken by people of European descent which has habitual do (it do be hot) or zero possessives (my daddy brother; Aceto and Williams 2003; Williams 2012). This “cline of creoleness” illustrated on p. 414–5 by listing different varieties can also be illustrated by looking at competing but functionally equivalent variants of a single variable, i.e. relatively basilectal, mesolectal, and acrolectal forms corresponding to each other. The first to do this in a post‐creole continuum framework was DeCamp (1971), and by and large we find such stratifications referred to in some of the literature (cf. Schneider 1990; Winford 1997–8). Schneider (1998) worked this out systematically for negative patterns as spoken in various Caribbean varieties. The cline of forms available and sociostylistically stratified (with individual positions on the cline varying by region) ranges from preverbal no/na (e.g. shiy no waan de byebiy ‘she did not want the baby’, Panama) via en/eh (the girl eh know ‘the girl didn’t know’, Trinidad), neva for single events (a neva evn siy it ‘I didn’t even see it’, Belize), doun/don (ai don nuo, Barbados) to upper‐mesolectal didn (it didn useta be that way ‘it didn’t use to be that way’, Gullah) and dialectal English multiple negation (non a di pikni‐dem neba si notn ‘none of the children didn’t see nothing’, Jamaica).

5.3  Dialect Input to the Caribbean The variability in language, which is taken to have characterized the early Anglophone Caribbean as outlined in the previous section, included variable input from dialect speakers of Britain and Ireland who first came in the ’homestead phase’ and largely remained in the region. This would have naturally led to an interaction between these early White settlers and the initial Black slaves, implying that the latter group was exposed to features of language in the speech of the former, features which would be regarded today as dialectal, given that the poor Whites who worked in the Caribbean colonies would have come from diverse regions of Britain and Ireland (Hickey 2004). The view of such scholars as Winford (1997–8: 123) is that creolization is a development which occurred somewhat after the initial settlement of the Caribbean and which was triggered by the establishment of a widespread plantation rural economy, something which was not present at the outset for the region.2 Supportive evidence for this stance is to be found in areas of the Caribbean where plantations were not established, e.g. on the Cayman Islands which retain distinctive traces of English regional input (Holm 1994: 332; Lacoste 2017: 393). The scenario in which approximation to English regional input precedes possible creolization has wide‐ranging implications for the interpretation of key structures in both present‐ day Caribbean creoles and African American English (Schneider 2004). It suggests that the first few generations were exposed sufficiently to regional British/Irish English input for

416  Edgar W. Schneider and Raymond Hickey structural features of the latter to be transferred to incipient varieties of English due to an unguided second language acquisition process among adults. Such a scenario is one where both syntactic transfer from the substrate languages and the adoption of salient grammatical features of the superstrate language are at a premium. In the present context the concern is with discerning the latter features and considering whether there is sufficient evidence to conclude that these were adopted into early forms of non‐native English in the Caribbean and that the presence of structural similarities in later attested forms of Caribbean English represent a case of historical continuity of regional input rather than a set of independent developments. It has been a received wisdom that during the ‘homestead phase’ non‐anglophones were exposed to regional forms of English. This point of view is accepted by all authors, e.g. Baker and Huber (2000, 2001) are of the opinion that non‐anglophones were exposed to foreigner talk versions of English varieties and that this explains world‐wide similarities.

5.3.1  Regional Input: Habitual Aspect in Dialects and Creoles Over the past few decades a number of scholars have referred to the likelihood of English dialect influence in the formative period of Caribbean forms of English. Various syntactic features of both Southwest British English and especially Southern Irish English show recognizable parallels with varieties in the Caribbean. This is perhaps not surprising seeing as how non‐anglophones were likely to have had most contact with the Whites with least status, i.e. the Irish. Possible influence from this group is mentioned by different authors. For instance, Bickerton (1975: 10) considers whether Gaelic is ‘is’ could be the source in is‐initial sentences in Guyanese creole (this kind of topicalization with is occurs in Barbados as well, Schneider 1990: 104). In present‐day Irish English, and in documentation for at least the last three centuries, habitual aspect can be expressed by using an unstressed form of do (inflected or not, and often phonetically reduced) as in She does/do [dʌz, dʊ, də] be worrying about the children. ’She is continually worrying about the children’ (see the detailed discussion in Hickey 2007: 213–37).3 Holm (1988/89) mentions Irish English at a few key points. One is when discussing do(es) be as a habitual marker and when dealing with embedded questions without inversion, both of which are features found in Irish English. Holm (1994) refers to unstressed does as a habitual aspect marker in British varieties of English in his discussion of it in Caribbean creoles and adds that in Ireland it may well have gained support from Irish (Holm 1994: 375). He also echoes Rickford’s view (Holm 1986: 257, 260, 263, cf. Rickford 1986) that the use of be in AAE and does be in Caribbean creoles may well reflect a differential influence of northern Irish English on the former and Southern Irish English on the latter. Le Page and Tabouret‐Keller (1985) refer to the settlement of various Caribbean islands, notably Barbados, by Irish immigrants. This can be traced to the deportation practice of Oliver Cromwell, who sent many Irish to the island in the late 1640s and early 1650s; but there are already official reports from the 1630s of Irish being press‐ganged into English service in the Caribbean. Le Page and Tabouret‐Keller also mention the West Indian do/does, used in an habitual sense, as deriving ‘to some extent’ from West of England and Irish usage. Roy (1986: 143) mentions the extent of Irish immigration to Barbados in the seventeenth century. He regards an Irish origin of certain aspectual distinctions such as do plus be as probable (Roy 1986: 150–1) and quotes established Irish sources to substantiate his arguments. The same is true of Roberts when talking of English on Barbados and Montserrat (Roberts 1988: 34). Williams (1988) begins by considering habitual aspect markers in Irish English and in South‐Western dialects of English (the latter with special forms of do‐support). He then

Contact and Caribbean Creoles  417 proceeds to consider the development of the aspectual systems of Anglophone creoles in the Caribbean area and concludes that the contribution of earlier varieties of English is in many cases sufficient to explain existing aspectual distinctions without having recourse to a creolization hypothesis. In his conclusion he assumes a considerable Irish English input for Saba Island (in the Netherlands Antilles), Bequia Island (in St Vincent and the Grenadines) and Barbados, three Caribbean locations he examined in detail. Bailey (1982) reports on a number of features of Irish English which link it up with African American English and points out that constructions with does + be could derive from transported Irish English in the Caribbean (Barbados and Jamaica) already in the seventeenth century. This would have happened before the slave trade got fully under way, Irish English features being then absorbed by the forms of English which arose with the displaced African population. Labov (1998), in his discussion of nonfinite be, remarks that “comparatively few creoles have a specialized marker for habituals.” He goes on to say that “Invariant be is rare in the Caribbean, but it is a common feature of Hiberno‐English” and recognizes the connection between a habitual use – with or without a form of do – and the existence of this category in Irish English and refers to Rickford (1986) who offered evidence for early contact between African slaves and Irish laborers (Beckles 1989, 1990). When examining the habitual with do in Caribbean Englishes one sees that for many varieties, this do, or a reduced form such as [də], is not necessarily present. Instead one may just find be. For instance, in his review of preverbal markers (Schneider 1990: 90) has an habitual marker doz/da in only five localities (of the 14 he examined): Grenada, Barbados, Trinidad (Youssef 1995), St Kitts/Nevis and the Bahamas, to which one could add Miskito Coast in Nicaragua. For those localities which do not show doz/da it may be possible to conclude that through phonetic attrition this form was lost from an earlier stage in which it existed (Rickford 1980).

5.3.2  Barbados in the Spread of English Throughout the Caribbean In the history of Anglophone settlement in the Caribbean the island of Barbados in the southeast (along with St Kitts somewhat to the north, see Baker and Bruyn (eds) 1998) plays a central role. It was among the first islands to be settled by the English in 1627 (Andrews 1984) and, given its small size and quick growth in population, there was later movement from here to other parts of the Caribbean, notably to Jamaica, Guyana, and to parts of the southern United States, chiefly to the coastal regions of South Carolina and Georgia, to the Sea Islands along the littoral of these states where the creole Gullah developed and is still to be found (Hancock 1980). There have been various reasons for the exploitation of Barbados by the English. Initially, the island functioned as a bridgehead for the English in the Caribbean which, by the beginning of the seventeenth century, was dominated by the Spanish. It was later to become important with the development of the cane sugar trade (Taylor 2001: 205–17; Dunn 1972), something which also came to be true of Jamaica after it was wrenched from Spanish control in the mid-seventeenth century (Taylor 2001: 217–21; Le Page 1960). Before the large‐scale importation of African slaves got under way in the later seventeenth century, the English had a system of indenture whereby settlers from the British Isles went to the Caribbean to work for a period, typically five to eight years, after which they were free to move at will, their circumstances at the overseas locations permitting. The settlement of Barbados is also linked to the deportation of Irish dissidents by Oliver Cromwell, as mentioned on p. 416, this element forming a significant proportion of early White settlers from the late 1640s onwards (O’Callaghan 2000: 65–76) and these would have been in contact with African slaves in a work context (Rickford 1986: 251). There was also a later deportation to Jamaica (O’Callaghan 2000: 77–88).4

418  Edgar W. Schneider and Raymond Hickey Given the size of Barbados and the relatively low social position of the Irish in the White community on the island there would have been fairly intensive contact between Africans and the Irish. A further aspect concerns the early African slaves. Campbell (1993: 98), when dealing with “the component parts of the population,” says that he found no mention of female Africans before the 1640s which means that for the first generation of English settlement on Barbados (from the late 1620s onwards) there was not only a preponderance of Whites but no Black families with children born there who could have initiated the creolization process, assuming that the social scenario was already suitable then. Campbell (1993: 148) mentions the late 1640s as the beginning of the sugar revolution with the switch from tobacco. This was also the switch from White indentured servants to Black slaves (Harlowe 1969 [1926]: 292–330) and the exodus of the former from Barbados. From 1650–80 upwards of 10,000 left Barbados (a conservative estimate). Settlers from other parts of the Caribbean left for the southeast of the North American mainland, chiefly to South Carolina, a movement which began in 1670 and which was largely completed by 1700 (Holm 1994: 342). With the later concentration of African slaves on Barbados settlement patterns arose which were conducive to creolization: Rickford and Handler (1994: 230) point out that “these slaves lived in compact village settlements located next to the plantation yard” and that “these are just the kinds of demographic and settlement patterns which would have produced and/or maintained creole‐speaking communities.” Yet another view is that proposed by Baker (1998: 346–7). He maintains that originally there would have been a “medium for inter‐ethnic communication” (his term) which the entire population would have participated in to begin with. The locally born Black population on St Kitts (the locality of Baker’s investigation) would have used this variety irrespective of whether they also acquired the language of the White community. Essentially what Baker is saying is that there was a split in the early Anglophone population of St Kitts between the superstrate language English and a mixed language among the Black population. He further maintains that “many of the features typical of A(tlantic) E(nglish) C(reole)s are likely to have been established in the embryonic M(edium for) I(nter‐ethnic) C(ommunication) of English St Kitts in the 1620s prior to other territories being settled from that island, and long before this became anyone’s first language” (Baker 1998: 347). Given that on both St Kitts and Barbados the newly imported slaves of the 1640s would have come in contact with the Whites on the islands it stands to reason that they would have adopted features of their speech and continued them. With the spread of Whites from Barbados in later decades this process would have been replicated providing further regional dialect input from Britain and Ireland to other locations throughout the Caribbean.

6  Contact in the Present‐day Caribbean The Caribbean has not lost its appeal to linguists as a rich source of data on language development on diverse levels and in many areas. Studies have continued to be made on Caribbean Englishes but often with different foci compared to literature from the late twentieth century and the early 2000s. One focus of recent research has been the ecologies of language contact (see the contributions in Yakpo and Muysken 2017). Another has been the question of translocality and contact on a supraregional level across various Anglophone locations (Deuber 2014), a concern which has parallels with Schneider’s notion of transnational attraction (Schneider 2014, 2017). Questions surrounding creoles and language variation in the media and the internet have been pursued in Moll (2015) and Westphal (2017). An emphasis on individual agency has also been the subject of attention, see Hinrichs and Farquharson (eds., 2011) and Faraclas (ed. 2012) as have questions of education and the urban‐rural divide, see Lacoste (2012) as well as Migge, Léglise and Bartens (eds., 2010).

Contact and Caribbean Creoles  419 Work by creolists has continued unbroken with the question of creoles and linguistic typology being a particular concern, see Bakker, Daval‐Markussen, Parkvall, and Plag (2011), Bhatt and Veenstra (eds., 2013) and McWhorter (2018).

7 Conclusion No doubt the Caribbean is a multifaceted region in many respects, not only culturally but also socially and linguistically, characterized by a multidimensional complexity of linguistic choices, symbolic representations, and negotiations that are actually difficult to fully understand and document. While much of the early effort in the field went primarily into investigating the dominant creole varieties as spoken, say, in Jamaica, Trinidad, or Guyana, the situation has now improved in that many other localities and situations have become objects of linguistic investigations (St‐Hilaire 2011; Meyerhoff and Walker 2013; Eberle and Schreier 2013; Prescod ed., 2015), though there is still much need for groundwork to be done. Theoretically speaking, there has been a tendency in recent decades for scholarship to step back from earlier strong claims, which saw creolization as a unique and highly exceptional process, steered by whatever causes different scholars considered decisive, and to embed the study of creoles and creolization more broadly in the discipline and the parameters of language contact studies. In terms of types and outcomes of language contact, almost the entire range of contact conditions and phenomena as outlined by Thomason and Kaufman (1988) and Thomason (2001) can be found and investigated there. Clearly, for scholars interested in language contact this is a prime region to turn to and investigate more closely.

NOTES 1 JamC = Jamaican Creole; GuyC = Guyanese Creole; TriC = Trinidadian Creole; TobC = Tobagian Creole. 2 According to Parkvall (1998: 69) the English slave trade to St Kitts only started ca. 1660 (English slaving on the Lower Guinea coast starts ca 1640 but in considerable numbers only about 20 years later), so this can be taken as the earliest date that new slaves arrived from Africa. For St Kitts the first mention of African slaves is 1626: 3 ships deliver 60 Africans to the island (Parkvall 1998: 66). 3 Bickerton is skeptical of tracing doz be for habitual aspect in Caribbean varieties to Irish influence. 4 The vicissitudes of the Civil War in England (1642–51) were also responsible for the emigration of English as well. A well‐known instance of this is Richard Ligon. He was probably one of the displaced persons who departed in 1647 due to war in England, claiming to have lost wealth and property in the Fens in East Anglia. Later he wrote a history of Barbados (Ligon 1657) which provides information on the early settlement of the island.

REFERENCES Aceto, Michael 2003. Barbudan Creole English: its history and some grammatical features. English World‐Wide 23: 223–50. Aceto, Michael 2004a. Eastern Caribbean English‐derived language varieties: phonology. In Edgar W. Schneider, Kate Burridge, Bernd Kortmann, Rajend Mesthrie, and Clive Upton

(eds.) 2004. A Handbook of Varieties of English, vol. 1: Phonology. Berlin: de Gruyter Mouton, pp. 481–500. Aceto, Michael 2004b. Eastern Caribbean English‐derived language varieties: morphology and syntax. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W.

420  Edgar W. Schneider and Raymond Hickey Schneider, and Clive Upton (eds.) 2004: A Handbook of Varieties of English, vol. 2: Morphology and Syntax. Berlin, New York: de Gruyter Mouton, pp. 439–53. Aceto, Michael and Jeffrey P. Williams (eds.) 2003. Contact Englishes of the Eastern Caribbean. Amsterdam: Benjamins. Alleyne, Mervyn C. 1980. Comparative Afro‐ American: An Historical‐Comparative Study of English‐Based Afro‐American Dialects of the New World. Ann Arbor, MI: Karoma. Alleyne, Mervyn C. 1988. Roots of Jamaican Culture. London: Pluto. Allsopp, Richard (ed.) 1996. Dictionary of Caribbean English Usage. Oxford: Oxford University Press. Andrews, Kenneth R. 1984. Trade, Plunder, and Settlement. Maritime Enterprise and the Genesis of the British Empire, 1480–1630. Cambridge: Cambridge University Press. Arends, Jacques 1993. Towards a gradualist model of creolization. In Francis Byrne and John Holm (eds.), Atlantic Meets Pacific. Amsterdam: Benjamins, pp. 371–80. Bailey, Charles‐James N. 1982. Irish English and Caribbean Black English: Another rejoinder. American Speech 57, 237–9. Baker, Philip and Magnus Huber 2000. Constructing new pronominal systems from the Atlantic to the Pacific. Linguistics 38: 833–66. Baker, Philip and Magnus Huber 2001. Atlantic, Pacific, and world‐wide features in English‐ lexicon contact languages. English World‐Wide 22.2: 157–208. Baker, Philip and Adrienne Bruyn (eds.) 1998. St Kitts and the Atlantic Creoles. The Texts of Samuel Augustus Mathews in Perspective. London: University of Westminster Press. Bakker, Peter, Aymeric Daval‐Markussen, Mikael Parkvall, and Ingo Plag 2011. Creoles are typologically distinct from non‐creoles, Journal of Pidgin and Creole Languages 26.1: 5–42. Beckles, Hilary 1989. White Servitude and Black Slavery in Barbados, 1627–1715. Knoxville: University of Tennessee Press. Available at www.jstor.org/journal/willmaryquar Beckles, Hilary 1990. A “riotous and unruly lot”: Irish indentured servants and freemen in the English West Indies, 1644–1713. The William and Mary Quarterly 47.4: 503–22. Bhatt, Parth and Tonjes Veenstra (eds.) 2013. Creole Languages and Lingusitic Typology. Amsterdam: Benjamins. Bickerton, Derek 1975. Dynamics of a Creole System. Cambridge: Cambridge University Press.

Bickerton, Derek 1981. Roots of Language. Ann Arbor, MI: Karoma. Blake, Renée 2002. Not as clear as black and white: a study of race, class and language in a Barbados community. Unpublished manuscript, New York University. Blake, Renée 2004. Bajan: phonology. In Edgar W. Schneider, Kate Burridge, Bernd Kortmann, Rajend Mesthrie, and Clive Upton (eds.) 2004. A Handbook of Varieties of English, vol. 1: Phonology. Berlin: de Gruyter Mouton, pp. 501–7. Boretzky, Norbert 1983. Kreolsprachen, Substrate und Sprachwandel. Wiesbaden: Harrassowitz. Campbell, P. F. 1993. Some Early Barbadian History. Barbados: Caribbean Graphics. Cassidy, Frederic G. 1961. Jamaica Talk: Three Hundred Years of the English Language in Jamaica. London: Macmillan. Cassidy, Frederic G. 1980. The place of Gullah. American Speech 55: 3–16. Cassidy, Frederic G. and Robert B. Le Page 1980 [1967]. Dictionary of Jamaican English. Cambridge: Cambridge University Press. Chaudenson, Robert 2001. Creolization of Language and Culture, revised in collaboration with Salikoko S. Mufwene. London and New York: Routledge. Christie, Pauline 2003. Language in Jamaica. Kingston: Arawak. DeCamp, David 1971. Toward a generative analysis of a post‐creole speech continuum. In Dell Hymes (ed.), Pidginization and Creolization of Languages. Bloomington: Indiana University Press, pp. 349–70. DeGraff, Michel 2003. Against creole exceptionalism. Language 79: 391–410. Deuber, Dagmar 2014. English in the Caribbean. Variation, Style and Standards in Jamaica and Trinidad. Cambridge: Cambridge University Press. Dunn, Richard S. 1972. Sugar and Slaves. The Rise of the Planter Class in the English West Indies, 1624–1713. Chapel Hill, NC: University of North Carolina Press. Eberle, Nicole and Daniel Schreier 2013. African Bermudian English and the Caribbean connection. English World‐Wide 34.3: 279–304. Faraclas, Nichlos (ed.) 2012. Agency in the Emergence of Creole Languages. Amsterdam: Benjamins. Fields, Linda 1995. Early Bajan: creole or non‐ creole? In Jacques Arends (ed.), The Early Stages of Creolization. Amsterdam: Benjamins, pp. 89–111.

Contact and Caribbean Creoles  421 Graham, Ross 2005. Partial creolization, restructuring and convergence in Bay Islands English. English World‐Wide 26: 43–76. Hackert, Stephanie 2004. Urban Bahamian Creole: System and Variation. Amsterdam: Benjamins. Hancock, Ian F. 1980. Gullah and Barbadian: origins and relationships. American Speech 55: 17–35. Hancock, Ian F. 1993. Creole language provenance and the African component. In Salikoko S. Mufwene (ed.), Africanisms in Afro‐American Language Varieties. Athens: University of Georgia Press, pp. 182–91. Harlowe, Thomas V. 1969 [1926]. A History of Barbados 1625–1685. New York: Negro University Press. Hickey, Raymond 2003. Rectifying a standard deficiency. Pronominal distinctions in varieties of English. In Irma Taavitsainen and Andreas H. Jucker (eds.), Diachronic Perspectives on Address Term Systems. Amsterdam: Benjamins, pp. 345–74. Hickey, Raymond 2004. English dialect input to the Caribbean. In Raymond Hickey (ed.) Legacies of Colonial English. Studies in Transported Dialects. Cambridge: University Press, pp. 326–59. Hickey, Raymond 2007. Irish English. History and Present‐day Forms. Cambridge: Cambridge University Press. Hinrichs, Lars and Joseph T. Farquharson (eds.) 2011. Variation in the Caribbean: From Creole Continua to Individual Agency. Amsterdam: Benjamins. Holm, John (ed.) 1983. Central American English. Heidelberg: Groos. Holm, John 1986. The spread of English in the Caribbean area. In Manfred Görlach and John A. Holm (eds.), Focus on the Caribbean. Amsterdam: Benjamins, pp. 1–22. Holm, John 1988/89. Pidgins and Creoles, vol. 1: Theory and Structure; vol. 2: Reference Survey. Cambridge: Cambridge University Press. Holm, John 1994. English in the Caribbean. In Robert Burchfield (ed.), The Cambridge History of the English Language, vol. 5: English in Britain and Overseas: Origins and Development. Cambridge: Cambridge University Press, pp. 328– 81. Holm, John 2004. Languages in Contact: The Partial Restructuring of Vernaculars. Cambridge: Cambridge University Press. Huttar, Mary 2008. Ndyuka (Creole English). In John Holm and Peter L. Patrick (eds.), Comparative Creole Syntax: Parallel Outlines of 18 Creole Grammars. London: Battlebridge, pp. 217–36.

James, Winford and Valerie Youssef 2002. The Languages of Tobago: Genesis, Structure and Perspectives. St. Augustine: University of the West Indies. Kortmann, Bernd, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider and Clive Upton (eds.) 2004. A Handbook of Varieties of English, vol. 2: Morphology and Syntax. Berlin: de Gruyter Mouton. Lacoste, Véronique. 2012. Phonological Variation in Rural Jamaican Schools. Amsterdam: Benjamins. Lacoste, Véronique 2017. The Caribbean, In Markku Filppula, Juhani Klemola, and Devyani Sharma (eds.), The Oxford Handbook of World Englishes. Oxford: Oxford University Press, pp. 389–408. Labov, William 1998. Co‐existent systems in African‐American vernacular English. In Salikoko Mufwene, John R. Rickford, Guy Bailey, and John Baugh (eds.), African‐American English. Structure, History and Use. London: Routledge, pp. 110–53. Lalla, Barbara and Jean D’Costa 1990. Language in Exile: Three Hundred Years of Jamaican Creole. Tuscaloosa: University of Alabama Press. Le Page, Robert B. 1960. An historical introduction to Jamaican creole. In Robert B. Le Page and David DeCamp (eds.), Jamaican Creole. London: Macmillan, pp. 1–24. Le Page, Robert B. and Andrée Tabouret‐Keller 1985. Acts of Identity: Creole‐Based Approaches to Language and Ethnicity. Cambridge: Cambridge University Press. Ligon, Richard 1657. A True and Exact History of the Island of Barbados (2nd edition 1673). McWhorter, John 1997. It happened at Cormantin: locating the origin of the Atlantic English‐based creoles. Journal of Pidgin and Creole Languages 12: 59–102. McWhorter, John 1998. Identifying the creole prototype: vindicating a typological class. Language 74: 788–818. McWhorter, John 2000. Defining “creole” as a synchronic term. In Ingrid Neumann‐ Holzschuh and Edgar W. Schneider (eds.), Degrees of Restructuring in Creole Languages. Amsterdam: Benjamins, pp. 85–123. McWhorter, John 2018. The Creole Debate. Cambridge: Cambridge University Press. Meyerhoff, Miriam and James A. Walker 2013. Bequia Talk (St Vincent and the Grenadines). London: Battlebridge Publications. Migge, Bettina 2003. Creole Formation as Language Contact. Amsterdam: Benjamins.

422  Edgar W. Schneider and Raymond Hickey Migge, Bettina (ed.) 2007. Substrate Influence in Creole Formation. Special Issue of Journal of Pidgin and Creole Languages 22.1. Migge, Bettina, Isabelle Léglise, and Angela Bartens (eds.) 2010. Creoles in Education: An appraisal of current programs and projects. Amsterdam: Benjamins. Moll, Andrea 2015. Jamaican Creole Goes Web: Sociolinguistic Styling and Authenticity in a Digital “Yaad.” Amsterdam: Benjamins. Mufwene, Salikoko S. 1986. The universalist and substrate hypotheses complement one another. In Pieter Muysken and Norval Smith (eds.), Substrata versus Universals in Creole Genesis. Amsterdam: Benjamins, pp. 129–62. Mufwene, Salikoko S. 1996. The Founder Principle in creole genesis. Diachronica 13: 83–134. Mufwene, Salikoko S. (ed.) 1993. Africanisms in Afro‐American Language Varieties. Athens: University of Georgia Press. Mufwene, Salikoko S. 2000. Language contact, evolution, and death: how ecology rolls the dice. In Gloria Kindell and M. Paul Lewis (eds.), Assessing Ethnolinguistic Vitality: Theory and Practice; Selected Papers from the Third International Language Assessment Conference. Dallas: SIL International, pp. 39–64. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Mufwene, Salikoko S. 2008. Language Evolution. London: Continuum. Mühleisen, Susanne 2002. Creole Discourse: Exploring Prestige Formation and Change across Caribbean English‐Lexicon Creoles. Amsterdam: Benjamins. Mühleisen, Susanne and Bettina Migge (eds.) 2005. Politeness and Face in Caribbean Creoles. Amsterdam: Benjamins. Muysken, Pieter and Norval Smith (eds.) 1986. Substrata versus Universals in Creole Genesis. Amsterdam: Benjamins. Neumann‐Holzschuh, Ingrid and Edgar W. Schneider (eds.) 2000. Degrees of Restructuring in Creole Languages. Amsterdam: Benjamins. O’Callaghan, Sean 2000. To Hell or Barbados. The Ethnic Cleansing of Ireland. Dingle: Brandon Books. Ó Siochrú, Mícheál 2008. God’s Executioner: Oliver Cromwell and the Conquest of Ireland. London: Faber and Faber. Parkvall, Mikael 1998. A short note on the peopling of English St. Kitts. In Philip Baker and Adrienne Bruyn (eds.), St Kitts and the Atlantic Creoles. The Texts of Samuel Augustus

Mathews in Perspective. London: University of Westminster Press, pp. 63–74. Parry, J. H. and Philip Sherlock 1981. A Short History of the West Indies, 3rd edition. Basingstoke: Macmillan. Patrick, Peter L. 1999. Urban Jamaican Creole: Variation in the Mesolect. Amsterdam: Benjamins. Patrick, Peter L. 2004. Jamaican Creole: morphology and syntax. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2: Morphology and Syntax. Berlin: de Gruyter Mouton, pp. 407–38. Patrick, Peter L. 2008. Jamaican Patwa (Creole English). In John Holm and Peter L. Patrick (eds.), Comparative Creole Syntax: Parallel Outlines of 18 Creole Grammars. London: Battlebridge, pp. 127–52. Prescod, Paula (ed.) 2015. Language Issues in Saint Vincent and the Grenadines, Amsterdam: Benjamins Reaser, Jeffrey and Benjamin Torbert 2004. Bahamian English: morphology and syntax. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2: Morphology and Syntax. Berlin: de Gruyter Mouton, pp. 391–406. Rickford, John R. 1980. “How does doz disappear?” In Richard R. Day (ed.) Issues in English Creoles. Papers from the 1975 Hawaii Conference. Varieties of English around the World, G2. Heidelberg: Groos, pp. 77–96. Rickford, John R. 1983. What happens in decreolization. In Roger Andersen (ed.), Pidginization and Creolization as Language Acquisition. Rowley, MA: Newbury House, pp. 298–319. Rickford, John R. 1986. Social contact and linguistic diffusion: Hiberno‐English and New World Black English. Language 62: 245–90. Rickford, John R. 1987. Dimensions of a Creole Continuum: History, Texts, and Linguistic Analysis of Guyanese Creole. Stanford: Stanford University Press. Rickford, John R. and Jerome S. Handler 1994. Textual evidence on the nature of early Barbadian speech, 1676–1835. Journal of Pidgin and Creole Languages 9: 221–55. Roberts, Peter A. 1988. West Indians and their Language. Cambridge: Cambridge University Press. Roy, John. 1986. Tense and aspect in Barbadian English creole. In Manfred Görlach and John A. Holm (eds.) 1986. Focus on the Caribbean. Amsterdam: Benjamins, pp. 141–56.

Contact and Caribbean Creoles  423 Schneider, Edgar W. 1990. The cline of creoleness in English‐oriented creoles and semi‐creoles of the Caribbean. English World‐Wide 11: 79–113. Schneider, Edgar W. 1998. Negation patterns and the cline of creoleness in English‐oriented varieties in the Caribbean. In Pauline Christie, Barbara Lalla, Velma Pollard, and Lawrence Carrington (eds.), Studies in Caribbean Language II: Papers from the Ninth Biennial Conference of the Society for Caribbean Linguistics 1992. St. Augustine: University of the West Indies, pp. 204–27. Schneider, Edgar W. 2004. The English dialect heritage of the southern United States In Raymond Hickey (ed.), Legacies of Colonial English. Studies in Transported Dialects. Cambridge: Cambridge University Press, pp. 262–309. Schneider, Edgar W. 2014. New reflections on the evolutionary dynamics of world Englishes. World Englishes 33.1: 9–32. Schneider, Edgar W. 2017. Models of English in the World. In Juhani Klemola, Markku Filppula, and Devyani Sharma (eds.), The Oxford Handbook of World Englishes. Oxford: Oxford University Press. Schneider, Edgar W., Kate Burridge, Bernd Kortmann, Rajend Mesthrie, and Clive Upton (eds.) 2004. A Handbook of Varieties of English, vol. 1: Phonology. Berlin: de Gruyter Mouton. Schneider, Edgar W. and Christian Wagner 2006. The variability of literary dialect in Jamaican Creole: Thelwell’s The Harder They Come. Journal of Pidgin and Creole Languages 21: 45–95. Singler, John 1988. The homogeneity of the substrate as a factor in pidgin/creole genesis. Language 64: 27–51. Smith, Norval and Vinije Haabo 2004. Suriname creoles: phonology. In Edgar W. Schneider, Kate Burridge, Bernd Kortmann, Rajend Mesthrie, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 1: Phonology. Berlin: de Gruyter Mouton, pp. 525–64. St‐Hilaire, Aonghas 2011. Kwéyòl in Postcolonial Saint Lucia: Globalization, Language Planning, and National Development. Amsterdam: Benjamins. Taylor, Alan 2001. American Colonies. The Settlement of North America. London: Allen Lane.

Thomason, Sarah G. 2001. Language Contact: An Introduction. Washington, DC: Georgetown University Press. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Walker, James A. and Miriam Meyerhoff 2006. Zero copula in the Eastern Caribbean: evidence from Bequia. American Speech 81: 146–63. Westphal, Michael 2017. Language Variation on Jamaican Radio. Amsterdam: Benjamins. Williams, Jeffrey P. 1988. Anglo‐Caribbean English: A Study of its Sociolinguistic History and the Development of its Aspectual Markers (University of Texas at Austin Ph.D. thesis). Williams, Jeffrey P. 2012. English varieties in the Caribbean. In Raymond Hickey (ed.), Areal Features of the Anglophone World. Berlin: de Gruyter Mouton, pp. 133–60. Winford, Donald 1997. Re‐examining Caribbean English creole continua. World Englishes 16: 233–79. Winford, Donald 1997–8. On the origins of African American Vernacular English – a creolist perspective, Diachronica 14.2: 305–44; 15.1: 99–154. Winford, Donald 2000. “Intermediate” creoles and degrees of change in creole formation: the case of Bajan. In Ingrid Neumann‐Holzschuh and Edgar W. Schneider (eds.), Degrees of Restructuring in Creole Languages. Amsterdam: Benjamins, pp. 215–46. Winford, Donald 2003. An Introduction to Contact Linguistics. Malden, MA: Blackwell. Winford, Donald and Bettina Migge 2004. Surinamese creoles: morphology and syntax. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider, and Clive Upton (eds.), A Handbook of Varieties of English, vol. 2: Morphology and Syntax. Berlin: de Gruyter Mouton, pp. 482–516. Yakpo, Kofi and Pieter C. Muysken (eds.) 2017. Boundaries and Bridges. Language Contact in Multilingual Ecologies. Berlin: de Gruyter Mouton. Youssef, Valerie 1995. Tense‐aspect in Tobagonian English: a dynamic transitional model. English World‐Wide 16.2: 195–213.

21 Contact and the Romance Languages JOHN CHARLES SMITH 1 Introduction Contact has played an important role in the history of Romance.1 The influence on Latin of a variety of other languages was one factor in its evolution and differentiation into the Romance languages (see Adams 2003, esp. 111–296), although it must also be stressed that much of this differentiation was due to internal changes which were not the result of contact. Subsequently, medieval Latin, as the “H” language in exoglossic or “Type Zero” diglossia (see Auer 2005), was itself a significant influence on the Romance languages of Western Europe. A variety of other contact influences have made themselves felt in Romance between the Middle Ages and the present day, with complex cultural and political factors within Europe playing an important role. Moreover, when “standard” Romance languages arose, they often involved an element of koineization, resulting from dialect contact; and, as they became more widespread, their interaction with local Romance speech often gave rise to new regional varieties. With the spread of Romance beyond Europe, through colonization and emigration, these languages have both influenced and been influenced by others, with consequences ranging from the occasional loanword to the emergence of fully‐fledged contact vernaculars. Finally, with their long history and great diversity, Romance languages can provide a useful test‐bed for recent theories of language contact. This short chapter cannot analyze all of these issues in detail; but it will attempt to provide an overview of some major themes.

2 The Differentiation of Latin and the Emergence of Romance A traditional view of the effects of language contact on the development of the Romance languages distinguishes substratum, superstratum, and adstratum (see, for instance, Elcock 1975: 183–311). A substratum (or substrate) language is one that is already spoken in an area when speakers of another language arrive and become dominant  –  socially, ­culturally, politically, and linguistically – but which nonetheless influences this new language. A superstratum (or superstrate) language is one that is spoken by new arrivals in an area who, however dominant socially, culturally, and politically, do not eradicate the

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

426  John Charles Smith language which is already spoken in the area (indeed, it may sometimes be the language of the new dominant group which fails to survive), but nonetheless influence it. An adstratum (or adstrate) language is one that influences another “from the side,” as it were, without there necessarily being any mass geographical displacement of speakers. Schematically: superstratum ↓ language “L” ← adstratum ↑ substratum The tripartite typology of substratum/superstratum/adstratum is useful as far as it goes, in as much as it identifies different sets of social relations which may result in contact influence on a language. However, it is relatively unsophisticated and says little about the actual mechanisms involved in change. We may distinguish two different types of language change through contact, according to whether the initiators of the change are first‐ or second‐language speakers (L1 vs L2). One of these is borrowing, which is defined as transfer due to recipient‐language agentivity. In this case, speakers of a language adopt elements of a foreign language into their own. This is perhaps most characteristic of adstratum, but is also found to an extent with superstratum influence. Contrasting with this mechanism is imposition, defined as transfer due to source‐language agentivity. Here, speakers of a language attempt to speak a foreign language; but, as L2 speakers, they are unable to master the new language fully, and so inevitably transfer elements of their native speech into it. This process is characteristic of both substratum and superstratum – in the Romance context, of the conquered learning Latin and of the conquerors learning Latin. Borrowing may be particularly implicated in the transfer of lexical items; imposition in phonological and grammatical change. Similar ideas are put forward in earlier work on language contact (see, for instance, Thomason and Kaufman 1988), but this bipartite typology of contact is especially associated with van Coetsem (1988, 2000) and Winford (2003, 2005, 2010). Another issue with the tripartite stratum distinction is that in reality we are dealing with a ­ continuum. At the extremes, substratum and superstratum involve mass movement of a population which comes to dominate and rule over a particular area, whilst adstratum involves little or no systematic movement of people, but rather a general cultural ­awareness seeping into a society as if by osmosis; substratum and superstratum are then distinguished by the directionality of the contact. However, the situation may often be much less clear cut. Empires may exercise their authority by sending a small number of administrators to manage a colony or province: if they impose their language, it nonetheless seems reasonable to speak of a superstratum. But they may also exert their influence through a network of client states, which retain some elements (often illusory) of autonomy (this was true, in part, of both the Roman and the Ottoman Empires, each of which is important, in different ways, in the history of language contact in Romance). And contact on the frontier may be difficult to characterize  –  for instance, Rome occupied some Germanic‐speaking areas, but recruited Germanic‐speaking soldiers into its legions from beyond. The frontier, although fortified and militarized, was highly porous (see Maas 2014: 11). The distinction between substratum, superstratum, and adstratum is nonetheless useful for purposes of exposition, and we shall make use of these terms as appropriate in what follows.

Contact and the Romance Languages  427

3 Substratum 3.1 Languages Latin was one of a number of languages spoken in central Italy in 753 bc (the traditional date of the founding of Rome), and was already in contact with some of them during the regal period and in the early days of the Republic (founded, again according to tradition, in 509 bce). Most of these languages belonged to the Italic family, of which Latin was also a member – thus Faliscan, closely related to Latin and spoken just north of Rome, and the more distantly related Sabellian languages (Drinka, this volume), including Umbrian (spoken to the north and east of Rome) and Oscan (spoken further south).2 A notable exception is Etruscan, a non‐Indo‐ European language spoken to the northwest of Rome. As Roman influence expanded, Latin came into contact with other languages – Greek, spoken in Magna Graecia (the Greek settlements of the southern Italian peninsula and Sicily); the speech of the indigenous peoples of Sicily (Siculi, Sicani, and Elimi, of whom at least the first spoke an Indo‐European language); and Messapic, an Indo‐European language spoken in southeastern Italy. North of the Apennines, there was contact with Venetic (an Indo‐European, probably Italic, language), Raetic and Ligurian (of uncertain origin), and Lepontic (which belonged to the Celtic family).3 Rome also conquered Semitic‐speaking Punic areas in Sardinia, North Africa, and the southern Iberian peninsula; further expansion in Iberia led to contact with the Indo‐European languages Lusitanian and Celtiberian4 and the non‐Indo‐European Iberian language (which may be related to Basque), whilst subsequent colonization of North Africa added contact with Libyan and Berber. Expansion into Greece brought further contact with Greek, whilst the conquest of Gaul and Britain exposed Latin to the influence of the Celtic languages Gaulish5 and British, and, in what is now southwestern France, the Aquitanian language, generally assumed to be related to, or even the ancestor of, Basque. Beyond Gaul, there was contact with Germanic. Across the Adriatic, Latin came into contact with the Indo‐European language Illyrian (plausibly related to modern Albanian), and further east, Thracian, Getic, and Sarmatian (although Adams 2003: 283 suggests that the last two languages were “not in any real sense ever in contact with Latin”). There was also contact with Aramaic and Hebrew in the Near East. For more detailed information, see Tagliavini (1972: 91–157), Neumann and Untermann (1980), and Adams (2003: 111–284). For a thorough discussion of regional variation within Latin and the extent to which it may be due to contact, see Adams (2007). Our knowledge of how these various substrates influenced Latin, and the extent to which such influence is subsequently manifest in Romance, is very variable. The major problem is that the languages invoked as influences were in most cases displaced by Latin; they are often sparsely attested and sometimes had no written form. We therefore have scant evidence of them, and often cannot assess the claims with any confidence. Sometimes, the only “knowledge” we have of a substratum language is its alleged influence on Latin – which in the worst case leads to circular arguments.6 Moreover, in areas of the former Roman Empire where Latin was displaced by the language of subsequent invaders (for instance, North Africa or Britain), no indigenous Romance languages survive to bear witness to possible influences. Sober rehearsals of the problems associated with substratum hypotheses in Romance can be found in Wanner (1977) and Tagliavini (1972: 150–3).

3.2 Lexicon The only substratum influences we can be reasonably certain of are a number of well‐attested lexical items. One example of an early Germanic loanword which had entered Latin by the first century ce and which survives in the contemporary Romance languages is saponem7 ‘hair dye’ (Glare 1982: 1690), which gives the word for ‘soap’ (Portuguese sabão, Spanish

428  John Charles Smith jabón, Catalan sabó, French savon, Italian sapone, Romanian săpun, etc.). (For the reasons noted on p. 426, it is actually uncertain whether this word should be regarded as a substrate or an adstrate borrowing.) In Spanish, we find words of Celtic origin, many of which refer to plants, animals, or utensils: thus beleño ‘henbane’, garza ‘heron’ (for further examples and fuller discussion, see Dworkin 2012: 29). Iberian and/or Basque is also an alleged source of loan words  –  vega ‘water‐meadow’ (Coromines 1954: IV, 685–7) being an often quoted example. However, Dworkin (2012: 26) notes: Words that can be securely related to documented Basque forms pose a major chronological problem. Basque and its ancestral variants have been in contact with spoken Latin and its Hispano‐Romance continuations for over two millennia. It may often be difficult to determine if a word of Basque origin entered the language during the period of the Roman occupation (a time when numerous Latin words entered varieties of Basque) or during the subsequent medieval period of Basque–Romance contact. […] Thus, with respect to Spanish, Basque has acted over the centuries both as a substratum and as an adstratum language.

Trask (1997: 415–21) also surveys the position, discussing 24 items which are commonly alleged to be of Basque origin, and notes, in an excoriating conclusion: In sum […], the attempts to derive elements of the Castilian vocabulary from Basque are unconvincing in the extreme. Few of them have been based upon anything more than the recognition of a problem, a Romance word with no etymology, and a vague hope that Basque might provide some kind of explanation. No more than two or three of the words in my list are even plausibly of Basque origin, and I have included all of the most promising cases. The effect of Basque upon the Castilian lexicon has been virtually nil.

Gaulish substrate items in French tend to refer to husbandry, cultivation, and domestic life, including names of animals (mouton ‘sheep’, bièvre ‘(European) beaver’) and trees (if ‘yew’, and plausibly chêne ‘oak’), and words such as soc ‘plowshare’, souche ‘tree‐stump’, suie ‘soot’, dru ‘dense, thick’ (for further discussion, see Nyrop 1914: 5–7; Wise 1997: 32–4). A number of Thracian substrate items have been proposed for Romanian, although many of these items are claimed to be substrate not on the basis of direct evidence (of which there is almost none), but by virtue of their having Albanian cognates which are not borrowed from Latin, or else by conjecture through comparison with a range of other Indo‐European languages – these are at best indirect methods of analysis. Sala (2013: 201) notes that “Thraco‐Dacian substrate terms generally denote flora or fauna specific to the relevant region or have a more specific sense when compared with synonymous terms inherited from Latin”; Poghirc (1969: 364–5) adds terms referring to the body, the family, the house, farming, and landscape. Examples are provided by Sala (1999: 91–9); for detailed discussion, see Poghirc (1969: 327–64) and Brâncuș (1983). Structural substratum influences on Romance have been proposed, but they are more controversial, for the reasons given on p. 426. Examples include the following.

3.3 Phonology 3.3.1  Nasal Assimilation The assimilation of voiced plosives to a preceding homorganic nasal /mb/ > /mm/, /nd/ > /nn/ takes place in most areas of Southern Italy (for instance, Rome: Vignuzzi 1997: 318; Naples: Ledgeway 2009: 102–4; southern Salento: Loporcaro 1997: 343; southern Lucania and northern Calabria: Trumper 1997: 357; most of Sicily: Ruffino 1997: 367), and is sometimes

Contact and the Romance Languages  429 claimed to be the result of influence from Oscan, in which a similar phenomenon is found (see Buck 1904: 8, 80, 84–5). But the area in which this development occurs is by no means coterminous with the area of Oscan speech – in particular, it stretches much further south. In any case, this type of assimilation also occurs in early Catalan (Sampson 1999: 165) and in Gascon (Rohlfs 1970: 154–6).8 It is also found outside Romance (for instance, in German – see Bithell 1952: 192–3). As Adams (2003: 119) notes, “the influence of Oscan on non‐standard Latin has been exaggerated, and alleged phonetic manifestations of such influence in particular are usually explicable in other ways.” On the basis of textual evidence, Ledgeway (2009: 103–4) rules out any Oscan influence on Neapolitan (the first attestations of the assimilation are late medieval). It is possible that the phenomenon spread southwards into Campania, but, again using textual evidence, Ledgeway suggests that it was instead a spontaneous change which originated in rural areas of Campania and spread into the mainstream urban variety. We seem to be dealing with a simple assimilation which is easily explicable in structural terms and is rather less likely to be the result of contact. For further discussion, see Tagliavini (1972: 101–4).

3.3.2  The Gorgia Toscana The spirantization of voiceless plosives in Tuscan varieties of Italian ([p] > [ɸ], [t] > [θ], [k] > [h]), known as the gorgia toscana, has been attributed to Etruscan influence. We know relatively little about Etruscan – for what is known about its influence on Latin, see Adams (2003: 159–84). The hypothesis of Etruscan influence was first put forward by Fernow (1808: 265–7), and the hypothesis was subsequently developed by Nissen (1883) and Merlo (1927, 1950), amongst others. However, the Etruscan hypothesis is roundly criticized by Izzo (1972) on grounds of geography, chronology, and general lack of evidence; see also Tagliavini (1972: 107–13), Agostiniani and Giannelli (1983), and Sala (2013: 197–9). The spirantization of voiceless plosives is in fact a natural process which is attested in a variety of languages (compare, for instance, the “Second Germanic Consonant Shift” of the fourth to fifth centuries ce – see Schwerdt 2000). Kaplan (2010: 13) examines the broad sample of 136 languages discussed by Gurevich (2004) and finds that 17 of them (12.5%) spirantize voiceless plosives. Once again, the change is at least likely to be internal rather than contact‐induced.

3.3.3  /f/ > /h/ in Spanish The debuccalization of initial /f/ in Spanish, and its subsequent disappearance (/f/ > /h/ > /Ø/) was for a time regarded as being the result of contact with a substrate language which is the ancestor of modern Basque. However, considerable skepticism has been expressed regarding this suggestion (see especially Izzo 1977; Blake 1989; Trask 1997: 424–8, and Sala 2013: 195–7), which is no longer accepted without question. Once again, we are dealing with a natural sound change which is widely attested in other languages: within Romance, in Gascon (Rohlfs 1970: 145–9) (where there was arguably also a Basque substrate) and (albeit only before /i/) in many northern dialects of Romanian (where there was clearly no Basque substrate), as shown by ALR 2167 (further Romance examples are given by Trask 1997: 426); but also in Japanese since the eighteenth century (Frellesvig 2010: 386–7). We should note that the change in Spanish is lexically diffuse and sometimes gives rise to doublets (e.g. Lat. forma(m) > Sp. forma ‘shape, form’, Sp. horma ‘(cobbler’s) last’).

3.3.4  Western Romance Voicing of Intervocalic Consonants Most western Romance varieties exhibit voicing of voiceless intervocalic plosives, and this change has been linked to Celtic lenition (Martinet 1952). However, yet again, there is no need to postulate external influence in order to account for the change; it could simply be

430  John Charles Smith due to internal factors – specifically, the assimilation of a voiceless intervocalic consonant to the voicedness of the vowels on either side of it. A similar process operates in Tamil (where it is known as “Caldwell’s Law”), although there is controversy over its date (see Krishnamurti 2001: 127–32 and Mahadevan 2004). Japanese rendaku may constitute another non‐Romance example of such a change – although it is highly morphologized and lexicalized and is not phonologically predictable; see Frellesvig (2010: 40–1). Kaplan (2010: 13) examines the broad sample of 136 languages discussed by Gurevich (2004) and finds that 26 of them (19.1%) lenite voiceless plosives into voiced plosives.

3.3.5  Fronting of /u/ > /y/ in Gallo‐Romance The fronting of /u/ to /y/ is characteristic of Gallo‐Romance (French, Francoprovençal, Occitan) and the northern Italian varieties known as “Gallo‐Italian.” It has also been seen as a Celtic substratum influence, although, once again, the evidence in support of this view is tenuous. Front rounded vowels do not exist in most modern Celtic languages (the exception is Breton – but here the influence may be from French, rather than the other way round), and it is not certain that they ever existed in earlier stages of these languages (although Late British had a high central rounded vowel [ʉ]; see Watkins 1993: 297). The only argument normally adduced concerns the fact that the area of intense Celtic settlement and the area of the /u/ > /y/ change in Romance seem to be approximately coterminous. But the issue is ultimately difficult to judge, and correlation does not prove causation. For discussion, see Nyrop (1914: 202–3) and Tagliavini (1972: 135–7). In fact, no external influence needs to be postulated, as similar changes are attested in a variety of languages – for instance, the back vowel [u] has moved forward in British English over the last few generations (so‐called “goose‐ fronting”: compare the pronunciation of, say, do in the 1930s ([duː]) and now ([dʉː], or even [dÿː]; see Henton 1983); so this type of change, too, can be a completely spontaneous development. The foregoing are amongst the major alleged substratum influences on the phonology of Romance. In each of these cases, the change in question is a “natural” sound change, attested in non‐Romance languages – it is therefore not necessary to attribute it to contact influence. However, it is not impossible that contact may sometimes have acted as a catalyst, facilitating or accelerating an internal structural development.

3.4 Morphology Standard French exhibits a partly vigesimal numeral system – thus soixante‐dix ‘seventy’ (lit. ‘sixty‐ten’), quatre‐vingts ‘eighty’ (lit. ‘four‐twenties’), quatre‐vingt‐dix ‘ninety’ (lit. ‘four‐ twenty‐ten’). (Vigesimal elements are also found sporadically elsewhere in Romance, in some varieties of Francoprovençal, Occitan, Sicilian, and Salentino, where the vigesimal numerals often have a specialized function, serving to count particular types of item – see Price 1992.) Because of limited parallels in some Celtic languages, vigesimality in French has often been laid at the door of a Celtic substratum, but Vennemann (2010: 385–8), noting the non‐Celtic areas where the language has a base‐twenty numeral system (such as Sicilian and Salentino, but also Danish), ascribes the pattern to (the ancestor of) Basque, which he assumes to have been widely spoken across Europe in pre‐Indo‐European times.

3.5  Decline of Infinitival Complementation in Romanian A striking fact about Romanian is the virtual disappearance of the infinitive from complementation structures, in contrast to the situation in Latin and the Romance languages of Western Europe. Thus ‘I want to come’ is vreau să vin (lit. ‘I want that I come’); contrast

Contact and the Romance Languages  431 French je veux venir, Italian voglio venire, Spanish quiero venir, all lit. ‘I want to come’. The one verb which may take an infinitival complement in the contemporary language is a putea ‘to be able’: thus pot veni ‘I can come’; the infinitive is also found in one type of future (voi veni ‘I shall come’). For discussion, see Joseph (1983: 149–78). There are clear structural parallels between the situation in Romanian and that found in Greek, Albanian, and South Slavonic languages (again, see Joseph 1983). This has led to speculation that infinitive loss is the result of substratum influence. Joseph (1983: 179–212) examines the evidence and arguments, and concludes that the substratum hypothesis is at best conjectural, and that the situation is more likely to have arisen at a much later stage as a result of language‐particular internal factors which plausibly interacted with contact between individual languages.

3.6  Absence of Infinitival Complementation in Southern Italy In parts of southern Italy (central‐southern Calabria, northeastern Sicily, and the Salento), there is little or no infinitival complementation (Ledgeway 2006: 112): complementation rather involves a finite clause. There are obvious parallels with the situation in Romanian described in the previous section; however, the explanation in this case is less complex. It is generally accepted that this is the result of influence from Greek, although whether this is a Greek which continues the language of Magna Graecia (and is hence a substrate), or the language spoken by Greeks who emigrated to the area from the Byzantine Empire (and is therefore a superstrate or an adstrate), or a conflation of the two is a matter of dispute (which has sometimes assumed nationalist overtones). (The infinitive disappeared progressively from Greek between the tenth and fifteenth centuries – for some discussion, see Joseph 1983: 81–4). Greek has also influenced the lexicon of these Romance varieties. For discussion of both the origins of southern Italian Greek and its influence on Italo‐Romance, see Fanciullo (2001).

4 Superstratum 4.1 Languages Between the late fourth and sixth centuries, the Western Roman Empire progressively collapsed in the face of Germanic migrations (often known as the “Germanic invasions,” although, despite some violence, much of this migration was peaceable). The earliest waves of Germanic settlers were speakers of East Germanic languages. The Vandals crossed Gaul in the early fifth century, and settled in Spain, before moving into North Africa and founding a kingdom there. They were followed by the Visigoths (western Goths), who sacked Rome in 410 ce, and subsequently founded kingdoms in southern Gaul and in Spain. The Ostrogoths (eastern Goths) held territory in the Balkans, but occupied Italy from the late fifth to the early sixth century. Some later groups of settlers spoke West Germanic languages. Langobards9 ruled much of Italy from the mid sixth to the late eighth century. They were displaced by the Franks, who had entered northern Gaul in the mid fifth century, where they remained most firmly established, although expanding into southern Gaul and Catalonia, as well as northern Italy. For a survey of the earlier period and fuller discussion, see Maas (2014: 12–25); for the Franks, see James (1988) and McKitterick (2008). One question which needs to be asked at the outset is: why did so many Germanic peoples, who were the socially dominant group and had their own language, adopt Latin when they migrated on to the territory of the Roman Empire? Latin was the language of an established culture and civilization, which could certainly be seen as more advanced than that of the Germanic tribes; but this sort of argument is ultimately subjective. The Franks, Langobards, and Goths had laws and literature, too. There is, however, one striking

432  John Charles Smith objective cultural difference between the two languages  –  unlike Germanic, Latin was ­written. Compare Toon’s comments on Old English (Toon 1983: 2): A preliterate people had conquered a new land and had established there a new social order. […] Members of a culture as pervasively literate as our own need to be reminded of the magnitude of that change. In oral societies, culture and history reside exclusively in memory; social and political structures depend entirely upon present and immediate ­balances of power. Spoken agreements are valid only as long as those who spoke them choose to remember them; material possessions are limited to what an individual can grasp and defend. Literacy makes possible a past independent of memory. Written documents outlive convenience and document ownership; bookkeeping makes possible more complex administrative structures; a king can make a permanent ally by granting a privilege for all time in a written charter.

However, this cannot be the whole story, as some Germanic languages that came into contact with Latin (such as the Old English that Toon is talking about) were not displaced by it, but rather evolved their own written form under its influence. There must have been other cultural and/or demographic factors at work, as well. For instance, it is probable that – like the Normans in England – the Germanic‐speakers in areas of the former Roman Empire where a Romance language survives, despite their power and influence, were not especially numerous, and that, unlike the Angles and Saxons who invaded Roman Britain, they were already well versed in Roman culture. Both of these factors may have favored the ultimate adoption of Latin. For a discussion of these and related issues, see Pohl (2014).

4.2 French There are words of Germanic origin in Spanish (see Dworkin 2012: 65–80) and Italian (see, for instance, Marazzini 2002: 151). However, the Romance language which was arguably most affected by its superstrate was French, and in what follows we shall use this language as a case study. The major superstratum influence on French is Germanic, specifically the West Germanic Frankish language spoken by the Germanic tribe which, as we have noted, occupied the northern part of what is now France from the mid fifth century onwards. We have very little direct knowledge of this language, although it is assumed to be the ancestor of modern Dutch. For discussion, see Quartermaine (2002). Other parts of Gaul were occupied by tribes speaking East Germanic languages – the Visigoths (albeit briefly) in the south and the Burgundians in the east; although Frankish influence spread into these areas, it seems to have been less intense. It has been argued that the tripartite linguistic division of Gaul/France into French (langue d’oïl), Occitan (langue d’oc) and Francoprovençal can be attributed to these different superstratum influences  –  see Jochnowitz (1973), who also points out that the linguistic boundaries correlate with non‐ linguistic differences, such as traditional legal systems, patterns of crop rotation, and the pitch of roofs on houses. Loanwords entering French from Germanic include nouns, adjectives, and verbs across a very wide range of semantic fields – a far from exhaustive list includes warfare (guerre ‘war’, heaume ‘(medieval) helmet’, brand ‘two‐handed sword’), rank (maréchal ‘marshal’, échevin ‘alderman’), emotions (haine ‘hatred’, orgueil ‘pride’, honte ‘shame’), parts of the body (hanche ‘hip’, échine ‘spine’), color adjectives (bleu ‘blue’, gris ‘gray’, blanc ‘white’). For further discussion, see Nyrop (1914: 7–15), Chaurand (1977: 25–36), Guinet (1982). Non‐lexical developments which have been ascribed to Germanic influence include the following.

Contact and the Romance Languages  433

4.2.1 Phonology The diphthongization of stressed vowels in tonic open syllables is a feature of the transition between Latin and many Romance languages. However, this process affected a greater range of vowels in French than in most other Romance varieties; in particular, mid‐high vowels were diphthongized, as well as (and almost certainly later than) mid‐low vowels. The greater range of diphthongization in northern Gaul has been linked to word stress – an increase in the length component of tonic stress in Late Latin/early Old French, which some have ascribed to Frankish contact influence (see Pope 1952: 102), although this external motivation is by no means proven. It is worth noting that a similar diphthongization of the vowels [e] and [o] is found in the history of Modern English (Lass 1999: 58) – so it is perfectly possible for this type of change to take place without any obvious external influence. Another, more plausible, Frankish influence on Old French phonology is the introduction of a voiceless compact fricative in certain loanwords (Nyrop 1914: 430–1; Wartburg 1969: 60). This may originally have had a velar (/x/) or a uvular (/χ/) articulation, but became a glottal fricative (/h/) in Old French (Pope 1952: 227). It is no longer a phoneme in the contemporary standard language, but many (for instance, Roegiest 2009: 110) have argued that it is the origin of the sandhi phenomenon known as h aspiré.

4.2.2 Syntax10 4.2.2.1 Verb‐second One feature of Old French which is often ascribed to Germanic influence is its verb‐second syntax (see Mathieu 2009). Moreover, the fact that French is in a minority of Romance languages which do not generally allow null subjects has been linked to the verb‐second syntax of this stage of the language. Wolfe (2018) demonstrates that verb‐second syntax is extremely widespread in medieval Romance, including in languages such as Sardinian, which are unlikely to have been influenced by Germanic. He also claims (Wolfe 2018: 133–4) that there is indeed a connection between verb‐second syntax on the one hand and, on the other, the nature of clitics and null realizations of certain categories, but that it is a complex and often indirect one. 4.2.2.2 Demonstratives Late Latin had a three‐term demonstrative system which broadly correlated with discourse‐ roles: iste speaker‐orientated; ipse interlocutor‐orientated; ille non‐participant‐orientated. Many – probably most – Romance languages have retained a reflex of this three‐term demonstrative system similar in function to the Latin one (see Ledgeway and Smith 2016). However, Old French reduced the number of demonstratives to two, omitting the middle term and recasting the relationship as one of proximity vs distance: cist ‘this’ and cil ‘that’. It has been argued that this was under the influence of Germanic, which has at most a two‐term demonstrative system. Once again, though, it is not obvious that this change is the result of contact influence. A similar change has affected standard Italian since the mid nineteenth century (again, see Ledgeway and Smith 2016), and is clearly not the result of contact. It is not impossible that contact was at work in Old French, but neither is it necessarily the case. 4.2.2.3  Indefinite Subject Pronoun French also has an indefinite subject clitic on ‘one’ (now also widely used as a first‐person plural ‘we’), which originates as the nominative case of the word for ‘man(kind)’ (< Latin homo). The existence of indefinite subject pronouns with a similar etymology in some Germanic languages (compare, for instance, German man) has led some scholars to claim that it results from the superstrate influence of Frankish). However, this cannot be

434  John Charles Smith demonstrated, and indeed a similar development is found in a variety of languages, including the Romance languages Occitan, Catalan, and some (mainly northern) dialects of Italian (where Germanic influence cannot totally be ruled out), but also Brazilian Portuguese (where a gente ‘people’ is found in a similar function, see Zilles 2002) and, outside Romance, contemporary London English (where man is acquiring a variety of pronominal values, including indefinite third person; see Cheshire 2013). For further examples and general discussion of the origin of such pronouns, see Heine and Song (2011). 4.2.2.4  Ordering of Noun and Adjective Whereas the majority of adjectives follow the noun in standard French, adjectives more commonly precede nouns in many northern oïl varieties. This is the normal order in Wallon (Hendschel 2012: 85–7), especially in eastern varieties; see also Bernstein (1991). In Norman, many adjectives, including color adjectives and monosyllabic adjectives, regularly precede the noun (Liddicoat 1994: 217–19); Jones (2001: 111–14) finds that in Jersey all color adjectives and four out of five other adjectives, including 35% of “long” adjectives, are preposed. Similar phenomena are found in Picard and Lorrain. ‘In all these cases, this word order has been ascribed to greater contact with Germanic at different stages of the history of the varieties concerned. 4.2.2.5 Case In general, the Latin nominal case system disappeared during the transition to Romance. However, a reduced nominal case system survived into Old French, although it became increasingly unstable, before disappearing completely (see Schøsler 1984). Whilst the case‐ forms themselves were not the result of Germanic influence (for discussion of a handful of apparent counter‐examples, see Pitz 2009), it has been suggested that the existence of a case system in Frankish may have inhibited the loss of the French case system (Jud 1907; Meillet 1931: 36–7). Ultimately, these claims are unproven; against them must be set the fact that a case system similar to that of Old French is also found in Occitan – which was subject to relatively little Germanic influence – where it survived longer than in French (Jensen 1994: 2–18).

4.3  Subsequent Germanic Influences Subsequent Germanic influences are more difficult to categorize. As a result of invasions ­between the late eighth and eleventh centuries, Norse has left a mark on the French of Normandy, but the influence is mainly toponymic (including the name of the region itself); it has also left a handful of loan words in standard French, of which the best‐known is vague ‘wave’. For discussion, see Nyrop (1914: 18–19) and Wartburg (1969: 74–5). Swiss Raeto‐Romance, which has been progressively losing ground to German over the last few centuries, has been heavily influenced by this language. For instance, Haiman (1988: 368) notes: “In the Swiss dialects [of Ræto‐Romance], almost certainly under massive and prolonged German influence, word order in principal assertive clauses is verb‐second.”

5  Superstratum and Adstratum in Romanian We consider Romanian separately when discussing superstratum and adstratum, since Romanian stands apart from the Romance languages of Western Europe, and it is not always easy (or, indeed, possible) to untangle the two influences. The linguistic history of Romania is complex, and can only be summarized very briefly here. The territory of what is now Romania broadly corresponds to historical Dacia, the western part of which became a short‐ lived Roman province (106–275 ce). This area was subsequently invaded by Ostrogoths,

Contact and the Romance Languages  435 Huns, and Avars, and then by speakers of early Slavonic during the seventh and eighth ­centuries. Many northern areas of what is now Romania came to be occupied during the Middle Ages by Hungarians, who originally practiced Catholicism, although many of them converted to Calvinism during the Reformation. Most urban centers were dominated by German‐speakers from Saxony, who became converts to Lutheranism. Although some high‐ class Romanian‐speakers adopted Hungarian ways, most ordinary Romanian‐speakers were members of the Slavonic Orthodox church, the language of which was Old Church Slavonic. Another medieval presence was that of the nomadic Turkic‐speaking Cumans. By the sixteenth century, most Romanian‐speakers lived under Ottoman Turkish suzerainty, although, by the eighteenth century, the economic and political life of the Ottoman Balkans was dominated by the Phanariots – ethnic Greeks, who spoke Greek and practiced Greek Orthodox Christianity. In the late eighteenth and early nineteenth century there was a growing awareness of the “Romanceness” of Romanian, along with the expansion of the language into new semantic domains, and this led to the adoption of some items from Latin and, in particular, a large number of French words. Thus, Latin, Hungarian, German, Slavonic, Turkish, Greek, and French have all been significant contact influences on Romanian. Exhaustive listing lies beyond the scope of this chapter (for fuller discussion, see Sala 1999: 109–15; Bolocan 1969; Mihăescu 1966, 1969; Isbășescu 1969; Tamás 1967; Wendt 1960; Gáldi 1939), but the following will serve as examples of common words from these sources. Early Latin ecclesiastical terms include cruce < crvcem ‘cross’ and biserică < basilicam (itself from Greek βασιλική) ‘church’ (although, in the absence of contemporary evidence, it is difficult to tell whether these are loans or internal developments from Latin). Other indicative loanwords are: bogat ‘rich’ (Slavonic бгатъ ‘rich’); bolnav ‘ill’ (Slavonic болнав ‘ill’); cartof ‘potato’ (German Kartoffel ‘potato’); fel ‘sort, type, kind’ (Hungarian féle ‘sort, type, kind’); vamă ‘customs’ (Hungarian vam ‘customs’); folos ‘advantage, benefit’ (Greek óφελος ‘benefit’); tavan ‘ceiling’ (Turkish tavan ‘ceiling’). Turkish, or Turkic, has also served as the conduit for words of Arabic or Persian origin – thus musafir ‘guest’ (Ottoman Turkish musafir, from Arabic musāfir ‘traveler’), dușman ‘enemy’ (Turkish (possibly Cuman) düşman, from Persian došman ‘enemy’). French is an exceptional case, accounting for a large proportion of the vocabulary of contemporary Romanian. Sala (1999: 113) points out that unambiguously French words constitute 7.47% (N=193) of his ‘representative’ Romanian word list; once multiple etymologies are taken into account (i.e. cognate words borrowed from French and Italian or Latin simultaneously), the figure rises to 22.12% (N=571). Examples include abajur ‘lampshade’ (abat‐jour), deputat ‘deputy, member of parliament’ (député), a declara ‘to declare’ (déclarer), the last two demonstrating adaptation to Romanian morphology. The list even includes adverbs, such as deja ‘already’ (déjà). Structural contact influences on Romanian have also been proposed. Daco‐Romanian exhibits two central vowel phonemes: mid /ə/ and high /ɨ/ or , the latter of which is not found in any other Romance language. These phonemes have been ascribed to both substratum and superstratum influence (see Nandriș 1963: 46–57, 1957; Poghirc 1969: 320): however, Sala (1999: 168–70) claims convincingly that they are internal developments from the Latin vowel system. And, alone amongst the Romance languages, Romanian exhibits a highly compositional numeral system above ‘ten,’ which is identical in structure to that of Slavonic and is therefore often assumed to be calqued on it. The teens are expressed by a numeral from ‘one’ to ‘nine’ followed by the preposition spre (nowadays ‘towards’, but here with an earlier meaning ‘on’) and the numeral zece ‘ten’: thus unsprezece ‘eleven’ (lit. ‘one on ten’), optsprezece ‘eighteen’ (lit. ‘eight on ten’) (the reduced forms cinsprezece ‘fifteen’, șaisprezece ‘sixteen’, and paisprezece ‘fourteen’ have all but ousted the full forms patrusprezece, cincisprezece, and șasesprezece). Meanwhile, multiples of ten are composed of a numeral from ‘two’11 to ‘nine’ followed by the plural of zece ‘ten’: thus douăzeci ‘twenty’ (lit. ‘two tens’). Once again, we generally find the reduced forms cinzeci for cincizeci ‘fifty’ and șaizeci for

436  John Charles Smith șasezeci ‘sixty’ – the full forms are rare. Another plausible contact influence is the feminine vocative ending ‐o, which cannot be derived from Latin, but which is identical to the vocative found in Slavonic. Its masculine counterpart ‐e may also be Slavonic, but may also derive from Latin. For discussion, see Rosetti (1968: 303–4). More generally, in an argument reminiscent of that proposed in respect of Germanic and Old French (see Section 4.1.2.5), it has been suggested that Slavonic influence may have acted as a ‘brake’ on the disappearance of the Latin nominal case system, which has not completely vanished from modern Romanian (see Niculescu 1981: 51, which contains further references; for a counterargument, see Rosetti 1968: 304). For detailed discussion of Slavonic influence on Romanian, see Densusianu (1901: 237–87).

6  Adstratum Further West 6.1 Latin As the language of the Catholic church, of scholarship, of certain types of literature, of the law, and, more generally, of educated exchange between speakers of different languages throughout the Middle Ages and for some time afterwards, Latin was a powerful adstratum influence on the Romance languages of Western Europe. A detailed survey is provided by Pountain (2011). The adoption of Latin words by the Romance languages of Western Europe gathered pace during the medieval period. For instance, many Latin elements already existed in thirteenth‐ century French, but it is commonly claimed that the adoption of such items thereafter is of a different order. “The borrowing of learned [Latin] loanwords, which had been an infiltration in Old French, became a stream in the fourteenth century and almost a deluge in the fifteenth and sixteenth centuries, when latinisation of language began to confer social distinction” (Pope 1952: §53; see also Brunot 1966: 566 and Chaurand 1977: 37–49, and, for comparable developments in Spanish, Dworkin 2012: 157–81). A key element in the adoption of Latin vocabulary was the progressive replacement of Latin by the vernacular in domains such as scholarship and legal proceedings from the sixteenth century onwards. A consequence of this development was that the vernacular languages had to acquire terminology in domains in which they had not hitherto been used, and one solution was simply to transfer the vocabulary of Latin. Nonetheless, these words were often accommodated to the phonology and morphology of the Romance language in question. For instance, a handful of Latin abstract nouns with the accusative ending ‐tatem had yielded French words ending in ‐té by regular sound change (compare bonté ‘goodness’ < bonitatem); the ending ‐té subsequently came to be used to form French equivalents of Latin ‐tatem words with no existing French counterpart (for instance, qualité ‘quality’ < qvalitatem). The learned origin of several such words is apparent: contrast, for instance, the stem‐vowel in “popular” sourd [suʁ] ‘deaf’ (Lat. svrdvm), which has been subject to regular sound change, and “learned” (borrowed) surdité [syʁdite] ‘deafness’ (Lat. svrditatem). The popular vs learned contrast is also in evidence elsewhere – for example, we find a number of doublets, with the same Latin etymon yielding a “popular” form, which, again, has undergone regular sound change, and a later, “learned” form, which has been borrowed with some accommodation. Examples, again from French, are frêle ‘frail, feeble’ (popular) vs fragile ‘fragile’ (learned), both from Latin fragilem ‘fragile’, and chose ‘thing’ vs. cause ‘cause’ (learned), both from Latin cavsam ‘business, cause’. Interestingly, we also find some “semi‐learned” words, in which some regular sound change has taken place, but other regular sound change has been blocked, arguably under the influence of the Latin cognate: thus French siècle ‘century’ < saecvlvm and peuple ‘people’

Contact and the Romance Languages  437 < popvlvm have both undergone diphthongization, but have not been subject to other expected changes (yielding **sieil and **peuble), possibly on account of their frequent usage in Latin texts, such as the liturgy (compare et in saecvla saecvlorvm ‘for ever and ever’). However, Latin influence was not restricted to the lexicon. Pountain (2011: 643–56) concludes that it is at least plausible that Latin also had an impact on the syntax of western Romance, discussing, inter alia, absolute constructions, the “accusative and infinitive”, hyperbaton, and adjectives preposed to the noun.

6.2  French and Italian As an important language of culture  –  including art, music, literature, and gastronomy  – during the Renaissance, Italian influenced many other languages (for its impact on Spanish, see Dworkin 2012: 139–56), but its influence on French was particularly striking, in part because of the marriage of King Henri II and the Florentine noblewoman Caterina de’ Medici, who was twice Queen Regent of France. Further discussion can be found in Hope’s lengthy and detailed survey of the reciprocal influence of French and Italian (Hope 1971).

6.3 Arabic As is well known, Spanish contains a number of Arabic loan words (García González 1993; Millar 1999; Viguera Molíns 2002)  –  about 4,000 items in all. However, most of these are nouns. Very few other categories are borrowed (but note the preposition hasta ‘until’ and the interjection ojalá ‘hopefully’), betokening a relatively low level of influence. However, the range of semantic fields covered by these nouns is great, including buildings (alcázar ‘citadel, fortress’, aljibe ‘cistern’), furnishings (alfombra ‘carpet’, almohada ‘pillow’), rank (alcalde ‘mayor’, alférez ‘ensign, second lieutenant’, alguacil ‘sheriff, master‐at‐arms’), and crops (albaricoque ‘apricot’, algodón ‘cotton’). Note that, as in the examples just given, the vast majority of Arabic nouns entering Spanish do so with the definiteness prefix al‐ (Odisho 1997), implying that the speakers adopting these items had little or no knowledge of the grammatical structure of Arabic. The evidence therefore points to these being adstratum items entering Castilian. For further discussion, see Dworkin (2012: 81–117). An Ibero‐ Romance language known as Mozarabic was spoken in the parts of Spain conquered by the Arabs, and here Arabic loanwords were arguably superstratum items (see Galmés de Fuentes 1983). Reflexes of many of the Arabic words which entered Spanish are also found in Portuguese; but, possibly as the result of the earlier completion of the Reconquista (1249 in Portugal, 1492 in Spain), there are fewer of them. Thus, the normal word for the referents of several Arabic loan words in Spanish is of Latin origin in Portuguese – compare ‘sewer’ (Sp. alcantarilla < Ar. al‐qanṭara ‘the bridge’ plus Spanish diminutive suffix; Port. esgoto, by back‐ formation from esgotar < *exgvttare ‘to drip out’). However, there are occasional examples of the reverse, for instance ‘lettuce’ (Port. alface < Ar. al-xass, Sp. lechuga < Lat. lactvcam). Arabic was also an influence on Sicilian during the period of Arab invasion and subsequent rule (827–1091) – see Caracausi (1983) and Pellegrini (1989). And, as an important language of science and philosophy, Arabic also had a more general adstrate influence in much of medieval Europe, yielding words such as ‘algebra’ (Sp. algebra, Fr. algèbre, It. algebra). A final wave of Arabic influence on Romance relates to the French colonization of the Maghreb, beginning with Algeria in 1830, followed by Tunisia in 1881 and Morocco in 1912. French became a significant influence on the Arabic spoken in the area, but was itself influenced by North African Arabic. Loan words that date from this period include toubib ‘doctor’ (at first ‘military doctor’, then, colloquially, just ‘doctor’) from North African Arabic ṭbīb (Classical Arabic ṭabīb) ‘doctor’ and chouia ‘a spot (small quantity) of’, again colloquial, from North

438  John Charles Smith African Arabic šuya (Classical Arabic šuwayya) ‘a little’. The migration of many Maghrebians to France during the twentieth and twenty‐first centuries led to more North African Arabic words entering the colloquial speech of the young. Recent work showing that the speech of young urban French‐speakers of immigrant descent in Strasbourg may have significant distinguishing characteristics, some of which are derived from North African Arabic (but also from Berber) is presented in Marchessou (2018).

6.4 English English has been a more recent adstrate influence on Romance, as it has been on other languages globally. In fact, the earliest English loanwords in Romance are the names of the cardinal points of the compass in French – nord ‘north’, sud ‘south’, est ‘east’ ouest ‘west’, which entered French from Old English in the Middle Ages. Otherwise, English loanwords begin to enter Romance in the late seventeenth and eighteenth century, often in the fields of political and social organization (compare French comité ‘committee’, partenaire ‘partner’). The rate of borrowing increased during the nineteenth century, and, with the internationalization of English from the mid twentieth century onwards, the influence of this language on Romance (as on other languages) has become pervasive. For English lexical influence on Spanish, see Dworkin (2012: 212–29); for French, see Wise (1997: 79–102); for Italian, see Klajn (1972). For some discussion of possible syntactic influences from English, see McLaughlin (2011; 2018).

6.5  New World Languages When Spanish, Portuguese, and French spread to the New World, they came under the influence of native American languages, and were especially prone to borrow lexical items for flora and fauna which speakers were encountering for the first time: thus Spanish llama ‘llama’ from Quechua and tomate ‘tomato’ and chocolate ‘chocolate’ from Nahuatl, but also French caribou ‘reindeer’ from Algonquian (contrast European French renne, borrowed from Norwegian or Swedish). However, items in other semantic fields were also borrowed. For Spanish, see especially Dworkin (2012: 200–11), who notes (2012: 203–4): “Almost all the New World words that have come to form part of the vocabulary of general Spanish (and which have spread subsequently into many European languages) come from the Arawak/ Taíno group, Nahuatl or Quechua. Borrowings from other indigenous languages turn up almost exclusively in those varieties of New World Spanish that are in contact with the donor language.” Structural influences have also been proposed, but often with little or no evidence; for a survey, see Lipski (1994: 63–92). For further discussion, see Sippola (this volume).

6.6 “Romancization” Stolz, Bakker and Salas Paloma (2008) coin the term “Romancization” to describe the contact influence of Romance on other languages. This is mainly the result of invasion and colonization, and is the reverse of the situation described in the preceding section: indigenous languages acted as an adstrate or substrate on Romance; but Romance also acted as an adstrate or superstrate on indigenous languages. Spanish influenced Meso‐ American and South American languages – for instance, Nahuatl in Mexico (San Giacomo and Peperkamp 2008), Quechua in Ecuador, Peru, Bolivia, and an area of northern Argentina (Bakker, Gómez Rendón and Hekking 2008), Aymara in Peru, Bolivia, and north‐east Chile (Briggs 1981), and Guaraní in Paraguay (Bakker, Gómez Rendón and

Contact and the Romance Languages  439 Hekking 2008; Gómez Rendón 2019). It also influenced languages of the Philippines, including Tagalog, which later served as the basis for the national language (Cuadrado Múñiz 1972). Portuguese influenced the indigenous languages of Brazil (Dienst 2008). French has been an influence on several native North American languages (Bakker and Papen 2008) and, at a later date, on the Arabic of the Maghreb (Sayahi 2014) and Lebanon (Abou 1962), and on many sub‐Saharan African languages (see, for instance, Dumont 1983). Even the short‐lived Italian occupation of Eritrea, Somalia, and (briefly) Ethiopia led to loan words entering Tigrinya, Somali, and Amharic (Tosco 2008). For further discussion, see the references cited. In another context, Spanish was arguably a substratum of US English; it was a language spoken in Florida and the southwestern states before they were conquered by English‐ speakers, and some words (pueblo, adobe) entered American English as a result. In practice, however, it is difficult to disentangle substratum and adstratum elements here. See Bentley (1932), Lodares (1996), and Klee and Lynch (2009: 193–262).

6.7  The Influence of French on English One of the most significant examples of contact between Romance and non‐Romance ­languages is the influence of French on English. England was conquered by the French‐ speaking Normans in 1066, and Norman French (or “Anglo‐Norman”) was the dominant official language of the country until the late fourteenth century. Although the ruling class eventually adopted English as their language, it was an English which had absorbed many Norman French words. As Blake (1992: 17) observes: “Many of the words borrowed at this stage have become so much part of the language that many modern speakers find it difficult to think of them as loan words.” Anglo‐Norman also influenced English phonology and morphology. In addition to Anglo‐Norman, arguably a superstrate, the nascent Parisian standard was a significant adstrate influence on English from the Middle Ages onwards. The role of French in England during the Middle Ages and the Renaissance is discussed by Kibbee (1991). Subsequent adstrate borrowings from French also abound  –  according to Chirol (1973: 30–2), the mythopœic function of these items, which tend to cluster in semantic fields related to high culture or prestige, is often as important as their meaning. For fuller discussion of French influence on English, see McColl Millar (this volume).

7  Contact, Sociolinguistics, and Structure We turn now to a consideration of Romance evidence for some recent theories concerning the relationship between contact, sociolinguistics, and language structure (also discussed in Smith, forthcoming).12

7.1  Simplification and Complexification An abiding debate in the history of linguistics has centered round the question: Do l­ anguages grow more simple or more complex? The debate on simplification vs. complexification has often been anecdotal, but it has recently been placed on a firmer footing by Trudgill (2010, 2011a, 2011b), who begins by pointing out a paradox. Sociolinguists claim that language contact leads to simplification (often through the process known as koineization or dialect leveling  –  see, for instance, Siegel 1985; Kerswill 2013), whereas linguistic typologists see contact as a cause of complexification. Who is right?

440  John Charles Smith Trudgill suggests that an answer to this question will vary according to sociolinguistic circumstances, and proposes the following typology: 1. Languages exhibiting isolation (little or no contact): these languages tend to retain existing complexity, and may become more complex through endogenous complexification. Icelandic is a good example, having retained its nominal case system, “quirky” subject constructions, and various other complexities. The mainland Scandinavian languages (Danish, Norwegian, Swedish), which derive from the same ancestor, but which have had much greater contact with other languages (especially Low German during the Hanseatic period), have lost nominal case and “quirky” subjects; in addition, many (but not all) of their dialects have reduced the number of genders from three to two. 2. Languages exhibiting short‐term contact involving L2 learning by adults: these languages will exhibit simplification, often drastic. Extreme examples are pidgins and creoles. 3. Languages exhibiting long‐term contact where territory is shared, and involving child bilingualism: these languages will exhibit complexification, mainly through additive borrowing (exogenous complexification). The extreme example is the Sprachbund or ‘linguistic area’, such as the Balkan Sprachbund (see Sandfeld 1930), where a variety of languages which are only very distantly related – Romanian, Greek, Albanian, Bulgarian – share a number of characteristic features which are not found in much more closely related ­languages (for discussion, see Joseph 2010 and this volume). How does this typology fare when applied to Romance? As far as the first category (isolation leading to retention of complexity and/or endogenous complexification) is concerned, the point raised by Trudgill (2011b: 238) in his discussion of Germanic languages applies equally well: “All the major standard language varieties in Europe today are relatively high‐contact koinés and creoloids which are the result in part of simplification resulting from dialect contact. […] [C]ases of spontaneous, non‐additive complexification should be looked for in relatively isolated non‐standardized low‐contact varieties of modern European languages in comparison with their local standards. We can also look at those languages which are spoken by small groups of speakers in tightly‐knit communities and compare them to related languages.” Trudgill goes on to present instances of endogenous complexification in Germanic; the following are some plausible examples of similar ­phenomena in Romance: 1. Cardinal numerals above ‘one’ (un (m), une (f)) do not vary for gender in French; exceptionally, we find separate gender forms for ‘two’ and ‘three’ (m [dø], [tɾwa]; f [døs], [tɾwas], respectively) in the Tracadie dialect of Acadian French (Cichocki 2012: 227). 13 2. In French, the second person plural form of the verb exhibits a single inflection ‐ez (/e/), regardless of whether it is functioning as a true plural or as a “polite” singular; in the oïl varieties of central Lorraine, it has two separate inflections, one (generally /o(w)/) indicating plurality, the other (generally /e(j)/) indicating “politeness” (ALLR 1979–88: 1054). 3. Most Romance languages have a single imperfect form of the verb. However, southern Lorraine varieties have innovated a so‐called “second” imperfect (cf. té chantéïe pan‐ Romance imperfect (< Lat. cantabas) vs. té chantézor “second” imperfect, both glossed as ‘you were singing’). It has been argued that this “second” imperfect results from the incorporation into the verb of a following adverb hora (Latin) or ores (Lorrain) ‘now’. Adam (1881) analyzes this form as a proximate or hodiernal imperfect; Russo (2017) suggests that the contrast between the pan‐Romance imperfect and this “second” imperfect corresponds to a distinction between “larger conceptual distance” and “smaller conceptual distance,” although this basic contrast is interpreted in a variety of ways, according to

Contact and the Romance Languages  441 area, corresponding to more remote past vs more recent past, evidentiality vs direct ­perception, or irrealis/non‐factive vs realis/factive. 4. In some Wallon varieties, the second person pronouns are té (“familiar” singular) and vos ‘vous’ (plural and “polite” singular), with vos also exhibiting a form vos‐ôtes lit. ’you others’ when plural; however, because of the use of vos as a ’polite’ singular, vos‐ôtes has been reinterpreted as a “polite” plural, and an analogical “familiar” plural pronoun tés‐ ôtes has been created from té (Remacle 1952: 243). 5. In the variety of Ripatransone in the Italian Marche, all finite verbs agree with their subject in gender, regardless of person or number (see Parrino 1967; Loporcaro 2018: 309–10). Loporcaro describes this agreement pattern as “spectacular.” 6. Exceptionally for Romance, the Lombard variety spoken in Monno, in the far north of the Italian province of Brescia, exhibits a type of “do‐support,” similar, but not identical, to the use of do as an auxiliary verb in English. See the study by Benincà and Poletto (2004), who note (2004: 56n.5) as a possible explanation for the existence of this phenomenon that “[u]ntil 1963 the village was reachable only by a footpath; contacts with people speaking other varieties were rare.” Short‐term contact, involving L2 learning, leads to simplification. The extreme example of creoles has already been mentioned. Other Romance examples include Angolan Vernacular Portuguese and Brazilian Vernacular Portuguese, both of which have undergone changes that “can be characterized as structural reduction: reduced morphological marking for person or tense on verbs, for number on nouns and other elements in the NP, or for case on personal pronouns” (Holm 2009: 119). Similar developments can be found in the French spoken in many African countries (Daff 2001; Ploog 2002, 2006; Ngamountsika 2007; Loussakoumounou 2009). The northern Sardinian variety of Sennorese has lost morphological distinction of gender in the plural through contact with Sassarese (Loporcaro 2018: 295– 7). In koineization (Siegel 1985; Kerswill 2013), the contact is between varieties of the same language rather than between different languages, although the outcome is the same – simplification and/or loss of marked variants. Probably the most celebrated example from Romance comes from Spanish – Tuten (2003) points to massive simplifications of the Castilian morphological system as a result of three broad phases of dialect contact, as a result of which, for instance, its verbal morphology is in many ways simpler than that of other Ibero‐Romance languages. But, as pointed out by Trudgill (2011b: 238 – see p. 440), virtually all standard languages are to some extent the result of koineization. Long‐term contact involving child bilingualism on shared territory, leads to complexification through additive borrowing. Romanian membership of the Balkan Sprachbund has already been mentioned. We should also consider the case of Istro‐Romanian, spoken by a community now numbering no more than 500 on the Istrian peninsula, which has long been bilingual in Croatian. Maiden (2006), drawing on work by Kovačec (1963, 1966, 1968), notes, “it is likely that there has been a stable coexistence of the two languages over several centuries, with Istro‐Romanian limited to village and family life, and Croatian being employed in wider public spheres.” In his discussion, he focuses on “what one might term ‘structured accommodation,’ whereby the penetration into Istro‐Romanian of a Croatian word or grammatical phenomenon is systematically attached to a particular semantic or structural context, giving rise to distributional patterns which are native neither to Istro‐Romanian nor to Croatian, but a product of the encroachment of the latter.” He gives three examples. 1. For numerals between ‘five’ and ‘eight,’ Istro‐Romanian uses both the “indigenous” Romance items and items borrowed from Croatian, but they are not interchangeable: the Croatian numerals appear in all and only “lexical measure phrases” (i.e. with nouns indicating time, weight, distance, etc.); Romance numerals appear in all and only other

442  John Charles Smith contexts. Thus (to quote one of Maiden’s examples) /ʃɑpte kɑse/ ~ */sedəm kɑse/ ‘seven houses’ vs. /sedəm let/ ~ */ʃɑpte let/ ‘seven years’. 2. Istro‐Romanian has developed the systematic aspect‐marking which is typical of Slavonic languages (and not usual in Romance), whereby each verb has both perfective and imperfective aspect forms. Sometimes, both forms are borrowed from Croatian; in other instances, a Romance root is common to both forms, but the perfective is formed by adding a Croatian prefix. However, in many cases, the Romance root provides the imperfective aspect form and the Croatian root the perfective: examples are ‘to sleep’ (imperfective /durmi/, perfective /zaspi/) and to drink (imperfective /bɛ/, perfective /popi/). 3. Finally, Istro‐Romanian, alone amongst Romanian varieties, has acquired a morphological distinction between adjectives and adverbs. In both Romanian and Croatian, adverbs are normally identical to some form of the adjective (the masculine singular in Romanian, the nominative/accusative of the neuter singular in Croatian). In addition to borrowing some Croatian neuter singular adjectival forms as adverbs corresponding to etymologically Romance adjectives (e.g. /teʃko/, ‘heavily’, alongside /ɣrev/ ‘heavy’), Istro‐Romanian has also taken the final ‐o which is characteristic of Croatian neuter singulars and added it to native Romance adjective stems to form an adverb (e.g. /plin/ ‘full’, /plino/ ‘fully’). Complex sociolinguistic circumstances may sometimes exist in which both types of contact effect seem to apply. For instance, Bolivian Spanish, almost certainly under the influence of Aymara and/or Quechua, has innovated an evidential form of the verb, which is absent from other varieties of Spanish (Laprade 1981: 222–5; Lipski 2014: 51–2). This represents complexification through additive borrowing, and tallies with the fact that there has been a long period during which substantial numbers of speakers have been bilingual in Aymara or Quechua and Spanish. However, this evidential is expressed not with a new form, but by the form which in other varieties of Spanish is the exponent of the pluperfect indicative (e.g. habían hecho lit. ‘they had done’). In some varieties of Bolivian Spanish, this form has lost its original function and can now only convey evidentiality – there is no pluperfect indicative tense. This represents simplification, at least of the tense system, and is consistent with the fact that, in addition to bilinguals, there have long been (and still are) many Bolivians who learn Spanish as a second language. So the system of moods is complexified, whilst the system of tenses is simplified, and the overall number of forms remains the same. It appears that both types of change (additive borrowing and simplification) are at work here simultaneously, arguably because we find both types of relationship between the languages (bilingualism and adult learning of Spanish) in the speech community. It is plausible (although speculative) to suggest that these developments were sequential, as follows. Spanish was originally acquired as an L2 by adult speakers of Aymara or Quechua. As there is no pluperfect in these languages, these learners simplified the Spanish tense system correspondingly. However, an awareness of the form persisted, and when long‐term bilingualism developed in at least part of the speech community, it was co‐opted in a new function (which already existed in Aymara and Quechua), according to the general principles of refunctionalization discussed by Smith (2011).

7.2  Code‐switching and Contact Vernaculars A phenomenon related to language contact is code‐switching, in which bilingual speakers switch language within and/or between sentences in the same discourse. Code‐switching is a complex subject, and once again, it is not the purpose of this chapter to provide a detailed synopsis: for surveys, see Myers‐Scotton (1997), Gardner‐Chloros (2009), and Gardner‐Chloros

Contact and the Romance Languages  443 (this volume). What concerns us here is that code‐switching may become institutionalized, yielding a contact vernacular, a stable speech variety in which elements of two languages coexist in broadly equal measure – for discussion, see Matras (2000) and Bakker (this volume). Striking examples of contact vernaculars involving Romance are the group of Ecuadorian varieties known as Media Lengua, which derive most of their lexicon from Spanish and most of their morphology and syntax from Quechua (Muysken 1997) and Michif, spoken mainly on the Canadian prairie, where the structure of the noun‐phrase follows French and the structure of the verb‐phrase Cree (Bakker 1997; Bakker and Papen 1997). In some contact vernaculars, both languages involved are Romance. For instance, a Spanish‐Portuguese contact variety known as fronterizo or fronteiriço is spoken in northern Uruguay, near the border with Brazil. Penny (2000: 164–6) notes the following characteristics of this variety: a transitional vocalic system, shading from a Spanish‐like system in the south to a Portuguese‐ like system in the north; a transitional sibilant system, with the same characteristics; the use of Portuguese definite articles with Spanish nouns (e.g. todo o día), or, occasionally the reverse; the use of tú, rather than você, as the informal second person singular pronoun; generalization of the third person singular inflection of the verb to other persons; plural marking on the first element of the noun‐phrase only, with the other elements unmarked; mixed Spanish and Portuguese lexicon. And, in the early twentieth century, the massive migration of Italian‐speakers to Buenos Aires (where in parts of the city they represented 50% of the population) and Montevideo led to the emergence of a Spanish‐Italian contact vernacular known as cocoliche. Some of the most salient characteristics of this variety, discussed by Meo Zilio (1955a, 1955b, 1955c, 1956, 1958, 1959, 1960, 1970, 1993) are: mixed Italian and Spanish phonetics and phonology; Italian noun lexemes with Spanish gender (la latte ‘milk’; contrast It. il latte, Sp. la leche); Italian noun lexemes with Spanish plural in ‐s; earlier or non‐standard Italian forms reintroduced under the influence of Spanish (alcun amico ‘some friend’, contrast It. qualche amico, Sp. algún amigo); simplification of the Italian system of articles; word order, use of prepositions, use of articles, use of verb forms, and derivational suffixes all calqued from Spanish using Italian forms; replacement of Italian lexemes by cognates of Spanish lexemes (passeri lit. ‘sparrows’ for ‘birds’, contrast It. uccelli, Sp. pájaros; largo lit. ‘wide’ for ‘long’; contrast It. lungo, Sp. largo).

7.3  Contact in Diaglossia and Advergence Contact may play a role in the formation of a standard language. For instance, on account of the cultural, commercial, and administrative importance of Paris during the Middle Ages, the variety of French spoken there came to serve as a lingua franca throughout northern France; but it is important to realize that the “language of Paris” was, by the time of its subsequent “selection” (for the use and meaning of this term see Haugen 1966) as a standard, the result of koineization (Siegel 1985; Kerswill 2013), resulting from massive in‐migration over several centuries (Lodge 1993, 2004; for discussion of some relevant data, see Smith and Sneddon 2002). Once a standard language has become established, contact between this language and local varieties may lead to the creation of intermediate codes, often known as “regional standards.” For instance, Pellegrini (1960) suggested that Italian‐speakers in fact had access to four different linguistic codes, rather than the two codes of diglossia – moving from local to standard, these were: local dialect, regional dialectal koiné, regional Italian, and standard Italian. In similar vein, Carton (1981), discussing the Picard‐speaking area of northern France, distinguished four levels of speech: local patois, local or dialectal French, regional French, and general French (although it is doubtful whether this four‐way distinction has  survived into the twenty‐first century  –  see Hornsby 2006). More recently, Hernández‐Campoy and Villena‐Ponsoda (2009: 192–4) have claimed that a similar situation

444  John Charles Smith exists in western Andalusia, involving local vernacular, Seville regional standard, common Spanish (a generalized form of Spanish which has absorbed some southern features), and standard (Castilian) Spanish. Auer (2005: 22), following Bellmann (1997), uses the term “diaglossia” to refer to a situation in which intermediate varieties intervene between the standard language and local speech, noting that these “intermediate forms often fulfil a sociolinguistic function by enabling their users to act out, in the appropriate contexts, an identity which could not be symbolised through the base dialects (which may have rural, backwardish or non‐educated connotations) nor through the national standard (which may smack of formality and unnaturalness and/or be unable to express regional affiliaton).” Historically, the development of diaglossia is characteristic of the late nineteenth or early twentieth century. Auer charts its rise and fall, from exoglossic diglossia (“Type Zero”), through a stage of endoglossic diglossia where the “H” language is largely confined to writing (“Type A”), to spoken endoglossic diglossia (“Type B”), diaglossia (“Type C”), and post‐diaglossia (“Type D”), in which there is little or no variation between local, regional, and national varieties. Post‐diaglossia tends to arise through the process of unidirectional convergence, or “advergence” (“Advergenz”: see Mattheier 1996: 34) of the local and regional varieties with the national language. It has been argued that contemporary France is a largely post‐diaglossic speech community (see Boughton 2013: 132; Armstrong and Pooley 2010: 150–204; Durand, Eychenne, and Lyche 2013; Jones and Hornsby 2013; and Mooney 2016). On the other hand, diaglossia continues to be prevalent in many parts of Italy (see Cerruti 2011: 19). A detailed study of diglossia and related issues with reference to Romance is Kabatek (2016).

8 Conclusion At the end of this brief and far from exhaustive survey, it will nevertheless be clear that contact influence from other languages has played a complex and important role in the evolution of Romance, and that the Romance languages, in turn, have come into contact with other languages, and influenced them. Nonetheless, it is also fair to claim that contact influence on Romance has often been overstated. For the most part, non‐lexical substrate influences on Latin and early Romance remain at best unproven and at worst implausible. Although lexical substrate influences are easier to adduce, some of those proposed are fanciful. We are on slightly firmer ground with early influences from superstrate languages, although here, too, structural influences can be difficult to demonstrate. But, whilst we can rarely be certain in ascribing structural phenomena in Romance simply to contact, and whilst, in many instances, internal structural change must be the default hypothesis, it is not impossible that, in some cases, contact may have played the role of catalyst, facilitating or accelerating an internal structural development. More recent contact influences are less controversial. Particularly interesting is the dual role of Latin as not only the parent language of Romance but also a subsequent contact influence. Moreover, observable contact change in Romance provides us with valuable evidence for the nature of structural and sociolinguistic change. As Romance spreads, contact vernaculars develop, involving a Romance language and a language from a different, often unrelated, family, or two Romance languages. Koineization, resulting from dialect contact following in‐migration to an important cultural and commercial center, often plays a role in the development of a lingua franca, and such varieties may sometimes form the basis for a subsequent “standard” Romance language. Once the “standard” language has taken root, contact with local vernaculars may lead to the emergence of regional varieties of this standard. We have also seen that Romance provides evidence to support Trudgill’s ­sociolinguistic typology of contact.

Contact and the Romance Languages  445

NOTES 1 I am grateful to John Coleman, Adam Ledgeway, Martin Maiden, Stephen Parkinson, Peter Trudgill, and Sam Wolfe for information and helpful comments. Errors and shortcomings are entirely my own. 2 Adams (2007: 113) notes: “[L]anguage contact influenced the Latin of rural parts of Italy in the Republic. In the transitional stage when speakers were shifting from (say) Oscan to Latin, some language mixing took place […], either because of imperfect learning of the new language, or because of fading knowledge of the original language and an inability to use it uncontaminated by Latin inflections, or because of conservative retention of old elements in the new language […]. The influence of Oscan or other Italic languages on Latin is limited, and most clearly identifiable in lexical items. There is no evidence for forms of Latin displaying structural modification under Italic influence.” 3 Lepontic may have been a variety of Gaulish, although this is not proven. 4 Celtiberian is a Celtic language which shows some similarities with modern Irish, although they seem not to be closely related (see Jordan Cólera 2004). The Celtic languages of this period also have some affinities with the Italic languages, of which Latin is one (see Eska this volume). 5 Our knowledge of Gaulish is limited to several hundred largely fragmentary inscriptions. Gaulish is related, at some distance, to modern (insular) Celtic languages, particularly Welsh, although the precise details of the relationship remain unclear. As stated in note 4, it also has some affinities with Italic. For discussion, see Lambert (1994). 6 Even superstratum and adstratum languages may sometimes present this type of difficulty. The fact that we often have modern languages that are related to the earlier ones does not always help us to disentangle the problems. 7 Unless otherwise stated, Latin noun etyma are quoted in the accusative singular, as this is the form which generally gives rise to the modern Romance item. 8 This fact did not stop Menéndez Pidal (1929) from claiming, quite implausibly, that it was the result of Oscan influence on the Latin spoken in and around the Pyrenees – see Rohlfs (1970: 155–6) 9 Langobards are also known as Lombards; the former term is preferable, in order to avoid confusion with the present‐day inhabitants of the Italian region or linguistic area of Lombardy. 10 For a general survey of Frankish influence on French syntax, see Hilty (1975). 11 ‘Two’ is in the feminine (două), agreeing with the word for ‘ten’; Romanian numerals apart from ‘one’ and ‘two’ do not vary for gender. 12 Some of the material in this section is copyright © Cambridge University Press 2020, and is used here with permission. 13 It should be stressed that, although distinct gender forms for ‘two’ exist in some other Romance languages, such as Portuguese and Romanian, they are inherited from Latin. The Acadian forms represent a complete innovation, and cannot be traced back to Latin.

REFERENCES Abou, Sélim 1962. Le Bilinguisme arabe–français au Liban: essai d’anthropologie culturelle. Paris: Presses universitaires de France. Adam, Lucien 1881. Les Patois lorrains. Paris: Maisonneuve. Adams, J. N. 2003. Bilingualism and the Latin Language. Cambridge: Cambridge University Press. Adams, J. N. 2007. The Regional Diversification of Latin 200 BC–AD 600. Cambridge: Cambridge University Press.

Agostiniani, Luciano and Luciano Giannelli 1983. Fonologia etrusca, fonetica toscana: il problema del sostrato. [Etruscan phonology, Tuscan phonetics: the problem of substrate] Florence: Olschki. ALLR – Lanher, Jean, Alain Litaize, and Jean Richard 1979–1988. Atlas linguistique et ethnographique de la Lorraine romane. Paris: CNRS. Armstrong, Nigel and Tim Pooley 2010. Social and Linguistic Change in European French. London: Palgrave Macmillan.

446  John Charles Smith Auer, Peter 2005. Europe’s sociolinguistic unity, or: a typology of European dialect/standard constellations. In Nicole Delbecque, Johan van der Auwera, and Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative,. Berlin: De Gruyter Mouton, pp. 7–42. Bakker, Dik, Jorge Gómez Rendón, and Ewald Hekking 2008. Spanish meets Guaraní, Otomí and Quichua: a multilingual confrontation. In Thomas Stolz, Dik Bakker, and Rosa Salas Paloma (eds.), Aspects of Language Contact: New Theoretical, Methodological and Empirical Findings with Special Focus on Romancisation. Berlin: De Gruyter Mouton, pp. 165–238. Bakker, Peter 1997. A Language of Our Own: The Genesis of Michif, the Mixed Cree‐French Language of the Canadian Métis. Oxford: Oxford University Press. Bakker, Peter and Robert Papen 1997. Michif: a mixed language based on Cree and French. In Sarah G, Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: John Benjamins, pp. 295–363. Bakker, Peter and Robert Papen 2008. French influence on the native languages of Canada and adjacent USA. In Thomas Stolz, Dik Bakker, and Rosa Salas Paloma (eds.), Aspects of Language Contact: New Theoretical, Methodological and Empirical Findings with Special Focus on Romancisation, Berlin: De Gruyter Mouton, pp. 239–87. Bellmann, Günter 1997. Between base dialect and standard language. Folia Linguistica 32: 23–34. Benincà, Paola and Cecilia Poletto 2004. A case of do‐support in Romance. Natural Language and Linguistic Theory 22: 51–94. Bentley, Harold W. 1932. A Dictionary of Spanish Terms in English, with Special Reference to the American Southwest. New York: Columbia University Press. Bernstein, Judy. 1991. DPs in French and Walloon: evidence for parametric variation in nominal head movement. Probus 3: 101–26. Bithell, Jethro 1952. German Pronunciation and Phonology. London: Methuen. Blake, Norman 1992. Introduction. In Norman Blake (ed.), The Cambridge History of the English Language, vol. II: 1066–1476. Cambridge: Cambridge University Press, pp. 1–22. Blake, Robert J. 1989. Sound change and linguistic residue: the case of [f‐] > [h‐] > [Ø]. In Thomas J. Walsh (ed.), Synchronic and Diachronic Approaches to Linguistic Variation and Change (Georgetown University Round Table on

Languages and Linguistics 1988)., Washington, DC: Georgetown University Press, pp. 53–62. Bolocan, G. 1969. Contactul românei cu slava. [Romanian contact with Slavonic] In Alexandru Rosetti et al. (eds.), Istoria limbii române: Volumul II. [History of the Romanian language. Volume II] Bucharest: Academia Republicii Socialiste România, pp. 372–4. Boughton, Zoë 2013. Locating variation in French: geolinguistic patterns, levelling and the “French exception.” In Mari C. Jones and David Hornsby (eds.), Language and Social Structure in Urban France. Oxford: Legenda, pp. 119–32. Brâncuș, Grigore 1983. Vocabularul autohton al limbii române. [Indigenous vocabulary in Romanian] Bucharest: Editura Științifică și Enciclopedică. Briggs, Lucy T. 1981. Aymarization: an example of language change. In M. J. Hardman, M. J. (ed.), The Aymara Language in its Social and Cultural Context: A Collection of Essays on Aspects of Aymara Language and Culture. Gainesville FL: University Presses of Florida, pp. 127–45. Brunot, Ferdinand 1966 [1905]. Histoire de la langue française des origines à nos jours, tome 1: De l’époque latine à la Renaissance (new edition). Paris: Armand Colin. Buck, Carl Darling 1904. A Grammar of Oscan and Umbrian, with a Collection of Inscriptions and a Glossary. Boston: Ginn. Caracausi, Girolamo 1983. Arabismi medievali di Sicilia. [Medieval Sicilian Arabisms] Palermo: Centro di Studi Filologici e Linguistici Siciliani. Carton, Fernand 1981. Les parlers ruraux de la région Nord–Picardie: situation sociolinguistique. International Journal of the Sociology of Language 29: 15–28. Cerruti, Massimo 2011. Regional varieties of Italian in the linguistic repertoire. International Journal of the Sociology of Language 210: 9–28. Chaurand, Jacques 1977. Introduction à l’histoire du vocabulaire français. Paris: Bordas. Cheshire, Jenny 2013. Grammaticalisation in social context: the emergence of a new English pronoun. Journal of Sociolinguistics 17: 608–33. Chirol, Laure 1973. Les “Mots français” et le mythe de la France en anglais contemporain. Paris: Klincksieck. Cichocki, Wladyslaw 2012. An overview of the phonetics and phonology of Acadian French spoken in northeastern New Brunswick

Contact and the Romance Languages  447 (Canada). In Randall Gess, Chantal Lyche, and Trudel Meisenburg (eds.), Phonological Variation in French: Illustrations from Three Continents. Amsterdam: Benjamins, pp. 211–33. Coromines, Joan 1954. Diccionario crítico etimológico de la lengua castellana. [A critical etymological dictionary of the Castilian language] Madrid: Gredos. Cuadrado Múñiz, Adolfo 1972. Hispanismos en el tagalo: diccionario de vocablos de origen español vigentes en esta lengua filipina [Hispanisms in Tagalog. A dictionary of words of Spanish origin current in the Philippine language], rev. Antonio M. Molina. Madrid: Oficina de Educación Iberoamericana. Daff, Moussa 2001. Le français mésolectal oral et écrit au Sénégal: approche sociolinguistique, linguistique et didactique (Université Cheikh Anta Diop, Dakar Thèse de Doctorat d’État en Sciences du langage), Densusianu, Ovide 1901. Histoire de la langue roumaine: tome premier, Les origines. Paris: Leroux. Dienst, Stefan 2008. Portuguese influence on Kulina. In Thomas Stolz, Dik Bakker, and Rosa Salas Paloma (eds.), Aspects of Language Contact: New Theoretical, Methodological and Empirical Findings with Special Focus on Romancisation. Berlin: De Gruyter Mouton, pp, 287–98. Dumont, Pierre 1983. Le Français et les langues africaines au Sénégal. Paris: ACCT / Karthala. Durand, Jacques, Julien Eychenne, and Chantal Lyche 2013. On levelling and counter‐levelling in French: a phonological perspective. In Mari C. Jones and David Hornsby (eds.), Language and Social Structure in Urban France. Oxford: Legenda, pp. 58–68. Dworkin, Steven N. 2012. A History of the Spanish Lexicon: A Linguistic Perspective. Oxford: Oxford University Press. Elcock, W. D. 1975. The Romance Languages (revised with a new introduction by John N. Green). London: Faber. Fanciullo, Franco 2001. On the origins of modern Greek in southern Italy. In Mark Janse, Brian D. Joseph, and Angeliki Ralli (eds.), Proceedings of the Second International Conference on Modern Greek Dialects and Linguistic Theory. Patras: University of Patras, pp. 67–78. Fernow, Carl Ludwig 1808. Römische Studien: dritter Theil. Zürich: Gessner. Frellesvig, Bjarke 2010. A History of the Japanese Language. Cambridge: Cambridge University Press.

Gáldi, László 1939. Les Mots d’origine néo‐grecs en roumain à l’époque des phanariotes. Budapest: Magyar‐görög Tanulmanyok. Galmés de Fuentes, Álvaro 1983. Dialectología mozarabe. Madrid: Gredos. García González, Javier 1993. El contacto de dos lenguas: los arabismos en el español medieval y en la obra alfonsí. [Contact between two languages: Arabisms in medieval Spanish and in Alfonsine texts] Cahiers de linguistique hispanique médiévale 18: 335–65. Gardner‐Chloros, Penelope 2009. Code‐switching. Cambridge: Cambridge University Press. Glare, P. G. W. 1982. Oxford Latin Dictionary (combined edition). Oxford: Clarendon Press. Gómez Rendón, Jorge 2019. Language contact in Paraguayan Guaraní. In Anthony Grant (ed.), The Oxford Handbook of Language Contact. Oxford: Oxford University Press. Guinet, Louis 1982. Les Emprunts gallo‐romans au germanique (du Ier à la fin du Ve siècle). Paris: Klincksieck. Gurevich, Naomi 2004. Lenition and Contrast: The Functional Consequences of Certain Phonetically Conditioned Sound Changes. London: Routledge. Haiman, John 1988. Rhæto‐Romance. In Martin Harris and Nigel Vincent (eds.), The Romance Languages London: Routledge, pp. 351–90. Haugen, Einar 1966. Language, dialect, nation. American Anthropologist 68: 922–35. Heine, Bernd and Kyung‐An Song 2011. On the grammaticalization of personal pronouns. Journal of Linguistics 47: 587–630. Hendschel, Lorint 2012. Li Croejhete walone: contribution à une grammaire de la langue wallonne (2nd edition). Available at http:// docplayer.fr/13387061-Li-croejhete-walonecontribution-a-une-grammaire-de-la-languewallonne.html Henton, Caroline 1983. Changes in the vowels of Received Pronunciation. Journal of Phonetics 11: 353–71. Hernández‐Campoy, Juan Manuel and Juan Andrés Villena‐Ponsoda 2009. Standardness and nonstandardness in Spain: dialect attrition and revitalization of regional dialects of Spanish. International Journal of the Sociology of Language 196–7: 181–214. Hickey, Raymond 2017. The scope of English historical linguistics. In Laurel Brinton (ed.), Approaches to English Historical Linguistics. Cambridge: Cambridge University Press, pp. 12–41.

448  John Charles Smith Hilty, Gerold 1975. Westfränkische Superstrateinflüsse auf die galloromanische Syntax. Romanische Forschungen 87, 413–26. Holm, John 2009. The genesis of the Brazilian vernacular: insights from the indigenization of Portuguese in Angola. Papia 19, 93–122. Hope, T. E. 1971. Lexical Borrowing in the Romance Languages: A Critical Study of Italianisms in French and Gallicisms in Italian from 1100 to 1900. Oxford: Blackwell. Hornsby, David 2006. Redefining Regional French: Koinéization and Dialect Levelling in Northern France. Oxford: Legenda. Hornsby, David and Mari C. Jones. 2013 Exception française? Levelling, exclusion and urban social structure in France. In Mari C. Jones, and David Hornsby (eds.), Language and Social Structure in Urban France. Oxford: Legenda, pp. 94–109. Isbășescu, M. 1969. Influența germanică. [Germanic influence] In Alexandru Rosetti, et al. (eds.), Istoria limbii române: Volumul II. [History of the Romanian language. Volume II] București: Academia Republicii Socialiste România, pp. 368–70. Izzo, Herbert J. 1972. Tuscan and Etruscan: The Problem of Linguistic Substratum Influence in Central Italy. Toronto: University of Toronto Press. Izzo, Herbert J. 1977. Pre‐Latin languages and sound changes in Romance: the case of Old Spanish /h‐/. In M. P. Hagiwara (ed.), Studies in Romance Linguistics: Proceedings of the Fifth Linguistic Symposium on Romance Languages. Rowley, MA: Newbury House, pp. 227–53. James, Edward 1988. The Franks. Oxford: Blackwell. Jensen, Frede 1994. Syntaxe de l’ancien occitan. Tübingen: Niemeyer. Jochnowitz, George 1973. Dialect Boundaries and the Question of Franco‐Provençal. The Hague: Mouton. Jones, Mari C. 2001. Jersey Norman French: A Linguistic Study of an Obsolescent Dialect (Publications of the Philological Society 34). Oxford: Blackwell Jones, Mari C. and David Hornsby (ed.). 2013. Language and Social Structure in Urban France. Oxford: Legenda. Jordan Cólera, Carlos 2004. Celtiberica. Zaragoza: Universidad de Zaragoza. Joseph, Brian D. 1983. The Synchrony and Diachrony of the Balkan Infinitive: A Study in Areal, General, and Historical Linguistics. Cambridge: Cambridge University Press.

Joseph, Brian D. 2010. Language contact in the Balkans. In Raymond Hickey (ed.), The Handbook of Language Contact. Malden MA: Wiley‐Blackwell, pp. 618–33. Jud, Jakob 1907. Recherches sur la genèse et la diffusion des accusatifs en ‐ain et en ‐on (première partie). Halle: Karras. Kabatek, Johannes 2016. Diglossia. In Adam Ledgeway and Martin Maiden (eds.), The Oxford Guide to the Romance Languages. Oxford: Oxford University Press, pp. 624–33. Kaplan, Abby 2010. Phonology shaped by phonetics: the case of intervocalic lenition (University of California Santa Cruz Ph.D. dissertation). Kerswill, Paul 2013. Koineization In J. K. Chambers and Natalie Schilling (eds.), The Handbook of Language Variation and Change, 2nd edition. Oxford: Wiley‐Blackwell, pp. 519–36. Kibbee, Douglas A. 1991. For to Speke Frenche Trewely: The French language in England 1000–1600. Amsterdam: Benjamins. Klajn, Ivan 1972. Influssi inglesi nella lingua italiana. [English influences on the Italian language] Florence: Olschki. Klee, Carol A. and Andrew Lynch 2009. El español en contacto con otras lenguas. [Spanish in contact with other languages] Washington DC: Georgetown University Press. Kovačec, A. 1963. Notes de lexicologie istro‐ roumaine. Sur la disparition des mots anciens et leur remplacement par des mots croates. Studia Romanica et Anglica Zagabriensia 15/16: 3–39. Kovačec, A. 1966. Quelques influences croates dans la morphosyntaxe istroroumaine. Studia Romanica et Anglica Zagabriensia 21/22: 57–76. Kovačec, A. 1968. Observations sur les influences croates dans la grammaire istroroumaine. La Linguistique 1: 79–115. Krishnamurti, Bhadriraju 2001. Comparative Dravidian Linguistics: Current Perspectives. Oxford: Oxford University Press. Lambert, Pierre‐Yves 1994. La Langue gauloise: description linguistique, commentaire d’inscriptions choisies. Paris: Errance. Laprade, Richard A. 1981. Some cases of Aymara influence on La Paz Spanish. In M. J. Hardman (ed.), The Aymara Language in its Social and Cultural Context: A Collection of Essays on Aspects of Aymara Language and Culture. Gainesville FL: University Presses of Florida, pp. 207–27. Lass, Roger 1999. Phonology and morphology. In Roger Lass (ed.), The Cambridge History of the

Contact and the Romance Languages  449 English Language: Volume III, 1476–1776. Cambridge: Cambridge University Press, pp. 56–186. Ledgeway, Adam 2006. The dual complementiser system in Southern Italy: spirito greco, materia romanza? In Anna Laura Lepschy and Arturo Tosi (eds.), Rethinking Languages in Contact: The Case of Italian. Oxford: Legenda, pp. 112–26. Ledgeway, Adam 2009. Grammatica diacronica del napoletano. [A diachronic grammar of Neapolitan] Tübingen: Niemeyer. Ledgeway, Adam and John Charles Smith 2016. Deixis. In Adam Ledgeway and Martin Maiden (eds.), The Oxford Guide to the Romance Languages. Oxford: Oxford University Press, pp. 879–96. Liddicoat, Anthony 1994. A Grammar of the Norman French of the Channel Islands. Berlin: De Gruyter Mouton. Lipski, John M. 1994. Latin American Spanish. London: Longman. Lipski, John M. 2014. The many facets of Spanish dialect diversification in Latin America. In Salikoko Mufwene (ed.), Iberian Imperialism and Language Evolution in Latin America. Chicago: University of Chicago Press, pp. 38–75. Lodares, Juan Ramón 1996. Hispanic tracks in English: cowboys and gold‐rushers in the Old West. In Félix Rodríguez González (ed.), Spanish Loanwords in the English Language: A Tendency Towards Hegemony Reversal. Berlin: De Gruyter Mouton, pp. 157–76. Lodge, R. Anthony 1993. French. From Dialect to Standard. London: Routledge. Lodge, R. Anthony 2004. A Sociolinguistic History of Parisian French. Cambridge: Cambridge University Press. Loporcaro, Michele 1997. Puglia and Salento. In Martin Maiden and Mair Parry (eds.), The Dialects of Italy. London: Routledge, pp. 338–48. Loporcaro, Michele 2018. Gender from Latin to Romance: History, Geography, Typology. Oxford: Oxford University Press. Loussakoumounou, Alain‐Fernand 2009. La personne verbale en français écrit et parlé au Congo: étude morphosyntaxique systématique et modélisations en fin de cycle de grammaire (Université de Yaoundé I and Université Marien Ngouabi, Brazzaville doctoral thesis). Maas, Michael 2014. Reversals of fortune: an overview of the Age of Attila. In Michael Maas (ed.), The Cambridge Companion to the Age of Attila. Cambridge: Cambridge University Press, pp. 3–25. Mahadevan, Iravatham 2004. Voicing of consonants in Old Tamil: new evidence from

Tamil–Brāhmı̄. In Jean‐Luc Chevillard and Eva Wilden (eds.), South Indian Horizons: Felicitation Volume for François Gros on the Occasion of his 70th birthday., Paris: École Française d’Extrême‐Orient, pp. 505–11. Maiden, Martin 2006. Endangerment as stimulus to innovation? Some observations on the Romance languages of the eastern Adriatic. Unpublished paper given to Third Oxford–Kobe Linguistics Seminar “The Linguistics of Endangered Languages,” Kobe Institute, Kobe, April 2006. Marazzini, Claudio 2002. La lingua italiana: profilo storico. [The Italian language. A historical profile] Bologna: Il Mulino. Marchessou, Agnès 2018. Strasbourg: another setting for sociolinguistic variation in contemporary French. Journal of French Language Studies 28: 265–89. Martinet, André 1952. Celtic lenition and western Romance consonants. Language 28: 192–217. Mathieu, Éric 2009. On the Germanic properties of Old French. In Paola Crisma and Giuseppe Langobardi (eds.), Historical Syntax and Linguistic Theory,. Oxford: Oxford University Press, pp. 344–57. Matras, Yaron 2000. Mixed languages: a functional‐communicative approach. Bilingualism, Language and Cognition 3: 79–99. Mattheier, Klaus J. 1996. Varietätenkonvergenz. Überlegungen zu einem Baustein einer Theorie der Sprachvariation. Sociolinguistica 10: 31–52. McKitterick, Rosamond 2008. Charlemagne: The Formation of a European Identity. Cambridge: Cambridge University Press. McLaughlin, Mairi 2011. Syntactic Borrowing in Contemporary French: A Linguistic Analysis of News Translation. Oxford: Legenda. McLaughlin, Mairi 2018. When Romance meets English. In Wendy Ayres-Bennett and Janice Carruthers (eds.), Manual of Romance Sociolinguistics. Berlin: De Gruyter, pp. 644–73. Meillet, Antoine 1931. Sur une période de bilinguisme en France. Comptes Rendus des Séances de l’Académie des Inscriptions et Belles Lettres 75: 29–38. Menéndez Pidal, Ramón 1929. Orígenes del español: estado lingüístico de la península ibérica hasta el siglo XI [The origins of Spanish. The linguistic situation of the Iberian peninsula up to the eleventh century], 2nd revised and expanded edition. Madrid: Hernando. Meo Zilio, Giovanni 1955a. Influenza dello spagnolo sull’italiano parlato nel Rio de la Plata. [The influence of Spanish on the Italian

450  John Charles Smith spoken in the Rio de la Plata] Lingua Nostra 16: 16–22. Meo Zilio, Giovanni 1955b. Fenomeni lessicali dell’italiano rioplatense. [Lexical phenomena in Rio de la Plata Italian] Lingua Nostra 16: 53–5. Meo Zilio, Giovanni 1955c. Contaminazioni morfologiche nel cocoliche rioplatense. [Morphological contamination in Rio de la Plata cocoliche] Lingua Nostra 16: 112–17. Meo Zilio, Giovanni 1956. Interferenze sintattiche nel cocoliche rioplatense. [Syntactic interfernce in Rio de la Plata cocoliche] Lingua Nostra 17: 54–9 and 88–91. Meo Zilio, Giovanni 1958. Un morfema italiano con funzione stilistica nello spagnolo rioplatense. [An Italian morpheme with a stylistic function in Rio de la Plata Spanish] Lingua Nostra 19: 58–64. Meo Zilio, Giovanni 1959. Una serie di morfemi italiani con funzione stilistica nello spagnolo nell’Uruguay. [A series of Italian morphemes with a stylistic function in Uruguayan Spanish] Lingua nostra 20: 49–54. Meo Zilio, Giovanni 1960. Sull’elemento italiano nello spagnolo rioplatense. [On the Italian element in Rio de la Plata Spanish] Lingua nostra 21, 97–103. Meo Zilio, Giovanni 1970. El elemento italiano en el habla de Buenos Aires y Montevideo. [The Italian element in the speech of Buenos Aires and Montevideo] Florence: Valmartina. Meo Zilio, Giovanni 1993. The acquisition of a second Romance language by immigrants in Latin America. In Rebecca Posner and John N. Green (eds.), Bilingualism and Linguistic Conflict in Romance. Berlin: De Gruyter Mouton, pp. 559–90. Merlo, Clemente 1927. Lazio sannita ed Etruria latina? Italia Dialettale 3: 84–93. Merlo, Clemente 1950. Gorgia toscana e sostrato etrusco. [The “gorgia toscana” and the Etruscan substrate] Italica 27: 253–5. Mihăescu, Haralambie 1966. Influența grecească asupra limbii române pînă in secolul al XV‐lea. [Greek influence on the Romanian language up to the fifteenth century] Bucharest: Academia Republicii Socialiste România. Mihăescu, Haralambie 1969. Influența veche greacă. [Early Greek influence] In Alexandru Rosetti et al. (eds.), Istoria limbii române, vol. II. [History of the Romanian language, vol. II] Bucharest: Academia Republicii Socialiste România, pp. 366–7. Millar, María Angélica 1999. Los arabismos en la lengua española. [Arabisms in Spanish] Boletín de filología de la Universidad de Chile 37: 781–801.

Mooney, Damien 2016. Southern Regional French: A Linguistic Analysis of Language and Dialect Contact. Oxford: Legenda. Muysken, Pieter 1997. Media Lengua. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 365–426. Myers‐Scotton, Carol 1997. Code‐switching. In Florian Coulmas (ed.), The Handbook of Sociolinguistics. Oxford: Blackwell, pp. 217–37. .Nandriș, Octave 1957. Survivances phonétiques préromanes en roumain. Mélanges de linguistique et de littérature romanes à la mémoire d’István Frank, offerts par ses anciens maîtres, ses amis et ses collègues de France et de l’étranger. Saarbrucken: Universität des Saarlandes. Nandriș, Octave 1963. Phonétique historique du roumain. Paris: Klincksieck. Neumann, Günter and Jürgen Untermann (eds.) 1980. Die Sprachen im römischen Reich der Kaiserzeit: Kolloquium von 8.–9. April 1974.Cologne: Rheinland‐Verlag. Ngamountsika, Édouard 2007. Le français parlé en République du Congo: étude morphosyntaxique. (Université de Provence (Aix–Marseille I) and Université Marien Ngouabi, Brazzaville doctoral thesis). Niculescu, Alexandru 1981. Outline History of the Romanian Language. Bucharest: Editura știinti̦ fică și Enciclopedică. Nissen, Heinrich 1883. Italische Landeskunde: Band 1: Land und Leute. Berlin: Weidmann. Nyrop, K. 1914. Grammaire historique de la langue française. Tome premier: Histoire générale de la langue française, phonétique. Copenhagen: Gyldendal. Odisho, Edward Y. 1997. Al‐prefixed Arabic loanwords in Spanish: linguistic implications. Zeitschrift für arabische Linguistik 33: 89–99. Parrino, Flavio 1967. Su alcune particolarità della coniugazione nel dialetto di Ripatransone. [On some distinctive characteristics of conjugation in the dialect of Ripatransone] L’Italia Dialettale 30: 156–66. Pellegrini, Giovan Battista 1960. Tra lingua e dialetto in Italia. [Between language and dialect in Italy] Studi Mediolatini e Volgari 8: 137–53. Pellegrini, Giovan Battista 1989. Ricerche sugli arabismi italiani con particolare riguardo alla Sicilia. [Research into Italian Arabisms with special reference to Sicily] Palermo: Centro di Studi Filologici e Linguistici Siciliani. Penny, Ralph 2000. Variation and Change in Spanish. Cambridge: Cambridge University Press

Contact and the Romance Languages  451 Pitz, Martine 2009. Le type Charles‐Charlon ou Eve‐Evain en ancien français – vraiment un reflet du superstrat germanique? Nouvelle Revue d’onomastique 51: 175–200. Ploog, Katja 2002. Le Français à Abidjan: pour une approche linguistique du non‐standard. Paris: CNRS. Ploog, Katja 2006. Appropriation du français en context multilingue: éléments sociolinguistiques pour une réflexion didactique à propos de situations africaines. Besançon: Presses universitaires de Franche‐Comté. Poghirc, Cicerone 1969. Influent ̦a autohtonă. [Indigenous influences] In Alexandru Rosetti, Alexandru et al. (eds.), Istoria limbii române, vol. II. [History of the Romanian language. Volume II] București: Academia Republicii Socialiste România, pp. 313–65). Pohl, Walter 2014. Migrations, ethnic groups, and state building. In Michael Maas, (ed.), The Cambridge Companion to the Age of Attila. Cambridge: Cambridge University Press, pp. 247–63. Pope, M. K. 1952. From Latin to Modern French with Especial Consideration of Anglo‐Norman: Phonology and Morphology, 2nd edition. Manchester: Manchester University Press. Posner, Rebecca and John N. Green (eds.) 1993. Bilingualism and Linguistic Conflict in Romance. Berlin: De Gruyter Mouton. Pountain, Christopher J. 2011. Latin and the structure of written Romance. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds.), The Cambridge History of the Romance Languages, vol. 1: Structures. Cambridge: Cambridge University Press, pp. 606–59. Price, Glanville 1992. Romance. In Jadranka Gvozdanović (ed.), Indo‐European Numerals. Berlin: De Gruyter Mouton, pp. 447–96. Quartermaine, Nicola Jane 2002. The language of the Franks in Gaul (University of Manchester unpublished Ph.D. thesis). Remacle, Louis 1952. Syntaxe du parler wallon de La Gleize – tome 1: noms et articles, adjectifs et pronoms. Paris: Les Belles Lettres (Bibliothèque de la Faculté de Philosophie et Lettres de l’Université de Liège, fascicule 139). Roegiest, Eugeen 2009. Vers les sources des langues romanes: un itinéraire linguistique à travers la Romania (2nd edition). Leuven: Acco. Rohlfs, Gerhard 1970. Le Gascon: études de philologie pyrénéenne. Tübingen: Niemeyer. Rosetti, Alexandru 1968. Istoria limbii române de la origini pînă în secolul al XVII–lea. [The history of the Romanian language from its origins up to the seventeenth century] Bucharest: Editura pentru Literatură.

Rosetti, Alexandru et al. 1969. Istoria limbii române, vol. II. [History of the Romanian language, vol. II] Bucharest: Academia Republicii Socialiste România. Ruffino, Giovanni 1997. Sicily. In Martin Maiden and Mair Parry (eds.), The Dialects of Italy. London: Routledge, pp. 365–75. Russo, Stéphanie 2017. The imparfait lorrain in the context of grammaticalization (University of Texas unpublished Ph.D. dissertation). Sala, Marius 1999. Du Latin au roumain. Paris: L’Harmattan; Bucharest: Univers Enciclopedic. Sala, Marius 2013. Contact and borrowing. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds.), The Cambridge History of the Romance Languages, vol. 2, Contexts, Cambridge: Cambridge University Press, pp. 187–236. Sampson, Rodney 1999. Nasal Vowel Evolution in Romance. Oxford: Oxford University Press. San Giacomo, Marcela and Sharon Peperkamp 2008. Presencia del español en náhuatl: estudio sociolingüístico de la adaptación de préstamos. [The presence of Spanish in Nahuatl. A sociolinguistc study of loan adaptation] In Maurice Westmoreland and Juan Antonio Thomas (eds.), Selected Proceedings of the 4th Workshop on Spanish Sociolinguistics. Somerville MA: Cascadilla Proceedings Project, pp. 149–56. Sandfeld, K. 1930. Linguistique balkanique: problèmes et résultats. Paris: Champion. Sayahi, Lotfi 2014. Diglossia and Language Contact: Language Variation and Change in North Africa. Cambridge: Cambridge University Press. Schøsler, Lene 1984. La Déclinaison bicasuelle de l’ancien français. Odense: Odense Universitetsforlag. Schwerdt, Judith 2000. Die zweite Lautverschiebung: Wege zu ihrer Erforschung. Heidelberg: Winter. Siegel, Jeff 1985. Koines and koineization. Language in Society 14: 357–78. Smith, John Charles 2011. Change and continuity in form–function relationships. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds.), The Cambridge History of the Romance Languages, vol. 1, Structures. Cambridge: Cambridge University Press, pp. 268–317. Smith, John Charles. forthcoming. Social factors in language change and variation. In Adam Ledgeway and Martin Maiden (eds.), The Cambridge Handbook of Romance

452  John Charles Smith Linguistics. Cambridge: Cambridge University Press. Smith, John Charles and Clive Sneddon 2002. Further evidence for a “Middle French” koine: the fusion of “preposition + article” sequences. University of Pennsylvania Working Papers in Linguistics 8: 265–80. Stolz, Thomas, Dik Bakker, and Rosa Salas Paloma 2008. Aspects of Language Contact: New Theoretical, Methodological and Empirical Findings with Special Focus on Romancisation. Berlin: De Gruyter Mouton. Tagliavini, Carlo 1972. Le origini delle lingue neolatine [The origins of the Neolatin languages], 6th edition. Bologna: Pàtron. Tamás, Lajos 1967. Etymologisch‐historisches Wörterbuch der ungarischen Elemente im Rumänischen (unter Berücksichtigug der Mundartwörter). The Hague: Mouton. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley CA: University of California Press. Toon, Thomas E. 1983. The Politics of Early Old English Sound Change. New York and London: Academic Press. Tosco, Mauro. 2008. A case of weak Romancisation: Italian in East Africa. In Thomas Stolz, Dik Bakker, and Rosa Salas Paloma (eds.), Aspects of Language Contact: New Theoretical, Methodological and Empirical Findings with Special Focus on Romancisation. Berlin: De Gruyter Mouton, pp. 377–98. Trask, R. L. 1997. The History of Basque. London: Routledge. Trudgill, Peter 2010. Contact and sociolinguistic typology. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐ Blackwell, pp. 299–319. Trudgill, Peter 2011a. Sociolinguistic Typology: Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. Trudgill, Peter 2011b. Social structure, language contact and language change. In Ruth Wodak, Barbara Johnstone, and Paul Kerswill (eds.), The SAGE Handbook of Sociolinguistics. Los Angeles: SAGE, pp. 236–48. Trumper, John 1997. Calabria and southern Basilicata. In Martin Maiden and Mair Parry (eds.), The Dialects of Italy. London: Routledge, pp. 355–64.

Tuten, Donald 2003. Koineization in Medieval Spanish. Berlin: De Gruyter Mouton van Coetsem, Frans 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. van Coetsem, Frans 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Winter. Vennemann, Theo 2010. Contact and prehistory: the Indo‐European Northwest. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐Blackwell, pp. 380–405. Vignuzzi, Ugo 1997. Lazio, Umbria and the Marche. In Martin Maiden and Mair Parry (eds.), The Dialects of Italy. London: Routledge, pp. 311–20. Viguera Molíns, María Jesús 2002. Lengua árabe y lenguas románicas. [The Arabic language and Romance languages] Revista de filología románica 19: 45–54. Wanner, Dieter 1977. The heuristics of substratum. In M. P. Hagiwara (ed.), Studies in Romance Linguistics: Proceedings of the Fifth Linguistic Symposium on Romance Languages. Rowley, MA: Newbury House, pp. 1–13. Wartburg, Walther von 1969. Évolution et structure de la langue française (neuvième édition). Bern: Francke. Watkins, T. Arwyn 1993. Welsh. In Martin J. Ball with James Fife (eds.), The Celtic Languages. London: Routledge. Wendt, Heinz F. 1960. Die türkischen Elemente im Rumänischen. Berlin: Akademie‐Verlag. Winford, Donald 2003. An Introduction to Contact Linguistics. Oxford: Blackwell. Winford, Donald 2005. Contact‐induced changes: classification and processes. Diachronica 22: 373–427; Winford, Donald 2010. Contact and borrowing. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐Blackwell, pp. 170–7. Wise, Hilary 1997. The Vocabulary of Modern French: Origins, Structure and Function. London: Routledge. Wolfe, Sam 2018. Verb Second in Medieval Romance. Oxford: Oxford University Press. Zilles, Ana M. S. 2002. Grammaticalization of a gente in Brazilian Portuguese. University of Pennsylvania Working Papers in Linguistics 8: 297–310.

22 Contact and Spanish in the Pacific EEVA SIPPOLA 1 Introduction This chapter examines different contact processes of Spanish in the Pacific in a socio‐historical framework. Although Spanish has a long history in the region, the Pacific is a largely forgotten area in Spanish contact linguistics. The Spanish influence is most visible in the Philippines and the Marianas, where we find very different situations and outcomes of Spanish in contact, including the maintenance of Spanish as a heritage language, heavy ­borrowing from Spanish to local languages (e.g. Tagalog in the Philippines and Chamorro in the Marianas), and creolization leading to the emergence of a new variety called Chabacano. Colonial contexts, as the Spanish presence in the Americas and the Pacific, often present ­situations of bilingualism where universal processes of grammatical change are involved in the contact and diffusion of linguistic features. This chapter takes as a starting point the view according to which language contact and change have communicative and/or sociolinguistic motivations where speakers function as agents of change within the limits set by the external ecology of the contact situation. They make the categories existing in the languages in contact mutually compatible and more easily translatable, thus determining the outcome of the contact situation (Heine and Kuteva 2003: 561). The present chapter focuses on these contact outcomes taking the socio‐historical background in the Pacific as a starting point. The chapter’s aim is to shed light on the reasons for the outcomes of different contact settings of Spanish in the Pacific. These include the motivations for lexical borrowing in Tagalog and Chamorro, the features attested in Philippine Spanish, as well as structural innovations in Chabacano in the first place, and highlighting the differences between creolization and other types of contact situations in the second place. As the varieties here examined are still spoken in the area, contemporary ­language ideological perspectives to the role of the colonial language are also discussed. Cross‐linguistically, contact‐induced effects are attested to varying degrees ranging from very similar to quite different results of the contact situations. Different regions of the world with high concentrations of contact settings offer interesting laboratories for examining contact phenomena and the structural and sociolinguistic motivations behind them to the benefit of the theory of language contact. The Pacific region is but one of these settings, yet despite its diversity and a relatively well documented history since the sixteenth century, it is not very well known. It therefore offers new information on the scenarios that complement the diverse picture of Spanish in contact in particular and processes of language contact in general.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

454  Eeva Sippola Contact with Spanish in the Pacific region started with the Magallanes‐Elcano’s expedition in the early sixteenth century. The Spanish entered the area via the Americas, because the world had been divided between the Spanish and the Portuguese Crowns leaving the Portuguese in control of the route to Asia via Africa. The main aim of these explorations was to find a route to the Spice Islands in today’s Eastern Indonesia. As the first Europeans who arrived in the Pacific, important parts of the region were under Spanish rule for centuries. The Philippines was a Spanish colony from the times of the conquistador Legazpi in 1565 until Philippine Independence and the Spanish‐American war in 1898. Although the contact with Spain was never very intense and did not involve large Hispanic population groups, the contact setting led to different linguistic outcomes in the Philippines. Similarly, the Marianas were on the sailing route from the Philippines to Mexico, and the missionary activities started in the islands in 1668. Like the Philippines, the Marianas was never a destination for Spanish‐speaking settlers, and the contact of the local population with Spanish came mainly through the missionaries. These historical developments, however, left a permanent impact on the lexicon and structure of Chamorro.

2  Overseas Varieties of Spanish Today, Spanish has over 400 million L1 speakers. It is the official language in 21 countries around the globe and in Puerto Rico, an unincorporated territory of the United States; in addition, it has a large speaker population elsewhere in the United States without an official status (Simons and Fenning 2018; Instituto Cervantes 2018). In the Iberian Peninsula, Spanish is in contact with regional languages, such as Basque, Catalan, Galician, Asturian, and historically also Arabic, but the overseas varieties of Spanish present an even greater variety of contact situations. Together with English, Portuguese, and Russian, Spanish is one of the languages of the former European empires that spread extensively over the world during the colonial period triggering a vast array of contact situations with typologically and sociolinguistically very diverse local and colonial languages. The spread of Spanish from the Iberian Peninsula started in 1492 with the expulsion of Sephardic Jews1 and Muslims from the Spanish Empire and the voyage of Christopher Columbus that led to the conquest and colonization of the Americas in the following decades. The Spanish empire extended over large areas of the Americas and the Pacific in the following centuries and, later on, also Africa. In Africa, Spanish is in contact with Arabic dialects in Morocco and in the Western Saharan communities and with Bantu languages and creoles in Equatorial Guinea (Lipski 1985; Yakpo 2009; Castillo‐Rodríguez and Morgenthaler García 2016). In the Americas, Spanish mainly came in contact with local indigenous languages, such as Nahuatl, Guaraní, Mayan, Aymara, Quechua, and Mapudungun as well as a number of smaller ones (cf. Escobar 2012). In addition, Spanish was also in contact with languages of forced2 and voluntary migration. On the one hand, forced migration and slavery brought Spanish in contact with a number of African languages, especially in the Caribbean region. This contact in unequal conditions of power between the population groups brought about the birth of creole and other restructured varieties (Lipski 2005; Perez et al. 2017). On the other hand, Spanish also came in contact with languages of voluntary migration in the Americas, among which English, Italian, Portuguese, and Cantonese were the most important ones (Lipski 2010). These varied contact situations have had an effect on the large‐scale adaptation of Spanish. As most of the population of the Hispanic Americas adopted Spanish as their main language, new Spanish dialects developed, yet contact‐induced change and bilingualism also led to language shift and language death in many indigenous, African‐ descendant, and migrant communities. As a comprehensive overview of Spanish contact situations and related processes of change is not possible within the limits of this chapter, the

Contact and Spanish in the Pacific  455 following will focus on the Pacific as a showcase region of fascinating contact settings involving Spanish that highlight central processes in language contact.3

3  Socio‐historical Background of Spanish in the Pacific Spanish arrived in the Pacific Islands from America in the sixteenth century as the first European colonial language when Spain extended its possessions from Mexico to the Philippines, Marianas, Carolinas, Palau, as well as other islands on isolated expeditions without permanent settlements. The Pacific administrative area was part of the Viceroyalty of New Spain with its capital in Mexico, and for a great part of the Spanish colonial period all contact with the Pacific possessions was through the Americas. The only contact with the metropolis was via a yearly galleon that carried American silver and Asian luxury products from Mexico to the Philippines and back, and again from Mexico to Spain. From the nineteenth century onwards and especially since the independence of Mexico, the Pacific territories were directly administered from Spain. That is why the Carolinas and Palau, which were colonized in the second half of the nineteenth century, have experienced greater influence of Peninsular Spanish. However, Spanish colonization in the Carolinas and Palau did not have any lasting linguistic consequences on the islands, since the inhabitants of the islands and the European settlers communicated mostly in an English pidgin (Lipski et al. 1996: 282). Settlers never arrived in large numbers in the region, and the most important Spanish‐ speaking population on the islands consisted of missionaries, soldiers, and administrators recruited to a great degree from the Americas. The Spanish footprint in the Philippines and Marianas shows clear Hispanic‐American traits, especially Mexican ones, as many spoke Latin‐American varieties of Spanish and not Peninsular dialects. This demographic situation meant that there was no significant mestizo group nor possibilities for demographic shifts among the native population to take place. As a consequence, Spanish never had a similarly strong impact in the Pacific as in the Americas that would have led to a total or near‐total adaptation of the colonial language by the native populations (Lipski et al. 1996; Quilis and Casado‐Fresnillo 2008). Accordingly, while the language of the colonial administration in the Philippines was Spanish, the use of Spanish was never widespread among the local population. Local laws and customs were largely maintained, although the legal code was codified in Spanish. For most of the Spanish period, the policy was for priests to interact with Filipinos in the local vernaculars rather than teach Spanish, and Spanish education was limited mostly to a small elite. For example, during Spanish colonial rule, there was no general education system and Spanish never spread widely through schools. As a consequence, only 2.5% of the population spoke Spanish at the end of the Spanish period (González 1998: 495). Although Spanish retained its position as an official language until 1987, the Spanish‐American War and the loss of the last Spanish colonies in 1898 marked the beginning of the dominance of English and American culture in the Philippines, and today English has replaced Spanish as a prestige language in all contexts. In the Marianas, Spanish military personnel’s and missionaries’ presence had a significant impact in the islands because the total number of inhabitants in the region was not very high in the eighteenth century. The Hispanic population came primarily from the Philippines and Mexico. During the first phases of colonization, diseases introduced to the islands and conflicts between the native population and the colonizers caused a sharp decline of the indigenous population. To remedy the situation, the Spanish resettled the population in the urban centers on the islands where they were under the direct control of the missions and the Crown. As a result of these resettlements, the intensified contact situation triggered mestizaje

456  Eeva Sippola between the local population and the colonizers and, consequently, a rapid process of Hispanization (cf. Pagel 2010: 35–6). The new mestizo group adapted the customs and ­religion of the colonizers, mixing them with Chamorro elements and other influences. New social, cultural, and linguistic practices were created and they replaced the traditional Chamorro systems. Although Spanish was spoken as the lingua franca of the Marianas, it never totally replaced Chamorro, which remained the everyday language, albeit in a new form. As a result of these social and demographic changes, Chamorro changed considerably and a new form emerged (Lipski et  al. 1996: 281, Rodríguez‐Ponga 2009). Today, new Chamorro is still widely spoken in the islands, while the numbers of Spanish speakers in the Marianas are low, and Spanish is spoken mainly by the Hispanic personnel employed in the US army and stationed on the islands.

4  Lexical and Structural Borrowing 4.1 Contact‐induced Change in Tagalog: Conjunctions and Discourse Markers Spanish has had extensive lexical influence on Tagalog and other Philippine languages. In addition to names, the most evident are the counting system, the calendar, the expression of time, and greetings. Estimates of the Spanish component in the Tagalog lexicon range ­between 10% and 30% (Bowen 1971; Wolff 2001), and it is most visible in nouns. Even for some core semantics that are seen as least borrowable in language contact, Tagalog has borrowings that have fully or partly replaced the original forms, e.g. braso ‘arm’ and kantá ‘song’ from Sp. brazo and cantar ‘to sing’ (Baklanova 2017). The lexical borrowings also include function words, such as the modal verb puwedi ‘can, be able to’ (< Sp. puede ‘can‐3SG.PRES’), other modal particles, such as siguro ‘probably’ and sigurado ‘certain’ (< Sp. seguro ‘certain’), and elements in comparative constructions, where the Spanish‐origin comparative más ‘more’ is used, as in (1). (1)

Tagalog (Schachter and Otanes 1972: 239) Mas maganda si Rosa kaysa more pretty FOC Rosa than ‘Rosa is prettier than her sister.’

sa POSS

kapatid sister

niya. her

Beyond the lexicon, Spanish has had minimal influence on the phonological and syntactic systems of Tagalog. The initial affricate /t͡ʃ/ and stop‐liquid clusters occur in loanwords, such as tsinelas ‘slippers’ (< Sp. chinelas), plato ‘plate’ (< Sp. plato), and krus ‘cross’ (< Sp. cruz). The Spanish contact also had an effect on the development of the three‐vowel system (with /i/, /a/, and /u/) into a five‐vowel system in Tagalog (including also /e/ and /o/). However, the change was probably already under way by the arrival of the Spanish, and Spanish just provided new minimal pairs for establishing the phonological contrasts in the five‐vowel system in emergence (Reid 1973). This outcome is not surprising, keeping in mind that slightly intense contact situations often present borrowing that affects mostly content words and only some function words (Thomason and Kaufman 1988). Other well‐known examples include the Hellenization of Latin and the Roman Empire, Sanskrit and Arabic borrowings in Malay, or French/Norman influence on English. Tagalog thus presents a large borrowed vocabulary from Spanish that includes some function words. To explain the motivation of the borrowings despite the low presence of L1 speakers of Spanish, let us have a closer look at one area of high‐frequency

Contact and Spanish in the Pacific  457 items where we find a relatively high presence of borrowed function words: conjunctions and discourse markers. In Tagalog, many conjunctions are borrowed from Spanish. These include, for example, the following (Schachter and Otanes 1972): (2)

o ‘or’ (< Sp. o) pero ‘but’ (< Sp. pero) porke ‘because’ (< Sp. porque) miyentras ‘while’ (< Sp. mientras) maski ‘even though’ (< Sp. más que) para ‘for/in order’ (< Sp. para) imbis + na/‐ng ‘instead of’ (< Sp. en vez de)

In addition, several discourse markers are found among the borrowings, as for example (Baklanova 2017): (3)

sige ‘okay, agreed, right; Go ahead!’ (< Sp. ¡Sigue! ‘Go ahead!’, seguir ‘to follow’) as a response bweno4 ‘so, well / well then’ (< Sp. bueno ‘good, nice; well; okay, alright’) medyo ‘somewhat, comparatively, a little, rather’ (< Sp. medio ‘half; average; medium, etc.’) and parang ‘looks like / it seems; apparently; somewhat’ (< Sp. para ‘for, towards’ + Tag. attribution particle ng) as hedges to sound less direct Grabe! ‘How sad, serious!’ (< Sp. grave ‘serious, severe, etc.’) and syempre ‘of course, sure’ (< Sp. siempre ‘always’) as attitudinal interjections

The main function of these items is discourse organization. A motivation for the borrowing of these elements can be explained because discourse strategies that organize communication in Spanish lead to better opportunities of participation in the colonial society and the distribution of power in it (Stolz 1996). This is further supported by the fact that in the colonial period, Spanish functioned as a marker of social status among Tagalog speakers, p ­ ossibly in mixed forms with local languages (Quilis and Casado‐Fresnillo 2008: 62–6; Wolff 2001: 234). Some influential members of the Tagalog speech community were bilingual in Tagalog and Spanish (of some kind) and used Spanish to mark their social position. Similar uses of loan conjunctions and discourse markers are also found in contemporary English‐Tagalog mixed codes used by the higher classes of the Philippine society. It thus seems probable that comparable processes were at play already during colonial times (Wolff 2001).

4.2 Contact‐induced Change in Chamorro: Irrealis Mood and Future Tense Intensive contact with Spanish from the eighteenth century onwards has led to notable ­modifications in the Chamorro language, creating a new Chamorro, whose status as a heavy borrower or a mixed language is debated (cf. Stolz 2003: 273, Rodríguez‐Ponga 2009). Most evidently, like Tagalog, Chamorro has a large loan vocabulary from Spanish. About 55% of the Chamorro lexicon is derived from Spanish (Stolz 2003: 290), although the figures might differ according to the text type. The semantic domains of the borrowed items include person names, counting system, calendar and time. The Chamorro sound system is maintained to a great degree, and borrowings are adapted to it, as in betde (Sp. < verde) ‘green’ and lepblo (Sp. < libro) ‘book’. In addition to this kind of standard lexical borrowing commonly found in colonial settings, functional elements, such as conjunctions and the comparative marker mas, are borrowed from Spanish, in the same way as in Tagalog. However, also elements derived

458  Eeva Sippola from contact with Spanish that are seen as borrowings pertaining to more intense contact situations are found. These include, among others, the superlative i mas ‘the most’, the indefinite article un, the pronouns yu’ ‘1SG’ and esti ‘DEM’, optional grammatical gender marking, habitual past marker estaba and the irrealis and future markers bai, debi di, para and syiempre (Bowen 1971; Pagel 2010: 146–7). These borrowings can be considered as pertaining to more core areas of grammar than lexical borrowings or other functional elements, such as the conjunctions or discourse markers. It is also worth noting that unlike direct borrowings from Spanish, some of these elements show considerable degrees of innovation. The irrealis or future markers deserve attention here because they have been interpreted as forming a new category of future tense (Topping 1973; Pagel 2008, 2010). In the Western‐Malayo‐Polynesian languages of the area, future tense does not exist as a grammatical category. Related meanings are expressed in aspectual and modal systems. When looking for the origins of the future markers in Chamorro, the influence of the contact with Spanish, a language that does have a grammaticalized expression of future tense, thus seems like an attractive option for explaining the origins of these items. The Chamorro future paradigm includes both autochthonous elements and elements derived from Spanish, presenting both relatively direct borrowings, such as bai, and elements, such as debi.di ‘FUT’, para ‘FUT’ and syiempre ‘FUT’ that have acquired new functions via grammaticalization in Chamorro (Topping 1973; Pagel 2008). The expression of future in Chamorro is done with a combination of autochthonous and Spanish‐derived markers (Pagel 2008). In Chamorro, bai from Spanish voy ‘go.PRES.IND.1SG’ is used to indicate future for 1SG and 1PL.EXCL, as in (4a). It can be combined with other future markers, such as debi.di and para in the same example, although in colloquial speech, the double marking is often seen as superfluous. Bai has clear correspondences in the analytic Spanish future construction ir a + verb, e.g. voy a comer ‘I will eat’, which can be identified as the source for this element, while similar markers are not found in the languages of the area thus downsizing a substrate explanation for this item (Stolz 2003).

(4) Chamorro (Pagel 2008: 186) a. debi.di bai hanao para Jerusalen para FUT FUT.1 go to Jerusalem FUT ‘I will have to go to Jerusalem to suffer many (things).’

bai FUT.1

famadesi suffer

meggai much

b.

klaru na yanggen ti ma sapotte i kinalam­ten‐ña that if NEG PASS support the activity‐POSS.3SG clear i lenguahi, kun tiempo syiempre u måfnas giya Guåhan the language with time FUT FUT.3 die in Guam ‘It is clear that if the language activities are not supported, surely [Chamorro] is going to die in Guam.’

c.

para u ma protehi FUT FUT.3 3PL.ERG protect ‘They (3+) will protect this resource.’

este this

na LK

ressource resource

Of the Spanish‐derived elements, para in Chamorro indicates a neutral, uncertain future, and is used, among others, for subjunctive and final meanings. The form is replicated from the Spanish preposition para ‘for, to’ with directional meanings. Future with debi.di, on the other hand, implies an obligation for the agent of the action, reflecting the modal origin of the marker in Spanish deber de ‘have to, must’.

Contact and Spanish in the Pacific  459 Pagel (2008, 2010) explains the contact‐induced origin of the future marker syiempre ‘FUT’ from a Spanish adverbial siempre ‘always’ to a future marker in Chamorro. The marker syiempre in Chamorro implies a strong certainty or determination of the speaker with regard to the future action. First, a semantic overlap can be found in the fact that the adverbial meaning in Spanish indicates security and determination. Siempre ‘always’ does not have any meaning concerning verbal tense in Spanish, but when used in a phrase in the future tense, this meaning can be emphasized. Second, in Mexican Spanish siempre can act as a discourse marker with meanings such as ‘in the end, finally, finally, after all’. The varieties of Spanish that were spoken by the military and missionaries in the islands during the early centuries of contact were speakers of American varieties of Spanish, and among them, Mexican varieties prevailed. According to Pagel (2010), the adverbial and discourse functions in Spanish were probably at the origin of the replication process of the form from Spanish to Chamorro. However, the grammaticalization of the item and the creation of the future tense category has happened in Chamorro, because in Spanish the same elements are not used for expressing future tense. The exact path of grammaticalization in Chamorro is yet to be explored based on the scarce historical documentation available, but the current studies have clearly established the starting point for the process. The contact‐induced changes in Chamorro are extensive, also when compared to a relatively similar Tagalog case. In both languages there has been vast lexical and structural borrowing from Spanish. However, in Chamorro, Spanish elements have continued a grammaticalization path to create new categories, such as the future tense. Despite these major changes in certain areas of grammar, Chamorro features typical traits of the West‐Malayo‐Polynesian languages (morphology, ergativity, grammatical focus) and can at best be described as a heavy borrower instead of a mixed language (Pagel 2008; Stolz 2003).

5 Creolization 5.1  The Chabacano Varieties Chabacano is the common name used for several creole varieties that have Spanish as the lexifier and Philippine languages as the adstrates.5 Today, Chabacano is spoken in Cavite City and Ternate in the Manila Bay region on the island of Luzon (the Manila Bay Creoles, MBC), and in Zamboanga on the southern island of Mindanao (Zamboanga Chabacano, ZC). The varieties are for the most part mutually intelligible, but there are socio‐historical circumstances and linguistic differences that distinguish them (Lesho and Sippola 2013, 2014). Zamboanga Chabacano is a widely spoken language, but Cavite and Ternate Chabacano are both endangered (Lesho and Sippola 2013). The Manila Bay varieties have Tagalog as the adstrate language, and Zamboanga Chabacano has Hiligaynon and Cebuano as the main adstrates. Tagalog, Hiligaynon, and Cebuano are part of the Central Philippine language family, and the latter two are more closely grouped together in the Visayan subfamily. Zamboanga Chabacano has become a regional lingua franca and is well supported at the community and institutional levels. Other languages spoken in the area include the national language, Filipino (a standard register based mainly on Tagalog); the country’s other official language, English; the creole’s main adstrates, Hiligaynon and Cebuano; and other local languages. In contrast, Cavite and Ternate Chabacano speakers are now a minority in their communities. Tagalog/Filipino has become the main language of these communities, and English is also widely used. Cavite Chabacano is severely endangered because most speakers

460  Eeva Sippola are of grandparental age or older (Lesho and Sippola 2013). Ternate Chabacano is still being learned by children, so it is threatened but relatively stable. Despite the proximity of Cavite City and Ternate, there is little interaction between these communities. Cavite is close to Manila and has historically had a strong cultural link to it, whereas Ternate has historically been more rural and isolated. Creoles are languages born in situations of intense contact, where people of diverse ethno‐cultural and linguistic backgrounds came together, often under conditions of slavery or indentured labor, and formed distinct communities and languages. The Chabacano varieties are typically creole in that they have a hybrid history, although forced labor conditions are not known from their history. Their lexicon is to a great extent from Spanish and is largely shared between the varieties. In fact, they are one of the only three known Spanish‐lexified creoles. For example, 95% of the entries in basic vocabulary is shared between them (Sippola 2011: 27). In contrast, their grammar differs from Spanish and presents many innovations. Some of these are clearly contact‐ induced changes from the adstrate languages, such as the Philippine verb initial word order, the plural particle manga, and the plural personal pronoun series of Zamboanga Chabacano with an inclusive‐exclusive distinction in the first person kita ‘we.INCL’, kamé ‘we.EXCL’, kamó ‘you.PL.FM’, ustedes ‘you.PL.FR’, silá ‘they’ (Lipski 2013). Beyond lexicon, other features show the continuity of Spanish grammatical elements, such as the indefinite and definite articles un ‘INDEF’ and (k)el ‘DEF’, numerals, and a small set of gender marked adjectives, such as gwápa ‘beautiful.F’, blánka ‘white.F’, etc. However, even if the form of the Spanish elements has been replicated rather directly, the functions and meanings often differ from Spanish. In other cases, like the marking of the object, the marking follows similar patterns, but the forms of the markers differ. In both Spanish and Chabacano, animacy is a defining factor for marking the objects, but in Chabacano the marker is kon instead of the Spanish a. Although the varieties show clear differences in the area of grammar, due to local developments and historical contact settings (Lesho and Sippola 2014; Sippola and Lesho 2020), they are here treated together to highlight the contact‐induced changes from the lexifier and the adstrate languages. There are no widely accepted structural definitions of creoles (cf. Bakker et al. 2017; Blasi et al. 2017). Velupillai (2015) resumes some common notions about the structural nature of creoles in relation to non‐creoles. It is common to assume that the features that set a given creole apart from its lexifier are typical creole features, and at the same time, it is indirectly taken as a given that lexifier languages are similar to non‐creole languages. The other prominent idea is that creoles are morphosyntactically less complex than non‐creoles, although complexity is a rather opaque term.

5.2  A Restructured Verbal System with Creole Traits The Chabacano verbal system has undergone heavy restructuring and shows features that are often defined as typical for creoles (Velupillai 2015: 54). The verbs are marked for tense and aspect with preverbal markers that cannot be combined. They are derived from the Spanish infinitives without the final /r/, as in kantá (< Sp. cantar) ‘sing’ which combine with the TMA particles presented in Table 22.1. Preverbal TMA markers that combine in a fixed order are well‐known from creoles around the world. In Chabacano, the markers cannot be combined. Some verbs are also derived from the third person singular present forms in Spanish that are frequent in discourse, as in tyéne ‘have’ (< Sp. tiene ‘have.3SG’), pwéde (< Sp. puede ‘can, be able to.3SG’), but these are not commonly used in combination with the TMA markers.

Contact and Spanish in the Pacific  461 Table 22.1  Tense‐Mood‐Aspect particles in Chabacano. Spanish etymon

Tense/aspect

Mood

generic, past and present time reference perfective, past time reference imperfective, past and present time reference di (MBC), ay (ZC) ha/he de ‘should/will’ contemplated, future time reference irrealis

Ø ya/a ta

ya ‘already’ está ‘be.PRES.3SG’

The use of the TMA particles is generally obligatory, although zero marking is found with generic reference and in narrative contexts where the TMA marker has been expressed previously. The use of the aspect markers is shown in examples in (5). The perfective forms are used for actions and states of affairs which have been completed at reference time. The imperfective aspect is used for actions and states of affairs which are ongoing at the reference time, including both habitual and progressive meanings. The contemplated aspect refers to states of affairs which have not yet taken place at reference time. (5) Chabacano (Sippola 2013: 145–6) a. Perfective aspect Ya huntá yo, PFV gather 1SG ‘I saved, I bought this.’ b. Imperfective aspect Ta trabahá éle IPFV work 3SG ‘S/he works in Las Piñas.’ Ta hirbí ya IPFV boil already ‘The water is boiling already.’

a PFV

komprá buy

na LOC

Las Pínyas. Las Piñas.

el DET

ágwa water

yo 1SG

ésti. this

.

c. Irrealis mood / contemplated aspect Si di yubá raw manyána na munisípyu, If CTPL take DM tomorrow LOC town.hall bo di indá húntu. 2SG CTPL go together ‘If I am taken to the town hall tomorrow, you will come with me.’ From these examples, it is clear that the verbal system of Chabacano has lost all inflectional marking that is characteristic of Spanish or the affix marking of the Philippine languages that use complex verbal affixes to mark TMA values and voice. Although the forms of the TMA markers are clearly derived from the Spanish constructions, the TMA meanings and functions are rather different from the lexifier, and are rather similar to the Philippine TMA system (Schachter and Otanes 1972: 368). The lexical forms, the matter, have thus been partially replicated from the lexifier, while Tagalog, the adstrate language, has provided the source for the replication of the TMA pattern. Matter replication often occurs with

462  Eeva Sippola lexifier forms, as seen from the prominent lexical presence of the European colonial languages in creoles around the world, while substrate ancestry is prominently related to features of the verbal domain, such as TMA and argument marking (Blasi et al. 2017). The TMA marking of Chabacano shows how in situations of intense language contact, new constructions are formed based on processes of replication from the lexifier and the adstrate languages, making the categories existing in languages in contact mutually compatible and more readily translatable (Heine and Kuteva 2003). When compared to the influence of Spanish to the indigenous languages of the region, the degree of restructuring in the TMA system (and in many other areas of grammar) in the creole varieties is more prominent. The particles depart clearly from the Spanish inflectional model and Philippine affixation and present a new creole ­innovation. Although the verbal system of Chabacano shows typical creole features, it should also be kept in mind that the basic questions about creolization are often difficult to answer with purely synchronic data. In the Chabacano case, in order to trace back the development of the verbal system in detail, we would need to provide historical documentation of different speech forms showing a continuum of TMA marking from more flexional and affixational systems to particles and their related meanings. So far, historical Chabacano texts only reach back a century and a half. They show that different codes presenting variation between Spanish verbal constructions with inflection and creole TMA marking were already part of the multilingual communication in the creole communities (Fernández and Sippola 2017).

6  Language Loss and L2 Features in Heritage Spanish Today, Spanish in the Philippines is a dying language with very few native speakers. It is spoken by a small number of older Filipinos that often represent the Spanish‐Filipino mestizo group. They are descended directly from Spanish settlers from the last wave of Spanish emigration to the Philippines in the nineteenth century and represent the higher socio‐ economic groups of the society (Lipski 1987). The number of native speakers of Spanish is at most a few thousand, but difficult to estimate with confidence (Simons and Fenning 2018). It is safe to assume that all of them are bilingual in Spanish and another language, mainly Tagalog or English. Lipski (1987) describes contemporary Philippine Spanish as an aristocratic, conservative language with Castilian traits, such as the use of the typically Peninsular second person plural pronoun vosotros and le as the third person object clitic, and which is maintained to a large part artificially among Spanish‐English/Tagalog bilinguals. The native Spanish of the Philippines shows evident traits of the influence of Filipino languages and typical features of communities where Spanish is a second language, setting them apart from relatively balanced bilinguals (Lipski 1987, 1993). This kind of traits are typical in situations where a community shifts away from a minority language to a majority language in a short time span. Here, a generation of transitional bilinguals have passive abilities that are very similar to those of native speakers but they never have learned a full form of the language (Dorian 1977; Lipski 1993). Language loss and endangerment can lead to regularization and partial loss of agreement patterns. Lipski (1987) interprets this as an indication of lack of constant practice in Spanish for the majority of Spanish speakers, similar to learners of Spanish as a L2. In the Philippine data, examples are found that represent the lack of verbal person agreement (6a) and (6b) and gender agreement in noun‐adjectives (6c) and pronouns (6d), etc. showing the instability of agreement patterns.

Contact and Spanish in the Pacific  463 (6)

Philippine Spanish

a.

Por aquí entra los PREP here enter.3SG the.PL ‘International ships enter here.’ (Lipski 1987: 43)

b.

No me gusta nada perros. NEG me please.3SG nothing dog.PL ‘I do not like dogs at all.’ (Quilis and Casado‐Fresnillo 2008: 130)

c.

Tenía sangre español had blood.F Spanish.M ‘He had Spanish blood.’ (Quilis and Casado‐Fresnillo 2008: 112)

d.

Las costumbres  nuestros the.F.PL custom.F.PL our.M.PL ‘Our customs.’ (Quilis and Casado‐Fresnillo 2008: 112)

barcos ship.PL

internacionales. international.PL

The majority languages for the Spanish‐speaking community, Tagalog (or other major Philippine languages) or English do not present gender agreement in verb or noun phrases. It is therefore difficult to determine conclusively if the above described patterns are influenced by the majority languages or if they show an instability of the agreement patterns due to l­ anguage loss. However, these examples follow hierarchies of morphological reduction that have been found for foreigner talk, where verb forms in third person singular and masculine forms in nouns are at the top of the hierarchy (Lipski 1993: 167). However, Quilis and Casado‐Fresnillo (2008: 130) claim that the agreement errors are not very frequent in Philippine Spanish, and that most of them appear together with nouns expressing collectives, such as perros ‘dogs’ in (6b). A habitual feature also representing L2 features is the lack of an article, as in (6b), where in Standard Spanish los perros ‘the dogs’ would be expected. Other examples include el dueño de casa [de la casa] ‘the house’s owner’ and de bajo de tierra [de la tierra] ‘from underground’. Other typical features identified in Lipski (1987) are the avoidance of embedded constructions, especially when subjunctive forms would be needed, non‐standard uses of prepositions, their omission, and also, to some degree, the categorical use of subject pronouns when standard varieties of Spanish do not require an overt subject pronoun. These traits are found in L2 and contact varieties of Spanish world‐wide, connecting the language loss situation in the Philippines with the lack of exposure to Spanish. Another peculiarity that is more difficult to connect with L2 varieties is the use of the personal pronoun tú ‘you.SG’ in formal and informal contexts, while in many Spanish varieties, tú is the familiar form and usted is used in formal contexts. This might be due to the disappearance of the politeness distinctions or to the fact that the Spanish‐speaking community is so small and socially closely‐knit (Lipski 1987).

7 Language Ideologies about Spanish in the Chabacano Communities The structural features described in this chapter show how language changes in response to its social context. One dimension that has not been explored in detail so far are the attitudes and ideologies that affect the social uses of language and, by consequence, might lead to or emerge as a response to structural changes. Linguistic systematicity is as much a function of historically rooted ideologies and of the ordering practices of social life, as of language per se (Gal and Irvine 1995).

464  Eeva Sippola Studies on language attitudes in the Chabacano communities show that several historical and prestige dimensions regarding languages and their social meanings are at play in the Philippines (Lesho and Sippola 2014; Sippola 2016a, 2016b). As a colonial language, Spanish still holds some prestige today, although it is no longer extensively spoken or even present in the everyday life of the Filipinos. However, some speakers distinguish a “pure” or “legitimate” Chabacano as distinct from Spanish (Lesho and Sippola 2014: 35), which echoes the desire to express independence and difference from the colonial language. In the same way, language attitudes about Chabacano are twofold, reflecting and reproducing both negative and positive ideologies. On the one hand, the creole language has symbolic value and is used in ceremonial contexts indicating pride in the Spanish heritage and the local identities. On the other, Chabacano is often seen as “broken Spanish” and the language of the poor, which contributes to negative language attitudes. Spanish also has a special role in language preservation purposes, as seen when studying the writing practices in the creole. In Cavite and Ternate, a growing production of written materials aimed at the preservation and promotion of the language has emerged in recent decades. In Cavite, the materials consist mostly of language‐learning materials and dictionaries, in addition to other texts, such as newsletters, blog texts, and newspaper columns. In Ternate, writings often deal with the promotion of local history and culture, and folk stories and accounts of local history have been published in Chabacano. As for other languages, the writing of previously uncodified varieties gains public visibility generally in different stages: first, writing happens in simple humorous texts, letters, riddles, and folksongs, and later it extends to newspapers, official documents, and other types of more formal writing (Mühleisen 2005). Orthographic choices in the materials show the importance of the Spanish heritage for the communities even today (Sippola 2016a). In language‐learning materials, the Spanish etymological spelling is preferred, including the graphemes , , , , which have no phonemic correspondences in Chabacano, and others, such as for /ɲ/ and for /r/, whose phonemic status is debatable. Notable exceptions to the preference of the Spanish etymological spelling are English adaptations for newly introduced items referring to modern technologies and for items that are (partially) shared between Spanish and English, such as nation, association, difference/diferencia, etc. The Spanish etymological orthography is used mostly by older speakers and language activists. However, among younger speakers, the phonemic Philippine spelling is more common, and only is used to represent the affricate /t͡ʃ/. Spanish spelling is more accessible and more easily reproduced in writing by older generations due to their access to education in Spanish in earlier decades. The use of Philippine spelling conventions among younger generations is a practically motivated choice, as younger speakers have Tagalog as their first language of literacy, accompanied by English, which is taught at school from an early age, and have only limited access to texts written in Spanish. Chabacano speakers have started to expand the use of the language to new domains, including not only internet and text messaging related uses but also more traditional domains, such as music including rap, which is a popular music genre in the Philippines today (Sippola 2016b). As a consequence of the young speakers’ own activity and language promotion programs at school and in the municipality of Ternate, Chabacano is occasionally used in rap lyrics, especially in peer meetings, local music contests, and town fiestas. Rap lyrics by young Chabacano speakers in Ternate make use of Chabacano to express their social realities and local identities. Rap lyrics are often drafted on paper or in an electronic format on mobile phones, prior to the performance. In the lyrics, although Tagalog spelling conventions prevail, there is a slightly higher number of etymological spellings than in the written texts by young Chabacano speakers in general. For example, there is an occasional use of , as in /kjen/ ‘who’ and /kel/6 ‘DEF’, as well as as in

Contact and Spanish in the Pacific  465 /iho/ ‘son’. Also, local features characteristic of the Ternate variety, such as syllable final /h/ and velarization of final /n/ are reflected in the texts. The marking of local features can be explained by the fact that rappers see themselves as promoters of local language and history (Sippola 2016b). In sum, orthography is the site of language practices where speakers express linguistic identities that are different from Tagalog, the main language of the communities. The changes in writing practices and new domains of language use for Chabacano, such as rap music, in the Manila Bay varieties, show how multilingual and multicultural resources are negotiated and appropriated by the speakers according to the local needs and the historical context. Language ideologies about the former colonial and current national languages Spanish, Filipino, and English shape the multilingual resources and their use. Naturally, the language endangerment situation of Chabacano is a major factor in these processes, in that it affects the knowledge of, skills in, and attitudes about the creole. It is well known that multilingual speakers often resort to linguistic and cultural items of the dominant languages and cultures. In the Philippines, Spanish has a special role in the language preservation purposes of the creole. Speakers may select Spanish elements and use them to index a historical connection to the former prestige language. At the same time, the stigma attached to the creole language can be reversed by its use in new domains. In the Tagalog and English dominated context of today, Spanish is also used to highlight aspects of local authenticity and identity. These practices are changing speakers’ notions of linguistic and sociocultural authenticity that has an impact on the assignation of overt and covert sociolinguistic prestige as well as language use and development.

8  Concluding Remarks These examples from the Philippines and Marianas show a variety of outcomes of Spanish contact situations in the Pacific. The Pacific settings differ from American contexts, where slavery and a greater number of Spanish settlers have influenced the outcomes of Spanish in contact. In the Pacific, local indigenous languages have been maintained, but they show heavy borrowing with important lexical components from Spanish differing in the degree of functional borrowings. Chabacano, on the other hand, departs from the heavy borrowing situations in that the degree of structural changes in core areas of grammar is much greater. Social factors are central to the structural changes and processes as seen from the various examples of Spanish in contact in the Pacific. The examples show how on the societal level, the types of community settings influence linguistic practices, along with demographic, economic, ideological, historical, and political factors. At the individual level, the speaker’s engagement, the communicative situation, and smaller‐scale social constraints are central for borrowing and the creative process of new innovative constructions (Heine and Kuteva 2003; Matras 2009), as seen in the prominence of elements of discourse organization that are affected by borrowing and in the creole innovations. In sum, linguistic change cannot be accounted for by one single factor, but rather involves complex and interconnected social and linguistic processes.

9 Abbreviations 1 = first person, 2 = second person, 3 = third person, CTPL = contemplative DEM = d ­ emonstrative, DET = determiner, DM = discourse marker, ERG = ergative, EXCL exclusive, F = feminine, FM = familiar, FOC = focus particle, FR = formal, FUT = future, INCL = inclusive, IND = indicative, IPFV = imperfective, L2 = second language, LK = linker, LOC = locative, M = masculine, MBC = Manila

466  Eeva Sippola Bay Creoles, NEG = negation, PASS = passive, PFV = perfective, PL = plural, POSS = possessive, PREP = preposition, PRES = present, SG = singular, TMA = tense‐mood‐aspect, ZC = Zamboanga Chabacano.

NOTES 1 2 3 4

Jews from the Iberian Peninsula. Forced migration is here used to indicate population movements under conditions of slavery. For an overview of contact situations involving Spanish see Lipski (2010). Baklanova (2017) notes that the word is getting obsolete in Tagalog and tends to be replaced by the English so. 5 The term adstrate is also used to cover possible substrate languages, as in Asian contexts, the ­substrate often continues to be spoken as an adstrate alongside the creole. 6 Here, the writer has chosen to include in the spelling although it is not a sound in the word when pronounced.

REFERENCES Bakker, Peter, Finn Borschenius, Carsten Levisen, and Eeva Sippola (eds.) 2017. Creole Studies. Phylogenetic Approaches. Amsterdam: Benjamins. Baklanova, Ekaterina 2017. Types of borrowings in Tagalog/Filipino. Kritika Kultura 28: 35–54. Blasi, Damián E., Susanne Maria Michaelis, and Martin Haspelmath 2017. Grammars are robustly transmitted even during the emergence of creole languages. Nature Human Behaviour 1: 723–9. Bowen, Donald 1971. Hispanic languages and influences in Oceania. In Thomas Sebeok (ed.), Current Trends in Linguistics 8: Linguistics in Oceania. The Hague: de Gruyter Mouton, pp. 938–53. Castillo‐Rodríguez, Susana and Laura Morgenthaler García (eds.) 2016. Exploring glottopolitical dynamics in Africa: the Spanish colonial past and beyond. Special issue of the International Journal of the Sociology of Language 239. Escobar, Ana María 2012. Spanish in contact with Amerindian languages. In José Ignacio Hualde, Antxon Olarrea, and Erin O’Rourke (eds.), The Handbook of Hispanic Linguistics. Oxford: Blackwell, pp. 65–88. Dorian, Nancy 1977. The problem of the semi‐speaker in language death. International Journal of the Sociology of Language 12: 23–32. Fernández, Mauro and Eeva Sippola 2017. A new window into the history of Chabacano: two unknown mid‐nineteenth century texts. Journal of Pidgin and Creole Languages 32.2: 304–38.

Gal, Susan and Judith T. Irvine 1995. The boundaries of languages and disciplines: how ideologies construct difference. Social Research 62.4: 967–1001. González, Andrew 1998. The language planning situation in the Philippines. Journal of Multilingual and Multicultural Development 19.5: 487–525. Heine, Bernd and Tania Kuteva 2003. On contact‐induced grammaticalization. Studies in Language 27.3: 529–72. Instituto Cervantes 2018. El español: una lengua viva. Informe de 2018 [Spanish: a Living Language. 2018 Report]. Available at https:// cvc.cervantes.es/lengua/espanol_lengua_ viva/pdf/espanol_lengua_viva_2018.pdf Lesho, Marivic and Eeva Sippola 2013. The sociolinguistic situations of Manila Bay Chabacano‐speaking communities. Language Documentation and Conservation 7: 1–30. Lesho, Marivic and Eeva Sippola 2014. Folk perceptions of variation among the Chabacano creoles. Revista de Crioulos de Base Lexical Portuguesa e Espanhola 5: 1–46 Lipski, John 1985. The Spanish of Equatorial Guinea: The Dialect of Malabo and Its Implications for Spanish Dialectology. Tübingen: Max Niemeyer Verlag. Lipski, John 1987. Contemporary Philippine Spanish: comments on vestigial usage. Philippine Journal of Linguistics 18: 37–48. Lipski, John 1993. Creoloid phenomena in the Spanish of transitional bilinguals. In John

Contact and Spanish in the Pacific  467 Lipski and Ana Roca (eds.), Spanish in the United States: Linguistic Contact and Diversity. Berlin: de Gruyter Mouton, pp. 155–82. Lipski, John 2005. A History of Afro‐Hispanic Language: Five Centuries and Continents. Cambridge: Cambridge University Press. Lipski, John 2010. Spanish and Portuguese in contact. In Raymond Hickey (ed.), Handbook of Language Contact. Oxford: Wiley‐Blackwell, pp. 570–600. Lipski, John 2013. Remixing a mixed language: the emergence of a new pronominal system in Chabacano (Philippine Creole Spanish). International Journal of Bilingualism 17.4: 448–78. Lipski, John, Peter Mühlhäusler, and F. Duthin 1996. Spanish in the Pacific. In Stephen Wurm, Peter Mühlhäusler, and Darrel Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas, II.1. Berlin and New York: de Gruyter Mouton, pp. 271–98. Matras, Yaron 2009. Language Contact. Cambridge: Cambridge University Press. Mühleisen, Susanne. 2005. Introduction: Creole languages in Creole literatures. Status and standardization. Creole Languages in Creole Literatures. Special issue, Journal of Pidgin and Creole Languages 20.1: 1–14. Pagel, Steve 2008. The old, the new and the in‐ between: comparative aspects of Hispanisation on the Marianas and Easter Island (Rapa Nui). In Rosa Salas Palomo, Dik Bakker, and Thomas Stolz (eds.), Hispanisation: The Impact of Spanish on the Lexicon and Grammar of the Indigenous Languages of Austronesia and the Americas. Berlin: de Gruyter Mouton, pp. 167–201. Pagel, Steve 2010. Spanisch in Asien und Ozeanien [Spanish in Asia and Oceania]. Frankfurt‐am‐ Main: Peter Lang. Perez, Danae, Eeva Sippola, and Sandro Sessarego 2017. Afro‐Hispanic varieties in comparison: new light from phylogeny. In Peter Bakker, Finn Borchsenius, Carsten Levisen, and Eeva Sippola (eds.), Creole Studies: Phylogenetic Approaches. Amsterdam: Benjamins, pp. 269–92. Quilis, Antonio and Celia Casado‐Fresnillo 2008. La lengua española en Filipinas: historia; situación actual; el chabacano; antología de textos [Spanish language in the Philippines: history, current situation; Chabacano: text anthology]. Madrid: CSIC. Reid, Lawrence 1973. Diachronic typology of Philippine vowel systems. In Thomas A. Sebeok (ed.), Current Trends in Linguistics

11: Diachronic, Areal, and Typological Linguistics. The Hague: de Gruyter Mouton pp. 485–506. Rodríguez‐Ponga, Rafael 2009. Del español al chamorro. Lenguas en contacto en el Pacífico [From Spanish to Chamorro. Languages in contact in the Pacific]. Madrid: Ediciones Gondo. Schachter, Paul and Fe T. Otanes 1972. Tagalog Reference Grammar. Berkeley: University of California Press. Simons, Gary F. and Charles D. Fenning (eds.) 2018. Ethnologue: Languages of the World, 21st edition. Dallas, Texas: SIL International. Sippola, Eeva 2011. Una gramática descriptiva del Chabacano de Ternate [A descriptive Grammar of Ternate Chabacano]. Helsinki: University of Helsinki. Sippola, Eeva 2013. Ternate Chabacano. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 2: Portuguese‐based, Spanish‐based, and French‐based Languages. Oxford: Oxford University Press, pp. 149–55. Sippola, Eeva. 2016a. Postcolonial language ideologies: writing in Chabacano. In Daniel Schmidt‐Brücken, Susanne Schuster, and Marina Wienberg (eds.), Aspects of (Post) Colonial Linguistics. Current Perspectives and New Approaches. Berlin: De Gruyter, pp. 151–76. Sippola, Eeva. 2016b. Rap and resistance in Chabacano. In Martin Pütz and Neele Mundt (eds.), Vanishing Languages in Context: Ideological, Attitudinal and Social Identity Perspectives. Frankfurt‐am‐Main: Peter Lang, pp. 157–76. Sippola, Eeva and Marivic Lesho 2020. Contact‐ induced grammatical change and independent development in the Chabacano creoles. Bulletin of Hispanic Studies 97: 103–21. Stolz, Thomas 1996. Grammatical hispanisms in Amerindian and Austronesian languages. Amerindia 21: 137–60. Stolz, Thomas 2003. Not quite the right mixture: Chamorro and Malti as candidates for the status of maixed language. In Yaron Matras and Peter Bakker (eds.), The Mixed Language Debate: Theoretical and Empirical Advances. Berlin: de Gruyter Mouton, pp. 271–316. Thomason, Sarah and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press.

468  Eeva Sippola Topping, Donald, with the assistance of Bernadita C. Dungca 1973. Chamorro Reference Grammar. Honolulu: University of Hawai‘i Press. Velupillai, Viveka 2015. Pidgins, Creoles and Mixed Languages. An Introduction. Amsterdam: Benjamins. Wolff, John 2001. The influence of Spanish on Tagalog. In Klaus Zimmermann and Thomas Stolz (eds.), Lo propio y lo ajeno en las lenguas austronésicas y amerindias: procesos interculturales en el contacto de lenguas

indígenas con el español en el Pacífico e Hispanoamérica [The own and the foreign in the Austronesian and Amerindian languages: intercultural processes in the contact of indigenous languages and Spanish in the Pacific and Latin America]. Madrid: Iberoamericana, pp. 233–53. Yakpo, Kofi 2009. Complexity revisited: Pichi (Equatorial Guinea) and Spanish in contact. In Nicholas G. Faraclas and Thomas Klein (eds.), Simplicity and Complexity in Creoles and Pidgins. London: Battlebridge, pp. 83–215.

23 Contact and Portuguese‐ Lexified Creoles HUGO C. CARDOSO 1 Introduction When, in August 1415, King João I of Portugal led an armada across the mouth of the Mediterranean Sea to lay siege to the city of Ceuta, he was in fact launching the first wave of a process of European overseas expansion which would impact the world for centuries to come. Over the course of the next century and a half, the Portuguese would establish themselves around the globe, with outposts in South America, the Atlantic islands, along the coast of East and West Africa, and in Asia and the Pacific, all the way to Japan and Timor. This historical process of political and economic expansion was necessarily accompanied by linguistic diffusion, which accounts for the global nature of Portuguese in the modern world, but also for the development of the Portuguese‐lexified creoles (henceforth PLCs), which occupy us here.1 PLCs have long been prominent in scholarship about creole languages. There are several reasons for this, not least the fact that, because of the chronology just mentioned, PLCs include some of the oldest of the European‐lexified creoles which developed out of the colonial expansion of the Early Modern Age. In addition, even though several members of this group have died out (especially in Asia) and some of the others remain among the most under‐studied of all creoles, PLCs also include some of the most vital, vibrant, and conspicuous creole languages anywhere. Besides, the extremely different social, historical, and political trajectories of individual PLCs mean that the group includes very diverse languages, which cover some of the most influential typologies of creole languages proposed in the literature. All three types in Bickerton’s (1988) socio‐historical classification2 are represented, not just fort creoles (such as Diu Creole), but also plantation creoles (like Santome), and maroon creoles (like Angolar). Chaudenson’s (1979) opposition between exogenous and endogenous creoles, based on whether or not their formation involved the relocation of the contributing populations, is also covered, with some PLCs (such as Capeverdean Creole) belonging to the first type, and others (including Malacca Creole) to the second. In addition, because of their wide geographical spread, the formation of the various PLCs involved contact with languages of many distinct families (including Niger‐Congo, Indo‐European, Dravidian, Austronesian, and Sino‐ Tibetan), producing very different linguistic encounters and significant typological diversity among them. Such diversity in terms of their formative and developmental settings makes this group of languages particularly adequate to test out theories of creole formation. In that

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

470  Hugo C. Cardoso respect, it is noteworthy that, in the late nineteenth century, two of the most important early creolists, Adolfo Coelho and Hugo Schuchardt, relied heavily on data from PLCs to develop their contrasting theories of language contact. There are, however, other dimensions to the diversity encapsulated by PLCs. One of them relates to the changing role of the lexifier language. In fact, while some PLCs (like those of Guinea‐Bissau or Daman) coexist with Portuguese to this day, others (including those of Sri Lanka or Korlai) eventually lost significant contact with it after their formation, making it possible to explore the impact of this variable on their diachronic trajectory, not only in terms of the source and/or development of particular linguistic features, but also of the prevalence of sociolinguistic variation, speakers’ attitudes, official recognition, vitality, and so forth. Another important dimension to these languages’ diversity, which opens up new possibilities of linguistic enquiry, pertains to their sociopolitical status within their host societies. While some PLCs, such as the creoles of Cape Verde and Guinea‐Bissau, have become stable languages with large speech populations and have made or are making significant progress towards standardization and official recognition, others never progressed past a fragile minority status within their wider contexts and have either died out (as in the case of the creole of Mangalore) or are currently facing linguistic contraction and potential death (as in the cases of Principense or Macau Creole). Evidently, such a diverse set of languages can be (and has been) approached by linguists from very different perspectives, and contributes many different insights with respect to contact‐induced language change. As it would be impossible to survey all of these here, we will focus on select studies and debates which in some way derive from one of the factors which make them stand out among the creole languages that resulted from European overseas expansion: their time depth. The fact that PLCs include, to the best of our knowledge, the first of that group of languages to form not only raises the possibility that they may have influenced other subsequent contact languages, but also enhances the potential for post‐ formative diachronic dynamics to have left a mark. Before embarking on these discussions, however, it is important to describe in a little more detail the current and past distribution of the PLCs.

2  The Portuguese‐lexified Creoles It has already been mentioned that the beginning of the Portuguese overseas expansion is symbolically placed in 1415, but this does not mean that all PLCs have their origin around that period. The progression of navigation and exploration made coastal West Africa the main locus of linguistic contact involving Portuguese throughout the fifteenth century – and, indeed, the formation of the West African PLCs has been located within that period – while linguistic diffusion to Asia and the Americas is a sixteenth‐century phenomenon. The geographical reach of the Portuguese between then and the end of the colonial era was ­substantial, stretching from Brazil to Japan. PLCs may have formed anywhere along that geographical span – and even beyond, as a result of the migration of previously influenced populations. In fact, those that either persist or have been recorded do stretch across the globe, even if, in some places  –  including Brazil and Japan, precisely, but also Goa or Angola – the formation of PLCs is a matter of conjecture, as no conclusive records are to be found.3 At present, PLCs are spoken in Western Africa (Cape Verde, Guinea‐Bissau, Senegal, São Tomé e Príncipe, Equatorial Guinea) and Asia (India, Sri Lanka, Malaysia, Singapore, Macau) (see Cardoso, Hagemeijer and Alexandre 2015 and Map 23.1) by populations of widely variable dimensions, which range from close to a million, in the cases of the creoles of Cape Verde and Guinea‐Bissau, to less than 10 speakers, in the case of the one spoken in Kerala

Contact and Portuguese‐Lexified Creoles  471

E F G H

A B C

D

I 8 9 10

12 3 5 6

18 17 19

K 14

L

12

4

11 15 16

7

20 21 13

J

Portuguese-lexified languages (existing) 1 Capeverdean Creole

6 Angolar

11 Cannanore Creole

2 Casamance Kriyol 3 Guinea-Bissau Kriyol

7 Fa d'Ambu 8 Diu Creole

12 Sri Lanka Portuguese 13 Papia Kristang

9 Daman Creole

14 Makista

4 Principense 5 Santome

10 Korlai Creole

Portuguese-lexified languages (extinct) 15 Mangalore Creole

18 Bengal Creole

16 Cochin Creole

19 Burma Creole

17 Nagapattinam Creole

20 Batavia and Tugu Creole

21 Flores Creole

Others A Hawaiian Pidgin

E Sranan

I Lingua Franca

B Palenquero

F Saramaccan

J Afrikaans

C Papiamentu

G Matawai

K Chinese Pidgin English

D Guyanese Creole

H Ndyuka

L Chabacano

Map 23.1  Location of contact languages mentioned in the chapter.

(South India). Researchers of these languages tend to group them in geographically‐defined clusters which also reflect significant linguistic continuities. These clusters operate at different levels, as demonstrated in Table 23.1 below: while, since at least Ferraz (1987), a higher‐order classificatory split is proposed between the Atlantic and the Asian creoles, lower‐order regional clusters such as e.g. Upper Guinea or Southern South Asia can also be recognized. The last column of Table 23.1 is entitled “Currently‐spoken languages” because, as a matter of fact, the current distribution of the Portuguese‐based creoles mentioned there represents only a portion of these languages’ past implantation (see Smith 1995). It is important to bear in mind not only that several of them were abandoned over time (in some cases, potentially, without any records), but also that, according to some hypotheses which will be explored in this chapter, some transformed or were absorbed into other, rather different contact languages, leaving behind linguistic traces which are sometimes evident and largely uncontroversial, and sometimes not. Not unlike creole languages in general, there are no substantial records for most of the Portuguese‐based creoles before the late nineteenth century.4 The one of Sri Lanka (then called Ceylon) is a notable exception, since, in the first half of the century, several publications in or about this language were produced, including grammars, dictionaries, religious texts, and translations.5 Nonetheless, because several early creolists developed a scholarly interest in the PLCs around the turn of the nineteenth century, we actually have records of a few of the extinct varieties, such as that of Batavia (modern Jakarta) and nearby Tugu, in Indonesia (Schuchardt 1891), or that of Nagapattinam, in Southeastern India (Dalgado 1917). In this

472  Hugo C. Cardoso Table 23.1  Proposed geographical groupings of the current PLCs (identified by several recurrent glossonyms); designations in italics correspond to the classification proposed by Schuchardt (1889a) for the Asian PLCs, based on the taxonomy of their substrate/adstrate languages. Geographical clusters

Currently‐spoken languages

Atlantic

Upper Guinea

• Capeverdean Creole/Kabuverdianu (Cape Verde) • Guinea‐Bissau Creole/Kriyol (Guinea‐Bissau) • Casamance Creole/Kriyol (Senegal)

Gulf of Guinea

• São Tomé Creole/Santome/Forro, Príncipe Creole/Principense/Lung’Ie, Angolar/Lunga Ngola (São Tomé e Príncipe) • Annobón Creole/Annobonese/Fa d’Ambu (Equatorial Guinea)

Asia

South Asia (Indo‐ Portuguese)

Northern (Gauro‐ Portuguese)

• Diu Creole, Daman Creole, Korlai Creole (India)

Southern (Dravido‐ Portuguese)

• Cannanore Creole (India) • Sri Lanka Portuguese (Sri Lanka)

Southeast + East Asia

Southeast Asia (Malayo‐ Portuguese)

• Malacca Creole/Papia Kristang (Malaysia + Singapore)

East Asia (Sino‐ Portuguese)

• Macau Creole/Makista/Patuá (Macau, China)

respect, Hugo Schuchardt stands out for the intensity of his scholarly production and the encompassing the encompassing nature of his descriptive studies (see Schuchardt 1882a, 1882b, 1883a, 1883b, 1888a, 1888b, 1888c, 1889b, 1998c). Adolfo Coelho (1880–6) also cast a global look at the PLCs to support his universalist theory of creole formation, as did Leite de Vasconcelos (1901) in his global dialectological survey of Portuguese. Other philologists of the time specialized in particular regions, such as – to name but a few – Joaquim Botelho da Costa and Custódio José Duarte (1886), and António de Paula Brito (1887), who documented the creole of Cape Verde, or Sebastião Rodolfo Dalgado (1900, 1902–3, 1906), who focused on those of South Asia.

3  Quest for the Origins Creole languages posit a number of questions, some of which stem from a certain recurrence of particular features among them, regardless of their formative settings. One of the proposed solutions to account for perceived similarities between creoles (in general or within particular clusters) has been to posit a common origin, and the inaugural role of Portuguese in the European language expansion of the Modern Age has made it a logical component of several such theories. The strongest formulations of a monogenetic theory6 proposed that all pidgin and creole languages resulting from European overseas expansion ultimately derived

Contact and Portuguese‐Lexified Creoles  473 from a Portuguese‐lexified pidgin (henceforth PLP) formed in West Africa in the fifteenth century – perhaps itself a transformation of the preceding Mediterranean Lingua Franca, or Sabir –, then spread around the globe through the networks of trade (including the slave trade) and adapted to new lexical bases (English, French, etc.) through relexification (Whinnom 1956; Taylor 1961; Thompson 1961). The presence of certain Portuguese‐derived lexemes in non‐Portuguese‐lexified creoles – some recurrent, such as pikinini/pikni/pequin/… ‘child’ (from Ptg. pequen(in)o ‘small’) or savi/savvy ‘to know’ (from Ptg. sabe ‘know(s)’), others circumscribed to particular (groups of) languages, such as palaver ‘dispute’ (from Ptg. palavra ‘word’) in West Africa or briga ‘to fight’ (from Ptg. briga(r) ‘(to) fight’) in Guyanese (see Baker and Huber 2001 for further examples) – were seen as remnants from this process. It is no longer widely held that a single origin is necessary to account for similarities among pidgins and creoles. One important challenge to the Lingua Franca and West African Portuguese Pidgin hypotheses comes from non‐European‐lexified pidgins and creoles (see e.g. Buchstaller, Holmberg, and Almoaily 2014), which could hardly have been part of the same process of diffusion and relexification, even though only some authors ever considered the scope of a strong form of monogenesis to extend beyond the European‐lexified creoles. However, the precedence of Portuguese in many former colonial settings and various instances of population displacement are historically attested facts; therefore, more circumscribed processes of linguistic transfer and speciation involving this language continue to be identified and discussed in the literature. In this section, we will explore some such cases. In Section 3.1, we survey diachronic proposals applicable within subgroups of PLCs, while in Section 3.2 we focus on those that imply the influence of a PLC or PLP on contact languages currently not classified (strictly) as Portuguese‐lexified.

3.1  Diffusion and Speciation The earliest PLCs – those of West Africa – developed during the fifteenth century, a period in which written records largely fail to document linguistic variation explicitly. As such, the actual process of language contact and creole formation occurred in something of a historiographical mist, which motivates many questions and different theories. However, there is one type of textual evidence from the period which is relatively p ­ lentiful, and whose relevance for the formation of PLCs has been the matter of debate. With the confluence in Portugal (especially Lisbon and the southern regions) of Africans from several parts of the continent, in the fifteenth century, a particular form of L2 Portuguese came to be associated with them. This sociolect, known then as Língua de Preto “Language of the Black” or fala Guiné ‘Guinea speech’, was often portrayed (mostly for comical effect) in Portuguese plays and literary works (see Lipski 2005; Pereira 2006: 69–73). Though certainly prejudiced and stereotypical, this highly variable corpus of attestations contains sentences such as the following, taken from Gil Vicente’s play O Clérigo da Beira [The clergyman of Beira], from the 1520s (transcribed in Kihm and Rougé 2013: 267): (1)

A mi bae furtá em tanto / camisa que sa na  muro. 1SG   go steal  meanwhile shirt REL COP on wall ‘Meanwhile, I’m going to steal the shirt that’s on the wall.’

Such passages are a testament to the social visibility of African communities in fifteenth‐ and sixteenth‐century Portugal, but they also interest linguists to the extent that they appear to reflect certain forms and constructions of the modern West African PLCs. With respect to the sentence in (1), for instance, pronominal a mi is reminiscent of Portuguese a mim ‘to me’ but, while in Portuguese this preposition‐pronoun combination can only occur in object position, here it appears as a subject; more significantly, however, it seems

474  Hugo C. Cardoso to reflect the same morphological and syntactic reinterpretation which produced the first person pronouns of the current West African PLCs (see Lipski 1991), either ami or (a)mi (also, for other persons and numbers, abó/(a)bo ‘2SG’; anós/(a)nos ‘1PL’; etc.; see Kihm and Rougé 2013: 260). Likewise, the copula sa, probably a derivation of Ptg. estar ‘to be’ or sam/são ‘[they] are’, is also identified in some modern West African PLCs, namely among those of the Gulf of Guinea (Lipski 2005: 53–4; Kihm and Rougé 2013: 267–8). For some authors (e.g. Naro 1978; Kihm and Rougé 2013)  –  although this is not without controversy – the contact varieties developed among Africans in Portugal and represented as Língua de Preto were then carried to West Africa on board the Portuguese ships and formed the seedlings of the local PLCs. Regardless of whether or not all West African PLCs had a common ancestry in the form of Língua de Preto, the split between Upper Guinea and Gulf of Guinea represented in Table  23.1 is linguistically as well as socio‐historically motivated. While the languages which make up both clusters do share a number of linguistic traits – e.g. the presence of distinct strong and weak pronouns or negative concord  –  there are also many features which separate both regions but are recurrent within each – e.g. discontinuous negation in the Gulf of Guinea creoles (Hagemeijer and Alexandre 2012) or a postverbal anterior morpheme (‐)ba (though with somewhat different morphological and syntactic constraints) in the Upper Guinea Creoles (see Holm 1988: 152). This has been interpreted as evidence of two distinct formative histories, leading to attempts at reconstructing the diachrony of each regional cluster. In the case of Upper Guinea, the internal consistency is relatively uncontroversial (see e.g. d’Andrade and Kihm 2000; Baptista, Mello, and Suzuki 2007). The bone of contention has been that of the ultimate origin of the proto‐language in the ancestry of these creoles – if indeed they do share a common ancestry. One theory posits that the original contact language was established on the continent, around the Portuguese outposts of Guinea, and was then carried over to Cape Verde; conversely, another theory posits an origin on the Cape Verdean island of Santiago and later diffusion both to the other islands of the archipelago and the West African coast.7 The directionality of spread is somewhat less obscure in the Gulf of Guinea. On the islands of the Gulf of Guinea, two PLCs are spoken in São Tomé (Santome and Angolar), one in Príncipe (Lung’Ie), and one in Annobón (Fa d’Ambu), and, while their mutual intelligibility is restricted, their relatedness has been determined on linguistic grounds since at least the work of Ferraz (1979). More recently, comparative studies of the four languages (Hagemeijer 2011, 2015) have not only strengthened the case for their genetic connection, but also suggested that the proto‐language received a substrate contribution constituted mostly of Niger‐ Delta languages, with Bantu influence kicking in later on (a fact supported also by a predominance of the Nigerian genotype among the population of São Tomé and Príncipe, see Hagemeijer and Rocha 2017). In the case of the Gulf of Guinea creoles, linguistic, sociodemographic, genetic and historical evidence converge to define the scenario of a proto‐creole forming on São Tomé towards the end of the fifteenth century, and then diffusing to the islands of Príncipe and Annobón in the sixteenth century. Angolar, on the other hand, appears to have broken away from the proto‐creole due to the constitution of a community of runaway slaves in Southeastern São Tomé. The PLCs of Asia present new challenges when it comes to reconstructing potential patterns of intra‐continental diffusion, but even their relationship with the creoles of the Atlantic is an important diachronic question. Since, historically, Portuguese overseas ­exploration proceeded southwards and eastwards around the Cape of Good Hope (as did the Asian‐bound vessels that sailed the Cape route), one could expect that whatever phenomena of language contact occurred on West Africa would be carried over to East Africa and Asia. However, an influential comparative study by Ferraz (1987) argued that the Asian

Contact and Portuguese‐Lexified Creoles  475 PLCs constitute a separate cluster (cf. Table 23.1) with a number of linguistic commonalities that set them apart from the West African PLCs, suggesting a fundamental genetic break between the two spheres (but see Clements 2000 and Holm 2009 for less radical views). Some of those “Luso-Asianisms” include the ones given in Table 23.2. It is important to note that these features are not necessarily universal among (modern) Asian PLCs, but they do cut across geographical clusters, which may be indicative of a common origin.8 Clements (2000) admits some carry‐over from West Africa but does posit an “Asian pidgin” which could be responsible for diffusing the linguistic features that crop up in various Asian PLCs. He further specifies that this Asian pidgin is likely to have been shaped in South Asia, and, more specifically, around the earliest Portuguese settlements in Asia (chiefly, Cochin), on the Malabar Coast of southwestern India (see also Clements 2009). The hypothesis that the initial language contact which occurred on the Malabar Coast (primarily involving Portuguese and Malayalam) determined the common features of Asian PLCs has not been fully worked out, but Baxter and Bastos (2012), in a study of the prototypical Asian‐Portuguese possessive construction (see Table 23.2, no. 1), find that the conditions of the initial encounter are consistent with the development of this particular structure. Several other instances of linguistic diffusion may be identified within the Asian PLCs, but two are especially interesting and will be described here: the Malacca‐Macau axis, and the establishment of a PLC on the island of Java. The first of these helps to explain the conflation of “Southeast + East Asia” as a second‐ order cluster in Table  23.1, and concerns the ultimately Malaccan origin of the creole of Macau. At least since the work of Batalha (1965–6), it has been a truism that the creoles of Malacca and Macau share a particular genetic bond (see also Holm 1989: 296–8; Baxter 2009a) regardless of the geographical distance between them and their different linguistic environments (Malay‐dominated in the case of Malacca, Cantonese in Macau). Since then, comparative studies have provided further evidence that certain features of Macau Creole are more easily explained by transfer from Malay than from Cantonese, including its causative‐facilitative serial verb construction (Baxter 2009b), and the structure of comparative constructions (Cardoso 2012). The following examples illustrate the reasoning in the case of the comparative constructions: (2)9

a.

Eles mais  pior que japonês. 3PL more bad CMPR Japanese ‘They [were] worse than the Japanese.’

b.

  mais  velha  de que  eu. E  ela    and  3SG.F more  old     CMPR   1SG ‘And she [was] older than me.’

c.

Iloutro ung pouco mais alto grau 3PL one little more high degree ‘They [were] a little higher‐class than us.’

de nós. CMPR 1PL

d.

eli más altu 3SG more tall ‘She is taller than Pio.’

Pio Pio

e.

Dia lebih tinggi dari  saya. 3SG more  tall CMPR 1SG ‘He is taller than me.’

(Malay (Bahasa Indonesia))

f.

Gāmyaht yiht  gwo   kàhmyaht. today  hot  CMPR  yesterday ‘Today is hotter than yesterday.’

(Cantonese)

di CMPR

(Macau Creole)

(Malacca Creole)

476  Hugo C. Cardoso Table 23.2  Some widespread features among Asian PLCs identified by Ferraz (1987). 1 2 3 4 5 6 7 8

A possessive construction with the structure: Possessor + su/sə/s(u)a + Possessed Syncretic dative‐accusative case‐marking, expressed by an adposition derived either from Portuguese para ‘for’ or por ‘by’, or from Portuguese com ‘with’ Preverbal Tense‐Aspect markers derived from Portuguese já ‘already’ (Past/Perfective), está/estar ‘is/to be’ (Non‐punctual aspect) and logo ‘immediately, later’ (Future/Irrealis) Syncretic existential/possessive/copular verbs A special future negator derived from Portuguese não há‐de ‘shall not’ Pluralizing reduplication Certain lexical items, such as ada/ade/adi ‘duck’ Multimorphemic words involving a derivation of Ptg. laia ‘fashion’ (e.g. question word kilai ‘how’)

Of the three slightly different comparative constructions in Macau Creole, which essentially differ in the form of the comparative marker, (2a/b) are very close to Portuguese (which uses the comparative marker que or do que), a language that has always played (and continues to play) an important role among the Macanese community; (2c), on the other hand, is practically equivalent to the construction in Malacca Creole (2d). Both of these have a very close model in the Malay construction in (2e), with respect to the order of the terms but also the presence of an index (lebih ‘more’, equivalent to más/mais) and the semantics of the comparative marker, which, in all these cases (dari in Malay, di in Malacca Creole, de in Macau Creole) also marks ablative. By contrast, even though we might expect Cantonese (or Hokkien, prominent in the earliest stages of colonial Macau, see Ansaldo and Matthews 2004) to have contributed to the structure of comparative constructions in Macau Creole, its comparative construction (2f) contains no index, and the comparative marker gwo is not an ablative marker, but rather a verbal element meaning ‘to cross/pass’. It is linguistic correspondences such as this, coupled with socio‐historical evidence of early population movements from Malacca to Macau (Batalha 1974; Baxter 2009a), that motivate the interpretation of Macau Creole as an offshoot of Malacca Creole, later transformed by speciation and the unavoidable linguistic influence of Sinitic languages (see e.g. Ansaldo and Matthews 2004, for an exploration of the Sinitic source of certain aspects of Macau Creole reduplication). The other instance of linguistic transfusion involving Asian PLCs to be discussed here concerns the implantation of one such creole in the city of Batavia (modern‐day Jakarta) and nearby Tugu, on the island of Java. Considering that this particular territory never had an official Portuguese presence, the use of a PLC until at least the late nineteenth century may come as a surprise. Batavia and Tugu Creole, now extinct, is known to us through a few written sources – most importantly, Schuchardt (1891) – which have been collated and analyzed in Maurer (2011).10 As it turns out, the establishment of this language on Java was the product of multiple migrations converging onto Batavia – earlier the center of the Verenigde Oost‐ Indische Compagnie (Dutch East Indies Company)  –  from across Asia, especially locations which the Dutch had conquered from the Portuguese around the mid seventeenth century. The phenomenon is clearly put in a 1708 memorial by two Protestant priests in charge of the “Portuguese Church” of Batavia: The Portuguese language is used daily and privately by the slaves of the families who come from Ceylon and the [Coromandel; HC] Coast; by all the owners of the slaves and their children in their daily interaction with the slaves and the indigenous Christians; by the families and people who come from Siam, Malacca, Bengal, the Coromandel Coast,

Contact and Portuguese‐Lexified Creoles  477 the island of Ceylon, the Malabar Coast, Surat and even Persia; and even the heathens who live in this city and trade with the Christians or their slaves learn to speak Portuguese. (Quoted in Lopes 1936: 67–8; trans. HC)

The multiple provenance of these PLC‐speaking populations appears to have created, among this community, a setting easily conducive to dialect leveling. It is true that, as noted by Baxter (1996: 310), Batavia and Tugu Creole data look remarkably like Malacca Creole, which is unsurprising considering the geographical proximity and similarities in terms of the linguistic ecologies of colonial Malacca and Batavia. However, when we dig deeper, we also find certain forms, lexical as well as functional, which establish links with South Asian PLCs (see Cardoso 2019), although this connection has not been systematically explored.

3.2  A Foundational Role? One of the most interesting effects of the inaugural role of Portuguese in the European global linguistic expansion of the Modern Age is that, for various reasons and to different degrees, Portuguese often came to impact the linguistic repertoire of the colonizers who followed and, on many occasions, occupied places where this language had already created roots. Even discarding strong formulations of a monogenetic nature, several theories have been proposed that involve Portuguese or a Portuguese‐lexified contact variety in the formation or development of a particular language which does not strictly fall under the PLC category. Here we will mention only a few of them.11 The various English‐lexified creoles of Suriname are a case in point. While there is a discernible Portuguese lexical element in all of them, including Sranan (Smith 1987) and Ndyuka (Huttar 1989), the Western Maroon Creoles, which consist of Saramaccan and Matawai, stand out in this respect. Consider the following excerpt of a story in Saramaccan (transcribed in Bakker, Smith, and Veenstra 1994: 165–6); the elements which can be ascribed a Portuguese etymon have been underlined: (3)

Hɛn totómbotí táa wɛ  a  o‐du  lúku tu. then woodpecker  said  well    3SG IRR‐do look  too ‘Then the woodpecker said that he would try too.’ ‘Gaamá, mi o‐gó náki lúku.’ Hɛn déé otowan táki  táa: granman,   1SG IRR‐go hit look then DEF.PL  other say that ‘“Chief, I am going to try to hit it.” Then the others said’: Kú ún‐búka,   i lánga bákahédi  ku dí  gaán taku fi‐i ɗɛ’? with which‐beak, 2SG long back.head  with DEF.SG big ugliness of‐2SG there ‘With what beak, you long back‐of‐the‐head, with your great ugliness?’

A study of the Saramaccan lexicon by Smith and Cardoso (2004) which pitched Portuguese‐derived lexemes against English‐derived lexemes (i.e. ignoring the other, minority lexical sources) found a nearly equal distribution in the domain of nouns (176 from Ptg., 179 from Eng.), a smaller but still highly significant proportion of Portuguese‐derived adjectives (27 from Ptg., 51 from Eng.), and a majority of Portuguese‐derived verbs (154 from Ptg., 121 from Eng.; see also Perl 2000). Such an even division of labor makes it possible to consider Saramaccan a mixed creole, although the distribution of the Portuguese‐ and English‐derived lexicon is not entirely equal: according to Smith (1987), while the Portuguese element is very solid among the basic vocabulary of Saramaccan, accounting for nearly 35% (but just under 4% in Sranan), it is much less well‐represented in the repertoire of function words, at just 16% (and 1.5% in Sranan). Smith’s interpretation is that this typifies a scenario in which, in the case of Saramaccan, an originally English‐lexified creole integrated the

478  Hugo C. Cardoso Portuguese element at a later stage, but the truth is that several different explanations have been proposed for the remarkable lexical impact of Portuguese in Suriname. Some of these run contrary to Smith’s scenario. Voorhoeve (1973), for instance, argues for a Portuguese‐ lexified origin for Saramaccan (and Sranan), and later relexification towards English. There has been some disagreement on the origin of the Portuguese (or Iberian, in some theories) lexical element in Suriname. Several possible routes have been proposed, including the diffusion of the Mediterranean Lingua Franca or a West African pidgin/creole, the impact of previously influenced African languages, or of a French‐lexified baraguoin circulating in the Caribbean from the sixteenth century onwards which had, however, absorbed several Ibero-Romance words (see Arends 2017: 39–43). However, according to some authors, the most important historical development in this respect was the influx of  (Judeo-)Portuguese-speaking Sephardic Jews from the short‐lived Dutch colony in Northeastern Brazil (1630–54) in the mid seventeenth century. When the Portuguese retook Pernambuco, the resident Sephardic Jews moved to Suriname (either directly, or by way of Guyana or Europe) and set up several plantations on the Upper Suriname River. Some authors believe that, in the process, they also brought over significant numbers of slaves from their earlier Brazilian plantations (e.g. Goodman 1987), but this is controversial.12 Be it as it may, by the eighteenth century, several historical sources mention a language known as Djutongo, which appears to have been a PLC used in these plantations and is thought to be the precursor of the West Maroon Creoles (see Arends 2017: 182–8). This particular mid‐seventeenth‐century migration of Portuguese‐speaking Jews from Brazil has also been seen as relevant to understand parallel linguistic developments in Aruba, Bonaire, and Curaçao. The creole spoken on these islands, Papiamentu, was traditionally classified as a Spanish‐lexified creole, but, here too, the Portuguese lexical element is evident. Whereas several lexemes can only be derived from Spanish, many may be either from Spanish or Portuguese, while others still can only be derived from Portuguese (e.g. ainda ‘still’, bai ‘to go’, bon ‘good’, forsa ‘strength’, nasementu ‘birth’, trese ‘to carry, wear’; see Grant 2008; Lipski 2008). First of all, as with Saramaccan, this has motivated the question of whether Papiamentu originated as a Portuguese‐ or Spanish‐lexified language, which has proven a bone of contention among scholars and speakers alike.13 In addition, the origin of such a significant Portuguese element has also attracted considerable debate. A number of explanatory theories have been proposed, including: (i) the transfer of a West African PLC or PLP; (ii) the influence of the Judeo‐Portuguese spoken by the Sephardic Jews relocating from Northeastern Brazil or from Europe; and (iii) even the influence of an Asian‐Portuguese creole (see Grant 2008). Recent comparative work has favored an Upper Guinea connection, on the basis that some essential characteristics of Papiamentu have very concrete correlates among the PLCs of Cape Verde and Guinea‐Bissau. Jacobs (2009) explores a few of these, including the fact that both Papiamentu and the Upper Guinea PLCs make a distinction ­between simple and emphatic forms within their pronominal system – related to the pronominal forms (a)mi, (a)bo, (a)nos discussed on p. 474 in connection with example (1) – but also correspondences concerning the distribution of mono‐ or bimorphemic question words, the form and semantics of prepositions, conjunctions, and so forth. The debate is still ongoing, but the recognition of the pervasiveness of the Portuguese element has now made it common to classify Papiamentu as an “Ibero‐Romance‐lexified” or “Spanish/Portuguese‐ lexified” creole. Over in Asia, similar proposals have also been advanced. One of the earliest – and perhaps the most contentious one – has to do with a putative Portuguese origin for Chabacano, the cluster of Spanish‐lexified creoles of the Philippines. The recognition of certain similarities between Chabacano and the Asian PLCs (especially those of Southeast Asia and Macau, see e.g. Batalha 1960) motivated the hypothesis  –  formulated particularly strongly by Whinnom (1956, 1965) – that the origin of Chabacano lay in the resettlement of a PLC‐speaking

Contact and Portuguese‐Lexified Creoles  479 community from the Moluccan island of Ternate to the Philippines, around 1660, followed by relexification towards Spanish. The applicability of this scenario to the history of all ­varieties of Chabacano has been questioned, although the possibility remains that it may be relevant for the formation of at least one: Ternateño (see Fernández 2012: 226). Less controversial than the proposed Portuguese past of Chabacano is the role that Portuguese – or rather, a PLC or PLP – played in the development of Chinese Pidgin English. For a long time, the Portuguese in Macau had a privileged role in the Canton trade, and, as such, a contact variety of Portuguese definitely circulated in the Pearl River Delta. It was only in the eighteenth century that the British became especially prominent in the region’s trade (van Dyke 2005). This transition in terms of economic influence led to the gradual replacement of a Portuguese‐based contact language with an English‐based one, which came to be known as Chinese Pidgin English. Interestingly, in a study of records of Chinese Pidgin English, Hall (1944) notices that the earliest ones were characterized by a large number of Portuguese-derived words (such as e.g. joss for ‘God’, a term preserved in the widespread English term joss‐stick), many of which, as time went by, were progressively replaced with English‐derived terms. In fact, one of the earliest known attestations of Chinese Pidgin English (Noble 1762: 240), collected in the 1740s during a trip to Canton, actually consists of a sentence that is mostly composed of Portuguese‐derived terms: (5)

Carei grandi hola, pickenini hola? want old   whore young whore ‘Do you want an old whore or a young whore?’

In this sentence, only hola has an English source; all the other terms are Portuguese‐derived: carei from Ptg. querer ‘to want’, grandi from Ptg. grande ‘big’, and pickenini from Ptg. pequenino ‘small’.14 Examples like this, not atypical in the 1700s, call into question the classification of the language as an English‐based pidgin. One hypothesis which has received some attention is that earlier contact varieties of Portuguese did not simply provide loanwords to Chinese Pidgin English, but were in fact its ancestors. According to this possibility, Chinese Pidgin English would have been formed, to some extent, through a process of relexification of a previous PLC or PLP. Matthews and Li (2012), for instance, interpret certain functional features of Chinese Pidgin English (such as the use of for as a complementizer or have as a locative copula) as retentions from a Portuguese‐based ancestor, since similar constructions are attested elsewhere among the Asian PLCs.15

4  Caught in the Middle One consequence of the PLCs’ relative antiquity is that the opportunity presents itself, more than perhaps in the case of younger languages, for sociopolitical and linguistic conditions to change over time and for various types of language change to operate. In fact, as already mentioned, among the PLCs there is a diversity of situations: some have had constant and substantial contact with the lexifier but not with the major substrate(s) (e.g. Capeverdean); some have remained in touch with their substrate language(s) but not the lexifier (e.g. Sri Lanka Creole); yet others have maintained contact with both (e.g. Guinea‐Bissau Creole); and others lost touch with both at some point (e.g. Fa d’Ambu). In addition, significant changes in these languages’ linguistic and sociopolitical contexts (including colonial takeovers and migrations) occurred at different times in history, which makes it possible to observe the long‐term consequences of variables such as the presence or absence of the

480  Hugo C. Cardoso lexifier, substrate languages, or adstrate languages, but also of the position of the creole and its speakers in their wider communities. Wherever PLCs have remained in close contact with Portuguese, a number of different trajectories can be identified. In places such as Cape Verde and Guinea‐Bissau, the creoles have gained wide currency but, for most of their history, have been subjugated to Portuguese in the formal and official domains. In fact, Portuguese remains the official language of these two countries, although the creoles are increasing their public visibility and taking steps towards an enhanced political status. Nonetheless, this particular co‐ existence has led to the development of considerable linguistic stratification, which, in the case of Guinea‐Bissau, has been interpreted as a creole continuum brought about by decreolization (Couto 2009). In other places, however, Portuguese has exerted a more pernicious pressure. In São Tomé and Príncipe, for instance, the creoles have been losing ground very quickly, to the point that their demise is now a possibility (Hagemeijer 2016). In Diu, it has been observed that, decades after Portuguese lost any official status in the territory, it still exerts some normative pressure on the creole, posing certain communicative constraints and defining language attitudes among the speech community (Cardoso 2007, 2016). As for Macau, the fact that the creole is now all but gone has also been attributed to a gradual process of shift – or rather, according to Baxter (2009), assimilation of the creole into a Macanese norm of Portuguese  –  motivated by the institutionalization of Portuguese during the late colonial period. In some former Portuguese outposts that went through drastic political changes, normally accompanied by the demotion or removal of Portuguese, the local PLCs have disappeared. Among those are the creoles of Cochin, Mangalore, Nagapattinam or Bengal (India), of Burma (i.e. Myanmar), and of Flores (Indonesia), to name but a few. In other places, however, these languages have stood their ground, but the removal of the lexifier from the ecology had profound effects on their linguistic development. The Asian PLCs provide a very interesting laboratory to observe the effects of such an event, since those which have lost contact with Portuguese did so at different points in time, and also because, as endogenous creoles (for the most part), they did remain in close contact with their former substrates. This situation has resulted in considerable convergence towards the adstrate(s)  –  or even, according to certain interpretations, metatypy. The creole of Sri Lanka is one of the most iconic examples of this process, as demonstrated by Smith (1979). The island was taken over by the Dutch in the mid seventeenth century, after one and half centuries of Portuguese presence, and, while this did not mean the complete eradication of Portuguese from the territory, it did bring about a very radical break. For the ensuing centuries, then, the PLC of Sri Lanka developed with little to no contact with the lexifier but in close proximity with the substrate/adstrate languages, Tamil and Sinhala, which the speakers themselves mastered. As a result, currently, the language  –  which is p ­ reserved mostly in Tamil‐majority areas  –  shows evident signs of grammatical convergence towards Tamil. The following is an example given by Smith (1979: 201), in both the creole (Sri Lanka Portuguese, SLP) and Batticaloa Tamil (BT), to illustrate the degree of isomorphism that obtains between the two: (6)

SLP BT

e:w eli‐pə diñe:ru na:n avan‐ukku calli 1SG 3SG.M‐DAT money ‘I gave him the money.’

‐ya ‐ACC

ja:‐ dá. kúTu  ‐tt  ‐an. PST give ‐PST‐CONC

Despite some differences, there are also important similarities, not only in terms of word order but also of case‐marking, which become even more significant once we compare the sentences in (6) with the equivalent in Portuguese (7):

Contact and Portuguese‐Lexified Creoles  481 (7)

Eu dei‐lhe 1SG give.1SG.PST‐3SG.OBL ‘I gave him the money.’

o DEF.M

dinheiro. money

Sri Lanka Portuguese is by no means the only Asian PLC to exhibit the structural influence of its substrate(s) or adstrate(s) (settings such as those that produce endogenous creoles call into question the very distinction between the two concepts), but the degree of convergence is perhaps only matched, among this group of languages, by the creole of the Malabar, which is understandable since the uprooting of Portuguese occurred in much the same period and circumstances as in Sri Lanka. In fact, recent ­comparative studies of the Asian PLCs have drawn attention to the long‐term impact of the removal of the lexifier from a creole’s ecology. Looking at word order features and comparative constructions, respectively, both Smith (2012) and Cardoso (2012) attempted to rank the various Asian PLCs (and also Chabacano, in the case of Smith’s study) according to their similarity with either the lexifier or the relevant substrate/adstrate. The results are given in Table 23.3: There are certain regularities in the two rankings which correlate neatly with the chronology of the break (if any) with the lexifier. At the end of the scale that indicates a closer similarity to the lexifier, we consistently find the Asian PLCs which remain in contact with Portuguese or only began to lose some connection in the twentieth century (Macau, Daman, and Diu); at the other end, signaling a strong resemblance with the relevant substrate(s)/ adstrate(s), are the ones for which contact with Portuguese was substantially severed in the mid seventeenth century (Sri Lanka, Malacca, Batavia/Tugu; also the Malabar, in the case of Cardoso’s study); the creole of Korlai, spoken in a region from which Portuguese was dislodged in the eighteenth century, hovers in the middle. This exercise provides a clear visualization of what Smith (2012: 141) calls the “tug‐of‐war” the lexifier and the substrate(s)/ adstrate(s) enter into as potential developmental models for the creole and sheds new light on the sociolinguistic condition of many creoles as languages caught in the middle of competing diachronic pulls.

Closer to substrate

Closer to lexifier

Table 23.3  Ranking of Asian creoles according to their similarities with the lexifier or the substrate/adstrate; adapted from Cardoso (2012) and Smith (2012). Cardoso (2012)

Smith (2012)

Daman

Macau

Korlai

Daman

Macau (or Diu)

Diu

Diu (or Macau)

Chabacano

Malabar

Korlai

Malacca =

Malacca

Batavia/Tugu =

Batavia/Tugu

Sri Lanka

Sri Lanka

482  Hugo C. Cardoso

5 Conclusion Earlier in this chapter, we made the point that PLCs were especially important for the early creolists, who developed their approaches to pidgins and creoles in the late nineteenth century and the beginning of the twentieth century. What is interesting is that, despite the fact that creole studies did develop a discernible focus on English‐ and French‐ lexified creoles (and especially those of the West Atlantic) throughout the twentieth century, PLCs have more than stood their ground within the field. In this chapter, we have discussed a number of prominent issues in scholarship about PLCs which help explain their relevance for creole studies – and for language contact as a whole – and also illustrate the diversity of debates that require a reference to them. Owing to space constraints, the range of possible issues to cover was restricted here to those that in some way relate to the circumstance that the PLCs are among the oldest European‐lexified creoles – which, in turn, derives from the fact that the Portuguese were the initiators of European overseas expansion in the Modern Age. Within this narrow domain, issues of historical diffusion and speciation become paramount, and, as we have seen, much research has gone into disentangling genetic connections between particular subgroups of PLCs but also, perhaps more unexpectedly, between PLCs/ PLPs and a number of English‐lexified and Spanish‐lexified contact languages (an exercise that could be extended to include Dutch‐lexified and French‐lexified ones as well). In addition, issues of diachronic change are also amplified by the relative ancientness of the PLCs among the creoles that resulted from European expansion. In this respect, we have highlighted the contribution of these languages to an understanding of how post‐formative stages of creoles can be impacted by the competing influences of the lexifier and of the adstrates, and to what extent changes in the status of one or the others can determine their evolution. However, evidently, these do not exhaust the contributions that PLCs have made and stand to make to linguistics. Untouched in this chapter are, for instance, the many ways in which synchronic studies of PLCs have benefited typology and formal linguistics, or how PLCs have also become prominent in sociolinguistics (e.g. language death and preservation, language policy, variation, and diaspora), language acquisition, or the theory and practice of language documentation. The field of PLC studies is, in fact, quite dynamic. To illustrate the history of creole studies as a discipline, Sousa, Mücke, and Krämer (2019: Fig. 1) charted the number of studies focusing on PLCs throughout the years, revealing a significant amount in the 1880s, and then a surge after the 1950s, with a peak in the 1980s. However, their survey ends in 1992. As a matter of fact, in the period since then, some important developments have taken place. One of the most significant of those, I would argue, concerns the documentation and description of some PLCs which were insufficiently known (or entirely unknown) until then – including, among others, those of Annobón, Diu, Daman, Korlai, and the Malabar – but also of particular varieties of better‐known creoles – such as some of the regional variants of Capeverdean Creole. The enhanced range and availability of PLC data, aided by the constitution of freely‐accessible corpora, open up fresh analytical possibilities. In addition, the inclusion of many PLCs or varieties of PLCs in comparative databases such as Comparative Creole Syntax (Holm and Patrick 2007) or The Atlas of Pidgin and Creole Language Structures (Michaelis et  al. 2013)16 is also highly relevant. If, to the growing availability of PLC data, we add the consolidation of a dedicated scientific society (the Association of Portuguese‐ and Spanish‐Lexified Creoles) and the establishment of focused publications (such as the Journal of Ibero‐Romance Creoles), these languages look set to retain their scholarly relevance.

Contact and Portuguese‐Lexified Creoles  483

Abbreviations 1 = first person (pronoun), 2 = second person (pronoun), 3 = third person (pronoun), CMPR = comparative marker, CONC = concord, COP = copula, DAT = dative (case), DEF = definite (determiner), F = feminine (gender), IRR = irrealis, M = masculine (gender), OBL = oblique (case), PL = plural, PLC = Portuguese‐lexified creole, PLP = Portuguese‐lexified pidgin, PST = past (tense), REL = relativizer, SG = singular.

NOTES 1 For published overviews of the Portuguese‐based creoles, spoken currently and in the past, see Holm (1989), Tomás (1992), Couto (1996), Baxter (1996), Tomás and Pereira (1998), Pereira (2006), Cardoso, Hagemeijer, and Alexandre (2015), as well as several survey chapters in vol. 2 of Michaelis et  al. (2013). For a comprehensive list of attested, referenced and probable Portuguese‐based ­pidgins and creoles, see Smith (1995). 2 Itself a reinterpretation of Reinecke’s earlier (1937) classification which, in addition to trade ­jargons (corresponding to pidgins), distinguished plantation creole dialects from settlers’ creole dialects. 3 For an overview of the issue with respect to Brazil, see Lucchesi, Baxter, and Ribeiro (2009), ­especially ch. 1. For the Goan case, see Dalgado (1921) and Cardoso (2016: 82–4). 4 For a survey of sources about the PLCs in general, see Cardoso, Hagemeijer, and Alexandre (2015). For Asia, see also Tomás (1992). 5 See Smith (2016) for a critical overview of early sources for Ceylon (Sri Lanka) Creole. 6 For overviews of monogenetic theories, see Holm (1988: 44–52), and den Besten, Muysken, and Smith (1995). 7 See Jacobs (2010) for an overview. 8 For an overview of debates surrounding the (dis)unity of the Asian PLCs, see Ansaldo and Cardoso (2009). 9 All sentences in example (2) given in Cardoso (2012: 101, 102, 105, 107). 10 See also Maurer (2013) for linguistic samples and grammatical analysis. One further source, not included in this lot, consists of a manuscript recently identified in the library of Lisbon’s Museu Nacional de Arqueologia (see Castro et al. 2019). 11 Further cases which will not be treated here involve Palenquero, for which a connection with the Gulf of Guinea PLCs has been proposed (cf. Schwegler 1999); Hawaiian Pidgin, on which the potential impact of Portuguese is explained by the presence of migrants from Portugal, from the late nineteenth‐century onwards (see Knowlton 1960; Siegel 2000; Cardoso 2016); and Afrikaans, for which an Asian‐Portuguese connection has been proposed (e.g. Hesseling 1899; den Besten 2012). 12 See Smith (2002: 137–41) and Arends (2017: 63–70) for a summary of this particular debate. 13 See Lipski (2008) and Jacobs (2009) for an overview. 14 Notice however that, as mentioned in Section  3, pickenini is part of a set of Portuguese‐derived words which occur in many non‐Portuguese‐lexified pidgins and creoles; therefore, the use of this term in (5) could be the result of diffusion through (a variety of) English. 15 For a summary of the issue, see also Li and Matthews (2016). 16 In Holm and Patrick (2007), 4 (or 5, if you include Papiamentu) of the 18 featured languages are PLCs: Angolar and the creoles of Cape Verde, Guinea‐Bissau, and Korlai. In Michaelis et al. (2013), out of 76 languages represented, 14 (or 15, once again depending on the inclusion of Papiamentu) are PLCs: 3 varieties of Capeverdean Creole (Santiago, Brava, and São Vicente), Angolar, and the Creoles of Guinea‐Bissau, Casamance, São Tomé, Príncipe, Annobón, Diu, Korlai, Sri Lanka, and Malacca, and even the extinct creole of Batavia.

484  Hugo C. Cardoso

REFERENCES Ansaldo, Umberto and Hugo C. Cardoso 2009. Introduction. In Hugo C. Cardoso and Umberto Ansaldo (eds.), Accounting for Commonalities among the Portuguese‐Lexified Creoles of Asia. Special issue of Journal of Pidgin and Creole Languages 8.2: 3–10. Ansaldo, Umberto and Stephen Matthews 2004. The origins of Macanese reduplication. In Geneviève Escure and Armin Schwegler (eds.), Creoles, Contact, and Language Change: Linguistic and Social Implications. Amsterdam: Benjamins, pp. 1–19. Arends, Jacques 2017. Language and Slavery: A Social and Linguistic History of the Suriname Creoles. Amsterdam: Benjamins. Baker, Philip and Magnus Huber 2001. Atlantic, Pacific, and world‐wide features in English‐ lexicon contact languages. English World‐Wide 22.2: 157–208. Bakker, Peter, Norval Smith, and Tonjes Veenstra 1994. Saramaccan. In Jacques Arends, Pieter Muysken, and Norval Smith (eds.), Pidgins and Creoles: An Introduction. Amsterdam: Benjamins, pp. 165–78. Baptista, Marlyse, Heliana Mello, and Miki Suzuki 2007. Cape Verdean, or Kabuverdianu, and Guinea‐Bissau, or Kriyol (Creole Portuguese). In John Holm and Peter Patrick (eds.), Comparative Creole Syntax. London: Battlebridge, 53–82. Batalha, Graciete N. 1960. Coincidências com o dialecto de Macau em dialectos espanhóis das Ilhas Filipinas [Coincidences between the dialect of Macau and the Spanish dialects of the Philippine islands]. Boletim de Filologia 19: 295–303. Batalha, Graciete N. 1965–66. A contribuição malaia para o dialecto macaense [Malay contribution to the Macanese dialect]. Boletim do Instituto Luís de Camões 1.1: 7–19; 1.2: 99–108. Batalha, Graciete N. 1974. Língua de Macau: o que foi e o que é [The language of Macau: what it was and what it is]. Macau: Imprensa Nacional de Macau. Baxter, Alan N. 1996. Portuguese and Creole Portuguese in the Pacific and Western Pacific Rim. In Stephen A. Wurm, Peter Mühlhäusler, Darrell T. Tryon (eds.), Atlas of Languages of Intercultural Xommunication in the Pacific, Asia and the Americas, vol. II.1. Berlin: de Gruyter Mouton, pp. 299–338.

Baxter, Alan N. 2009a. O português em Macau: contacto e assimilação [Portuguese in Macau: contact and assimilation]. In Ana M. Carvalho (ed.), Português em contacto [Portuguese in contact]. Madrid: Iberoamericana, pp. 277–312. Baxter, Alan N. 2009b. Causative and facilitative serial verbs in Asian Ibero‐Romance Creoles – a convergence of substrate and superstrate systems? Journal of Portuguese Linguistics 8.2: 65–90. Baxter, Alan N. and Augusta Bastos 2012. A closer look at the post‐nominal genitive in Asian Creole Portuguese. In Hugo C. Cardoso, Alan N. Baxter, and Mário Pinharanda Nunes (eds.), Ibero‐Asian Creoles: Comparative Perspectives. Amsterdam: Benjamins, pp. 47–79. Bickerton, Derek 1988. Creole languages and the bioprogram. In Frederick J. Newmeyer (ed.), Linguistics: The Cambridge Survey, vol. 2: Linguistic Theory, Extensions and Implications). Cambridge: Cambridge University Press, pp. 268–84. Brito, António de Paula 1887. Dialectos crioulos‐ portugueses. Apontamentos para a gramática do crioulo que se fala na ilha de Santiago de Cabo Verde [Creole‐Portuguese dialects. Notes for a grammar of the creole spoken on the Cape Verdean island of Santiago]. Reprinted in Jorge Morais‐Barbosa (ed.) 1967. Estudos Linguísticos Crioulos [Creole linguistic studies]. Lisbon: Academia Internacional da Cultura Portuguesa, pp. 329–404. Buchstaller, Isabelle, Anders Holmberg, and Mohammad Almoaily (eds.) 2014. Pidgins and Creoles beyond Africa‐Europe Encounters. Amsterdam: Benjamins. Cardoso, Hugo C. 2007. Linguistic traces of colonial structure. In Eric Anchimbe (ed.), Linguistic Identity in Postcolonial Multilingual Spaces. Newcastle‐upon‐Tyne: Cambridge Scholars Publishing, pp. 164–81. Cardoso, Hugo C. 2012. Luso‐Asian comparatives in comparison. In Hugo C. Cardoso, Alan N. Baxter, and Mário Pinharanda Nunes (eds.), Ibero‐Asian Creoles: Comparative Perspectives. Amsterdam: Benjamins, pp. 81–123. Cardoso, Hugo C. 2016. O português em contacto na Ásia e no Pacífico [Portuguese in contact in Asia and the Pacific]. In Ana Maria Martins and Ernestina Carreira (eds.), Manual de Linguística Portuguesa [Handbook of

Contact and Portuguese‐Lexified Creoles  485 Portuguese Linguistics]. Berlin: de Gruyter Mouton, pp. 68–97. Cardoso, Hugo C. 2019. The synchrony and diachrony of an Asian‐Portuguese causal morpheme. Journal of Ibero‐Romance Creoles 9.1: 27–54. Cardoso, Hugo C., Tjerk Hagemeijer, and Nélia Alexandre 2015. Crioulos de base lexical portuguesa [Portuguese‐lexified creoles]. In Maria Iliescu and Eugeen Roegiest (eds.), Manuel des anthologies, corpus et textes romans [Handbook of Romance anthologies, corpora, and texts]. Berlin: de Gruyter Mouton, pp. 670–92. Castro, Ivo, Hugo C. Cardoso, Gijs Koster, Alexander Adelaar, and Alan Baxter 2019. The Lisbon book of pantuns. O Arqueólogo Português series V.6/7 (2016–2017): 313–317. Chaudenson, Robert 1979. Les créoles français [The French Creoles]. Paris: Nathan. Clements, J. Clancy 2000. Evidência para a existência de um pidgin português asiático [Evidence for the existence of an Asian Portuguese pidgin]. In Ernesto d’Andrade, Dulce Pereira, and Maria Antónia Mota (eds.), Crioulos de Base Portuguesa [Portuguese‐based Creoles]. Braga: Associação Portuguesa de Linguística, pp. 185–200. Clements, J. Clancy 2009. The Linguistic Legacy of Spanish and Portuguese; Colonial Expansion and Language Change. Cambridge: Cambridge University Press. Coelho, Francisco Adolfo 1886. Os dialectos romanicos ou neo‐latinos na Africa, Asia e América [The romance or neo‐Latin dialects in Africa, Asia, and America]. Boletim da Sociedade de Geografia de Lisboa 2nd series (3), 129–96; 3rd series (8), 451–78; 6th series (12), 705–55. Costa, Joaquim V. Botelho daand Custódio José Duarte 1886. O crioulo de Cabo Verde. Breves estudos sobre o crioulo das ilhas de Cabo Verde oferecidos ao Dr. Hugo Schuchardt [The Creole of Cape Verde. Short studies about the Creole of the islands of Cape Verde offered to Dr. Hugo Schuchardt]. Reprinted in Jorge Morais‐Barbosa (ed.) 1967. Estudos linguísticos crioulos [Creole linguistic studies]. Lisbon: Academia Internacional da Cultura Portuguesa, pp. 235–328. Couto, Hildo Honório do 1996. Introdução ao estudo das línguas crioulas e pidgins [Introduction to the study of creole and pidgin languages]. Brasília: Editora UNB. Couto, Hildo Honório do 2009. O português e o crioulo na Guiné‐Bissau [Portuguese and Creole in Guinea‐Bissau]. In Ana M.

Carvalho (ed.), Português em contacto [Portuguese in contact]. Madrid: Iberoamericana, pp. 53–66. Dalgado, Sebastião Rodolfo 1900. Dialecto indo‐português de Ceilão [Indo‐Portuguese dialect of Ceylon]. Lisbon: Imprensa Nacional. Dalgado, Sebastião Rodolfo 1902–3. Dialecto indo‐português de Damão [Indo‐Portuguese dialect of Daman]. Ta‐Ssi‐Yang Kuo: Archivos e Annaes do Extremo Oriente Portuguez 3: 350–67; 4, 515–23. Dalgado, Sebastião Rodolfo 1906. Dialecto indo‐português do norte [Indo‐Portuguese dialect of the north]. Revista Lusitana 9: 142–66, 193–228. Dalgado, Sebastião Rodolfo 1917. Dialecto indo‐português de Negapatão [Indo‐ Portuguese dialect of Nagapattinam]. Revista Lusitana 20: 40–53. Dalgado, Sebastião Rodolfo 1921. Berço duma cantiga em indo‐português [Birth of a song in Indo‐Portuguese]. Revista Lusitana 22, Offprint. d’Andrade, Ernesto and Alain Kihm 2000. Stress in three creoles. In Ernesto d’Andrade, Dulce Pereira, and Maria Antónia Mota (eds.), Crioulos de base portuguesa [Portuguese‐based creoles],. Lisbon: Associação Portuguesa de Linguística, pp. 97–109. Den Besten, Hans 2012. Roots of Afrikaans: Selected Writings of Hans den Besten. (Ton van der Wouden, ed.). Amsterdam: Benjamins. Den Besten, Hans, Pieter Muysken, and Norval Smith 1995. Theories focusing on the European input. In Jacques Arends, Pieter Muysken, and Norval Smith (eds.), Pidgins and Creoles; An Introduction. Amsterdam: Benjamins, pp. 87–98. Fernández, Mauro 2012. Nenang, nino, nem não, ni no; Similarities and differences. In Hugo C. Cardoso, Alan N. Baxter, and Mário Pinharanda Nunes (eds.), Ibero‐Asian Creoles: Comparative Perspectives. Amsterdam: Benjamins, pp. 205–37. Ferraz, Luiz Ivens 1979. The Creole of São Tomé. Johannesburg: Witwatersrand University Press. Ferraz, Luiz Ivens 1987. Portuguese creoles of West Africa and Asia. In Glen Gilbert (ed.), Pidgins and Creole Languages: Essays in Memory of John E. Reinecke. Honolulu: University of Hawaii Press, pp. 337–60. Grant, Anthony 2008. The Portuguese elements in Papiamentu. In Nicholas Faraclas, Ronald Severing, and Christa Weijer (eds.), Linguistic Studies on Papiamentu. Willemstad: Fundashon pa Planifikashon di Idioma, pp. 47–72.

486  Hugo C. Cardoso Goodman, Morris 1987. The Portuguese element in the American Creoles. In Glen Gilbert (ed.), Pidgin and Creole Languages: Essays in Memory of John E. Reinecke. Honolulu: University of Hawaii Press, pp. 361–405. Hagemeijer, Tjerk 2011. The Gulf of Guinea creoles: genetic and typological relations. Journal of Pidgin and Creole Languages 26.1: 111–54. Hagemeijer, Tjerk 2015. The Gulf of Guinea creoles: a case‐study of syntactic reconstruction. In Carlotta Viti (ed.), Perspectives on Historical Syntax. Amsterdam: Benjamins, pp. 293–316. Hagemeijer, Tjerk 2016. O português em contacto em África [Portuguese in contact in Africa]. In Ana Maria Martins and Ernestina Carreira (eds.), Manual de linguística portuguesa [Handbook of Portuguese linguistics]. Berlin: de Gruyter Mouton, pp. 43–67. Hagemeijer, Tjerk and Nélia Alexandre 2012. Os crioulos da Alta Guiné e do Golfo da Guiné: uma comparação sintática [Upper Guinea and Gulf of Guinea creoles: a syntactic comparison]. PAPIA 22.2: 233–51. Hagemeijer, Tjerk and Jorge Rocha 2017. Creole languages and genes: the case of São Tomé and Príncipe. Faits de Langues 49: 167–82. Hall, R. A., Jr. 1944. Chinese Pidgin English grammar and texts. Journal of the American Oriental Society 64: 95–113. Hesseling, D. C. 1899. Het Afrikaansch: Bijdrage tot de Geschiedenis der Nederlandsche Taal in Zuid‐ Afrika. [Afrikaans: a contribution to the history of the Dutch language in South Africa] Leiden: Brill. Holm, John A. 1988. Pidgins and Creoles, vol. I: Theory and Structure. Cambridge: Cambridge University Press. Holm, John A. 1989. Pidgins and creoles, vol. II: Reference Survey. Cambridge: Cambridge University Press. Holm, John A. 2009. Atlantic features in Asian varieties of Creole Portuguese. Journal of Portuguese Linguistics 8.2: 11–22. Holm, John A. and Peter L. Patrick (eds.) 2007. Comparative Creole Syntax: Parallel Outlines of 18 Creole Grammars. London: Battlebridge. Huttar, George L. 1989. The Portuguese contribution to the Ndjuka lexicon. In (no ed.), Estudios sobre español de América y lingüística afroamericana: ponencias presentadas en el 45 Congreso Internacional de Americanistas [Studies about the Spanish of America Afro‐American linguistics: talks given at the 45th International Congress of

Americanists]. Bogotá: Instituto Caro y Cuervo, pp. 263–79. Jabocs, Bart 2009. The Upper Guinea origins of Papiamentu. Diachronica 26.3: 319–79. Jacobs, Bart 2010. Upper Guinea Creole; evidence in favor of a Santiago birth. Journal of Pidgin and Creole Languages 25.2: 289–343. Kihm, Alain and Jean‐Louis Rouge 2013. Língua de Preto, the basic variety at the root of West African Portuguese Creoles: a contribution to the theory of pidgin/creole formation as second language acquisition. Journal of Pidgin and Creole Languages 28: 203–98. Knowlton Jr., Edgar C. 1960. Portuguese in Hawaii. Kentucky Foreign Language Quarterly Review 7: 212–18. Leite de Vasconcelos, José 1901. Esquisse d’une dialectologie portugaise [Portuguese dialectology sketch]. Paris: Aillaud and Cie. Li, Michelle and Stephen Matthews 2016. An outline of Macau Pidgin Portuguese. Journal of Pidgin and Creole Languages 31.1: 141–83. Lipski, John M. 1991. On the emergence of (a)mí as subject in Afro‐Iberian pidgins and creoles. In Ray Harris‐Northall and Thomas Cravens (eds.), Linguistic Studies in Medieval Spanish. Madison: Hispanic Seminary of Medieval Studies, pp. 39–61. Lipski, John M. 2005. A History of Afro‐Hispanic Language: Five Centuries, Five Continents. New York: Cambridge University Press. Lipski, John M. 2008. Spanish‐based Creoles in the Caribbean. In Silvia Kouwenberg and John Victor Singler (eds.), The Handbook of Pidgin and Creole Studies. Oxford: Wiley‐Blackwell, pp. 543–64. Lopes, David 1936. A expansão da língua portuguesa no oriente nos séculos XVI, XVII e XVIII [The expansion of the Portuguese language in the Orient in the sixteenth, seventeenth, and eighteenth centuries]. Barcelos: Portucalense Editora. Lucchesi, Dante, Alan Baxter, and Ilza Ribeiro (eds.) 2009. O português afro‐brasileiro [Afro‐Brazilian Portuguese]. Salvador: EDUFBA. Matthews, Stephen and Michelle Li 2012. Portuguese Pidgin and Chinese Pidgin English in the Canton trade. In Hugo C. Cardoso, Alan N. Baxter, and Mário Pinharanda Nunes (eds.), Ibero‐Asian Creoles: Comparative Perspectives. Amsterdam: Benjamins, pp. 263–87. Maurer, Philippe 2011. The former Portuguese Creole of Batavia and Tugu (Indonesia). London: Battlebridge. Maurer, Philippe 2013. Batavia Creole structure dataset. In Susanne M. Michaelis, Philippe

Contact and Portuguese‐Lexified Creoles  487 Maurer, Martin Haspelmath, and Magnus Huber (eds.), Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/43 Michaelis, Susanne Maria, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) 2013. The Survey of Pidgin and Creole Languages, 3 vols. Oxford: Oxford University Press. Naro, Anthony J. 1978. A study on the origins of pidginization. Language 54: 314–47. Noble, C. F. 1762. A Voyage to the East Indies in 1747 and 1748. London: T. Becket and P. A. Dehondt. Pereira, Dulce 2006. Crioulos de base portuguesa [Portuguese‐based creoles]. Lisbon: Caminho. Perl, Matthias 2000. The Portuguese origin of the Saramaccan vocabulary reconsidered. In Ernesto d’Andrade, Dulce Pereira, and Maria Antónia Mota (eds.), Crioulos de base portuguesa [Portuguese‐based creoles]. Braga: Associação Portuguesa de Linguística, pp. 143–65. Reinecke, John 1937. Marginal languages: a sociological survey of the Creole languages and trade jargons (Yale University Ph.D. dissertation). Schuchardt, Hugo 1882a. Kreolische Studien I. Über das Negerportugiesische von S. Thomé [Creole Studies I. On the Black‐Portuguese of São Tomé]. Sitzungsberichte der Kaiserlichen Akademie des Wissenschaften zu Wien 101.II: 889–917. Schuchardt, Hugo 1882b. Kreolische Studien II. Über das Indoportugiesische von Cochim [Creole Studies II. On the Indo‐Portuguese of Cochin]. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien (Philosophisch‐historische Klasse) 102: 799–816. Schuchardt, Hugo 1883a. Kreolische Studien III. Über das Indoportugiesische von Diu [Creole Studies III. On the Indo‐Portuguese of Diu]. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien (Philosophisch‐historische Klasse) 103: 3–18. Schuchardt, Hugo 1883b. Kreolische Studien VI. Über das Indoportugiesische von Mangalore [Creole Studies VI. On the Indo‐Portuguese of Mangalore]. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien (Philosophisch‐historische Klasse) 105.III: 882–904. Schuchardt, Hugo 1888a. Beiträge zur Kenntnis des kreolischen Romanisch III. Zum Negerportugiesischen der Kapverden

[Contributions for the knowledge of the romance creoles III. On the Black‐Portuguese of Cape Verde]. Zeitschrift für Romanische Philologie 12: 312–22. Schuchardt, Hugo 1888b. Beiträge zur Kenntnis des kreolischen Romanisch II. Zum Negerportugiesischen Senegambiens [Contributions for the knowledge of the romance creoles II. On the Black‐Portuguese of Senegambia]. Zeitschrift für Romanische Philologie 12: 301–12. Schuchardt, Hugo 1888c. Kreolische Studien VII. Über das Negerportugiesische von Annobom [Creole Studies VII. On the Black‐Portuguese of Annobom]. Sitzungsberichte der Kaiserlichen Akademie des Wissenschaften zu Wien 116.I: 193–226. Schuchardt, Hugo. 1889a. Beiträge zur Kenntnis des creolischen Romanisch. V. Allgemeineres über das Indoportugiesische (Asioportugiesische) [Contributions for the knowledge of the romance creoles V. On Indo‐Portuguese (Asian‐Portuguese) in general]. Zeitschrift für Romanische Philologie 13: 476–516. Schuchardt, Hugo 1889b. Beiträge zur Kenntnis des kreolischen Romanisch IV. Zum Negerportugiesischen der Ilha do Príncipe [Contributions to the knowledge of the Romance Creoles IV. On the Black‐Portuguese of the island of Príncipe]. Zeitschrift für romanische Philologie 13: 463–75. Schuchardt, Hugo 1889c. Beiträge zur Kenntnis des kreolischen Romanisch: VI. Zum Indoportugiesischen von Mahé und Cannanore [Contributions for the knowledge of the Romance Creoles VI. On the Indo‐ Portuguese of Mahé and Cannanore]. Zeitschrift für Romanische Philologie 13: 516–24. Schuchardt, Hugo 1891. Kreolische Studien IX. Über das Malaioportugiesische von Batavia und Tugu [Creole Studies IX. On the Malayo‐ Portuguese of Batavia and Tugu]. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien (philosophisch‐historische Klasse) 122(IX): 1–256. Schwegler, Armin 1999. Monogenesis revisited. The Spanish perspective. In John R. Rickford and Suzanne Romaine (eds.), Creole Genesis Attitudes and Discourse: Studies Celebrating Charlene J. Sato. Amsterdam: Benjamins, pp. 235–64. Siegel, Jeff 2000. Substrate influence in Hawai’i Creole English. Language in Society 29: 197–236. Smith, Ian R. 1979. Convergence in South Asia: a creole example. Lingua 48: 193–222.

488  Hugo C. Cardoso Smith, Ian R. 2012. Measuring substrate influence: word order features in Ibero‐Asian Creoles. In Hugo C. Cardoso, Alan N. Baxter, and Mário Pinharanda Nunes (eds.), Ibero‐ Asian Creoles: Comparative Perspectives. Amsterdam: Benjamins, pp. 125–48. Smith, Ian R. 2016. The earliest grammars of Sri Lanka Portuguese. Papia 26.2: 237–81. Smith, Norval 1987. The genesis of the Creole languages of Suriname (University of Amsterdam unpublished Ph.D. dissertation). Smith, Norval 1995. An annotated list of creoles, pidgins, and mixed languages. In Jacques Arends, Pieter Muysken, and Norval Smith (eds.), Pidgins and Creoles: An Introduction. Amsterdam: Benjamins, pp. 331–74. Smith, Norval 2002. The history of the Surinamese Creoles II. In Eithne Carlin and Jacques Arends (eds.), Atlas of the Languages of Suriname. Leiden: KITLV. Smith, Norval and Hugo C. Cardoso 2004. A new look at the Portuguese element in Saramaccan. Journal of Portuguese Linguistics 3: 115–47. Sousa, Silvio Moreira de, Johannes Mücke and Phillipp Krämer 2019. A history of creole studies. In Oxford Research Encyclopedia of Linguistics. Available at https://oxfordre. com/linguistics/view/10.1093/ acrefore/9780199384655.001.0001/ acrefore‐9780199384655‐e‐387

Taylor, Douglas R. 1961. New languages for old in the West Indies. Comparative Studies in Society and History 3.3: 277–88. Thompson, R. W. 1961. A note on some possible affinities between the creole dialects of the Old World and those of the New. In Robert B. Le Page (ed.), Proceedings of the 1959 Conference on Creole Language Studies. London: Macmillan, pp. 107–13. Tomás, Maria Isabel 1992. Os crioulos portugueses do oriente: uma bibliografia [The Portuguese creoles of the Orient: a bibliography]. Macau: Instituto Cultural de Macau. Tomás, Maria Isabel and Dulce Pereira 1998. Os espaços do crioulo. Textos nos crioulos de base portuguesa [The spaces of the creole. Texts in the Portuguese‐based creoles]. Lisbon: Comissão Nacional para as Comemorações dos Descobrimentos Portugueses. Van Dyke, Paul A. 2005. The Canton Trade: Life and Enterprise on the China Coast, 1700–1845. Hong Kong: Hong Kong University Press. Voorhoeve, Jan 1973. Historical and linguistic evidence in favor of the relexification theory in the formation of creoles. Language in Society 2: 133–45. Whinnom, Keith 1956. Spanish Contact Vernaculars in the Philippine Islands. Hong Kong: Hong Kong University Press; Oxford: Oxford University Press. Whinnom, Keith 1965. Contacts de langues et emprunts lexicaux: the origin of the European‐ based creoles and pidgins. Orbis 14: 509–27.

24 Contact and the Celtic Languages JOSEPH F. ESKA The Celtic languages, once spoken across much of Central and Western Europe and even into Asia Minor, display evidence of contact with many languages from their first attestation in the late seventh century bce. The non‐typical features of the Insular Celtic languages vis‐à‐vis other Indo‐European languages, perhaps most famously basic verb‐initial clausal configuration, as attested in Irish1 and in the earliest records of the Brittonic languages (Old Welsh and Old Southwest Brittonic),2 conjugated prepositions, and polypersonal verbs which code agreement for both subject and object, have occasioned a considerable body of scholarship on the question of whether they are to be attributed to prehistoric contact with a substratal language and, if so, what this unattested language must have been like. As minority languages in recent centuries, the contemporary Celtic languages show the continuing effects of contact with English in Ireland (Stenson 1993) and Britain (Parry‐Williams 1923) and French in Brittany (Piette 1973). This chapter will summarize the evidence for contact in the early history of the Celtic languages and then will focus upon the argument for contact in the prehistory of the Insular Celtic languages.3

1  Possible Contact with (an) Unknown Language(s) There are two developments in the early history of Celtic that could be the result of contact with unknown languages. The first is the development of stress on the initial syllable of a word, away from the moveable pitch accent that is reconstructed for proto‐Indo‐European and attested, for example, in Vedic Sanskrit and Ancient Greek. Such a development is also attested in Germanic and Italic. It is uncertain whether this development is the result of contact with (an) unknown language(s) or an internal development. Salmons (1992: 135–85) provides a general orientation to the question. The second feature is the change in contrast in the plosive series from one of voicing, i.e. /p t k/ vs. /b d ɡ/, as reconstructed for proto‐Indo‐European and generally presumed to be continued in Celtic, to one of aspiration, i.e. /ph th kh/ vs. /p t k/. This is demonstrated for the modern Celtic languages by Voice Onset Time measurements, which show a long period of aspiration after initial /ph th kh/ and a shorter period of aspiration after initial /p t k/, i.e. pre‐voicing is entirely lacking; see Welby, Ní Chiosáin, and Ó Raghallaigh (2017: 139) for data from Irish, Nance and Stuart‐Smith (2013: 137) for data from Scottish Gaelic, and Ball (1984: 15) for data from Welsh. Pedersen (1909: 242) states that the voiceless plosives of proto‐Celtic were aspirated, but offers no opinion as to how they came to be so. Koch (1990:

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

490  Joseph F. Eska 198) argues that they developed as fortis allophones of lenis unaspirated voiceless plosives in Old Celtic. This is very possible at so‐called strong positions, including word‐initially (Keating, Wright, and Zhang 1999; Keating, Cho, Fougeron, and Hsu 2003), though Eska (2018a: 324–6) would date the change to proto‐Celtic4 and allow that contact with (i) non‐ Indo‐European substrate language(s) should not be ruled out as a contributing factor.5

2  Contact with Known Languages 2.1  Continental Europe The ancient Celtic languages of continental Europe are known to have been spoken in the Iberian Peninsula (Celtiberian), France (Transalpine Celtic), northern Italy and Switzerland (Cisalpine Celtic), and Asia Minor (Galatian). There is also evidence of Celtic languages having been spoken in Eastern Europe and the Balkans (Eastern Celtic and Noric, respectively). All of these languages are only fragmentarily attested, but in sufficient quantity to know that their linguistic structures were similar to those of other ancient Indo‐European languages such as Latin and Greek. Since the Continental Celtic languages are only fragmentarily preserved, we can get no more than a partial idea of the extent of the contact phenomena in which they participated. There is considerable evidence for lexical borrowing, especially, though not exclusively, with Latin, in both directions. A significant amount of borrowing from Celtic into Germanic is attested in the lexical fields of legal and social relations, military matters, technology, etc.6 Schmidt (1983) studies language contact in Roman Gaul in general terms. The evidence for borrowing into the Continental Celtic languages, of course, is minimized by the size of the attested corpus, though we do find lexical items such as Cisalpine Celtic leKaTos ← Latin lēgātus ‘ambassador’ and Transalpine Celtic πραιτωρ ← Latin praetor ‘the name of a magistrate’.7 A number of scholars have especially studied the evidence for Celtic lexical borrowings into Latin (and Greek), e.g. Schmidt (1967), Gernia (1981), and André (1985),8 and many others have focused upon Celtic substrate words in the Romance languages, e.g. Corominas (1956, 1976) for Spanish, Lambert (2003: 187–203) for French, and Campanile (1983) for Italian. Evidence for borrowing or interference in the areas of phonology, morphology, and syntax is much more difficult to identify, especially in the case of possible contact with Latin, because the linguistic features of the Celtic languages of ancient France and Italy were often very similar to those of Latin.9 There clearly must have been a protracted period of bilingualism (Adams 2003: 184–200) as evidenced, for example, by the thoroughly bilingual, complete with code‐switching, accounting records kept by the potters at the site of the ceramic factory at La Graufesenque in southern France (Adams 2003: 687–724). In some instances, Transalpine Celtic inscriptions seem to provide evidence for the late development of a mixed language in which elements of Latin and Greek are found (e.g. Meid 1980; Dröge 1989).

2.2  Ireland and Britain As on the continent, there is a lot of evidence for lexical borrowing, especially from Latin into the Celtic languages.10 Linguistic contact in Roman Britain is examined in detail by Evans (1983). The phonology of the Latin borrowings has been studied in detail for the Celtic languages, in general by Pedersen (1909: 189–242), for Old Irish by McManus (1983) and Uhlich (2004), and for the medieval Brittonic languages by Jackson (1953; see also Lewis 1943). Evidence for borrowing or interference in the areas of phonology, morphology, and syntax is rather thinner on the ground. Some examples include pre‐aspiration of plosives in Scottish

Contact and the Celtic Languages  491 Gaelic, e.g. cat ‘cat’ [khaht], which is usually taken to be the result of contact with Old Norse (e.g. Marstrander 1932: 298; Oftedal 1968: 257; Borgstrøm 1974);11 the pluperfect tense in the Brittonic languages, formed by the affixation of imperfect desinences to the preterite stem, which has been argued to be modeled on the Latin pluperfect (Mac Cana 1976); cf. Middle Welsh 1.sg. carass‐wn to Latin amāveram ‘I had loved’; and compound tenses of Breton, which are formed with an auxiliary verb and the past participle, which clearly are based upon French (Hemon 1975: 245–6), e.g.:12 (1) a. French: Vous 2.PL

avez have.2.PL.PRES

fait … make.PST‐PTCPL

b. Breton: C’hui och’eus 2.PL have.2.PL.PRES ‘You have made …’

great … make.PST‐PTCPL

3  Contact in the Prehistory of the Insular Celtic Languages 3.1  The Phonological Approach Archeological data demonstrate that Ireland and Britain had been settled by humans for many millennia before speakers of Celtic languages could have arrived there. There is no attested documentation of these languages, but some scholars have attempted to identify evidence for their survival into the first millennium ce via the examination of lexical items which bear phonological features which, they argue, cannot be Celtic, e.g. Schrijver (2000, 2005) and Mac Eoin (2007). Similarly, Hamp (e.g. 1990, 1991) has sought to identify the vowel inventory of a pre‐Indo‐European language with which various Indo‐European languages were in prehistoric contact in northern Europe.

3.2  More Comprehensive Approaches Far more attention, however, has been devoted by scholars to identifying a substratal language to which many of the non‐Indo‐European‐looking features of the Insular Celtic languages can be attributed via contact. Hewitt (2007, 2009) delineates 39 such features which have been proposed as diagnostic of contact between proto‐Insular Celtic and a language of Afro‐Asiatic type, among which are: • • • • • • • • • • •

Conjugated prepositions VSO clausal configuration Invariant relative particle Copying as a relativization strategy Special relative form of the finite verb Polypersonal verbs Infixing/suffixing alternation of personal object affixes Position of the article in genitive embeddings Lack of subject agreement with full plural NPs Verbal noun rather than infinitive Predicative particle identical to a local preposition

492  Joseph F. Eska • • • • • • • • •

Prepositional periphrastic tenses Periphrastic tenses formed with “do” + verbal noun Subordinating use of “and” Verbal noun used instead of finite verb in main clause Syntactically governed word‐initial phonological change Idiomatic genitive kinship constructions Amplification of negative by a noun after the verb Numerals followed by a singular noun Prepositional expression of “have”

3.3 Principal Advocates of the Substratum Theory and Their Views Although earlier scholars had noted parallels between the grammars of Welsh and Hebrew, Morris Jones (1900) is the first to focus upon a number of syntactic parallels between Welsh and Egyptian (and Berber to some extent). He does not propose that an Afro‐Asiatic substratum was necessary to explain the Insular Celtic facts, but does raise the question. Julius Pokorny (1926–9), on the other hand, believes firmly in the substratal explanation of a large number of features of the Insular Celtic languages, drawn not only from Afro‐Asiatic languages, but also from Bantu, Basque, Caucasian, Finno‐Ugric, and Eskimo‐Aleut. Heinrich Wagner (1959: 152–240), following in Pokorny’s footsteps, focuses most of his work upon the structure of the verbal system, but posits that a language of Afro‐Asiatic typology, but not necessarily Afro‐Asiatic genetically, must have been spoken in Ireland and Britain and deeply influenced the evolution of the Insular Celtic languages. George Brendan Adams (1975) stresses the significant portion of the Insular Celtic lexicon for which Indo‐European etymologies are not available, while also addressing some of the parallel linguistic features which have been posited as listed at the end of section 3.2. He also considers how archeological evidence may bear upon theories of an Afro‐Asiatic substratal influence on Insular Celtic. In his unpublished 1993 doctoral dissertation (extracts of which were published in 2007), Orin Gensler (1993, 2007) compares the structures of Insular Celtic and Afro‐Asiatic along with 58 other languages from around the world and identifies 17 “exotic” features shared by Insular Celtic and Afro‐Asiatic, but which are otherwise uncommon cross‐linguistically. He does not argue that this array of exotic parallels proves that the Insular Celtic languages were affected by an Afro‐Asiatic substratum, but the implication is that there could hardly be another explanation. Karel Jongeling (2000) posits a variation of the substratum hypothesis somewhat similar to that of Wagner whereby an unknown language (group) influenced the development of both Insular Celtic and Afro‐Asiatic, thus explaining the linguistic features that they share. Their geographic discontinuity is to be explained as a result of the Indo‐European conquest of the European continent, which erased any influence of this substratum in the intervening area. Theo Vennemann (e.g. 2002, 2003, 2010, 2016) views the theory that an Afro‐Asiatic substratum influenced the evolution of Celtic, and, through it, English, as effectively proven beyond all reasonable doubt.

3.4  Some Recent Critics The Afro‐Asiatic substratum theory has never found much favor with scholars of the Celtic languages. Much of the criticism, however, has been of an impressionistic nature, though recent work by two scholars are on a much higher plane.

Contact and the Celtic Languages  493 Steve Hewitt (2007; 2009) reviews the scholarship in favor of finding a substratum behind the exotic features of Insular Celtic and evaluates 39 linguistic features proposed in the literature that purport to show parallels between Insular Celtic and Afro‐Asiatic (see the list of the main features at the end of Section 3.2 above). He finds that over one third of these features are common cross‐linguistically, that at least five are not directly comparable between Insular Celtic and Afro‐Asiatic, that over 10% are only marginally attested in either Insular Celtic or Afro‐Asiatic, and that over 25% are found in only part of Insular Celtic or part of Afro‐Asiatic. If the presumed substratal effect occurred when the Insular Celtic languages were still a cohesive speech community, the last point causes one to wonder why a supposed substratal feature should appear only in Irish, or Welsh, or Breton, etc. Hewitt also calls attention, with reference to Gensler’s statistical approach, to the fact that proponents of the Afro‐Asiatic substratum theory focus exclusively on shared features (Gensler 1993, 2007). No mention is made of all of the features that are not shared. Failure to do so does not allow one to gauge the overall similarity of the two language families. And, again with reference to Gensler, Hewitt calls attention to the fact that the 17 exotic features identified as common to both Insular Celtic and Afro‐Asiatic are actually “more consistently present” in Insular Celtic than in Afro‐Asiatic, perhaps an unexpected outcome if the argument is that these shared features developed in Insular Celtic as the result of contact. Graham Isaac (2007) presents a vigorous attack against the substratum theory, noting, like Hewitt, that many of the proposed parallels with Afro‐Asiatic are common cross‐linguistically. He also pulls together many observations of earlier scholarship which argue that the exotic features of the Insular Celtic languages could have developed internally, without a  push or even guiding force of an external linguistic entity. In the end, Isaac concludes not  that the Afro‐Asiatic substratum theory is “unproven or unprovable,” but that “it is simply wrong.”

4  The Evidence of Ancient Celtic 4.1  Linguistic Features of Continental Celtic Virtually all of the scholarship on the question of an Afro‐Asiatic substratum underlying the Insular Celtic languages has failed to take note of the evidence of the Celtic languages of the ancient European continent. As noted in Section 2.1, these languages, mostly attested in the Mediterranean and the Alps where the presence of an Afro‐Asiatic substratum presumably is out of the question, resemble other ancient Indo‐European languages in their linguistic structures. It would be very important to the discussion to know whether any of the exotic features of the Celtic languages of Ireland and Britain can be found in those of ancient continental Europe. What we find is that, despite the fragmentary attestation of these languages, a number of the most striking exotic features are, indeed, present in some form.

4.2  Conjugated Prepositions It is clear that the conjugated prepositions of the Insular Celtic languages, as exemplified by Old Irish do ‘to’, continue as prepositions with attached clitic pronouns: (2)

1.sg 2.sg 3.sg

m. f. n.

dom duit dó dí dó

1.pl 2.pl 3.pl

dún dúib duaib

494  Joseph F. Eska Elsewhere in the Indo‐European languages, such structures are known to occur in the Anatolian language Hittite, e.g. katti=tti ‘with you’. Though precisely such a structure is not yet attested in the Continental Celtic languages, Transalpine Celtic does possess the connective form du=ci, composed of the preposition ‘to’ plus the clitic locative singular form ̑ , thus literally ‘to here’, and prepositional o’nda, composed of the of the deictic pronominal *kei̯ preposition ‘from’ plus a demonstrative with elided initial consonant and vowel (< *sindā). The presence of such structures guarantee that at least the later‐attested Continental Celtic languages had structures of the type that later evolved into the conjugated prepositions of the Insular Celtic languages.

4.3  Verb‐initial Clausal Configuration It is clear that the unmarked clausal configuration of the later‐attested Continental Celtic languages, late Cisalpine Celtic and Transalpine Celtic, was SVO with pro‐drop (Eska 2007), e.g.: (3)

a.

Martialis  Dannotali  ieuru     Ucuete  sosin M.NOM.SG D.GEN.SG  give.3.SG.PRET U.DAT.SG this.ACC.SG celicnon edifice.ACC.SG ‘Martialis, son of Dannotalos, gave this edifice to Ucuetis.’

b. reguc cambion straighten.1.SG.PRES crooked.ACC.SG ‘I straighten the bent thing.’ As in many SVO and SOV languages, however, the verb could be drawn to initial position for a variety of discourse purposes, e.g.: (4)

a.

albanos   pannai   extra sioxt =ii add.3.SG.PRET 3.ACC.PL.NEUT  A.NOM.SG vessel.ACC.PL beyond tud ̄(d ̄on) ccc allotment.ACC.SG    300 ‘Albanos added 300 vessels beyond the allotment.’

b.

kote Akisios Arkatoko{k}materekos to =  śoi = A.NOM.SG A.NOM.SG  PV 3.ACC.SG.MASC give.3.SG.PRET teuo‐xtonion atomi border.ACC.SG        god‐man.GEN.PL ‘Akisios Arkatokomaterekos, he gave the boundary (stones) of gods and men.’

It is likely that this initial placement of the verb, presumably into a Topic (5a) or Focus (5b) position in the left periphery of the clause, became sufficiently frequent that it was re‐analyzed as unmarked by a new generation of language learners, who constructed a grammar in which the verb moved to the Finite head and the language became VSO (6). (5)

a. [FrameP [ForceP [AboutTopP [AboutTop Verb] [ContrastTop [FocusP [FamiliarTop [FiniteP [TenseP Subject] [Tense Verb] …]]]]]]]] b. [FrameP [ForceP [AboutTopP [ContrastTop [FocusP [Focus Verb] [FamiliarTop [FiniteP [TenseP Subject] [Tense Verb] …]]]]]]]]

Contact and the Celtic Languages  495 (6)

[FrameP [ForceP [AboutTopP [ContrastTop [FocusP [FamiliarTop [FiniteP [Finite Verb] [TenseP Subject] [Tense Verb] …]]]]]]]]

This would hardly be an unusual occurrence and, in fact, has been proposed for other verb‐initial Indo‐European languages, e.g. Lycian of the Anatolian family (e.g. Garrett 1994: 37–40) and Late Latin and Old Sardinian (e.g. Wolfe 2016: 483–90).

4.4  Special Relative Form of the Verb Old Irish possessed in the first and third person plural special relative forms of the verb which continue a verbal form to which an uninflected clitic relative particle is attached, e.g. 3.pl. pres. bertae ‘who bear’ < *beronti=i̯o. The only vestiges in the Brittonic languages are Middle Welsh yssyd and Middle Breton so/zo ‘who/which is’ < *esti=i̯o. Forms of the ­identical structure as those which underlie the Insular Celtic forms are attested in Transalpine Celtic by dugijonti=jo ‘who serve’ and toncsijont=jo ‘who will destine’.

4.5  Polypersonal Verbs Just as the personal affixes of the conjugated prepositions of the Insular Celtic languages continue clitic pronominals which have suffered phonological attrition, so also do the so‐called infixed and suffixed pronouns of those languages continue clitic pronominals which have evolved into personal affixes which are exponents of object agreement, cf. the structures of the Old Irish verbs with personal object agreement affix with that of the Continental Celtic verbs with personal object agreement clitic:13 (7)

Old Irish: a. do‐s=mbeir PV‐3.SG.OBJ.FEM=give.3.SG.PRES ‘S/he gives it.’ b.

beirth‐i bear.3.SG.PRES‐3.SG.OBJ.MASC/NEUT ‘S/he bears it.’ Continental Celtic:

c.

to=śo=kote PV=3.ACC.SG.MASC=give.3.SG.PRET ‘He gave it.’

d.

sioxt=i add.3.SG.PRET=3.ACC.PL.NEUT ‘He added them.’

4.6  Infixing/Suffixing Alternation of Object Markers The evidence presented in Section 4.5 demonstrates that the infixing/suffixing alternation of object agreement markers also existed in the later‐attested Continental Celtic languages.

496  Joseph F. Eska

5  Some Conclusions The discussion in Sections 3.4 and 4 seems to undermine much of the argument for a specifically Afro‐Asiatic substratum being ultimately responsible for some or all of the exotic features of the Insular Celtic languages. Still, there surely were human languages spoken in Ireland and Britain prior to the arrival of speakers of Celtic language(s). Equally surely, these pre‐Celtic languages likely had some impact on the speakers of proto‐Insular Celtic. We just do not know what form that impact may have taken. Even though the fragmentary evidence of the Continental Celtic languages makes it clear that purely Celtic‐internal factors are responsible for the development of some of the exotic features of the Insular Celtic languages, external linguistic contact may have helped push the evolution of these features forward.14 Indeed, the presence of a particular configuration of a linguistic structure, however incipient in Insular Celtic, but also present in a substratal language, may have been a necessary precondition for such influence to take place. As Mithun (1992: 89) has remarked, it is often difficult to tease internal and external forces apart: Causes of linguistic change have traditionally been classified into two types: internal and external. Frequently cited internal causes of change include such factors as speakers’ preferences for simple and transparent systems, which can prompt learners to remodel apparently irregular or opaque paradigms. The most commonly cited external cause of change is language contact. While the distinction between internal and external causation may be clear‐cut in some cases, the separation of these factors is not always straightforward or even desirable, particularly in the area of syntactic change. Much syntactic development is driven by an interplay between internal and external factors. Grammaticization may seem to reflect a purely internal process: the cognitive routinization of patterns of expression. Yet structures are not grammaticized randomly. Speakers automate those structures they use the most often. Similarly, syntactic borrowing may seem to represent a purely externally caused development: it is dependent on external contact with other languages, under appropriate conditions of relative prestige and bilingualism. Yet aspects of the internal structure of the borrowing language can affect the facility with which a prospective loan is integrated. Internal and external factors can be difficult to untangle because syntactic change is so often the result of their interaction. For the same reason, examining the effects of either in isolation can be a mistake if we are to make progress in understanding the causes of syntactic change.

We seem to be left, then, in a position in which it is difficult, if at all possible, to draw any firm conclusions. Given the quite large number of Insular Celtic lexical items which, as yet, do not appear to have an Indo‐European etymology, it seems inevitable that we must conclude that the pre‐Celtic language(s) of Ireland and Britain had a significant impact upon proto‐Insular Celtic. Whether this language (group) guided the evolution of the syntax of the Insular Celtic languages in any significant way is very hard to know. However, in view of the fact that the fragmentarily attested Continental Celtic languages show evidence for the development of some of the “exotic” features of Insular Celtic in geographical regions in which a substratum, especially an Afro‐Asiatic one, is extremely unlikely, the probability of demonstrating such influence would appear to be so low as to be virtually nil.15

Contact and the Celtic Languages  497

NOTES 1 And its descendants, Scottish Gaelic and Manx. 2 Old Southwest Brittonic subsequently split into Old Breton and Old Cornish in the eleventh century (Schrijver 2011: 2–5). There are no verbal sequences in the Old Cornish corpus. 3 Hickey (2019) goes into detail about borrowings into early Irish. 4 This upon the basis of the differential treatment of proto‐IE initial */p/‐ and */sp/‐ in proto‐Celtic. Initial */p/‐ > */ph/‐ > /ɸ/‐ > /h/‐ > 0̸, while */p/ was preserved after */s/‐ because they shared a single [spread glottis] gesture which was extinguished prior to the release of occlusion (Kim 1970), it, hence, was never aspirated, and, therefore, was preserved. 5 Iverson and Salmons (2003: 55) raise the same possibility for the rise of aspiration in voiceless ­plosives in Germanic. 6 See the useful list compiled by Don Ringe available at www.ling.upenn.edu/~kroch/ courses/lx310/handouts/handouts‐09/ringe/celt‐loans.pdf. Eska (2018b) is based upon this list. 7 I ignore here the area of onomastics, in which considerable borrowing took place in both directions. See Stüber (2007) for a recent analysis of personal names. 8 Even for the meagerly attested Galatian, there is evidence for language mixing in the hybrid idionym Aρμɛδυμνoϛ, a compound of Lycian arma‐ ‘moon’ and Galatian dumno‐ ‘world,’ attested in a Greek inscription of coastal Asia Minor (Freeman 2007). 9 The only linguistic features borrowed into or from the ancient Celtic languages which have been proposed are the patronymic suffix ‐alo/ā‐ of early Cisalpine Celtic, which has been linked to the genitive singular exponent ‐al of the Etruscoid language Raetic (Pedersen 1921), and the thematic genitive singular ‐ı ̄, which may appear as early as the late seventh century in Cisalpine Celtic, but no earlier than ca. 300 bce in Latin, and is likely an areal phenomenon shared also by Messapic and Venetic (Eska and Wallace 2001: 92). 10 See the articles in Ureland and Broderick (1991) on language contact in Britain and Ireland in general. 11 Though Ní Chasaide and Ó Dochartaigh (1984: 154–5) view it as an intra‐Celtic development. 12 Grammatical abbreviations: ACC = accusative; DAT = dative; GEN = genitive; MASC = masculine; NEUT = neuter; NOM = nominative; OBJ = object; PL = plural; PRES = present; PRET = preterite; PROX = proximate; PST‐PTCPL = past participal; PV = preverb; SG = singular. 13 On object agreement in Insular Celtic, see Eska (2009–10) and Griffith (2011, 2015). 14 Such an approach is adopted by Hickey (2002a, 2002b). 15 There are other approaches which have been advocated to explain the evolution of some of the exotic features of the Insular Celtic languages. Matasović (2007), for example, proposes that mutual interference among Irish, the Brittonic languages, and Vulgar Latin in the period 350–550 ce was the cause, citing as a comparandum the languages of the Balkan Sprachbund, in which some “exotic” features – from the perspective of other European languages – arose which cannot be attributed to substratal languages of the region.

REFERENCES Adams, George Brendan 1975. Hamito‐Semitic and the pre‐Celtic substratum in Ireland and Britain. In James Bynon and Theodora Bynon (eds.), Hamito‐Semitica. The Hague: de Gruyter Mouton, pp. 233–47.

Adams, James N. 2003. Bilingualism and the Latin Language. Cambridge: Cambridge University Press. André, Jacques 1985. Noms de plantes gaulois ou prétendus gaulois dans les textes grecs et latins. Études celtiques 22: 179–98.

498  Joseph F. Eska Ball, Martin J. 1984. Phonetics for phonology. In Martin J. Ball and Glyn E. Jones (eds.), Welsh Phonology: Selected Readings. Cardiff: University of Wales Press, pp. 5–39. Borgstrøm, Carl H. 1974. On the influence of Norse on Scottish Gaelic: preaspiration of stops and pitch patterns. Lochlann 6: 91–103. Campanile, Enrico 1983. I Galli nella Cisalpina e i dialetti galloitalici [The Galli in Cisalpine Gaul and the Gallo‐Italic dialects]. In Enrico Campanile (ed.), Problemi di sostrato nelle lingue indoeuropee. [Problems of substrate in the Indo‐ European languages] Pisa: Giardini, pp. 27–35. Corominas, Joan 1956. New information on Hispano‐Celtic from the Spanish etymological dictionary. Zeitschrift für celtische Philologie 25: 30–58. Corominas, Joan 1976. Elementos prelatinos en las lenguas romances hispánicas [Pre‐Latin elements in the Hispanic Romance languages]. In Francisco Jordá, Javier de Hoz, and Luís Michelena (eds.), Actas del I coloquio sobre lenguas y culturas prerromanas de la Península Ibérica. [Acts of the first colloquium on pre‐ Roman languages and cultures of the Iberian Peninsula] Salamanca: Universidad de Salamanca, pp. 87–164. Dröge, Christoph 1989. Vulgärlatein, Griechisch und Gallisch in der Inschrift von Poitiers [Vulgar Latin, Greek, and Gaulish in the Inscription of Poitiers]. Zeitschrift für celtische Philologie 43: 207–13. Eska, Joseph F. 2007. On basic configuration and movement within the Gaulish clause. In Pierre‐Yves Lambert and Georges‐Jean Pinault (eds.), Gaulois et celtique continental. Genève: Droz, pp. 215–29. Eska, Joseph F. 2009–10. Where have all the object pronouns gone? The growth of object agreement in earlier Celtic. Zeitschrift für celtische Philologie 57: 25–47. Eska, Joseph F. 2018a. Laryngeal realism and the prehistory of Celtic. Transactions of the Philological Society 116: 320–31. Eska, Joseph F. 2018b. Celto‐Germanic lexis in light of Laryngeal Realism. In David M. Goldstein, Stephanie W. Jamison, and Brent Vine (eds.), Proceedings of the 29th Annual UCLA Indo‐European Conference. Bremen: Hempen, pp. 29–45. Eska, Joseph F. and Rex E. Wallace 2001. Remarks on the thematic genitive singular in ancient Italy and related matters. Incontri linguistici 24: 77–97. Evans, D. Ellis 1983. Language contact in pre‐ Roman and Roman Britain. In Wolfgang Haase and Hildegard Temporini (eds.), Aufstieg und

Niedergang der römischen Welt, Part II: Principat, vol. 29.2: Sprache und Literatur. Berlin: de Gruyter, pp. 949–87. Freeman, Philip 2007. Lycian/Galatian Aρμɛδνμνoϛ. Zeitschrift für celtische Philologie 55: 56. Garrett, Andrew 1994. Relative clause syntax in Hittite and Lycian. Die Sprache 36: 29–69. Gensler, Orin D. 1993. A typological evaluation of Celtic/Hamito‐Semitic syntactic parallels (University of California Berkeley Ph.D. dissertation). Gensler, Orin D. 2007. Extracts from A Typological Evaluation of Celtic/Hamito‐Semitic syntactic parallels. In Raimund Karl and David Stifter (eds.), The Celtic World, Vol. 4: Celtic Linguistics. London: Routledge, pp. 151–229. Gernia, Maria Luisa Porzio 1981. Gli elementi celtici del latino [The Celtic elements of Latin]. In Enrico Campanile (ed.), I celti d’Italia. [The Celts of Italy] Pisa: Giardini, pp. 97–122. Griffith, Aaron 2011. Old Irish pronouns: agreement affixes vs. clitic arguments. In Andrew Carnie (ed.), Formal Approaches to Celtic Linguistics. Newcastle‐upon‐Tyne: Cambridge Scholars Publishing, pp. 65–93. Griffith, Aaron 2015. Degrees of agreement in Old Irish. In Jürg Fleischer, Elisabeth Rieken, and Paul Widmer (eds.), Agreement from a Diachronic Perspective. Berlin: de Gruyter Mouton, pp. 165–88. Hamp, Eric P. 1990. The pre‐Indo‐European language of northern (central) Europe. In Thomas L. Markey and John A. C. Greppin (eds.), When Worlds Collide: The Indo‐Europeans and the Pre‐Indo‐Europeans. Ann Arbor, MI: Karoma, pp. 291–305. Hamp, Eric P. 1991. Language contact in the prehistory of English. In Vladimir Ivir and Darir Kalogjera (eds.), Languages in Contact and Contrast: Essays in Contact Linguistics. Berlin: de Gruyter Mouton, pp. 221–3. Hemon, Roparz 1975. A Historical Morphology and Syntax of Breton. Dublin: Dublin Institute for Advanced Studies. Hewitt, Steve 2007. Remarks on the Insular Celtic/Hamito‐Semitic question. In Raimund Karl and David Stifter (eds.), The Celtic World, vol. 4: Celtic Linguistics. London: Routledge, pp. 230–68. Hewitt, Steve 2009. The question of a Hamito‐ Semitic substratum in Insular Celtic. Language and Linguistics Compass 3: 972–95. Hickey, Raymond 2002a. Language change in early Britain: the convergence account. In David Restle and Dietmar Zaefferer (eds.),

Contact and the Celtic Languages  499 Festschrift for Theo Vennemann on the Occasion of his 65th Birthday. Berlin: de Gruyter Mouton, pp. 185–203. Hickey, Raymond 2002b. Internal and external forces again: changes in word order in Old English and Old Irish. Linguistic Sciences 24: 261–83. Hickey, Raymond 2019. Language contact in Celtic and early Irish. In Anthony Grant (ed.), Oxford Handbook of Language Contact. Oxford: Oxford University Press, 331–49. Isaac, Graham R. 2007. Celtic and Afro‐Asiatic. In Hildegard L. C. Tristram (ed.), The Celtic Languages in Contact. Potsdam: Potsdam University Press, pp. 25–80. Iverson, Gregory K. and Joseph Salmons 2003. Laryngeal enhancement in early Germanic. Phonology 20: 43–72. Jackson, Kenneth H. 1953. Language and History in Early Britain. Edinburgh: Edinburgh University Press. Jongeling, Karel 2000. Comparing Welsh and Hebrew. Leiden: Research School of Asian, African, and Amerindian Studies Keating, Patricia, Taehong Cho, Cécile Fougeron, and Chai‐Shune Hsu 2003. Domain‐initial strengthening in four languages. In John Local, Richard Ogden, and Rosalind Temple (eds.), Phonetic Interpretation. Cambridge: Cambridge University Press, pp. 143–63. Keating, Patricia, Richard Wright, and Jie Zhang 1999. Word‐level asymmetries in consonant articulation. UCLA Working Papers in Phonetics 97: 157–73. Kim, Chin‐Wu 1970. A theory of aspiration. Phonetica 21: 107–16. Koch, John T. 1990. *Cothairche, Esposito’s theory, and neo‐Celtic lenition. In Alfred Bammesberger and Alfred Wollmann (eds.), Britain 400–600. Language and History. Heidelberg: Carl Winter, pp. 179–202. Lambert, Pierre‐Yves 2003. La langue gauloise: description linguistique, commentaire d’inscriptions choisies [The Gaulish language: linguistic description, commentary on select inscriptions], 2nd edition. Paris: Errance. Lewis, Henry 1943. Yr Elfen Ladin yn yr Iaith Gymraeg [The Latin element in the Welsh language]. Caerdydd: Gwasg Prifysgol Cymru. Mac Cana, Proinsias 1976. Latin influence on British: the pluperfect. In John J. O’Meara and Bernd Naumann (eds.), Latin Script and Letters ad 400–900: Festschrift Presented to Ludwig Bieler on the Occasion of his 70th Birthday. Leiden: Brill, pp. 194–206.

Mac Eoin, Gearóid 2007. What language was spoken in Ireland before Irish? In Hildegard L. C. Tristram (ed.), The Celtic Languages in Contact. Potsdam: Potsdam University Press, pp. 113–25. Marstrander, Carl J. S. 1932. Okklusiver og substrater [Occlusives and substrates]. Norsk Tidskrift for Sprogvidenskap 5: 258–314. Matasović, Ratko 2007. Insular Celtic as a language area. In Hildegard L. C. Tristram (ed.), The Celtic Languages in Contact. Potsdam: Potsdam University Press, pp. 93–112. McManus, Damian 1983. A chronology of the Latin loan‐words in early Irish. Ériu 34: 21–71. Meid, Wolfgang 1980. Gallisch oder Lateinisch? Soziolinguisische und andere Bemerkungen zu populären gallo‐lateinischen Inschriften [Gaulish or Latin? Sociolinguistic and other notes on popular Gallo‐Latin Inscriptions]. Innsbruck: Institut für Sprachwissenschaft der Universität. (Reprinted 1983 in Wolfgang Haase and Hildegard Temporini (eds.), Aufstieg und Niedergang der römischen Welt, Part II: Principat, vol. 29.2: Sprache und Literatur. Berlin: Walter de Gruyter, pp. 1019–44.) Mithun, Marianne 1992. External triggers and internal guidance in syntactic development: coordinating conjunction. In Marinel Gerritsen and Dieter Stein (eds.), Internal and External Factors in Syntactic Change. Berlin: de Gruyter Mouton, pp. 89–129. Morris Jones, J. 1900. Pre‐Aryan syntax in Insular Celtic. Appendix B of John Rhŷs and David Brynmor‐Jones, The Welsh People. London: T. Fisher Unwin, pp. 617–41. (Reprinted 2007 in Raimund Karl and David Stifter (eds.), The Celtic World, vol. 4: Celtic Linguistics. London: Routledge, pp. 103–22.) Nance, Claire and Jane Stuart‐Smith 2013. ­Pre‐aspiration and post‐aspiration in Scottish Gaelic stop consonants. Journal of the International Phonetic Association 43: 129–52. Ní Chasaide, Ailbhe and Cathair Ó Dochartaigh 1984. Some durational aspects of preaspiration. Occasional Papers in Linguistics and Language Learning 9: 141–57. Oftedal, Magne 1968. North European geography of sounds. Scottish Gaelic Studies 11: 248–58. Parry‐Williams, T. H. 1923. The English Element in Welsh: A Study of English Loan‐Words in Welsh. London: Honourable Society of Cymmrodorion. Pedersen, Holger 1909. Vergleichende Grammatik der keltischen Sprachen I: Einleitung und Lautlehre [Comparative Grammar of the Celtic

500  Joseph F. Eska languages I: introduction and phonology]. Göttingen: Vandenhoeck and Ruprecht. Pedersen, Holger 1921. The Lepontian personal names in ‐alo‐s and some remarks on the Lydian inscriptions. Philologica 1: 38–54. Piette, J. R. F. 1973. French Loanwords in Middle Breton. Cardiff: University of Wales Press. Pokorny, Julius 1926–9. Das nicht‐ indogermanische Substrat im Irischen [The non‐Indo‐European substrate in Irish]. Zeitschrift für celtische Philologie 16: 95–144, 231–66, 363–94; 17: 373–88; 18: 233–48. Salmons, Joe 1992. Accentual Change and Language Contact: Comparative Survey and a Case Study of Early Northern Europe. Stanford: Stanford University Press. Schmidt, Karl‐Horst 1967. Keltisches Wortgut im Lateinischen [The Celtic Lexicon in Latin]. Glotta 44: 151–74. Schmidt, Karl‐Horst 1983. Keltisch‐lateinische Sprachkontakte im römischen Gallien der Kaiserzeit [Celtic‐Latin language contacts in Roman Gaul of the Empire]. In Wolfgang Haase and Hildegard Temporini (eds.), Aufstieg und Niedergang der römischen Welt, Part II: Principat, vol. 29.2: Sprache und Literatur. Berlin: Walter de Gruyter, pp. 988–1018. Schrijver, Peter 2000. Non‐Indo‐European surviving in Ireland in the first millennium ad. Ériu 51: 195–202. Schrijver, Peter 2005. More on non‐Indo‐ European surviving in Ireland in the first millennium ad. Ériu 55: 137–44. Schrijver, Peter 2011. Old British. In Elmar Ternes (ed.), Brythonic Celtic – Britannisches Keltisch: From Medieval British to Modern Breton. Bremen: Hempen, pp. 1–84. Stenson, Nancy 1993. English influence on Irish: the last 100 years. Journal of Celtic Linguistics 2: 107–28. Stüber, Karin 2007. Effects of language contact on Roman and Gaulish personal names. In Hildegard L. C. Tristram (ed.), The Celtic Languages in Contact. Potsdam: Potsdam University Press, pp. 81–92.

Uhlich, Jürgen 2004. Weiteres zur Chronologie der lateinischen Lehnwörter im Irischen [Further to the chronology of the Latin loanwords in Irish]. In Erich Poppe (ed.), Keltologie heute: Themen und Fragestellungen. Münster: Nodus, pp. 57–79. Ureland, P. Sture and George Broderick (eds.) 1991. Language Contact in the British Isles. Tübingen: Max Niemeyer. Vennemann, Theo 2002. Semitic > Celtic > English: the transitivity of language contact. In Markku Filppula, Juhani Klemola, and Heli Pitkänen (eds.), The Celtic Roots of English. Joensuu: University of Joensuu, pp. 295–330. Vennemann, Theo 2003. Syntax und Sprachkontakt: mit besonderer Berücksichtigung der indogermanischen Sprachen des Nordwestens [Syntax and language contact: with special attention to the Indo‐European languages of the Northwest]. In Alfred Bammesberger and Theo Vennemann (eds.), Languages in Prehistoric Europe. Heidelberg: Carl Winter, pp. 333–64. Vennemann, Theo 2010. Contact and prehistory in the Indo‐European Northwest: lexical influences. Sprachwissenschaft 35: 247–90. Vennemann, Theo 2016. Celtic as Vasconized Indo‐European? Three structural arguments. In John T. Koch and Barry Cunliffe (eds.), Celtic from the West 3: Atlantic Europe in the Metal Ages: Questions of Shared Languages. Oxford: Oxbow, pp. 503–32. Wagner, Heinreich 1959. Das Verbum in den Sprachen der britischen Inseln [The verb in the languages of the British Isles]. Tübingen: Max Niemeyer. Welby, Pauline, Máire Ní Chiosáin, and Brian Ó Raghallaigh 2017. Total eclipse of the heart? The production of eclipsis in two speaking styles of Irish. Journal of the International Phonetic Association 47: 125–53. Wolfe, Sam 2016. A comparative perspective on the evolution of Romance clause structure. Diachronica 33: 461–502.

25 Contact and the Slavic Languages LENORE A. GRENOBLE 1  The Slavic Languages and Contact The speakers of Slavic languages have expanded over a considerable amount of territory, stretching from Sorbian communities in Lusatia (in the German states of Saxony and Brandenburg) in the West to the far eastern border of Russia on the Pacific ocean in the East, from the Russian borders of the Arctic Ocean in the North to the Macedonian borders with Greece in the South, thus spanning from the heart of Western Europe all across Asia. The Slavic languages are generally classified in terms of three branches which capture not only their genetic classification but also, roughly, their geographic distribution: East, West, and South. The modern East Slavic languages are Belarusian, Russian, and Ukrainian; the West Slavic ones are Czech, Slovak, Polish, Kashubian (classified by some as a dialect of Polish), Lower Sorbian, Upper Sorbian (also called Wendish or Lusatian); and the South Slavic languages are Bulgarian, Macedonian, Bosnian‐Croatian‐Serbian (or BCS), and Slovenian.1 To this group we can add Rusyn,2 an East Slavic variety (Grenoble 2015) with strong influence from Slovak whose speakers live in Ukraine, Slovakia, Serbia, and Croatia. Following other criteria, however, it is possible to classify Slavic into North and South groups, a classification which places the East and West Slavic languages together in distinction to the South Slavic languages. Many of the criteria used for the North–South grouping stem from contact‐related phenomena. In addition to these living languages is Old Church Slavonic, which is now extinct as a vernacular language. It is a liturgical language created in the late ninth century based on the South Slavic variety spoken in the area of Moravia where the Slavs were first Christianized. Because of its importance in the spread of Christianity and literacy, it has played a significant role in the development of a number of Slavic languages, and has served as a vehicle for introducing borrowings from non‐Slavic languages, especially Greek. Polabian and Slovincian, both West Slavic, are also now extinct. The range and extent of contact‐induced phenomena vary according to time and language and are often difficult to assess. By and large cases of lexical borrowing are relatively clear, in terms of what is the source and what is the target. But in other areas of potential contact‐induced change, e.g. phonological, morphological, or syntactic phenomena, it can be impossible to prove without question that a given phenomenon or feature is the result of contact and not independent innovation or shared inheritance. This is perhaps particularly true for the impact of one Slavic variety upon the other, where the genetic and typological

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

502  Lenore A. Grenoble properties of both are extremely close to one another. Additional ambiguities are introduced by the fact that some important contact phenomena occurred during the prehistoric period.

2  Prehistoric Contact The Slavic prehistoric era can be divided into two phases based on a series of sound changes: the early period, from approximately 3000 to 1000 bce; and the more recent period, roughly from the 300s to the 900s ce. During this more recent time we find changes common to all Slavic dialects as well as changes which indicate the break‐up of Slavic around 900 ce, when we have the first written documentation of Slavic. Historically this period is flanked by the expansion of the Goths into the Black Sea region at its beginning and with Christianization of the Slavs and the development of a Slavic written language (Andersen 2003: 46). The region also served as a corridor for the migration of a variety of tribes (Turkic Huns, Bulgars, Pechenegs, and the Altaic Magyars) into Europe; there are few traces of lexical borrowings from these groups into Slavic but they may have had an impact on Slavic phonology (see Galton 1994). There are two sources of loanwords in prehistoric times which deserve special discussion – Germanic and Iranian (especially Scythian).

2.1 Iranian Iranian contact, in particular with Scythian and Sarmatian tribes, appears to have taken place in the area of what is now southern Russia from approximately 700 bce to approximately 300 ce, although some apparent loanwords may actually be cognates. The correspondences between Early Common Slavic and Iranian phonology make it hard to distinguish whether some similar items are cognate or borrowings (Andersen 2003; Zaliznjak 1963). A large percentage of these borrowings can be organized into four semantic categories: religion (e.g. bogъ ‘god’, divъ ‘demon’, gatati ‘to divine’, rajъ ‘paradise’, svętъ ‘holy’); law, society, and morality, broadly defined (kajati sę ‘to repent’, zъlъ ‘bad, evil’, cъstъ ‘honor’); and health and body parts (*sъdorvъ ‘healthy’, *pъrsi ‘breast’, *goldъ ‘hungry’). Such broad groupings constitute approximately 45% of all Iranian borrowings; the remaining include such words as radi ‘for the purpose of’, sobaka ‘dog’, and xvala ‘glory’. Many Iranian loanwords are not found in all Slavic varieties but are limited to certain areas, despite the fact that they appeared to have been borrowed early on (Trubačev 1977). A singular example is the word vurdoljakъ ‘werewolf’, which was borrowed into most or all Slavic languages but has undergone change to varying degrees in the different Slavic dialects (Nichols 1993: 387).

2.2 Germanic There are a large number of Germanic loanwords in the prehistoric period which can be seen as coming in separate waves. The absolute total number of these is unclear but in the range of 150, although there are only about 50 “secure” examples (Andersen 2003). From the standpoint of Balto‐Slavic contacts, Germanic loanwords can be divided into three groups: those found only in Baltic, those found only in Slavic, and those found in both. The first Germanic– Slavic contact affected only some Proto‐Slavic tribes dwelling in the Sub‐Carpathian zone, while later contacts were more extensive and lasted longer. Contacts with the Goths affected practically all the prehistorical Slavic tribes. In many instances it is difficult to pinpoint the date and precise source of a borrowing due to the duration and prehistoric nature of these contacts. There are numerous borrowings in technical terminology from Germanic (as well as Italic and Celtic) into Slavic but not Baltic. It should first be noted that the existence of

Contact and the Slavic Languages  503 loanwords (or cognates) in Baltic and Slavic is often invoked in favor of a shared heritage or, alternatively, in favor of extensive contact. The earliest evidence of Germanic contact comes from an inscription and appears to date to the first centuries ce. The loanwords found on the inscription cannot be attributed to any specific Germanic group. The second layer corresponds to contact between the Goths and Slavs when the former settled in the region north of the Black Sea, in approximately 200–300 ce. With the westward expansion of the Slavs, from the 400s onwards we find West Germanic loanwords, primarily from High German. Representative early Germanic borrowings are duma ‘thought’, Goth dōms ‘judgment’; gotoviti ‘to prepare’, Goth gataujan; kupiti ‘to buy’, Goth kaupōn; or šelmъ ‘helmet’, PGmc *helmaz. Later loanwords are not found in all Slavic dialects: bl’udo ‘dish’, Goth biuþs; buky ‘script’, Goth bōka ‘letter’ (Schenker 1995: 159). See in particular Gołąb (1991) and Kiparsky (1934) for a more detailed discussion of loanwords and more examples.

3 Finno‐Ugric Contact and the Finno‐Ugric Substrate in Russian The question of a Finno‐Ugric substrate in Russian is a matter of some debate, but it is fairly clear that at least some features of Modern Russian are the result of contact with Finno‐Ugric. That there have been long‐standing contacts between the different speakers is not in doubt. The regions of central and northern Russia were populated by a number of Finno‐Ugric tribes; speakers of these languages spread as far west as Finland and Estonia. The Slavs expanded into the east and north into Finno‐Ugric territories in the fifth century and to this day a large number of toponyms in central and northern Russian are Finno‐Ugrian, as in the city of Vologda (with stress on the first syllable, as in Finnish), or Lake Ilmen or Lake Ladoga. As with other issues, the question of Finno‐Ugric impact can be divided into those instances where the evidence of the substrate is clear, and where it is probable but not as clear. In both sets of cases, there is a difference between standardized Russian (CSR) and the dialects. (In northern dialects where variation has not been eliminated due to the influence of the standard, the number of clear influences is even greater.) Veenker (1967) cites three clear cases of Finno‐Ugric contact in CSR: 1. akan’e, or the so‐called “reduction” of unstressed vowels, as in moloko ‘milk’ [məlʌˈko]; the spelling retains the etymologically correct vowel /o/ in all three syllables; 2. the so‐called nominal sentence, or the loss of the copula in the present tense, as in ja čelovek ‘I am a man’, lit. ‘I – man’, a pattern found throughout all Finno‐Ugric languages except those in the Baltic subgroup; 3. loss of the verb ‘to have’ (imet’) except in scientific prose and a few set phrases. In this regard Russian is distinct from all other Slavic languages, where it is regularly used. The Russian normal verbal construction for possession is one with the possessed as the nominative subject and the possessor in a prepositional phrase, with the preposition u ‘at, by’ which governs the genitive case, e.g. u menja novoe paljto ‘I have a new overcoat’. More recent research has challenged the idea that the possessive construction is due to contact as opposed to language‐internal change. The hypothesis stems from the fact that Slavic in general is a ‘have’‐language (Isačenko 1974), while Russian uses a location schema as its basic possessive construction, in keeping with Finnic languages spoken in the Circum‐ Baltic, including Finnish, Estonian, Olonec and North Karelian, and Votic (Mazzitelli 2017: 13); this distribution supports a contact‐based explanation. However, the use of u + the genitive to express predicate possession is attested in Old Church Slavonic, as well as Old

504  Lenore A. Grenoble Czech, Old Serbian, and Middle Bulgarian, although it was a peripheral construction (Grković‐Major 2011: 51; McAnallen 2011: 156). This suggests that the construction itself was not calqued, but rather that areal influence supported its use and helped pave the way for it to become the primary construction for predicative possession in Russian. Finno‐Ugric influence also is most probable in CSR in the use of a partitive genitive and the use of a locative (‐u) in masculine and neuter singular paradigms of some nouns, which is in distinction to the prepositional case ‐e for these nouns (e.g. o lese ‘about the forest‐PREP’ versus v lesu ‘in the forest‐LOC’), while in most nouns the endings are homophonous. Probable Finno‐Ugric influence in CSR is found in the use of comitative constructions, such as my s vami ‘you and I’, lit. ‘we with you‐INST.PL’ or my s ženoj ‘my wife and I’, lit. ‘we with wife‐INST.PL’. Impact of Finno‐Ugric in some northern Russian dialects is seen in certain sound changes: (i) diphthongization of the vowels o > oa and e > ia; (ii) cokan’e or the collapse of the palatals c and č into one phoneme; and (iii) stress on first syllable (dialect póašla, CSR pošlá ‘[she] went’. There are also morphological changes: (i) comparative of substantives (e.g. berežee ‘closer to the shore’, from the noun bereg ‘shore’); (ii) the nominative object; and (iii) certain loaned suffixes, such as the causative ‐tta, ‐itta from Finnish. Finno‐Ugric influence on the northern dialects is probably also seen in the development of postpositive definite articles (e.g. ‐ot, ‐ta, ‐to, ‐ti, ‐te), which are arguably used under the influence of Komi‐Zyrian definite‐possessive suffixes (Tiraspol’skij 1998). Another construction frequently attributed to Finno‐Ugric contact is use of the construction of the type u nego uexano, with the preposition u plus genitive of the logical subject and a past passive participle in the neuter singular nominative instead of a finite past tense verb form, e.g. on uexal ‘he has departed’. In north Russian dialects this construction was prevalent with both transitive and intransitive verbs. It has also been argued that this list could be expanded by other changes, such as use of [p] instead of [f], and the development of perfect and pluperfect tenses. Even the clear evidence of a Finno‐Ugric substrate is not entirely straightforward. As mentioned on p. 503, one of the more contested claims is the absence of the verb have in modern Russian. Its lack has alternatively been explained as contact‐induced change or as a relic of the earlier Common Slavic state. Although Indo‐European is generally believed to have been lacking a have verb, the modern Slavic languages  –  with the exception of Russian – have one as a subsequent development. The question is whether Russian also had the verb and lost it or if it represents a more conservative stage of Slavic, arguably as part of a core–periphery change, with Russian removed from the center of innovation. Evidence in favor of this latter argument is the fact that in OCS texts all instances of iměti ‘have’ are loan translations from the Greek. In favor of the former theory, the existence of a pan‐Slavic verb iměti ‘have’, which was lost through contact with Finno‐Ugric tribes, comes from the Baltic languages, as Latvian – which had Finno‐Ugric contacts – also uses a be construction for possession, while Lithuanian – which did not – uses a have verb (Isačenko 1974; Kiparsky 1969; see Dingley 1995 for a review of the arguments). For further discussion of contact in Slavic prehistory and detailed arguments, see Nichols (forthcoming).

4  Contact in the Early History of the Slavs We have no written records of Slavic prior to the mid ninth century, and the earliest data are only brief inscriptions. The first manuscripts which have survived come from the early tenth century and are almost exclusively translations from the Greek gospels. The language used in these manuscripts is Old Church Slavonic, a liturgical language created as part of a mission

Contact and the Slavic Languages  505 from Byzantium to Christianize the Slavs which sent two missionaries, the brothers Constantine (later known as Cyril) and Methodius, to Moravia in 862–3. Bilingual in Greek and the South Slavic variety spoken in Salonika at the time, Cyril and Methodius created a largely artificial, liturgical language and begun their work by creating translations from Greek, and the oldest OCS texts are translations of the gospels. Thus Greek had a major impact on the structure of OCS. As Christianity spread among the Slavs, so too did OCS and its use in the liturgy. The modern Slavic languages can be divided according to the orthographic system they use today, Cyrillic or Roman. This division maps almost perfectly onto religious divisions, and those languages which use the Roman alphabet are primarily spoken by Catholics, and those using the Cyrillic alphabet are primarily Orthodox Christians (with the notable exception of some Muslim groups, in particular in South Slavic territory). The former group initially used Latin for its liturgy and the latter Old Church Slavonic. This dichotomy more or less corresponds to the initial source of religion and thus liturgical texts for each group, although not perfectly. The languages using the Roman alphabet are all of the West Slavic languages, and Croatian and Slovenian in South Slavic territory. Those using the Cyrillic alphabet are the East Slavic languages, Belarusian, Russian, and Ukrainian, along with the South Slavic languages Bulgarian, Macedonian, and Serbian. Old Church Slavonic played a particular role in the development of Russian and had a major impact on its structure. It was less central in the development of Belarusian and Ukrainian, which show a heavier Polish and German influence, due to their respective ­histories. Ukraine and Belarus were taken over by Lithuania and Poland, beginning with the annexation of Polotsk in the tenth century and continuing well into the fourteenth century. Thirteen eighty‐six marks the inception of the Lithuanian Union with Poland, and the time when the Lithuanian ruling class effectively became Polonized. As a result, the impact of OCS was less strong for Belarusian and Ukrainian than Russian, which show a heavier Polish and German influence, due to their respective histories. The East Slavs received this written language in the tenth century with the conversion in 988–9 of Prince Vladimir in Kiev to Christianity. A century later, it is clear from historical records that Church Slavonic had lost its unity and was replaced by regional variants, but it continued to be a primarily South Slavic language. Moreover, as the ecclesiastical language it was seen as sacred, in terms of both linguistic and orthographic form. It was thus very conservative and major changes in Church Slavonic are only seen as deliberate attempts to rid it of regional or “debased” elements and bring it to its “pure,” more Hellenistic form (although perceptions of what constituted “pure” Church Slavonic varied over time). At this stage, the phonemic and morphosyntactic differences between OCS and East Slavic were not so great as to prohibit comprehension,3 although OCS had a large, specialized lexicon that was lacking in East Slavic. OCS was also characterized by several South Slavic sound features, such as the use of ra < *ar, la < *al, in distinction to the pleophonic forms of ESl (‐oro, ‐olo), and the consonant clusters št j, žd j (versus the ESl forms šjčj or žž j). These differences have had a profound impact on Russian and can be readily found in paired lexical items, where the ESl term refers to the ordinary, everyday object and the Church Slavonic form to a more elevated or scientific term, as in R moloko ‘milk’ versus mlekopitajuščij ‘mammal’ or rovnyj ‘even’ versus ravnyj ‘equal’, ravenstvo ‘equation’. The biggest differences between the two languages were found more at the level of macrosyntax and involved differences in clause combining, e.g. coordination and subordination, the use of infinitival and participial constructions, and the dative absolute. Such differences are not surprising given that spoken East Slavic was not used for writing and that OCS was not only a liturgical language, but one built on the syntactic forms of Greek.

506  Lenore A. Grenoble Church Slavonic continued to dominate as the written language in first Rus’ and then Russia for many centuries. In fact its significance went through a revival in the Muscovy period in what is known as the “Second South Slav influence.” By this time, spoken Russian had diverged so much from Church Slavonic that it is appropriate to consider it an entirely different language. The period of high Muscovy (in the fifteenth and sixteenth centuries) can be considered a period of diglossia, with Russian used as the spoken language and in all everyday domains and Church Slavonic as the written language used in formal, administrative, and ecclesiastical domains and increasingly inaccessible to all but the most educated elite. The Second South Slavic influence, an attempt to purify Church Slavonic as used in Muscovy, brought with it a highly artificial literary style based on the elaborate and equally artificial Greek style of the time. Etymologically correct Church Slavonicisms replaced East Slavicisms which had crept into the language, including forms in ‐žd‐ such as prežde ‘formerly’ which replaced perež e and prež e, or pobeždën ‘conquered’ for pobež en; or the alternation between vremja ‘time’ and veremja is lost, with only the former, Slavonic form used. These changes remain in Russian to this day. Church Slavonic was not the only language to have an impact on Russian in the historical period. Poland, and thus the Polish language, became the primary link between Western Europe and Russia, and Polish served as the vehicle for introducing a great many lexical borrowings from Latin, French, and Greek at this time. French had a direct and profound impact on Russian in the eighteenth century, when both written and spoken forms of French provided models for the Russian aristocracy. The overall impact of French on Russian is hard to assess, as it extends beyond matters of the lexicon and goes into the question of style.

5  Western European Languages and Slavic Several Western European languages have had an impact on the different Slavic languages. At varying times throughout history, Latin and French, and Italian to a lesser extent, have primarily affected the lexicon of individual languages. Germanic contact in prehistoric times was significant, and a number of modern Slavic languages are notable for the imprint that sustained contact with German has left on their structures: Czech, Upper Sorbian, and Polish. In addition, German influence is clear on Polabian (now extinct) and on some dialects of Croatian.

5.1  Czech and German Czech–German contact can be established as early as the late ninth century, when both Latin and Greek liturgical terminology was borrowed into Czech through German. In the twelfth century, German became the court language in the Bohemian state and during the thirteenth and fourteenth centuries was the dominant language in urban areas due to the immigration of German merchants and others. A number of calques and borrowings from both Latin (karta ‘map’; figura ‘figure’) and German (hrabě ‘count’; řiše ‘realm, empire’) date to this time. In the eighteenth century, German had high prestige in Czech society, playing a significant and dominant role in public life, the schools, and administration. From the end of the eighteenth century, the prestige of German only increased as it was used in administration and education in the Habsburg Empire. Up until 1918 and the formation of the new Czechoslovak state, German served as the written, and often the spoken, language in much of Czech‐inhabited territory, especially among urban, educated Czechs and many Slovaks (Townsend 1981).

Contact and the Slavic Languages  507 One of the most frequently cited effects of Czech–German contact is the diphthongization in Old Czech of the long vowels (ú > ou, ý > ej) and monophthongization (ie > i, uo > u) of diphthongs on the one hand, and similar changes in Middle High German on the other (u > au, i > ei; ie > i, uo > u). Yet a causal relationship is not definitively proven; these may well be parallel developments. There are, however, a large number of loanwords, calques, and phraseological translations which have clearly entered Czech from German. Some are older borrowings and are found in all or nearly all Slavic languages, while others are exclusive to Czech or Slovak (e.g. Cz ramsl, Sk ramsl’a < G Ramschel ‘a card game’). It has also been claimed that the frequent use of Cz žádný ‘no, none’ is due to influence from G kein (Schuster‐Šewc 1996: 16). German has had some impact on the interpretation of aspect in calques, as seen in Cz vyznat se v nečě, a calque of the G sich in etwas auskennen ‘to be familiar with something; to know one’s way around something’ (literally, ‘to know oneself out in something’). In Czech, the verb vyznat’ is morphologically perfective, but is used as an imperfective under influence of the German calque (Townsend 1981: 7). By the late eighteenth and early nineteenth centuries, Czech had been largely eclipsed by German and, in an attempt to revitalize the language, a policy of re‐Slavicizing the lexicon was adopted, a policy largely associated with Josef Jungmann’s work. As a result, many borrowings from German (as well as from Greek and Latin) were replaced by words from Czech dialects or new lexical items based on Slavic roots, such as Cz hudba ‘music’ (versus R, P muzyka) or Cz kni‐hova ‘library’ (R, P biblioteka), while a number of calques from German using Czech roots have survived (zeměpis ‘geography’; krasopis ‘calligraphy’).

5.2  Upper and Lower Sorbian Two modern Slavic languages, Upper Sorbian and Lower Sorbian, are surrounded by German‐speaking territory and lack an autonomous region of their own. Upper Sorbs live mostly in upper Lusatia (Oberlausitz) in Saxony and Lower Sorbs in the lower Lusatia (Niederlausitz) in Brandenburg; they have long been an ethnic minority in Germany. Sorbian monolingualism ceased before World War II; nearly all Upper Sorbian speakers are fully bilingual in German and use primarily German; Lower Sorbian is seriously endangered. Both languages show significant impact of contact with German, change which has been almost entirely unidirectional (Toops 2006). A number of calques from German occur in Sorbian and Czech, as in the literal translation of G gern haben ‘to like’ (lit. ‘to have gladly’) as Cz mít rád; Upper Sorbian rady měć. Changes in the use of the instrumental case can also be ascribed to German influence. The pan‐Slavic pattern is to make a distinction between the instrumental or absolute use of the instrumental case, which does not occur with a preposition, and the instrumental of accompaniment, which is used with the preposition ‘with’. In Upper Sorbian these two usages are collapsed and are found only with the preposition, as seen in the following examples (Lötzsch 1996: 56) which both use the preposition z ‘with’: Upper Sorbian Ja dź ělam z ruku ‘I work with my hand’ (instrument); and Upper Sorbian Ja rěču z prěćelom ‘I speak with my friend’ (accompaniment). This is one area where Czech has not been influenced by German and has maintained the more Slavic pattern (Toops 1999: 275) and does not use a preposition: Cz Mluví chraplavým hlasem ‘He speaks with a hoarse voice’. Other cases of German influence are less unambiguous. One example is the German construction würde + infinitive which can function as a (non‐past) conditional or as a future‐ in‐the‐past. Upper Sorbian shows an analogous use of the (Slavic) conditional construction consisting of by and the ł‐participle as a future‐in‐the‐past, a use not found in other Slavic languages. That said, the remainder of an iterative preterite in many Sorbian dialects could also be the cause of or, more likely, simply the supporter of this change (Toops 2006: 153–4).

508  Lenore A. Grenoble

5.3 Polabian Polabian was spoken on the left bank of the Elbe River and so was the westernmost variety of the Lechitic sub‐branch of West Slavic. It was subject to heavy German contact, or more specifically Low German, from the Middle Ages until it became extinct in the mid eighteenth century. The last fluent speaker died in 1756 (Szydłowska‐Ceglowa 1987: 612). Relatively little documentation for Polabian has survived, and what there is consists primarily of word lists or lexicons with a few texts, for a total of approximately 2,800 lexical items. Some 20% of this extant lexicon consists of German borrowings which were phonologically assimilated and morphologically adapted to Polabian. German also had an impact on Polabian morphosyntax. Examples include the development of separable compound verbs, using both German and Polabian prefixes and particles, as in Pb to < LGmc *to, seen in Pb to‐vist (< *to‐ vesti), alternatively vizĕ‐to (< *vezĕtъ‐to) ‘to drive to’; to‐zinĕ or zinĕ‐to (< *to‐ž enetъ) ‘to drive to’; or, using Slavic roots, vånau dojĕ (< *vъnu dajetъ) ‘gives out’ (Polański 1993: 819). Other examples of German influence are the new perfect tense forms of the verbs ‘to be’ (Pb båit < *byti) and ‘to have’ (Pb met < *jъměti); and the second person plural pronoun jai (from Middle Low German jı̄). As with other Slavic languages in long‐standing contact with German (e.g. Upper Sorbian), in Polabian we see the spread of the use of the preposition ‘with’ with the instrumental case where other Slavic languages use only the instrumental.4 Other constructions arguably but less certainly stem from German influence: (i) the use of a subject pronoun with what would be impersonal constructions in Slavic: ‘it thunders’ Pb gramĕ (< *grъmitъ) versus tü gramĕ, G es donnert and the use of kå plus the dative of a verbal substantive, presumably on the model of G zu plus the infinitive, as in Pb nema ̆m nic kå våidońĕ versus G ich habe nichts auszugeben ‘I have nothing to give away’ (Polański 1993: 796).

5.4 Polish By and large the impact of language contact on Polish has been in the area of the lexicon, with the exception of Latin, which also had an impact on morphology. Two languages stand out in this regard: Czech and German. Czech missionaries brought Christianity and the Latin language and alphabet to Poland in 966. Thanks to the close cultural and political ties of Czech Bohemia with Poland, Czech had a major influence on the language, in particular in the realm of vocabulary. These borrowings come from a large number of semantic fields, such as education, science, religion, and cultural life, all reflecting the sense of Czech as the language representing the dominant culture, the one to be emulated. Owing to German colonization, some regions of Polish were German‐speaking and some, such as Cracow, are known to have even conducted business in German until some point in the sixteenth century. But by and large the relationship of the three languages – Czech, German, and Polish – is so intertwined that it is often difficult to determine whether a borrowing came directly from German or came into Polish via Czech. A significant number of German borrowings are, not unexpectedly, in the realm of government, commerce, and town planning, such as burmistrz ‘mayor’; borg ‘credit’ or gaska ‘street’. (See Schenker 1985: 198–9 for examples of both Czech and German borrowings.) Despite the long dominance of Latin in religious spheres, it did not have a real impact on Polish until the time of the Renaissance, when it came to be viewed as the language of high culture and was acquired by the nobility as a means of communication. Latin influence is seen in modern Polish in the phonology of a few borrowed words which have stress on the antepenultimate syllable, not the penultimate, as is the norm (e.g. muzyka ‘music’ or publika ‘public’). In morphology it is seen in the declension type of neuter nouns in ‐um (liceum ‘high school’); the nominative plural ‐a of some masculine nouns (koszta ‘costs’); the prefixes arcy‐, super‐; and the derivational suffixes ‐acja, ‐ysta/‐ista, ‐yzm/‐izm. In addition, a number of fixed phrases follow

Contact and the Slavic Languages  509 noun‐adjective order (e.g. wojna domowa ‘civil war’), again from Latin. A large number of lexical borrowings and calques also came from Latin. Later, in the late eighteenth century and into the nineteenth century, a number of borrowings came from French; these can generally be traced to the strong influence of court life at Versailles and include such words as dama ‘lady’; kotylion ‘cotillion’; and krawat ‘tie’, to name just a few (see Schenker 1985). Silesian (also called Upper Silesian or, pejoratively, Wasserpolnisch ‘watered‐down Polish’), is worthy of special mention. It is spoken in the region of Upper Silesia in territory that extends from Poland into the Czech Republic. Its status as a separate language or a dialect of Polish has been debated. Structurally it is very similar to Polish in terms of phonology and morphology but with some differences, including the loss of nasal vowels. Moreover, it shows greater influence in particular of Czech, especially in the lexicon and syntax,5 and also of German and Slovak (see Hannan 1996).

6  Slavic Languages in Contact with Each Other 6.1  Czech and Slovak The Czech language was codified for centuries prior to Slovak and so was well positioned, linguistically and sociopolitically, to have a significant impact on it. In fact Czech functioned as the language of literature and culture in Slovak‐speaking regions during the fifteenth to nineteenth centuries, continuing up to the end of the twentieth century in some groups. The notion of a unified “Czechoslovak” emerged during the eighteenth century in the Slovak Protestant environment. Czech was seen as the standard literary language for both Czechs and Slovaks but took on a number of Slovak features in the Slovak‐speaking communities, and for some Protestants was even understood to be a synthesis of Czech and Slovak. Even after standardization of Slovak in 1843 by L’udovít Štúr, “Czechoslovak,” or bibličtina ‘Biblical language’ as it was frequently called, continued to be the liturgical language for Slovak Protestants, and retained some of its functions until the end of the twentieth century. Ultimately the result was diglossia, with Czech used as the liturgical and sacred language and Slovak in all secular domains. The tendency toward merging Slovak with Czech only increased in the first Czechoslovak Republic (1918–39), where official language policy moved the two toward unification. The overall result was a destabilization of standard Slovak and a tendency for speakers to mix varieties (Nábělková 2007).

6.2  East Slavic I: Russian, Ukrainian, and Belarusian Russian, Ukrainian, and Belarusian have long been in contact and have mutually influenced one another. The East Slavic languages constitute a language–dialect continuum, and the boundaries which define “languages” as opposed to “dialects” are more political than linguistic. (The same is true in the West, and the border region between Belarus and Poland is a transitional zone, linguistically and culturally.) It is clear that an East Slavic variety, distinct from other varieties of Slavic, emerged during the seventh to eighth centuries but it is impossible to pinpoint the time when the three East Slavic varieties could be considered discrete languages and a pan‐East Slavic that was mutually intelligible ceased to exist. The use of Church Slavonic as the written language for the region into the fifteenth century further complicates the picture. Finally, although the impact of Russian on Ukrainian and Belarusian predates the Soviet period, it is important to note that all three were spoken in territory within the former Soviet Union. Language policies aimed at assimilation to Russian as well as the social, political, and economic dominance of Russian, have certainly had an impact on the direction of change.

510  Lenore A. Grenoble Two mixed languages have been identified as the result of language contact among the Slavs: Trasjanka, a mix of Belarusian and Russian, and Surzhyk, a mix of Ukrainian and Russian. Both terms are pejorative and their literal meanings referring to a lower grade grain: trasjanka refers to a mixture of wheat and straw, and surzhyk of wheat and rye. Further research, based on large corpora of naturally occurring connected discourse in both mixed varieties, is needed to determine just how regular such mixtures are. For more details, see Kittel et al. (2018), as well as the articles in Hentschel et al. (2014), and Stern (2017) for an assessment thereof.

6.3  East Slavic II: Russian, Ukrainian, and Surzhyk Since Ukraine became a part of Russia with the Treaty of Perejaslav in 1654, there has been ongoing contact between the Ukrainian and Russian languages, a contact shaped by varying social and political pressures over time. In the seventeenth and early eighteenth centuries the language influence was bi‐directional and mutual. During this time period, Ukrainian had an impact on Russian in both religious and secular circles, serving as the intermediary for loanwords, in particular from Polish. Ukrainian clergy held high positions in Moscow, and so Ukrainian pronunciation had an impact on Church Slavonic, and Ukrainian grammatical and rhetorical traditions were adapted to the liturgy. But by the end of the eighteenth century, the influence of Russian, and in particular the Russified variant of Church Slavonic, was so strong that it began penetrating all parts of Slavic orthodoxy, including even Transcarpathia and Bukovyna. Russification of Ukraine intensified during the periods of the Russian Empire and the Soviet Union. By the end of the Soviet era, it is possible to speak of diglossia in Ukraine, with Russian as the High variety used in formal, administrative, and educational domains, and Ukrainian in less formal, home settings. One result of this diglossia is Surzhyk, a ‘hybrid sociolect’ (Taranenko 2007: 125) with Ukrainian as a matrix language and certain inserted Russian features. More research is needed to determine to what extent the mixture of Russian and Ukrainian found in Surzhyk is regular and predictable and to what extent it is idiosyncratic, although we can identify a major division between those speakers who use fused‐lect Surzhyk as their native tongue and are not fluent in any other language variety, and those who mix Ukrainian and Russian due to incomplete knowledge of just one of them, and fluency in the other. In addition, there are speakers who know both languages, but who mix them in speech because that is perceived to be the norm for their speech community, or at least the norm in a given setting (Bilaniuk 2004). Many studies to date have had a prescriptive focus, arguing against language mixture. The most comprehensive study so far is Bilaniuk (2005) which looks primarily at Eastern Ukraine; differences might be found in the western, less Russified regions. The linguistic characteristics of Surzhyk include the following (Flier 2000): 1. Borrowings and lexical calques of Russian words not normally found in Ukrainian (e.g. R nakonec, Sur nakonic’ versus Ukr narešti ‘finally’; R/Sur pokupateli, Ukr pokupci ‘shoppers’; or R stolovaja, Sur stolova versus Ukr ïdal’nja ‘dining room’, where the Surzhyk adapts the Russian substantized adjective to Surzhyk/Ukrainian adjectival morphology). There appear to be no constraints on lexical borrowing and calques. 2. Syntactic calques, primarily in prepositional and numeral phrases, as well as head phrases, where government patterns differ between Ukrainian and Russian, as in R soveščanie po problemam, Sur narada po problemam ‘a conference on issues’, using the preposition po ‘on’, ‘concerning’ with the dative case, versus Ukr narada z problem, using z plus the genitive. (Use of the preposition po does not otherwise govern the dative in Ukrainian.) Another example is in time expressions, following the Russian model of the preposition v

Contact and the Slavic Languages  511 ‘at’ plus the accusative of a cardinal numeral which itself then governs the genitive (R v desjat’ časov, Sur v desjat’ godyn ‘at ten o’clock’) versus the Ukrainian pattern using o plus the locative case of a cardinal numeral and noun (Ukr o desjatij godyniı̆). 3. Pronouns and adjectives follow Ukrainian morphology. In nouns, the vocative follows the Russian pattern of being identical to the nominative, and the genitive singular of masculine nouns is generally ‐a, not ‐u, unlike Ukrainian, where the reverse is true. The dative singular of masculine animate nouns also typically follows the Russian ‐u, as opposed to the Ukrainian form in ‐ovi. 4. Verb paradigms largely follow Ukrainian forms and the third person non‐past ending in t is not replaced by Russian nonpalatalized ‐t; the preterite uses the Ukrainian ending ‐v and not the Russian ending ‐l; but Surzhyk tends to form a first person plural imperative using the non‐past first person form on the Russian model. Surzhyk cannot be considered a single, homogeneous linguistic variety but is rather a set of varieties, with differences depending on several parameters: rural versus urban; level of education; and time period (pre‐Soviet, Soviet, post‐Soviet). These have been analyzed as five different categories (Bilaniuk 2004, 2005) but are probably more accurately seen as different sociolects which have changed over time but can be united in one larger category of Surzhyk.

6.4  East Slavic III: Russian, Belarusian, and Trasjanka Over the course of history Belarusian has been spoken in a region which has been under a variety of different political controls and, subsequently, under different linguistic influences and pressures. Linguistically Belarusian is close to both Russian and Polish and can be seen as transitional between the two. The East Slavic predecessors of the Belarusians moved into the territory north of what is now Ukraine and east of Russia; in the tenth to eleventh centuries the region became part of Kievan Rus. The Mongol invasion of Rus in the thirteenth century did not extend to their territory, thus politically (and linguistically) separating them from their East Slavic neighbors, to be incorporated into the Grand Duchy of Lithuania in the fourteenth century. An estimated two thirds of the population of the Grand Duchy was Slavic and the local version of Church Slavonic was adopted as the official language. It was called Rusky during this period, although Russians in Moscow referred to it as Lithuanian, while in Ukraine the written form was seen as Russian and the colloquial (spoken) language as Lithuanian. In both cases the term “Lithuanian” refers to the territory, not the language, yet its use is symptomatic of the overall social situation. With the Union of Lublin treaty in 1569, Belarus became part of the Polish‐Lithuanian Commonwealth, Latin and Polish were used as official administrative languages, and the nobility became increasingly Polonized. In 1697 use of Belarusian was officially banned; Polish became the official language. Immigration of Polish gentry to the region resulted in the emergence of a variety known as polszczyzna kresowa ‘borderland Polish’, Polish with a significant Belarusian substrate. During this time the greatest number of Polish loanwords entered the language. These fall into a number of different semantic categories, ranging from everyday vocabulary to political and military terminology, such as vjandlina ‘ham’, ksëndz ‘prince’, or zbroja ‘weapons’. With the partitions of Poland (1772–95), the region was reincorporated into the Russian empire. An uprising in January1831 against tsarist rule resulted in restriction on the activities and influence of the Catholic Church and restrictions on the use of the vernacular. The name of the region was officially changed from Belorussija to Severo‐ zapadnyj kraj ‘Northwestern region’ and Russian replaced Polish in all public spheres, including education, government, and the courts. Russian loanwords from the nineteenth century reflect the political climate at the time: ssylka ‘exile’ and perevarot ‘revolution’.

512  Lenore A. Grenoble Russification only intensified during the Soviet period and loanwords from all domains entered the lexicon, although the influx of technical, political, and scientific terminology is particularly noteworthy. Belarusian has long existed under the shadow of Russian and is seen by many, even today, as a substandard form of CSR. In fact, in western and northwestern Belarus, Belarusians themselves have considered Russian to be a standardized variety of their own language, not a separate language (Gustavsson 1997: 1922). Trasjanka thus stems from a combination of the intense Russification policies, the prestige of Russian, and the linguistic similarity of Belarusian and Russian. In current politicized, nationalist discourse in Belarus, the term “Trasjanka” is often used to refer to any kind of speech which in some way deviates from standardized Belarusian or Russian languages. For linguists the term refers specifically to a mixed language that combines elements of both languages. As with Surzhyk, there is some disagreement as to how regular the combinations are and to what extent they are idiosyncratic, and the amount of Russianisms may vary from speaker to speaker. Trasjanka typically has Belarusian phonetics and intonation, a mixed morphology and mixed lexicon. It is possible to find examples where the preposition is Belarusian and the nominal morphology Russian, as in ab cheloveke ‘about the person’. This mixture is illustrated in the following example, where the Russian elements are underlined and the Belarusian in bold: (1)

ščas  pagljaž u iakie sapožki pradajuc’ now I’ll look what boots   are selling ‘I’ll take a look to see which boots are for sale’

Phonetically there are Belarusian elements, such as the pronunciation of pagljaž u (versus a more R pagljižu) with lexical elements from both. The verb pradajuc’ could be found in either language, but the final consonant (part of the third person plural suffix) ‐c’ is BR, instead of the expected R ‐t. Although Trasjanka is typically defined as a jargon or as being limited to casual speech, it is widespread in spoken speech of all levels; Gustavsson (1997) claims that there are few people who can speak without some admixture of Russian. It also occurs in written form, in particular in texts purporting to report actual speech. One of the challenges in studying Belarusian, Russian, and Ukrainian contact is that the three languages are closely related and share many elements; in particular the dialect varieties in contact (which may differ significantly from the standard languages) may be quite close. This means that speakers of Surzhyk or Trasjanka have a pool of potentially very similar resources to choose from. In his analysis of Trasjanka, Hentschel (2014:117) proposes that these are hierarchically organized, from features most likely to come from Belarusian to those most likely to come from Russian: phonic elements < inflectional endings (with pronouns < with lexical words) pronominal stems < functional words < lexical stems < discourse markers. That said, this hierarchy captures general tendencies, and not absolute rules, i.e. Trasjanka speakers are most likely to use Belarusian phonetics but Russian discourse markers, but could use Belarusian discourse markers and Russian phonetics.

7  Trade and Contact Historically, Slavs, in particular Russian speakers, moved and traded with different people giving rise to different trade pidgins. Russenorsk, a Russian‐Norwegian pidgin spoken for approximately 150 years, is notable for the relatively balanced social status of both sets of speakers, with Russian traders coming to northern Norway seasonally to exchange grain for fish. Both languages contributed nearly equally to the lexicon, with some items from other languages (e.g. Dutch/Low German, French, and Swedish). Russenorsk exhibits some Sámi

Contact and the Slavic Languages  513 substrate effects; in particular certain fish names appear to have been borrowed from Sámi. Notable traits are a nominal suffix ‐a; a verbal marker in ‐om, which does not obviously derive from Norwegian or Russian; case relations are marked by the preposition på [po]; and some marking of tense with auxiliaries skulle and ville (all from Norwegian). There is relatively fixed word order in sentences with adverbials (verb‐final word order); and the negator (N ikke/R njet) is in second position (Jahr 1996: 115–17). Broch and Jahr 1984 provide a succinct overview of the grammar and history. Taimyr Pidgin Russian (or Govorka), a pidgin with two major substrate languages, Dolgan (Turkic) and Nganasan (Samoyedic, Uralic) was spoken well into the twentieth century, with remnants found today in communication between grandparents and their grandchildren (Stern 2005). Speakers of these languages lived in contact with speakers of other Indigenous languages, Enets and Nenets (also Samoyedic) and Evenki (Tungusic), with relatively low levels of proficiency in these languages, so that Taimyr Pidgin Russian emerged as the local lingua franca. The lexicon derives almost solely from Russian, with noticeable substrate effects in phonetics. Gender and number in nominals is generally neutralized to forms that would be masculine singular in Russian and show a loss of morphological case. Stern (2014) traces the grammaticalization of mesto (< Russian ‘place’) as a case marker in Taimyr Pidgin Russian, arguing that the substrate languages supplied the grammatical construction, which was copied into the pidgin. A major difference between Taimyr Pidgin Russian and other Russian‐lexifier pidgins is that the former uses the genitive‐accusative Russian form of pronouns (menja 1SG and tebja 2SG), while the latter have generalized the possessive form, as in moja 1SG (vs. R moja ‘my’, 1SG.POSS.FEM.SG.NOM) tvoja 2SG (vs. R tvoja ‘your’, 2SG.POSS. FEM.SG.NOM), in Russenorsk and Sino‐Russian. Verbs are also atypical, with what Stern refers to as acrolectal speakers using approximately the same system as in Standard Russian, while in the variety of basolectal speakers the second person (Russian) forms have been replaced by third person forms; first and third person are largely intact (2005: 309–11). Sino‐Russian contact is long‐standing due to contacts along the Chinese‐Russian border. The earliest accounts of a trade pidgin in Kyakhta (or Maimachen on the Chinese side) date to the mid eighteenth century, and Sino‐Russian contacts spread as far east as Vladivostok. A pidgin emerged in Harbin at the end of the nineteenth century and thrived in the first decades of the 1900s, when a significant number of Russians emigrated there during the Bolshevik Revolution, with the local population of Russians rising to some 165,000. Harbin Chinese‐Russian was spoken until the 1950s, as the local Russian population did not learn Chinese. It has some features characteristic of other Sino‐Russian pidgins, such as a past tense marker le [la] or [lo] (< Mandarin), and both languages contributing roughly equally to the lexicon. (Bakich 2011 gives a general discussion and detailed examples.) In recent years, relatively freer movement across the Chinese‐Russian border has opened up new opportunities for trade and for service contacts, with the latter primarily on the Chinese side of the border as Russians seeking tourism and health opportunities come to China. One result is the rise of new Russian‐lexifier pidgins with (re)newed Sino‐Russian contact (see Frajzyngier, Gurian, and Karpenko, this volume).

8  Contact and the Slavic Diaspora The term diaspora encompasses a wide range of people in different scenarios in today’s world. Since the break‐up of the Soviet bloc and the formation of the European Union, there has been significant movement of Slavic speakers. Diaspora in this broad context can refer to any number of different categories of people: (i) ethnic Slavs who have left one Slavic homeland for another country, either recently (in the last decade) or generations ago; (ii) L1 speakers of a given Slavic language who have returned to their homeland that is not a Slavic

514  Lenore A. Grenoble country, such as ethnic Germans who are L1 speakers of Russian and moved back to Germany after 1992; or (iii) ethnic Russians who, as Soviet citizens had been living in a non‐Slavic Republic which withdrew from the USSR, and thus found themselves to be foreigners without having moved (see also discussion in Verschik 2009: 303). This latter category of Russian diaspora in post‐Soviet successor states comprise a somewhat distinct category and are discussed in Section 8.2.

8.1  Slavic Diaspora and Heritage Languages One effect of the opening of borders in the last 20 years has been the emigration of Slavic speakers away from traditional Slavic homelands to other countries, often far away in Australia, Canada, and the United States. This migration is not new, and waves of migrations from Slavic territories to other areas have occurred in the past in conjunction with major political or economic events and upheavals. Yet the last few decades have seen a marked increase in migration, with the collapse of the Soviet Union and the opening of borders in the European Union. There are also large numbers of Poles, Russians, and Ukrainians, now living in English‐speaking areas such as Australia, Canada, the United Kingdom and the United States, resulting in contact between those languages and English. During the late Soviet period, Russians began emigrating in larger numbers to the United States (Kagan and Dillon 2010; Isurin 2011 provides details on Russian migrations; Ryaznova‐ Clarke 2014 includes a range of studies of Russian diasporic speech, see also Polinsky 2006). These extensive migrations have led to research into the heritage languages of second‐ generation speakers. The term heritage speakers is understood here in a restricted sense as “bilingual speakers of an ethnic or immigrant minority language, whose first language often does not reach native‐like attainment in adulthood” (Benmamoun et al. 2013: 129). Heritage speakers of Russian in the United States have been the subject of considerable research: Dubinina and Polinsky (2013) provide an overview, with examples of relatively free borrowing of English lexical items, more rigid SVO word order, reduction in register variation, in addition to such focused research as anaphora resolution (Ivanova‐Sullivan 2014) and implications for education (Kagan et al. 2019). Widespread contact between English and Polish has led to the coining of the term Ponglish (or Poglish) for the variety of English that has emerged (Meyer 2015). Ponglish, like other diasporic Slavic contact varieties, exhibits widespread loan translations of idioms into the matrix language. Another view on English‐Polish contact is Polish heritage speech, which has been less studied than Russian; Seretiny and Lipińska (2016) provide a broad overview. Focusing specifically on Heritage Polish in contact in Chicago, Kozminska (2013) shows changes in keeping with Heritage Russian in the United States, including simplified nominal morphology and reconfiguration in verbal morphology. Russian‐English contact in Alaska dates to the second half of the eighteenth century with the rise of the Russian‐American Company; today there are traces of the contact variety that emerged today in Ninilchik, Alaska. The majority of the lexicon comes from Russian. Ninilchik Russian exhibits a number of features common to heritage varieties with incomplete acquisition, such as loss of grammatical gender in nouns, with reconfiguration of three genders in pronouns that correlate with semantic categories (Bergelson and Kibrik 2010). There was some bi‐directional contact, and Russian lexical borrowings can be found identified in Alaskan native languages as well as in local English varieties, and speakers from Ninilchik are reported to speak English with a local “accent.” Israeli Russian is a Russian‐Hebrew contact variety with extensive lexical borrowings from Hebrew into Russian, as well as changes in phonology, morphology, and syntax. A sizable number of Russian speakers have emigrated to Israel, with counts over a million in the 1990s from the former Soviet Union. Israeli Russians make up 20% of the Israeli

Contact and the Slavic Languages  515 population, and Russian is a third language in the country, after Hebrew and Arabic. One point that distinguishes Israeli Russian from other diasporic varieties in Israel is relatively loose allegiance (if at all) to Hebrew; Soviet Jews did not have access to the language while growing up, and Hebrew does not constitute a core part of their Jewish identity (Perelmutter 2018). Contact with Hebrew has had an impact on Israeli Russian: Hebrew elements are borrowed into a Russian morphosyntactic frame; Hebrew intonational contours are used; there is considerable use of Hebrew lexicon in Russian, and Hebrew word order is found with borrowed adjectives following the noun, instead of preceding it, as is the case with native Russian adjectives. Syntax is affected in a small set of constructions: indirect questions use the interrogative R esli ‘if’, and relative clauses use the subordinator R čto ‘that’, instead of the relative pronoun kotoryj ‘which’, calques from the Hebrew model (Naiditch 2004, 2008).

8.2  Russian in Post‐Soviet Successor States Soviet language policies resulted in the spread of Russian, and Russian speakers, throughout the territory of the former Soviet Union and the Eastern bloc. Some of these territories have remained as part of the Russian Federation, while others are independent countries. One immediate effect of the collapse of the USSR was that Russians living in some of the Soviet Republics (such as the Baltic states) suddenly found themselves living in foreign countries when these territories regained their independence. Some emigrated back to Russia, their homeland. Many Russians chose to stay, and overnight became diaspora, living in foreign countries. Ties between Russia and these regions continued even after disintegration of the Soviet empire. Language contact with Russian has continued to be intense in most of these successor states; Pavlenko (2008) describes the overall sociolinguistic situation and status of Russian in each. Estonian‐Russian contact has been notably prolonged and intense, leading to code‐mixing and to a number of changes, in Estonian and in the local Russian variety (Verschik 2004; Zabrodskaja 2009, 2013).

9 Conclusion This chapter has provided a brief survey of some of the key contact phenomena in the history and development of the Slavic languages. A major area not discussed in detail here is the impact of Slavic on minority languages and resulting language shift, the phenomenon when speakers replace the use of one language with another. More than any other Slavic language, Russian has spread further geographically and linguistically, encompassing more speech communities. Language shift is widespread across minority Indigenous languages in the Russian Federation, where massive shift results in language endangerment and loss. All minority Siberian and Arctic languages are endangered (Pakendorf, this volume; Vakhtin 2001), and even medium‐sized languages such as Buryat (Mongolic) and Kazakh (Turkic) exhibit not only contact effects, but also shift. Research has begun focusing on the impact of contact‐induced change and shift on these languages and others in contact with other Slavic languages, such as in Polish and Wymysiöeryś (also known as Wymorsys or Vilamovian), and endangered German variety spoken in Poland (Wicherkiewicz et al. 2017). Although I have provided only the briefest of overviews of some of the modern Slavic languages, it is fair to say that there is no Slavic language which has been unaffected by contact. In fact, many of the contact effects date to prehistoric times and the structure of modern Slavic languages can only be understood with some understanding of earlier and current contact.

516  Lenore A. Grenoble

NOTES 1 The following abbreviations are used here for these and other languages with which they come into contact: BCS = Bosnian‐Croatian‐Serbian; BR = Belarusian; CSR = Contemporary Standard Russian; Cz = Czech; ESl = East Slavic; G = German; Gk = Greek; Goth = Gothic; LGmc = Late Germanic; OCS  = Old Church Slavonic; P = Polish; Pb = Polabian; PGmc = Proto‐Germanic; R = Russian; Sk = Slovak; Sur = Surzhyk; Ukr = Ukrainian. 2 Rusyn is alternatively referred to as Ruthenian, Carpatho‐Rusyn, or Rusnak. The Rusyn population lives in Sub‐Carpathian Ukraine, in the Lemko region of Poland, in the Prešov region of Slovakia, in Vojvodina in Serbia, and into Croatia (Vaň ko 2007). 3 See Vlasto (1988: 10–23) for a succinct survey of the differences in phonology and morphology. 4 This has similarly been noted in Croatian dialects spoken in Italy, e.g. Molisean Croatian under Italian influence s nož em, Italian con un cotello ‘with a knife’ (Breu 1996: 26). 5 See the online dictionary Silesian–Polish (Słownki Śląski) at www.slownik_slaski.itatis.pl.

REFERENCES Andersen, Henning 2003. Slavic and Indo‐ European migrations. In Henning Andersen (ed.), Language Contacts in Prehistory: Studies in Stratigraphy. Amsterdam: Benjamins, pp. 45–76. Bakich, Olga 2011. Did you speak Harbin Sino‐ Russian? Itinerario. (Special issue: Ethnic Ghettos and Transcultural Processes in a Globalised City: New Research on Harbin) 35.3: 23–36. Benmamoun, Elabbas, Silvina Montrul, and Maria Polinsky 2013. Heritage languages and their speakers: opportunities and challenges for linguistics. Theoretical Linguistics 39.3–4: 129–81. Bergelson, Mira B. and Andrej A. Kibrik. 2010. The Ninilchik variety of Russian: linguistic heritage of Alaska. Slavic Helsingiensia 40: 299–313. Bilaniuk, Laada 2004. A typology of surzhyk: mixed Ukrainian–Russian language. International Journal of Bilingualism 8.4: 409–25. Bilaniuk, Laada 2005. Contested Tongues: Language Politics and Cultural Correction in Ukraine. Ithaca, NY: Cornell University Press. Breu, Walter 1996. Überlegungen zu einer Klassifizierung des grammatischen Wandels im Sprachkontakt (am Beispiel slavischer Kontaktfälle). Sprachtypologie und Universalienforschung 49.1: 21–38. Dingley, John 1995. Iměti in the Laurentian Redaction of the Primary Chronicle. In Henrik Birnbaum and Michael S. Flier (eds.), The Language and Verse of Russia: In Honor of Dean S. Worth on his 65th Birthday. Moscow: Vostochnaya Literatura, pp. 80–7.

Dubinina, Irina and Maria Polinsky 2013. Russian in the USA. In Michael Moser and Maria Polinsky (eds.), Slavic Languages in Migration. Vienna: University of Vienna, pp. 1–28. Flier, Michael S. 2000. Surzhyk: the rule of engagement. In Zvi Gitelman, Lubomyr A. Hajda, John‐Paul Himka, and Roman Solchanyk (eds.), Cultures and Nations of Central and Eastern Europe: Essays in Honor of Roman Szporluk. Cambridge, MA: Harvard University Press, pp. 113–36. Galton, Herbert 1994. The phonological influence of Altaic on Slavic. Journal of Slavic Linguistics 2.1: 77–91. Gołąb, Zbigniew 1991. The Origins of the Slavs: A Linguist’s View. Columbus, OH: Slavica. Grenoble, Lenore A. 2015. Language contact in the East Slavic contact zone. Balkanistica 28. 225–50. Grković‐Major, Jasmina. 2011. The development of predicative possession in Slavic Languages. In Motoki Nomachi (ed.), The Grammar of Possessivity in South Slavic Languages. Synchronic and Diachronic Perspectives. Sapporo: Slavic Research Center, Hokkaido University, pp. 35–54. Gustavsson, Sven 1997. Byelorussia. In Hans Goebl, Peter H. Nelde, Zdenèk Starý, and Wolfgang Wölck (eds.), Kontaktlinguistik: ein internationales Handbuch zeitgenössischer Forschung, vol. 2: 1919–1926. Berlin: de Gruyter, pp. 1919–26. Hannan, Kevin 1996. Borders of Language and Identity in Teschen Silesia. Frankfurt‐am‐Main: Peter Lang.

Contact and the Slavic Languages  517 Hentschel, Gerd 2014. Belarusian and Russian in the mixed speech of Belarus. In Juliane Besters‐Dilger (ed.), Congruence in Contact‐ Induced Language Change. Berlin: de Gruyter, pp. 93–121. Hentschel, Gerd, Oleksandr Taranenko, and Sjarhej Zaprudski (eds.) 2014. Trasjanka und Suržyk – gemischte weißrussisch‐russische und ukrainisch‐russische Rede. Sprachlicher Inzest im Weißrussland und der Ukraine? Frankfurt‐am‐ Main: Peter Lang. Isačenko, Alexander 1974. On “have” and “be” languages. In Michael S. Flier (ed.), Slavic Forum. The Hague: de Gruyter Mouton, pp. 43–77. Isurin, Ludmila 2011. Russian Diaspora: Culture, Identity, and Language Change. Berlin: de Gruyter. Ivanova‐Sullivan, Tanya 2014. Theoretical and Experimental Aspects of Syntax‐Discourse Interface in Heritage Grammars. Leiden: Brill. Jahr, Ernst Håkon 1996. On the pidgin status of Russenorsk. In Ingvild Broch and Ernst Jahr Håkon (eds.), Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: de Gruyter Mouton, pp. 107–22. Kagan, Olga and Kathleen Dillon 2010. Russian in the USA. In Kim Potowski (ed.), Language Diversity in the USA. Cambridge: Cambridge University Press, pp. 179–94. Kagan, Olga E., Maria M. Carreira and Claire Hitchens Chik (eds.) 2017. The Routledge Handbook of Heritage Language Education: From Innovation to Program Building. New York: Routledge. Kiparsky, Viktor 1934. Die gemeinslavischen Lehnwörter aus dem Germanischen (Annales Academiae Fennicae 32.2). Helsinki: Finnische Literaturgesellschaft. Kiparsky, Viktor 1969. Gibt es ein finnougrisches Substrat im Slavischen? Helsinki: Suomalainen Tiedeakatemia. Kittel, Bernhard, Diana Lindner, Mark Brüggermann, Jan Patrick Zeller, and Gerd Hentschel 2018. Sprachkontakt – Sprachmischung – Sprachwahl – Sprachwechsel. Eine sprachsoziologische Untersuchung der weißrussisch‐russisch gemischten Rede, “Trasjanka” in Weißrussland. Frankfurt‐am‐ Main: Peter Lang. Kozminska, Kinga 2013. Language contact in the Polish‐American community in Chicago. International Journal of Bilingualism 19.3: 239–58. Lötzsch, Renate 1996. Interferenzbedingte grammatische Konvergenzen und Divergenzen zwischen Sorbisch und Jiddisch.

Sprachtypologie und Universalienforschung 49.1: 50–9. Mazzitelli, Lidia Federica 2017. Predicative possession in the languages of the Circum‐ Baltic area. Folia Linguistica 51.1: 1–60. McAnallen, Julia 2011. The history of predicative possession in Slavic: internal development vs. language contact (University of California, Berkeley Ph.D. dissertation). Meyer, Anna‐Maria 2015. “Thanks from the mountain!”: humorous calques in Ponglish as an output of language contact and language creativity. Journal of Slavic Linguistics 23.1: 33–49. Nábělková, Mira 2007. Closely related languages in contact: Czech, Slovak, “Czechoslovak.” International Journal of the Sociology of Language 183: 53–73. Naiditch, Larissa 2004. Russian immigrants of the last wave in Israel. Patterns and characteristics of language usage. Wiener Slawistischer Almanach 53: 291–314. Naiditch, Larissa 2008. Tendencies in the development of the Russian language abroad. Russian Linguistics 32.1: 43–57. Nichols, Johanna 1993. The linguistic geography of the Slavic expansion. In Alan Timberlake (ed.), American Contributions to the Eleventh International Congress of Slavists. Columbus, OH: Slavica, pp. 377–91. Nichols, Johanna forthcoming. Slavic prehistory. In Victor A. Friedman and Lenore A. Grenoble (eds.), The Slavonic Languages, 2nd edition. London: Routledge. Pavlenko, Aneta 2008. Multilingualism in post‐ Soviet countries: language revival, language removal, and sociolinguistic theory. International Journal of Bilingual Education and Bilingualism 11.3–4: 275–314. Perelmutter, Renée 2018. Israeli Russian in Israel. In Benjamin Hary and Sarah Bunin Benor (eds.), Languages in Jewish Communities, Past and Present. Berlin: de Gruyter, pp. 520–43. Polański, Kazimierz 1993. Polabian. In Bernard Comrie and Greville G. Corbett (eds.), The Slavonic Languages. London: Routledge, pp. 795–824. Polinsky, Maria 2006. Incomplete acquisition: American Russian. Journal of Slavic Linguistics 14: 191–262. Ryazanova‐Clarke, Lara (ed.) 2014. The Russian Language Outside the Nation: Speakers and Identities. Edinburgh: Edinburgh University Press. Schenker, Alexander M. 1985. Polish. In Alexander M. Schenker and Edward

518  Lenore A. Grenoble Stankiewicz (eds.), The Slavic Literary Languages: Formation and Development. New Haven: Yale Concilium on International and Area Studies, pp. 195–210. Schenker, Alexander M. 1995. The Dawn of Slavic. New Haven: Yale University Press. Schuster‐Šewc, Heinz 1996. Grammar of the Upper Sorbian Language: Phonology and Morphology trans (Gary H. Toops, trans). Munich: LINCOM Europa. Seretiny, Anna and Ewa Lipińska 2016. Polish as a heritage language: somewhere in between. Studia Migracyjne – Przegląd Polonijny 2. 177–201. Szydłowska‐Ceglowa, B. 1987. Das Dravänopolabische des 17. und 18. Jahrhunderst als Zeugnis eines untergehenden Slawenstammes. Zeitschrift für Slawistik 32.4: 606–12. Stern, Dieter 2005. Taimyr Pidgin Russian. Russian Linguistics 29.3: 289–318. Stern, Dieter 2014. Grammaticalization in Russian lexifier pidgins. In Motoki Nomachi (ed.), Grammaticalization in Slavic Languages: From Areal to Typological Perspectives. Slavic Eurasian Studies 23. Sapporo: Slavic Research Center, Hokkaido University, pp. 49–90. Stern, Dieter 2017. Review of Hentschel, Gerd, Oleksandr Taranenko and Sjarhej Zaprudski (eds.) 2014. Acta Slavica Iaponica 38: 1–12. Taranenko, Oleksandr 2007. Ukrainian and Russian in contact: attraction and estrangement. International Journal of the Sociology of Language 183: 119–40. Tiraspol’skij, Gennadij I. 1998. Finno‐ugorskoe vlijanie na russkij jazyk kak prognostičeskij material [Finno‐Ugric influence on the Russian language as prognostic material]. Obščie problemy prepodavanija jazykov: tezisy meždunarodnoj naučno‐metodičeskoj konferencii MAPRJaL [Common problems of language teaching: theses of the internantional scientific‐ methodological conference MAPRJaL], p. 602. Toops, Gary H. 1999. Syntactic calquing: reflections of German verb governance in Czech and Upper Sorbian. Die Welt der Slaven 44.2: 271–84. Toops, Gary H. 2006. A contrastive perspective on several morphosyntactic features of Upper Sorbian. Canadian Slavonic Papers 48.12: 137–55. Townsend, Charles 1981: Czech through Russian. Columbus OH: Slavica.

Trubačev, Oleg N. 1977. Lingvističeskaja periferija drevenjšego slavjanstva: indoarijcy v severnom pričernomor’e [The linguistic periphery of the ancient Slavs: Indo‐Aryans in the northern Black Sea area]. Voprosy jazykoznanija 6: 13–29. Vakhtin, Nikolai B. 2001. Jazyki narodov severa v XX veke: ocherki jazykovogo sdviga. [Languages of the peoples of the North in the twentieth century: notes on language shift]. St. Petersburg: Dmitrii Bulanin. Vaň ko, Juraj 2007. The Rusyn language in Slovakia: between a rock and a hard place. International Journal of the Sociology of Language 183: 75–96. Veenker, Wolfgang 1967. Die Frage des finnougrischen Substrats in der russischen Sprache (Uralic and Altaic Series 82). Bloomington: Indiana University Press. Verschik, Anna 2004. Russian‐Estonian language contacts, linguistic creativity, and convergence: new rules in the making. Multilingua: Journal of Cross‐Cultural and Interlanguage Communication 24.4: 413–29. Verschik, Anna 2009. Introduction. Special issue on Language contacts in the post‐Soviet space. International Journal of Bilingualism 13.3: 299–307. Vlasto, Alexis P. 1988. A Linguistic History of Russia to the End of the Eighteenth Century. Oxford: Clarendon Press. Wicherkiewicz, Tomasz, Tymoteusz Król, and Justyna Olko 2017. Awakening the language and speakers’ community of Wymysiöeryś. European Review 26: 179–91. Zabrodskaja, Anastassia 2009. Evaluating the Matrix Language Frame Model on the basis of a Russian‐Estonian codeswitching corpus. International Journal of Bilingualism 13.3: 357–77. Zabrodskaja, Anastassia 2013. Morphosyntactic contact‐induced language change among young speakers of Estonian Russian. In Isabelle Léglise and Claudine Chamoreau (eds.), The Interplay of Variation and Change in Contact Settings. Amsterdam: Benjamins, pp. 77–106. Zaliznjak, Andrej A. 1963. O xaraktere jazykovogo kontakta mezdu slavjanskimi i skifo‐sarmatskimi plemenami [On the character of linguistic contact between the Slavic and Scytho‐Sarmatian tribes]. Slavjanskoe jazykoznanie. Moscow: Akademija nauk SSR, pp. 3–22.

26 Contact and the Finno‐Ugric Languages JOHANNA LAAKSO

1  Introduction: The Finno‐Ugric/Uralic Language Family The Uralic language family, consisting of 20–40 languages spoken in Eastern Europe and western Siberia, is traditionally described in terms of a binary family tree, starting from the first binary split of Proto‐Uralic into Proto‐Finno‐Ugric and Proto‐Samoyedic. However, this model is not unanimously accepted. In the alternative, more “bush‐like” models, Samoyedic is simply one of the three or more main branches, which means that the terms Uralic and Finno‐Ugric can be used as synonyms. This practice is adopted also in this chapter. The Finno‐Ugric/Uralic language family, irrespectively of the structure of the postulated family tree, consists of six main branches, two of which (Finnic‐Saami and Ugric) are contested: 1. The Finnic‐Saami branch. The Finnic (aka “Baltic Finnic” or “Fennic”) languages include the state languages Finnish and Estonian and a number of minority languages, all more or less severely endangered. Of these, Karelian has traditionally been spoken both in Russia and in Finland, Veps and the almost extinct Ingrian and Votic in Northwestern Russia; Livonian, spoken in Latvia, is practically extinct in its traditional form but revitalized to some extent by activists. Furthermore, two Finnish varieties of the Far North have recently been ­officially recognized as minority languages: Meänkieli (lit. ‘our language’) in Sweden in 2000 and Kven in Norway in 2005. In the southeast of Estonia, the Võro and Seto varieties, traditionally regarded as dialects of Estonian but actually representing a deeply different variety of Finnic, have since the late 1980s developed their own written standards. 2. The Saami (Sámi, ‘Lapp’) languages form a dialect continuum, usually divided into ten languages, stretching from Sweden and Norway through Finnish Lapland to the Kola Peninsula in northwest Russia. The border between Finnic and Saami is steep, clearly beyond mutual intelligibility, and the existence of a Saami‐Finnic branch is sometimes contested. 3. Mordvin, spoken in European Russia. The Mordvins are divided into two linguo‐ethnic groups, Erzya and Moksha, both having a standard language of their own. 4. Mari (aka Cheremis), spoken in European Russia. There are two standard language vari­ eties, East (Meadow) and West (Hill) Mari. For cultural‐geographic reasons, Mari and Mordvin are sometimes bundled together as the “Volgaic branch,” but there is no linguistic evidence for “Proto‐Volgaic.”

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

520  Johanna Laakso 5. The Permic branch includes two or three languages of European Russia: Komi (also known as Zyryan), with two standardized varieties, Komi‐Zyryan and Komi‐Permyak, and Udmurt (aka Votyak). 6. The Ugric branch is the most heterogeneous subgroup, as the relatedness between Hungarian and the two Ob‐Ugric languages in western Siberia is rather distant. The Ob‐Ugric languages, Mansi (aka Vogul) and Khanty (aka Ostyak) both fall into deeply different varieties which might be regarded as separate languages. 7. The Samoyed branch now includes four languages spoken in western Siberia: Nenets (aka Yurak, with two deeply different varieties: Tundra and Forest Nenets), Enets (aka Yenisey Samoyed, also with a deep division between the Tundra and the Forest variety), Nganasan (also known as Tawgy Samoyed), and Selkup (also known as Ostyak Samoyed). Some extinct Samoyed languages are also known; the most extensively documented is Kamass, the last speaker of which died in 1989. All Uralic languages except Finnish, Estonian, and Hungarian in their respective nation‐ states are more or less endangered. This applies also to the Mordvin, Mari, and Permic lan­ guages, which have hundreds of thousands of speakers and nominally a co‐official status in their titular republics (Mordvin, Mari, Udmurt, Komi) in the Russian Federation. The Uralic languages are spoken in vastly different ecological and sociopolitical environ­ ments, and the genetic and typological diversity within Uralic is comparable to that between the branches of Indo‐European. Therefore, an exhaustive description of all language contact situations involving Finno‐Ugric is an impossible task. In what follows, I will merely present a quick survey on the diversity of relevant language contacts, followed by some exemplary cases. In particular, I would like to draw attention on the rich Finno‐Ugristic research ­tradition which, being published in “less accessible” languages such as German, Russian, Hungarian, or Finnish, is often overlooked in today’s linguistic research.

2 Language Contact Situations Involving Finno‐Ugric: An Overview 2.1  Finno‐Ugric Minorities and the Majority Language Most Finno‐Ugric languages of today – i.e. all except Hungarian, Finnish, and Estonian in their nation‐states  –  are endangered minority languages that are mainly used among family members and friends or in small communities, in connection with the traditional way of living or the ethnic heritage in general. Practically all speakers of these languages are bilingual, and for most of them, in particular for younger speakers, the majority lan­ guage is the dominant or the sole language of higher education (or even all education), professional communication, literary culture, media, and urban life in general. This applies for the Finno‐Ugric minorities in Russia and also for the Saami in general and for the last Livonians in Latvia. It is also true of the Finnish minorities in Sweden and Norway, of the remainders of post‐World War II Estonian and Hungarian emigrant groups west of the former Iron Curtain, and, to some extent, of some the old Hungarian minorities in the neighbor states of Hungary. Of course, the degree of endangerment and the disruption of language transmission varies greatly. In some traditionally Hungarian‐speaking areas outside Hungary, Hungarian can still be the majority language, used in professional communication, media, primary schools, and higher education. The Finno‐Ugric peoples in Russia, in contrast, are a minority even in all their titular republics and “autonomous” areas, and

Contact and the Finno‐Ugric Languages  521 despite language laws and institutions (such as national theaters, museums, publications, and research institutes) the presence of these languages in urban life, in the media, or in the education system is often very marginal. For the Finno‐Ugric minorities in Russia, the dominant majority language today is Russian. However, there are clear differences in the age and depth of the impact of Russian: In the Volga region, the cultural and political dominance of Turkic‐speaking peoples only gradually gave way to Russian from the sixteenth century on, and for the Finno‐Ugrians of this region (Mari, Udmurt, to some extent also Mordvin), the most important contact language even until the twentieth century could still be Tatar, Chuvash, or Bashkir. Siberia was colonized only after the Middle Ages, and the dominance of Russian administration and culture remained rather superficial until the twentieth century; instead, there were contacts between the indig­ enous Uralic and non‐Uralic (Yeniseic, Turkic, Tungusic) languages of Siberia. In contrast, the eastern Finnic peoples (Karelians, Veps, Votes, Ingrians), in the vicinity of Novgorod, belonged to the core area of emerging Russian nationhood already in the Middle Ages. Also the Komi in northern Russia (see e.g. Leinonen 2002) and, to some extent, the Mordvin were incorporated by the Russian state early on. This long history of close contacts with Russian manifests itself in a plethora of loanwords, parallel structural developments, and even Sprachbund‐like phenomena such as those between northwest Russian dialects and eastern Finnic (cf. Sarhimaa 1992, 1999; Helimski 2003).

2.2  “Western” Finno‐Ugric Languages and Their Neighbors From the Middle Ages on, the Finnic and Saami languages  –  already deeply marked by intensive contacts with (Pre‐)Germanic and (Pre‐)Baltic and also showing some traces of early Slavic contacts  –  were divided by the great border between East and West, also the border line between eastern and western Christianity and coinciding with the western border of the emerging Russian Empire. West of this boundary, the politically dominant lan­ guages were Scandinavian (Swedish or Norwegian) or, south of the Gulf of Finland, German. Lapland outside the Russian power sphere was gradually divided between the states of Sweden and Norway, while today’s Finland arose from the part of the Finnic area which was annexed to Sweden. The border between Sweden and Russia divided the daughter dialects of Proto‐Karelian into today’s Karelian and East Finnish and joined the East Finnish dialects with the West Finnish dialects, thus gradually weakening the originally deep dialect border between East and West Finnish. In this area later known by the name of Finland, Swedish (beside and after Latin) became the dominant language of administration, higher education (although elementary literacy in Finnish was promoted by the Reformation from the sixteenth century on), and higher social strata. From the seventeenth to the nineteenth century, Finnish was the language of peasants and servants, and upward social mobility inevitably led to language shift. This superstrate influence resulted in hundreds of Swedish loanwords, loan transla­ tions, and other influences (cf. de Smit 2006). Partly hidden under the Swedish influences, there are also elements from West European culture languages in Finnish, such as Latin (whether these words were conveyed by Swedish or borrowed directly cannot always be unambiguously assessed) or German (Low German loanwords in particular are sometimes indistinguishable from the Swedish ones, cf. Bentlin 2008). In Estonia, the situation was largely a mirror image of that in Finland: The dominant lan­ guge in the colonized Baltic countries until the nineteenth century was German (Low German, later ousted by High German as in the northern parts of the German‐speaking area in general; cf. Hinderling 1981), Estonian being mainly the language of serfs and servants, was cultivated since the Reformation only sparsely for the goals of the Lutheran church and elementary education. According to the often‐quoted statistics of Rätsep (1983), 24% of the

522  Johanna Laakso underived words in Standard Estonian (modern internationalisms excluded) are of either Low German or High German origin. There were also influences from Swedish (sometimes difficult to distinguish from the ample Low German elements; cf. Raag 1987, 1997), Latvian (cf. Vaba 1977), and other languages. Livonian, since the Middle Ages rapidly giving way to Latvian, was deeply marked by the Latvian language, superseded by the dominant German (Winkler 2014; Grünthal 2015). The impact of Russian on Finnish and Estonian was clearly weaker than that of Swedish or German. However, quite a few Russian loanwords were adopted to the easternmost dia­ lects (for Estonian, see Must 2000) and, in Estonia in particular, from the language of Russian administration in the eighteenth and nineteenth centuries. Some of these words, sometimes spreading by way of gradual diffusion and intertwining with native descriptive words (Jarva 2001), made their way into the standard languages (Plöger 1973; Blokland 2009). Unlike most of their linguistic relatives, the linguistic ancestors of the Hungarians, adopt­ ing a mobile, half‐nomad way of life, passed through various language contact situations, which are partly reconstructible only on the basis of early loanword strata. Arriving in the Pannonian basin toward the end of the first millennium ce, they ousted and assimilated Slavic speakers (cf. Kniezsa 1955; Décsy 1988; Kallio and Laakso, forthcoming). In addition to Slavic and Romanian minorities and neighbors as well as some Turkic‐speaking groups such as the Cumans whose language was probably spoken in Hungary throughout the Middle Ages (Lyublyanovics 2011), Hungarian had intensive contacts with German – also the language of diverse minority groups in historical Hungary and, later on, a dominant lan­ guage of culture and education – and other languages. The influence of Latin, the language of administration and education in Hungary even until the nineteenth century, was remark­ ably strong. Prompted by Romantic Nationalism from the nineteenth century on, all three Finno‐Ugric nation‐state languages experienced an intensive phase of puristic language planning often explicitly attempting to reverse contact‐induced change. In addition to lexical purism result­ ing in numerous native‐based neologisms, there were attempts to get rid of identifiable foreign influences in phonology and orthography or, for instance, word order models (such as verb‐final subordinated clauses in Estonian) or word‐formation strategies (such as the adjectives in ‐rikas ‘rich’ or ‐vapaa ‘free’ in Finnish, mirroring Swedish adjectives in ‐rik or ‐fri) which were perceived as foreign. In all three Finno‐Ugric nation‐states, there are considerable language minorities. Hungarian in particular played a role uniquely dominant among the Uralic languages, vis‐à‐vis the numerous German‐speaking, Slavic, and Romanian minorities of historical Hungary until the end of the Austro‐Hungarian Empire, see, for instance, Gregor (1994) for Hungarian loanwords in Slovak. In post‐World War I Hungary, which lost two thirds of its area to the new neighboring states in the Treaty of Trianon (simultaneously, one third of ethnic Hungarians became minorities in the new neighboring countries), the numbers of remaining minorities, now ranging from a few thousand (Slavic, Romanian) to more than 30,000 (German) or some 50,000 speakers (Romani), have generally been receding, while the most numerous minority, the Roma, suffers from stigmatization and social problems. Estonia lost most of her “old” minorities (Baltic Germans, Swedes, Jews, and Roma, part of the “old” Russian minority) as a result with World War II, but the massive immigration from other parts of the Soviet Union in the postwar decades created a new, large, and mostly Russian‐speaking minority. During the Soviet period, very little was done to promote knowledge of Estonian among this minority, but now it is increasingly exposed to Estonian, the only official state language since the restoration of independence in 1991; for Estonian‐ Russian language contacts and new bilingual practices, see especially Verschik (2010). In Finland, the Finland variety of Swedish, spoken traditionally in southern and western coastal areas, is an official national language beside Finnish and has a strong institutional

Contact and the Finno‐Ugric Languages  523 basis. However, Finland Swedish is only spoken by less than 6% of Finland’s population. It is receding especially in the originally Swedish‐speaking Helsinki region and subject to the triple pressure of the Finnish majority, the Sweden‐Swedish standard, and global English (Saari 2000; Östman 2006). Beside other “old” minority languages (three Saami varieties, Karelian, Romani, etc.), there are languages spoken by rapidly growing immigrant groups in Finland since the 1980s and 1990s; the immigrant languages with the most speakers are Russian and Estonian.

2.3  Contact within Finno‐Ugric With the exception of the quasi‐isolate Hungarian, most Finno‐Ugric languages have (or have until recently had) contacts with related languages, sometimes resulting in compli­ cated, Sprachbund‐like networks, in which not only lexical borrowing but also transmission of morphological elements is possible. The contacts between Finnic (Finnish / Karelian) and Saami are an excellent example. The distance between Saami and Finnic is well beyond mutual intelligibility, but for bilingual speakers, structural similarities and systematic corre­ spondences are identifiable and psychologically real. This has led to borrowing of bound morphs and nativization of loanwords (Aikio 2007), which makes it difficult to distinguish borrowings from common inheritance. Extrapolating from this situation to the prehistory of Finnic and Saami, one can claim that a substantial part of the Finnic‐Saami common features stem from a long‐time contact between Proto‐Finnic and Proto‐Saami, and the existence of a Saami‐Finnic protolanguage can be contested (Aikio 2012). In Siberia, similarly close relationships have developed between different varieties of Mansi and Khanty (Honti 1998: 352), the speakers of which were culturally close to each other and sometimes connected by interethnic social networks. Of the Permic languages, Udmurt belongs to the same “Volgaic” cultural sphere as Mari and Mordvin, characterized by intensive contacts with Turkic, and as Mari and Mordvin in particular are spoken in scat­ tered language islets across the whole Volga region, there are local contacts between individual varieties of these languages (see e.g. Bereczki 2007). Also the Hungarians on their way to the west may have had contacts with speakers of Permic varieties (Rédei 1964b). Komi, on the other hand, bears traces of contacts with Finnic or “Para‐Finnic” (Northwest‐ Finno‐Ugric) languages probably spoken in today’s northern Russia before the East Slavic expansion; there are some good loanword etymologies, but the hypotheses cautiously for­ mulated by Hausenberg (1998) about the possible role of Finnic contacts for the morphosyn­ tactic divergence of Komi from its sister language Udmurt still call for further research, also in the light of the developing substrate language research (cf. Section 3.2). The Komi also came into contact with the westward‐spreading Nenets in the utmost northeast of European Russia (Wichmann 1902; Rédei 1962), and as merchants, middlemen, and colonists in west­ ern Siberia, with the Mansi and Khanty (Rédei 1964a, 1970).

2.4  Typical(?) Outcomes of Language Contact The comparative linguistic study of Finno‐Ugric languages is characterized by a strong tra­ dition of etymology and loanword research. Older loanwords from different Indo‐European branches in Finnic have given rise to a strong research tradition, starting with the seminal work of Vilhelm Thomsen on the old Baltic and Germanic loanwords in Finnic (Thomsen 1869, 1870, 1890); the current stand of research is summarized by Junttila (2012) and Kallio (2012). A similarly strong tradition of loanword research has developed around the prehis­ tory of the Hungarian language. The migrations of the linguistic ancestors of the Hungarians from the East to the Pannonian basin are practically undocumented, so that almost the only way of attaining evidence from those times has been research into the early loanword strata

524  Johanna Laakso from Iranic (for a summary, see Korenchy 1988) and Turkic (Róna‐Tas and Berta 2011). Loanword research in connection with Indo‐European and Turkic studies has been a central part of Uralic etymology and historical linguistics; in recent decades, substrate studies (see Section 3.2) have emerged as a new avenue of research. In addition to etymological dictionaries (mainly of Finnish and Hungarian), there are numerous studies of specific loanword strata, also in the minor Finno‐Ugric languages. In contrast, research into contact‐induced “Indo‐Europeanization” (or “Turkicization”) in other subsystems of language than the lexicon has been much less intensive, and there are contro­ versies symptomatic of fundamental problems of language contact research; some of these will be dealt with in more detail in what follows. In the most intensive contact situations, even adoption of morphological elements, such as Latvian verbal prefixes for aspectual or Aktionsart meanings in Livonian (or, corre­ spondingly, Russian prefixes in Karelian; Kiefer and Honti 2003), has been attested. In contacts within Finno‐Ugric, examples of transmission of inflectional or derivational mor­ phology are known at least between Finnic and Saami. In syntax, a classic example of bor­ rowed elements triggering deep‐going structural changes is the borrowing of conjunctions: in many Finno‐Ugric languages of Russia, conjunctions borrowed from Russian have replaced inherited means of expressing causal, temporal, etc. relations with converbs or other nonfinite verb forms (for an interesting case study in Khanty, see Csepregi 1997; cf. also Thomason 2001: 62). There are also examples of systematic adoption of phonological features and phonotactic constraints: Already in the early twentieth century, younger Livonian speakers substituted the ü and ö vowels (unknown in Latvian) with i and e. Livonian also has taken over the accent/intonation system of Latvian, including the often‐mentioned stød (Tuisk 2014). The aspiration of voiceless word‐initial stops as in Scandinavian has spread from Swedish (and Norwegian) to the language of the old Finnish minority in northern Sweden and to some Saami varieties. In some cases, the exposure of the Finno‐Ugrians to the majority or culturally dominant language was weak enough to allow for the development of a pidginized variety: examples include Govorka or the Taimyr Pidgin Russian in northernmost Siberia (Wurm 1996), Halbdeutsch spoken by uneducated Estonians until the nineteenth century (Lehiste 1965), and the poorly documented gavppe‐daro (“Trade Norwegian”) and borgarmålet (pidginized Swedish) spoken by some Saami groups in the eighteenth or nineteenth century (Jahr 1996). Compared with the research on contact‐conditioned changes in Finno‐Ugric, there is much less literature on the possible impact of Finno‐Ugric languages on other lan­ guages – with the exception of substrate studies (see Section 3.2) and the study of majority‐ language impact on the minority languages in the Finno‐Ugric nation‐states, such as the Fennisms in Finland Swedish (Saari 2000) or Finnish Romani (Borin and Vuorela 1998). This obvious imbalance has many reasons. First, it may go back to factual power and prestige relations in the contact situation. In today’s Russia, for instance, practically all Finno‐Ugric minority speakers are bilingual in Russian, while the knowledge of Finno‐Ugric languages among the Russian majority is very rare. Even in areas where the contacts between Russians and Finno‐Ugrians have traditionally been intensive and more balanced (such as the con­ tacts between northwest Russian dialects and Karelian or Veps), the Finno‐Ugric influences are often restricted to local substandard dialects; there are only a few Finnic loanwords that have managed to spread to Standard Russian. Second, the non‐Finno‐Ugric contact partner may not be accessible to research any more. It is generally assumed, for instance, that the numerous (Pre‐)Baltic and (Pre‐)Germanic loanwords in Finnic were at least partly acquired from Indo‐European (IE) speaker popula­ tions who lived among the Proto‐Finnic speakers in present‐day Finland (and possibly Estonia) and were later assimilated, so that their IE language, probably marked by contacts

Contact and the Finno‐Ugric Languages  525 with Finnic, did not survive. Third, there may simply not be enough interest and expertise: Kallio (2000b), presenting a Finnic etymology for Germanic *maþōn (> English moth; the Finnic word, in turn, could be an Indo‐Iranic loan), suspects that more loanwords from Finn(o‐Ugr)ic into IE could be found by systematic searches. Junttila (2015b) and Holst (2015), in contrast, show that the hypotheses of Proto‐Finnic loans or substrates in Balto‐ Slavic are largely unfounded and based on outdated or misunderstood data. Recently, Hyllested (2014) has claimed that many more ancient Finnic or Finno‐Ugric loanwords can be found in Germanic and even in Celtic, but so far, his new etymologies have not met with unanimous acceptance.

3  Some Central Questions 3.1  Contact or Relatedness? On the Earliest Loanword Strata The well‐known similarity of certain Uralic basic vocabulary items to IE, such as Proto‐Uralic (PU) *nimi ‘name’, *weti ‘water’ or *ku‐ ‘who’, has been interpreted in many ways. While most Finno‐Ugrists are very wary of categorical statements in this question, there have been more or less cautiously formulated versions of the “Indo‐Uralic” hypothesis, and Helimski (2001) prefers to regard these words as evidence for a Nostratic relatedness. Others have attempted to explain all similarities in terms of very early borrowing; the most prominent representative of this approach is the Finnish Indo‐Europeanist Jorma Koivulehto (for a syn­ thesis of his work, see Koivulehto 1999, 2001a; for sympathizing views, see also e.g. Anttila 2000; Kallio 2002). Koivulehto began by discovering very early Germanic loanwords in Finnic (and, partly, in the neighboring branches of Finno‐Ugric as well) and thus revolutionizing the established chronologies of Finnic–Germanic contacts. Delving deeper into Pre‐Germanic and Northwest Indo‐European, he found more and more loanwords representing an even more archaic, practically Proto‐Indo‐European (PIE) level of reconstruction – for instance, reflexes of PIE laryngeals in Finnic (Finnish kaski ‘slash‐and‐burn’ PIE *Hazg‐ ‘ashes’) or even in Uralic. Inspired by the reconstruction of Proto‐Uralic phonology by Janhunen and Sammallahti (Sammallahti 1988), which introduced a mystery consonant *x, functionally not unlike the PIE laryngeals, Koivulehto discovered cases of Uralic *x substituting a PIE laryngeal, such as PU *näxi ‘woman’ PIE *gwneH‐, or PU *tuxli ‘wind’ PIE *dhuH‐li‐. All of these loan etymol­ ogies have not been unanimously accepted, and the criticism highlights some central prob­ lems in the research of reconstructed language contact. While acknowledging the technical brilliance of Koivulehto’s etymologies, his critics have pointed out that he sometimes operates with IE roots not attested firmly enough or extended with suffixes of a questionable status (Helimski 2001), that he postulates semantic shifts difficult to motivate (e.g. ‘put into motion’ > ‘row’, ‘pour (a libation?)’ > ‘drink’), that it is not realistic to assume such a great number of loanwords in the basic vocabulary (Rédei 2002) and that, in general, he seems to over‐exploit his model, relying on sound substitutions (cf. Abondolo 1998a: 7; Janhunen 1999). The great differences bet­ ween the consonant systems of PIE and PU make it theoretically possible to find many kinds of IE originals; for instance, an initial k‐ in Uralic might have been used to substitute PIE *k‐, *sk‐, *kw‐, *k’‐, *g‐, *g’‐, *gh‐, *gw‐ or *H‐. Koivulehto also ingeniously exploits reconstructible intermediate stages of a change in progress; for example, a whole set of new Indo‐Iranic etymologies is based on the idea that the satemization in Indo‐Iranic (for instance, PIE *k’ > *ć > s) proceeded through a depalatalized affricate phase and that this nonpalatalized *c (or *dz), lacking an exact counterpart in Uralic, could have been substituted by s‐ or ‐ks‐ (Koivulehto 2001a: 252–7).

526  Johanna Laakso As Koivulehto’s critics see it, he has pushed the exploitation of reconstructed phonologies (of different stages of reconstruction) and sound substitution to its limits. Koivulehto may be right in referring to the well‐known fact that in intensive language contact anything can be borrowed, but as long as there is hardly any other evidence for the intensive character of early Uralic–IE language contacts (for instance, convergent developments in Uralic and IE morphosyntax), his etymologies, although technically flawless, remain vulnerable. As the methods of historical linguistics are based on the interaction of historical phonology and ­lexicology (etymology), finding unambiguous evidence for language contact outside the vocabulary means a serious methodological challenge for historical language contact research – and, so far, this challenge remains unanswered. Similarly to Indo‐European, the relationship between Uralic and Yukaghir, an indigenous language in Siberia, has triggered various speculations about possible relatedness versus very early loan contacts. As shown by Aikio (2014), the putative Uralic‐Yukaghir vocabulary shows no regular correspondences and therefore no evidence of relatedness, and a large part of seemingly common words may be due to chance resemblance, although there are some probable old borrowings as well.

3.2  Substrate Studies In historical contact linguistics, there seems to be a growing awareness of the complexity of language contact situations in pre‐historic Europe. These may well have involved extinct languages of unknown descent (cf. Schrijver 2001), also in the northernmost parts of Europe, where the linguistic map has radically changed due to the northward spread of both Uralic and Indo‐European languages. The role of a substrate (“Proto‐Lapp”) component in the genesis of the Saami languages has been a persistent question. The linguistic relatedness between Saami and Finnic is unmistak­ able, but the great differences in culture, identity, anthropology as well as certain vocabulary items and features of unknown origin in Saami have provoked diverse speculations about the Saami as a Palaeo‐Arctic or even Asiatic people who only secondarily adopted their language from their Finnic neighbors. Aikio (2004) has shown that – purged of its racist foundations in early Finnish nationalism – the hypothesis of a non‐Uralic substrate in Saami can be supported by a systematic analysis of potential substrate vocabulary. Elaborating on the criteria presented by Salmons (1992), he lists an impressive number of probable substrate words, identifiable on the basis of their semantics (animals and plants, nature, topography, and weather conditions of the North), structure (un‐Uralic phonotactics), and/or irregular sound correspondences bet­ ween different dialects (pointing at multiple, separate borrowings). In central and northern Russia, the expansion of Slavic only began at the end of the first millennium ce and led to the assimilation of many presumably Finno‐ Ugric language vari­ eties. The question of a Finno‐Ugric substrate in Russian has time and again been dealt with in linguistic literature, most notably by Veenker (1967). Many linguists have paid attention to features of Russian that deviate from other Slavic and Indo‐European languages and resemble certain Finno‐Ugric languages (such as the loss of the habeo verb in favor of the mihi est construction, or the abundance of unipersonal constructions, or, in general, “anti‐analytism” – Weiss 2004). However, the substrate interpretation often competes with other explanations (for instance, the mihi est construction does have IE roots as well; see also McAnallen 2011), and of the statements circulating in literature, at least the idea of the (Moksha) Mordvin origin of the Russian akanje (reduction of unstressed o) must be considered unfounded (Ravila 1973). In the last few years, the question of a Finno‐Ugric substrate in Russian has been taken up again, now concentrating on the toponymy of northern and central Russia. There are inter­ esting research results (see e.g. Saarikivi 2006; Nuorluoto 2006) which point at a complex

Contact and the Finno‐Ugric Languages  527 language situation in northern Russia before Slavicization, involving diverse interrelated and interacting Finnic, Para‐Finnic, Saamic, Para‐Saamic, or perhaps even Para‐Permic ­language varieties (as well as languages of unknown descent). As Helimski (2006) states, on the basis of recent research the family tree of the Finno‐Ugric languages could be partly re‐drawn, expanding the Finnic‐Saamic group to a Northwest Finno‐Ugric branch. Recently, Frog and Saarikivi (2015) have attempted to chart the locations of Finnic, Saami, Baltic, Slavic, Germanic and other speech communities in Northeast Europe around 1000 ce. As for Baltic, more recent studies do not support the once‐popular hypothesis of a Uralic substrate in Baltic or Balto‐Slavic (see Kallio 2005; Holst 2015). However, it is obvious that Latvian bears some more recent traces of the Finnic (Livonian and/or other varieties) lan­ guages once spoken in parts of present‐day Latvia (cf. Zeps 1962). There are Finnic topo­ nyms and also some loanwords, especially in the so‐called Livonian dialects of Latvia; it is also frequently stated that certain characteristic innovations in Latvian such as word‐initial stress (in contrast to the more archaic prosody of Lithuanian) are due to Finnic influence.

3.3  Competing Explanations – Multiple Causation? In etymological studies – the most intensively cultivated part of contact‐linguistic studies involving Finno‐Ugric – at the turn of the millennium a heated debate arose between Finnish etymologists, concerning competing etymologies for numerous Finnish lexemes and partly prompted by the preparation and publishing of the new etymological dictionary SSA (Suomen sanojen alkuperä: Etymologinen sanakirja [The origin of the Finnish words: an etymo­ logical dictionary]). This debate is also an indirect consequence of two different approaches in etymology: as loanword researchers often depart from a certain mechanism of sound substitution or from certain phono‐morphological or semantic features typical of loanwords, they typically end up discovering whole clusters of new loan etymologies – and risk ignoring the language‐internal mechanisms of lexical morphology and analogy. Editors of etymolog­ ical dictionaries, on the other hand, must deal with the word stock of a language as a whole and pay attention to the internal relations, irregular and/or secondary networks of word formation, analogy, and association within the lexicon – and, tantalized by the rich deriva­ tional morphology and diverse systems of forming ideophones, sound‐symbolic and expres­ sive vocabulary in Finnic (cf. Mikone 2001), they may lose sight of the possibility of borrowing. Actually, already earlier in the twentieth century Finnish etymological research started to focus on the internal mechanisms of explanation, which led to what Junttila (2015a: 20–7) has called ‘the era of disfavor for loanword research’. To give but one example: is Finnish puhdas ‘clean’ (Pre‐Finnic *puštas) an early IE loan (‘cleansed by sifting or winnowing’?, cf. Pre‐Germanic *powH‐eye/o‐, OHG fewen ‘to sift’), as stated by Koivulehto, or does it belong to the Finnic family of descriptive words for ‘blow, puff’ etc. (cf. puhu‐, puhalta‐ ‘to blow’ and also poh‐ta‐ ‘to sift, to winnow’), as suggested by Eino Koponen, who is also one of the authors of SSA? (Note that also the Finnish verb puhu‐ has lost its original expressive motivation and is now the neutral word for ‘to speak, to talk’; this ‘fading’ of sound symbolism plays a central role in Koponen’s model.) In his summary of the debate, Koivulehto (2001b) sharply criticizes Koponen’s technique of (allegedly) operating with monosyllabic “descriptive roots,” which allows for unsystematic and arbi­ trary vowel changes and “stem extensions.” However, Koivulehto’s expert criticism misses one important point: Koponen’s “root method” does not exclude the possibility of words being primarily loanwords which are only secondarily attached to a family of expressive words. In the same vein, Vesa Jarva (2001) has investigated the process of intertwining between loanwords and native expressive words or elements. In effect, this would mean introducing the idea of multiple causation into loanword research (cf. also Laakso 2001a).

528  Johanna Laakso Similar questions of native versus borrowed also arise in morphology, phonology, and syntax. In particular, the contacts between the westernmost Finno‐Ugric and Germanic (or Standard Average European) languages are often mentioned as the primary explanation for certain (morpho)syntactic “Europeanization” phenomena. As for Finnic, typical morpho­ syntactic examples recurring in literature are the development of a perfect tense with the auxiliary BE (reflecting the HAVE perfect in many SAE languages), the agreement of adjective modifiers (as in Finnish iso‐ssa talo‐ssa ‘big‐INESSIVE house‐INESSIVE’ ‘in a big house’; cf. Hungarian (a/egy) nagy ház‐ban ‘(the/a) big house‐INESSIVE’) or the word order change from SOV (in most Uralic languages) to SVO. For Hungarian, similar examples are the gram­ maticalization of definite and indefinite articles (definite a(z) from the demonstrative pro­ noun az, indefinite egy from the numeral egy ‘1’) and the debated verbal prefixes or “preverbs” used for adverbial, aspectual, or Aktionsart meanings and thus functionally resembling the verbal prefixes in German and Slavic (Kiefer and Honti 2003). In phonology, the most famous case is probably the radical simplification of the consonant system and the development of the consonant gradation (a morphophonological consonant alternation) in Finnic. In his often‐cited paper, Posti (1954) explained all these consonant changes with a Germanic super­ strate, assuming that Proto‐Finnic speakers would have imitated the prestigious accent of their Germanic neighbors. In all these cases, the similarity between possibly contact‐induced phenomena and their purported models in neighboring languages – between the Finnic BE perfect and the SAE HAVE perfect, for instance, or between the Finnic consonant gradation and Verner’s Law in Germanic – seems obvious to an outsider. Contact explanations for phonological or morpho­ syntactic phenomena can be connected with Sprachbund hypotheses or, in any case, with the amply attested lexical contacts, the main direction of loanwords often being from the neigh­ boring IE languages to Finno‐Ugric. However, many of these phenomena also have internal and/or general explanations. For example, the consonant gradation in Finnic‐Saami is also clearly connected with the archaic word architectonics best preserved in these languages (Helimski 1995), and the simplifying or reductive consonant changes in Pre‐Finnic (such as *š > h, *mt > nt) are “natural” and do not necessarily require any external explanations (Kallio 2000a). The Finnic BE perfect, employing the copula and a past participle of the main verb (‘he is gone’), can also be compared with past tense categories developed from past partici­ ples in other Finno‐Ugric languages, and the agreement of adjective modifiers could be regarded as an extension of the agreement of modifying pronouns (as in Hungarian ab‐ban a ház‐ban ‘in that house’). The “verbal prefixes” in Hungarian are not genuine verbal prefixes but separable preverbs with native etymologies and sometimes even functionally similar cognates in related languages. In a synthesis of critical responses to diverse contact hypotheses, Honti (2007) sharply criticizes contact explanations in Finno‐Ugric morphosyntax and ends by quoting Peter H. Nelde: ‘Language contact research has no methodology yet’ (Nelde 1992: 241). In Honti’s “insider” view, obviously, the internal explanation is always to be preferred, other things being equal  –  external influences may of course contribute to a greater frequency of a construction already marginally present in a language, or individual constructions may be calqued from another language. This is in line with the traditions of historical linguistics: (systematic) language change is internally motivated by default, and external explanations are only needed when all else fails. Honti is also certainly right in criticizing outsiders for jumping to conclusions: in typology and areal linguistics, unfounded or misinformed state­ ments about Finno‐Ugric languages are not rare. Part of the problem, however, seems to be that we have very little knowledge of what really happens in language contact and very little means of predicting the outcome, as it also depends on conscious actions and choices of language users. For these, the identification of similarities between contacting languages, “docking” (Laakso 2001b), may play a crucial

Contact and the Finno‐Ugric Languages  529 role: speakers choose words, elements, and structures that match both languages. Considering this, exploring multiple causation and interaction of various factors might be a promising avenue for further research.

3.4 Modernization, Globalization, and the Changing Character of Language Contact In the early twentieth century, Finno‐Ugric speakers in the three nation‐states were largely monolingual, and among many Finno‐Ugric minorities, at least a few older speakers could be found whose command of the majority language was very weak. (True, there are great differences between Finno‐Ugric minorities in this respect; in some Mari speaker commu­ nities, for instance, bilingualism in some Turkic language has been common for a long time already, and North Saami speakers a hundred years ago often had a good command of Norwegian, Swedish, and/or Finnish.) Traditional Finno‐Ugric studies could thus operate with idealized monolingual speaker communities, and field linguists chose their informants so as to represent as “pure” a language variety as possible. Today, the situation has changed radically. In the Finno‐Ugric nation‐states, the school system now aims at providing everybody with a practical command of at least one “world” language, and according to the Eurobarometer survey of 2012, 87% of the population of Estonia, 75% of the population of Finland, and 35% of the population of Hungary can speak at least one language other than their mother tongue at the level of being able to have a conversation (Special Eurobarometer 386). In Finland, younger generations are generally expected to know English, and the dominance of English in business, science, and entertain­ ment is already strong enough to make experts concerned about English interference in young people’s written Finnish or the status of the Finnish language in science and professional communication (see Leppänen et  al. 2011). Whether this kind of language contact, involving language use in an unprecedented diversity of new modalities, styles, and genres (“superdiversity”), is different from language contact situations traditionally researched in contact linguistics (in connection with minorities or migrant groups) remains to be investigated. For many Finno‐Ugric minorities, the school system (at least at higher levels) only exists in the majority language, the majority language dominates in most domains of language use outside home and family, and practically all speakers of today are bilingual. For these speakers, the grammar of the majority language is psychologically real and operates, for instance, within the abundant code‐switches. Gender assignment in Russian words (Finno‐ Ugric languages have no grammatical genders) or the inflection of Russian numerals (in colloquial speech, years and dates, for instance, are often inserted in Russian) are mastered by modern Finno‐Ugrians of Russia without difficulty. For quite a few younger speakers, even if they claim to be bilingual and identify themselves with the heritage language speaker community, the majority language might well be their “first,”: i.e. primary or “matrix” language. The endangerment and obsolescence of Finno‐Ugric minority languages have been sub­ ject to diverse sociolinguistically oriented studies, the most famous example probably being Susan Gal’s (1979) investigation on language shift among the Hungarian minority in Burgenland, Austria. These studies typically concentrate on the social conditions of lan­ guage use, language choices, and language shift or on attitudes of speakers toward the main­ tenance or revitalization of minority languages (cf. Huss 1999). There is much less research on what the Finno‐Ugric minority languages in their present‐day condition really are like. Although it is generally acknowledged that today’s Finno‐Ugric minority languages often clearly differ from the “classical” varieties, i.e. texts recorded from old and conservative

530  Johanna Laakso informants in the most fruitful period of Finno‐Ugric linguistic fieldwork before World War I, systematic comparisons between modern and “pure” language varieties seem to be lack­ ing. Language revitalization poses further challenges. While there are some studies on the language of new speakers (in particular, immigrants) in the Finno‐Ugric nation‐states, much less is known about the current state of revitalized languages such as Livonian (Ernštreits 2016) or Inari Saami (Olthuis et al. 2013), many speakers of which have only learned the lan­ guage in adult age. In any case, what is known about the current state of Finno‐Ugric minority languages indicates a wide spectrum of multilingualism, different degrees of language skills, and a diversity of attitudes influencing the choice of language. Among the Karelian informants of Sarhimaa (1999), there were terminal speakers with a restricted command of Karelian but also speakers who still could speak fluent “Traditional Karelian” with little or no code‐ switchings into Russian; at the same time, they had mixed varieties or codes at their disposal, employing Russian (Sarhimaa calls this code “Karussian”) or even Finnish elements. For modern bilingual speakers, alternating between codes can be a conscious “act of identity.” Conversely, modern speakers may choose not to mix codes. In her study of the language of two Hungarian‐speaking families in Burgenland, Dávid (2008) notes that her informants, despite some signs of insecurity in their command of Hungarian and despite abundant calques and syntactic interference from German, hardly ever switched into German in the interview situation. During the 40 years since Gal’s study, there have been changes in the status of Hungarian in Burgenland, including the introduction of Hungarian into the curricula of some primary and secondary schools. Standardization together with the increasing use of minority lan­ guages in education and media adds a new, important dimension to the traditional research of Finno‐Ugric language contacts. Expatriate varieties of Finnish and Hungarian must choose between creating a standard of their own (like the old autochthonous Finnish minority in north Sweden, now developing their own Meänkieli) and using the homeland standard (as the Hungarian varieties spoken in the neighboring countries of Hungary, or the numerous post‐World War II Finnish emigrants in Sweden do). In Hungarian‐speaking areas especially, there is an increasing tension between the puristic tradition of language planning in the homeland and the reality of multilingualism in expa­ triate speaker communities. In the 1990s, this tension brought about a debate on “linguistic treason vs. rescue of language” (Kontra and Saly 1998; summarized e.g. by Maráz 2006; see also Fenyvesi 2005). To support a more pluricentric idea of the Hungarian language, a “de‐ trianonization” project has been launched: collecting words used in expatriate Hungarian varieties and including them in new dictionaries of Standard Hungarian. Another development increasing pluralism concerns Finnish and Estonian: as the large immigrant Russian minority in Estonia and the relatively small but very rapidly growing immigrant communities in Finland must be integrated, there are probably more people now than ever before learning Finnish or Estonian as a foreign language. This could mean that Finnish and Estonian are gradually losing their character as ethnic in‐group languages and that native speakers will have to develop a greater tolerance towards their language as used by non‐natives.

4 Conclusion The Uralic/Finno‐Ugric language family covers a large geographic area with considerable typological and language‐sociological diversity. It is relatively well researched, and although for most of its languages, proper documentation only begins in the nineteenth century, there is a strong comparative‐historical tradition of reconstructions covering several millennia of

Contact and the Finno‐Ugric Languages  531 language development. In this tradition, contacts with better documented neighboring lan­ guages have played a central role. In particular, loanwords adopted into Uralic from various Indo‐European and Turkic languages, and even to some extent other languages have been extensively studied. These investigations have raised interesting questions about the recon­ struction of language contact, the assessment of native vs borrowed elements, and about multiple causation, i.e. the intertwining of internal and external factors in language change. Loanword research provides practically the only sources to the prehistory of the Hungarians before the ninth century ce. Similarly, recent studies on possible Paleo‐European substrates in Saami as well as Finno‐Ugric substrates in the toponymy of northern and central Russia shed light on the ethnolingual prehistory of these areas. Today’s Finno‐Ugric languages, with the exception of Finnish, Estonian, and Hungarian in their respective states, are all minority languages. Practically all speakers today are bilingual, and their language is characterized by influences from the majority language at all levels. Even the three Finno‐Ugric state languages have experienced periods of political and cultural dominance of other languages (in particular, Swedish in Finland and German in Estonia). In connection with the ethnic emancipation prompted by Romantic Nationalism from the nineteenth century onwards, this was countered by puristic language planning which was explicitly aimed at perceived foreign influences. At present the situation is chang­ ing again, as globalization (in particular, the presence of global forms of English) and new forms of international mobility and media are increasingly challenging the traditional mono­ lingual nation‐state models.

REFERENCES Abondolo, Daniel 1998a. Introduction. In Daniel Abondolo (ed.). The Uralic Languages. London: Routledge, pp. 1–42. Aikio, Ánte 2004. An essay on substrate studies and the origin of Saami. In Irma Hyvärinen, Petri Kallio, and Jarmo Korhonen (eds.), Etymologie, Entlehnungen und Entwicklungen: Festschrift für Jorma Koivulehto zum 70. Geburtstag (Mémoires de la Société Néophilologique de Helsinki 63). Helsinki: Société Néophilologique, pp. 5–34. Aikio, Ánte 2007. Etymological nativization of loanwords: a case study of Saami and Finnish. In Ida Toivonen and Diane Nelson (eds.), Saami Linguistics. Amsterdam: Benjamins, pp. 17–51. Aikio, Ánte (Luobbal Sámmol Sámmol Ánte) 2012. An essay on Saami ethnolinguistic prehistory. In Riho Grünthal and Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe (Mémoires de la Société Finno‐Ougrienne 266). Helsinki: Finno‐Ugrian Society, pp. 63–117. Aikio, Ánte (Luobbal Sámmol Sámmol Ánte) 2014. The Uralic‐Yukaghir lexical correspondences: genetic inheritance, language contact, or chance resemblance? Finnisch‐ Ugrische Forschungen 62: 7–76.

Anttila, Raimo 2000. The Indo‐European and the Baltic‐Finnic interface: time against the ice. In Colin Renfrew, April McMahon, and Larry Trask (eds.), Time Depth in Historical Linguistics, vol. 2. Cambridge: The McDonald Institute for Archaeological Research, pp. 481–528. Bentlin, Mikko 2008. Niederdeutsche Lehnwörter im Finnischen (Universität Greifswald Ph.D. dissertation). Bereczki, Gábor 2007. Der Sprachbund des Wolga‐ Kama‐Gebietes. Incontri Linguistici 30: 11–28. Blokland, Rogier 2009. The Russian Loanwords in Literary Estonian. (Veröffentlichungen der Societas Uralo‐Altaica 78.) Wiesbaden: Harrassowitz. Borin, Lars and Katri Vuorela 1998. Finnish Romani. In Ailbhe Ó Corráin and Séamus MacMathúna (eds.), Minority Languages in Scandinavia, Britain and Ireland (Studia Celtica Upsaliensia 3). Uppsala: Acta Universitatis Upsaliensis, pp. 51–76. Csepregi, Márta 1997. Russische Einflüsse in der südostjakischen Syntax. In Sirkka‐Liisa Hahmo, Tette Hofstra, László Honti, Paul van Linde, and Osmo Nikkilä (eds.), Finnisch‐ ugrische Sprachen in Kontakt: Vorträge des Symposiums aus Anlaß des 30‐jährigen Bestehens

532  Johanna Laakso der Finnougristik an der Rijksuniversiteit Groningen 21.–23. November 1996. Maastricht: Shaker, pp. 71–7. Dávid, Ágnes 2008. Ungarisch in Österreich: eine neue Sprechergeneration des Ungarischen. In Johanna Laakso (ed.), Ungarischunterricht in Österreich: Probleme und Vergleichspunkte [Teaching Hungarian in Austria: problems and points of comparison (Finno‐Ugrian Studies in Austria 6). Vienna: LIT Verlag, pp. 154–78. de Smit, Merlijn 2006. Language Contact and Structural Change: An Old Finnish Case Study (Studia Fennica Stockholmiensia 9). Stockholm: Stockholms universitet. Décsy, Gyula 1988. Slawischer Einfluss auf die uralischen Sprachen. In Denis Sinor (ed.), The Uralic Languages: Description, History, and Foreign Influences. Leiden: Brill, pp. 618–37. Ernštreits, Valts 2016. Livonian in recent years. Journal of Estonian and Finno‐Ugric Language Studies 7.1: 257–74. Fenyvesi, Anna (ed.) 2005. Hungarian Language Contact outside Hungary: Studies on Hungarian as a Minority Language. Amsterdam: Benjamins. Frog and Janne Saarikivi 2015. De situ linguarum fennicarum aetatis ferreae, Pars I. [The situation of the Finnic languages of the Iron Age. Part I] The Retrospective Methods Network Newsletter 9: 64–115. Gal, Susan 1979. Language Shift: Social Determinants of Linguistic Change in Bilingual Austria. New York: Academic Press. Gregor, Ferenc 1994. Einige Fragen der ungarischen Elemente im Slowakischen. Wiener Slawistischer Almanach 34: 239–59. Grünthal, Riho 2015. Livonian at the crossroads of language contacts. In Santeri Junttila (ed.), Contacts between the Baltic and Finnic Languages (Uralica Helsingiensia 7). Helsinki: Finno‐ Ugrian Society, pp. 97–150. Hausenberg, Anu‐Reet 1998. Language affinity and language contacts in the history of settlement of Northeastern Europe. In Jussi Niemi, Terence Odlin, and Janne Heikkinen (eds.), Language Contact, Variation, and Change (Studies in Languages 32). Joensuu: University of Joensuu, pp. 111–24. Helimski, Eugen 1995. Proto‐Uralic gradation: continuation and traces. In Heikki Leskinen (ed.), Congressus Octavus Internationalis Fenno‐ Ugristarum: Orationes Plenariae et Conspectus Quinquennales, Pars I. Jyväskylä: Kumerus Kirjapaino, pp. 17–51. Helimski, Eugen 2001. Early Indo‐Uralic linguistic relationships: real kinship and

imagined contacts. In Christian Carpelan, Asko Parpola, and Petteri Koskikallio (eds.), Early Contacts between Uralic and Indo‐European: Linguistic and Archaeological Considerations (Mémoires de la Société Finno‐Ougrienne 242). Helsinki: Société Finno‐Ougrienne, pp. 187–205. Helimski, Eugen 2003. Areal groupings (Sprachbünde) within and across the borders of the Uralic language family: a survey. Nyelvtudományi Közlemények 100: 156–67. Available at www.nytud.hu/nyk/100/ helimski.pdf Helimski, Eugen 2006. The “Northwestern” group of Finno‐Ugric languages and its heritage in the place names and substratum vocabulary of the Russian North. In Juhani Nuorluoto (ed.), The Slavicization of the Russian North: Mechanisms and Chronology (Slavica Helsingiensia 27). Helsinki: University of Helsinki, pp. 109–27. Available at www.slav. helsinki.fi/nwrussia/eng/Conference/papers. htm Hinderling, Robert 1981. Die deutsch‐estnischen Lehnwortbeziehungen im Rahmen einer europäischen Lehnwortgeographie. Wiesbaden: Harrassowitz. Holst, Jan Henrik 2015. On the theory of a Uralic substratum in Baltic. In Santeri Junttila (ed.), Contacts between the Baltic and Finnic Languages (Uralica Helsingiensia 7). Helsinki: Finno‐Ugrian Society, pp. 151–73. Honti, László 1998. Ob‐Ugrian. In Daniel Abondolo (ed.), The Uralic Languages. London: Routledge, pp. 327–57. Honti, László 2007. Syntaktische Konstruktionen fremden Ursprungs: eine skeptische Stellungnahme. Incontri Linguistici 30: 49–70. Huss, Leena 1999: Reversing Language Shift in the Far North: Linguistic Revitalization in Northern Scandinavia and Finland (Studia Uralica Upsaliensia 31). Uppsala: Uppsala University. Hyllested, Adam 2014. Word exchange at the gates of Europe: five millennia of language contact (University of Copenhagen. Ph.D. dissertation). Jahr, Ernst Håkon 1996. On the pidgin status of Russenorsk. In Ernst Håkon Jahr and Ingvild Broch (eds.), Language Contact in the Arctic: Northern Pidgins and Contact Languages (Trends in Linguistics: Studies and Monographs 88). Berlin: de Gruyter Mouton, pp. 107–22. Janhunen, Juha 1999. Critical remarks on current Indo‐Uralic comparisons. In Cornelius Hasselblatt and Paula Jääsalmi‐Krüger (eds.),

Contact and the Finno‐Ugric Languages  533 Europa et Sibiria: Beiträge zu Sprache und Kultur der kleineren finnougrischen, samojedischen und paläosibirischen Völker. Gedenkband für Wolfgang Veenker (Veröffentlichungen der Societas Uralo‐Altaica 51). Wiesbaden: Harrassowitz, pp. 211–15. Jarva, Vesa 2001. Some expressive and borrowed elements in the lexicon of Finnish dialects. In F. K. Erhard Voeltz and Christa Kilian‐Hatz (eds.), Ideophones (Typological Studies in Language 44). Amsterdam: Benjamins, pp. 111–19. Junttila, Santeri 2012. The prehistoric context of the oldest contacts between Baltic and Finnic languages. In Riho Grünthal and Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe (Mémoires de la Société Finno‐ Ougrienne 266). Helsinki: Finno‐Ugrian Society, pp. 261–96. Junttila, Santeri 2015a. Tiedon kumuloituminen ja trendit lainasanatutkimuksessa: kantasuomen balttilaislainojen tutkimushistoria [The accumulation of information and tendencies in the study of loanwords: the research history of Baltic loans in Proto‐Finnic] (University of Helsinki Ph.D. dissertation). Junttila, Santeri 2015b. Proto‐Finnic loanwords in the Baltic languages? An old hypothesis revisited. In Santeri Junttila (ed.), Contacts between the Baltic and Finnic Languages (Uralica Helsingiensia 7). Helsinki: Finno‐Ugrian Society, pp. 12–37. Kallio, Petri 2000a. Posti’s superstrate theory at the threshold of a new millennium. In Johanna Laakso (ed.), Facing Finnic: Some Challenges to Historical and Contact Linguistics (Castrenianumin toimitteita 59). Helsinki: Finno‐Ugrian Society. Kallio, Petri 2000b. Germanic “maggot” and “moth.” In Dirk Boutkan and Arend Quak (eds.), Language Contact: Substratum, Superstratum, Adstratum in Germanic Languages (Amsterdamer Beiträge zur älteren Germanistik 54). Amsterdam/Atlanta: Rodopi, pp. 117–22. Kallio, Petri 2002. Prehistoric contacts between Indo‐European and Uralic. In Karlene Jones‐Bley, Martin E. Huld, Angela Della Volpe, and Miriam Robbins Dexter (eds.), Proceedings of the Thirteenth Annual UCLA Indo‐European Conference: Los Angeles, November 9–10, 2001 (Journal of Indo‐ European Studies Monograph 44). Washington, DC: Institute for the Study of Man. Kallio, Petri 2005. A Uralic substrate in Balto‐ Slavic revisited. In Gerhard Meiser and Olav

Hackstein (eds.), Sprachkontakt und Sprachwandel: Akten der XI. Fachtagung der Indogermanischen Gesellschaft. Wiesbaden: Reichert, pp. 275–83. Kallio, Petri 2012. The prehistoric Germanic loanword strata in Finnic. In Riho Grünthal and Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe (Mémoires de la Société Finno‐Ougrienne 266). Helsinki: Finno‐Ugrian Society, pp. 225–38. Kallio, Petri and Johanna Laakso forthcoming. Uralic and Slavic. In Marc L. Greenberg and René M. Genis (eds.), Encyclopedia of Slavic Languages and Linguistics. Leiden: Brill. Kiefer, Ferenc and László Honti 2003. Verbal “prefixation” in Uralic languages. Acta Linguistica Hungarica 50.1–2: 137–53. Kniezsa, István 1955. A magyar nyelv szláv jövevényszavai [The Slavic loanwords in Hungarian]. Budapest: Akadémiai Kiadó. Koivulehto, Jorma 1999. Verba mutuata (Mémoires de la Société Finno‐Ougrienne 237). Helsinki: Suomalais‐Ugrilainen Seura/Société Finno‐Ougrienne. Koivulehto, Jorma 2001a. The earliest contacts between Indo‐Europeans and Uralic speakers in the light of lexical loans. In Christian Carpelan, Asko Parpola, and Petteri Koskikallio (eds.), Early Contacts between Uralic and Indo‐ European: Linguistic and Archaeological Considerations (Mémoires de la Société Finno‐Ougrienne 242). Helsinki: Société Finno‐Ougrienne, pp. 235–63. Koivulehto, Jorma 2001b. Etymologie und Lehnwortforschung: Ein Überblick um 2000. Finnisch‐Ugrische Forschungen 56: 42–78. Kontra, Miklós and Noémi Saly (eds.) 1998. Nyelvmentés vagy nyelvárulás? Vita a határon túli magyar nyelvhasználatról [Language rescue or language treason? The debate on the use of the Hungarian language outside Hungary]. Budapest: Osiris. Korenchy, Éva 1988. Iranischer Einfluss in den finnisch‐ugrischen Sprachen. In Denis Sinor (ed.), The Uralic Languages: Description, History, and Foreign Influences. Leiden: Brill, pp. 665–81. Laakso, Johanna 2001a. Native or borrowed, or both – is it possible to have many mothers? Word 52.2: 197–212. Laakso, Johanna 2001b. Reflections on the verb suffix ‐om in Russenorsk and some preliminary remarks on “docking” in language contact. In Sándor Maticsák, Gábor Zaicz, and Tuomo Lahdelma (eds.), Ünnepi könyv Keresztes László tiszteletér (Folia Uralica Debreceniensia 8).

534  Johanna Laakso Debrecen: Debreceni Egyetem, Finnugor Nyelvtudományi Tanszék, pp. 315–24. Lehiste, Ilse 1965. A poem in Halbdeutsch and some questions concerning substratum. Word 21: 55–69. Leppänen, Sirpa, Anne Pitkänen‐Huhta, Tarja Nikula, Samu Kytölä, Timo Törmäkangas, Kari Nissinen, Leila Kääntä, Tiina Räisänen, Mikko Laitinen, Päivi Pahta, Heidi Koskela, Salla Lähdesmäki, and Henna Jousmäki 2011. National Survey on the English Language in Finland: Uses, Meanings and Attitudes (Varieng: Studies in Variation, Contacts and Change in English 5). Available at www.helsinki.fi/ varieng/series/volumes/05/ Leinonen, Marja 2002. Influence of Russian on the syntax of Komi. Finnisch‐Ugrische Forschungen 57: 195–358. Lyublyanovics, Kyra 2011. The Cumans in medieval Hungary and the question of ethnicity. Annual of Medieval Studies at CEU 17: 153–169. Maráz, Gabriella 2006. Sprachrettung oder Sprachverrat: Zur Diskussion über die Norm des Ungarischen. Trans: Internet‐Zeitschrift für Kulturwissenschaften. Available at www.inst.at/ trans/15Nr/06_1/maraz15.htm McAnallen, Julia 2011. The history of predicative possession in Slavic: internal development vs. language contact (University of California, Berkeley Ph.D. dissertation). Mikone, Eve 2001. Ideophones in the Balto‐Finnic languages. In F. K. Erhard Voeltz and Christa Kilian‐Hatz (eds.), Ideophones (Typological Studies in Language 44). Amsterdam: Benjamins, pp. 223–33. Must, Mari 2000. Vene laensõnad eesti murretes [Russian loanwords in Estonian dialects]. Tallinn: Eesti Keele Sihtasutus. Nelde, Peter H. 1992. Mehrsprachigkeit und Kontaktlinguistik. In Werner Roggausch (ed.), Dokumentation der Tagungsbeiträge / Germanistentreffen Belgien – Niederlande – Luxemburg – Deutschland 29.9.–3.10.1992 (DAAD‐Dokumentationen and ‐Materialien 21). Bonn: DAAD, pp. 231–47. Nuorluoto, Juhani (ed.) 2006. The Slavicization of the Russian North: Mechanisms and Chronology (Slavica Helsingiensia 27). Helsinki: University of Helsinki. Olthuis, Marja‐Liisa, Suvi Kivelä, and Tove Skutnabb‐Kangas 2013. Revitalising Indigenous Languages: How to Recreate a Lost Generation. Clevedon: Multilingual Matters. Östman, Jan‐Ola 2006. Finlandssvenskan under dubbelpress [Finland Swedish under double

pressure]. Paper presented at the conference “Nordiske språkklima under engelsk press,” Denmark 2005. Available at www.norden.org/ sprak/forside/6.%20finlandssvensk.pdf. Plöger, Angela 1973. Die russischen Lehnwörter der finnischen Schriftsprache (Veröffentlichungen der Societas Uralo‐Altaica 8). Wiesbaden: Harrassowitz. Posti, Lauri 1954. From Pre‐Finnic to Late Proto‐Finnic. Finnisch‐Ugrische Forschungen 31: 1–91. Raag, Raimo 1987. Mittelniederdeutsche und skandinavische Lehnwörter im Estnischen und Livischen. In P. Sture Ureland (ed.), Sprachkontakt in der Hanse: Aspekte des Sprachausgleichs im Ostsee‐ und Nordseeraum. Akten des 7. internationalen Symposions über Sprachkontakt in Europa, Lübeck 1986. Tübingen: Max Niemeyer, pp. 317–46. Raag, Raimo 1997. Criteria for establishing lexical borrowings in Estonian. In Sirkka‐Liisa Hahmo, Tette Hofstra, László Honti, Paul van Linde, and Osmo Nikkilä (eds.), Finnisch‐ ugrische Sprachen in Kontakt: Vorträge des Symposiums aus Anlaß des 30‐jährigen Bestehens der Finnougristik an der Rijksuniversiteit Groningen 21.–23. November 1996. Maastricht: Shaker, pp. 185–90. Rätsep, Huno 1983. Eesti kirjakeele tüvevara päritolu [The origins of the word stems in Estonian]. Keel ja Kirjandus 10: 539–48. Ravila, Paavo 1973. Zur Kritik der mordwinischen Substrathypothese des russischen Akanje. In Commentationes Fenno‐ Ugricae in honorem Erkki Itkonen sexagenarii (Mémoires de la Société Finno‐Ougrienne 150). Helsinki: Suomalais‐Ugrilainen Seura, pp. 296–310. Rédei, Károly 1962. Jurák‐szamojéd jövevényszók a zürjén nyelvben [Nenets loanwords in Komi]. Nyelvtudományi Közlemények 64: 71–95. Rédei, Károly 1964a. Obi‐ugor jövevényszók a zürjén nyelvben [Ob‐Ugric loanwords in Komi]. Nyelvtudományi Közlemények 66: 3–15. Rédei, Károly 1964b. Vannak‐e az előmagyar‐ permi érintkezésnek nyelvi nyomai? [Are there linguistic traces of contacts between Pre‐ Hungarian and Permic?] Nyelvtudományi Közlemények 66: 253–61. Rédei, Károly 1970. Die syrjänischen Lehnwörter im Wogulischen. Budapest: Akadémiai Kiadó. Rédei, Károly 2002. Etymologie nach den großen Wörterbüchern. In Rogier Blokland and Cornelius Hasselblatt (eds.), Finno‐Ugrians and Indo‐Europeans: Linguistic and Literary Contacts.

Contact and the Finno‐Ugric Languages  535 Proceedings of the Symposium at the University of Groningen, November 22–24, 2001 (Studia Fenno‐Ugrica Groningana 2). Maastricht: Shaker, pp. 303–18. Róna‐Tas, András and Árpád Berta. 2011. West Old Turkic – Turkic loanwords in Hungarian. Wiesbaden: Harrassowitz. Saari, Mirja 2000. Schwedisch als die zweite Nationalsprache Finnlands: Soziolinguis­ tische Aspekte. Linguistik Online 7: 3/00. Available at www.linguistik‐online.de/ 3_00/saari.html Saarikivi, Janne 2006. Substrata Uralica: Studies on Finno‐Ugrian Substrate in Northern Russian Dialects. Tartu: Tartu Ülikooli Kirjastus. Available at http://ethesis.helsinki.fi/ julkaisut/hum/suoma/vk/saarikivi Salmons, Joseph C. 1992. Northwest Indo‐ European vocabulary and substrate phonology. In Roger Pearson (ed.), Perspectives on Indo‐European Language, Culture and Religion: Studies in Honor of Edgar C. Polomé, vol. 2 (Journal of Indo‐ European Studies Monograph Series 9). Washington, DC: Insititute for the Study of Man, pp. 265–79. Sammallahti, Pekka 1988. Historical phonology of the Uralic languages with special reference to Samoyed, Ugric, and Permic. In Denis Sinor (ed.), The Uralic Languages: Description, History, and Foreign Influences. Leiden: Brill, pp. 478–554. Sarhimaa, Anneli 1992. Karelian Sprachbund? Theoretical basis of the study of Russian/Baltic Finnic contacts. Finnisch‐Ugrische Forschungen 50: 209–19. Sarhimaa, Anneli 1999. Syntactic Transfer, Contact‐ induced Change, and the Evolution of Bilingual Mixed Codes: Focus on Karelian–Russian Language Alternation (Studia Fennica Linguistica 9). Helsinki: Suomalaisen Kirjallisuuden Seura. Schrijver, Peter 2001. Lost languages in Northern Europe. In Christian Carpelan, Asko Parpola, and Petteri Koskikallio (eds.), Early Contacts between Uralic and Indo‐European: Linguistic and Archaeological Considerations (Mémoires de la Société Finno‐Ougrienne 242). Helsinki: Société Finno‐Ougrienne, pp. 417–25. Special Eurobarometer 386: Europeans and their Languages. Report. Available at http://ec. europa.eu/commfrontoffice/publicopinion/ archives/ebs/ebs_386_en.pdf SSA Suomen sanojen alkuperä: etymologinen sanakirja [The origin of Finnish words: an etymological dictionary], vols. 1–3. (Erkki

Itkonen (1) and Ulla‐Maija Kulonen (1–3), ed.). Helsinki: Suomalaisen Kirjallisuuden Seura, 1992–2000. Thomason, Sarah G. 2001. Language Contact. Washington, DC: Georgetown University Press. Thomsen, Vilhelm 1869. Den gotiske sprogklasses indflydelse på den finske. En sproghistorisk undersøgelse. Copenhagen: Den Gyldendalske boghandel. Thomsen, Vilhelm 1870. Über den einfluss der germanischen sprachen auf die finnisch‐lappischen: eine sprachgeschichtliche untersuchung. Halle: Verlag der Buchhandlung des Waisenhauses. Thomsen, Vilhelm 1890. Beröringer mellem de finske og de baltiske (litauisk‐lettiske) sprog: en sproghistorisk undersøgelse. Copenhagen: Blanco Lunos. Tuisk, Tuuli 2014. Survey of research on Livonian prosody. Journal of Estonian and Finno‐Ugric Linguistics 5.1: 261–92. Vaba, Lembit 1977. Läti laensõnad eesti keeles [Latvian loanwords in Estonian]. Tallinn: Valgus. Veenker, Wolfgang 1967. Die Frage des finnougrischen Substrats in der russischen Sprache (Indiana University Publications: Uralic and Altaic Series 82). Bloomington: Indiana University. Verschik, Anna 2010. Contacts of Russian in the post‐Soviet space. Applied Linguistics Review 2010: 85–127. Weiss, Daniel 2004. Das moderne Russisch als antianalytische Sprache. In Uwe Hinrichs (ed.), Die europäischen Sprachen auf dem Wege zum analytischen Sprachtyp (Eurolinguistische Arbeiten 1). Wiesbaden: Harrassowitz, pp. 263–83. Wichmann, Yrjö 1902. Samojedisches lehngut im syrjänischen. Finnisch‐Ugrische Forschungen 2: 165–83. Winkler, Eberhard 2014. Loanword strata in Livonian. Journal of Estonian and Finno‐Ugric Linguistics 5.1: 215–27. Wurm, Stephen A. 1996. The Taimyr Peninsula Russian‐based pidgin. In Ernst Håkon Jahr and Ingvild Broch (eds.), Language Contact in the Arctic: Northern Pidgins and Contact Languages (Trends in Linguistics: Studies and Monographs 88). Berlin/New York: de Gruyter Mouton, pp. 79–90. Zeps, Valdis J. 1962. Latvian and Finnic Linguistic Convergences (Indiana University Publications: Uralic and Altaic Series 9). Bloomington: Indiana University.

27 Language Contact in the Balkans BRIAN D. JOSEPH 1 The Languages and Their Convergent Character: Introducing the “Sprachbund” Southeastern Europe is the home of an intense contact zone that takes in a number of distantly related languages and then some unrelated languages (see p. 538). The Balkans, as this area is generally known as, is the most thoroughly studied contact area in the world, and thus occupies a special place in contact linguistics as one of the most important regions for understanding the mechanisms and results of language contact. A rough and mountainous region that forms a peninsula bounded by the Adriatic and Ionian Seas on the west and the Black and Aegean Seas on the east and south, respectively, the Balkans have constituted a crossroads for speakers of many different languages since at least the second millennium bce. The interrelations among speakers in the Balkans in ancient times are of considerable interest since clearly various sorts of cross‐language transfer showing the effects of language shift (substrata) and borrowing must have occurred. These effects are especially evident in the lexicon – for example, there appears to be a layer of Indo‐European but non‐Greek words in Ancient Greek (e.g. aleiphō ‘rub, anoint’ where the #a‐ and the ‐ph‐ are unexpected and the ‐lei‐ derives from Indo‐European *li‐p‐, seen in genuine Greek forms such as lip‐os ‘fat’) – but other sorts of effects involving various ancient languages could be (and have been) imagined. The languages in question include the following: Continental Celtic (in some form), Dacian (Daco‐Mysian), Gothic, Greek, Illyrian, Latin, Macedonian, Phrygian, Pelasgian (“Pre‐Greek”), and Thracian. But these pre‐historic contacts present a number of challenges to language historians, as some of these languages, quite frustratingly for scholars, are only very sparsely attested or known just from brief mention in ancient sources. And some may not even be identifiable as individual languages.1 Thus much about their interactions must remain speculative.2 For all the intrinsic allure of the study of the ancient pre‐historic situation and the speculations about contact that these languages offer, it is the more modern situation, dating from about 1000 ce, that has attracted the most attention among researchers in contact linguistics. The range of languages relevant in the Balkans in this more recent period extends over a number of branches of Indo‐European and even beyond that family as well. They include the following (excluding languages, such as Tagalog or Arabic in Greece, that have entered the

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

538  Brian D. Joseph Balkans quite recently as the result of employment‐ or war‐related modern immigration trends), listed with some explanatory annotations where deemed appropriate: (1)

Albanian (both major dialects: Geg (north) and Tosk (south)) Armenian (spoken in Bulgaria) Circassian (Adygey variety; spoken in Kosovo) Bulgarian German (spoken in Romania) Greek (including the very divergent dialects like Tsakonian and Pontic, the latter only in the Balkans proper via the postwar population exchanges of the 1920s) Hungarian (spoken in Romania) Italian (spoken in Istria area of Croatia) Judezmo (also known as Ladino or Judeo‐Spanish) Macedonian (the Slavic language, thus different from Ancient Macedonian, mentioned on p. 537) Romani (the Balkan variety of the Indic language of the Roms) Romanian (more accurately listed as four separate languages: Aromanian (Vlach), Istro‐Romanian, Megleno‐Romanian, and Romanian (the national language of Romania (and Moldova))) Ruthenian (also known as Rusyn, spoken in Vojvodina area of Serbia) “Serbo‐Croatian” (now, after the break‐up of Yugoslavia, generally considered to be four separate (“successor”) languages: Bosnian, Croatian, Montenegrin, and Serbian)3 Slovak (in a small enclave in Vojvodina area of Serbia) Slovenian Turkish

Depending in part on just how the Balkans are defined geographically, especially as to the northern border,4 these languages may be called “languages of the Balkans,” a purely geographic designation.5 A further distinction, important for contact linguistics, needs to be made here. In particular, one needs to recognize further a class of “Balkan languages,” referring to those languages of the Balkans that show considerable structural and lexical convergence due to centuries of intense, intimate, and sustained contact involving multilaterally multilingual speakers. The effect is what one of the first commentators on Balkan convergence, Kopitar (1829: 86), described as an area in which “nur eine Sprachform herrscht, aber mit dreyerley Sprachmaterie” (“only one language‐form dominates but with threefold language material”). Under this more restricted designation, the following languages can be included as “Balkan languages,” listed again with appropriate annotation as needed for clarification concerning the extent of convergence shown: (2)

Albanian Aromanian Bulgarian Greek (most dialects, including Tsakonian (but excluding Asia Minor dialects)) Judezmo (mostly only at the phonological and lexical levels) Macedonian Megleno‐Romanian6 Romani Romanian Serbian (with Torlak dialects of southeast Serbian being most relevant, much less so both the Croatian and Bosnian standards) Turkish (not a “full” structural participant but crucial nonetheless).

Language Contact in the Balkans  539 These Balkan languages represent several different genetic affiliations: Albanian is its own branch within Indo‐European, as is Greek; Bulgarian, Macedonian, and (Bosnian‐Croatian‐ Montenegrin‐)Serbian are all South Slavic languages, again within Indo‐European; the Romanian group and Judezmo belong to the Romance languages (of the Italic branch of Indo‐European); Romani belongs to the Indic branch of the Indo‐Iranian subgroup of Indo‐ European; and Turkish is part of the Turkic language grouping, generally believed to be part of a larger Altaic family. It is convenient to refer to these languages more generally as Balkan Albanian, Balkan Greek, Balkan Romance, Balkan Romani, and Balkan Turkish, to distinguish the Balkan varieties of these languages or language groups from their relatives outside the Balkans, inasmuch as the non‐Balkan varieties generally do not show the structural and lexical properties that their Balkan relatives do or show them to a lesser degree. These languages offer some diversity to be sure, but more significantly, also cross‐ language similarities that have led to the recognition of a key construct for language contact studies – the “sprachbund” (French union linguistique, Russian jazykovoj sojuz, English linguistic league or linguistic area or even just sprachbund) – that has come to be applied to geographically based convergence zones, including South Asia (Emeneau 1956; Masica 1976), Meso‐America (Campbell, Kaufman and Smith‐Stark 1986), and the Pacific Northwest of the United States and Canada (Beck 2000), to name just a few. A sprachbund can be defined as any group of languages that due to intense and sustained multilingual contact share linguistic features, largely structural in nature but possibly lexical as well, that are not the result of shared inheritance from a common ancestor nor a matter of independent innovation in each of the languages involved.7 Taking note of, and ruling out, common inheritance is an important part of recognizing a sprachbund, inasmuch as this concept, when first formulated by Trubetzkoy in 1923 (with his 1928 pronouncement at the First International Congress of Linguists being the better known and more widely cited source), was explicitly contrasted with a Sprachfamilie (‘language family’), both being types of Sprachgruppen (‘language group(ing)s’). In a language family, the languages are genetically related (in the technical linguistic sense of deriving by a direct lineal descent from a common source). Trubetzkoy’s new concept, by contrast, was for languages that are geographically related, being located in the same region and often co‐existing side by side in the same territory, but are not genetically related, and which yet, due to prolonged contact, show resemblances in form and structure. For Trubetzkoy, the Balkans were the prime example of this new type of grouping. Others before him had made similar observations about the Balkans, though without generalizing to a new construct, let alone offering a handy label for it, or contrasting it clearly with genetic groupings of languages: Kopitar (1829), as noted on p. 538, drew attention to a single feature, the postposed definite article, Miklosich (1861) noted several convergent features involving Balkan languages, among which he included Modern Greek, and Sandfeld (1926), more widely known through the French translation of 1930, elaborated in a systematic way on a large number of such features in phonology, morphology, syntax, and lexicon, with the majority being lexical and phraseological in nature.

2  The Convergent Features Themselves: Balkanisms It is appropriate at this point to flesh out the somewhat abstract references to structural and lexical convergences with some concrete details. While there are several convergent features – which may be called “Balkanisms” – that have attracted considerable attention in the rather large literature on Balkan linguistics in the period since Sandfeld, appearing for instance in most of the handbooks (Schaller 1975; Banfi 1985; Feuillet 1986, 2012; Asenova 1989/2002; Demiraj 1994/2004; and Steinke and Vraciu 1999) but also in specialized studies

540  Brian D. Joseph (e.g. Joseph 1983; Friedman 2003), largely because they are widely realized in the Balkan ­languages, there are actually dozens of features that link small clusters of languages and dialects in the Balkans (see Friedman and Joseph 2021 for details and see also the discussion on p. 546). The following is a representative list of those that are widespread, and a sampling of those that are more localized, thus overall giving a feel for the most significant relevant features and types of features shared by various of these languages; they cover phonology (a–f), morphology (g–j), syntax (k–p), and lexicon (q–r): (3)

a. the presence of a (stressed) mid‐to‐high central vowel; this feature is found in Albanian, Romanian, Bulgarian, some dialects of Macedonian and Bosnian‐ Croatian‐Montenegrin‐Serbian, some Romani dialects, and Turkish; b. the presence of i‐e‐a‐o‐u in the vowel inventory without phonological contrasts in quantity, openness, or nasalization; this feature is found in Greek, Tosk Albanian, Romanian, Macedonian, Bulgarian, Torlak Serbian, and Romani; c. devoicing of word‐final stops; this feature is found in Bulgarian, Macedonian, Megleno‐Romanian, Modern Greek (dialectally only in one part of northern Greece, as reported in early twentieth century), some Romani dialects, South Montenegrin and Torlak Serbian, and Turkish (somewhat generally but with greater consistency in West Rumelian Turkish); d. development of nasal + voiced stop clusters (e.g. [mb]) out of nasal + voiceless stop combinations, so that the former clusters are rare or nonexistent, or present only in loanwords; this feature is found in Albanian, Aromanian, and Greek; e. presence of ð/θ (voiced/voiceless interdental spirants); this feature is found in Greek, Albanian, Aromanian, and (mostly in loanwords) dialectally in Macedonian; f. realization of /mj/ as [mɲ]; this feature is found in Greek and dialectally in Arvanitika (Tosk Albanian dialects spoken in Greece); g. a reduction in the nominal case system, especially a falling together of genitive and dative cases; this feature is found in Greek, Albanian, Romanian, Bulgarian, and Macedonian (though note that the latter two have eliminated other case distinctions as well); h. the formation of a future tense based on a reduced, often invariant, form of the verb ‘want’; this feature is found in Greek, Tosk Albanian, Romanian, Macedonian, Bulgarian, Bosnian‐Croatian‐Montenegrin‐Serbian, and Romani; i. the use of an enclitic (postposed) definite article, typically occurring after the first word in the noun phrase; this feature is found in Albanian, Romanian, Macedonian, Bulgarian, and Torlak Serbian; j. analytic comparative adjective formations; this feature is found in Greek, Albanian, Romanian, Bulgarian, Macedonian, and Romani, as well as in Turkish; k. marking of personal direct objects with a preposition; this feature is found in Aromanian, Megleno‐Romanian, and Romanian (via inheritance) and in southern Macedonian dialects; l. double determination in deixis, that is a demonstrative adjective co‐occurring with a definite article and a noun (thus, “this‐the‐man”); this feature is found in Greek, southern Macedonian, and to a limited extent in Albanian and Bulgarian too; m. possessive use of dative enclitic pronouns; this is found in South Slavic and in Greek;

Language Contact in the Balkans  541 n. the use of verbal forms to distinguish actions on the basis of real or presumed information‐source, commonly referred to as marking a witnessed/reported distinction but also including nuances of surprise (admirative) and doubt (dubitative); this feature is found in Albanian, Bulgarian, Macedonian, and Turkish, and to a lesser extent in Romani, Serbian, Aromanian (dialectally), and Romanian (the presumptive); o. the reduction in use of a nonfinite verbal complement (generally called an “infinitive” in traditional grammar) and its replacement by fully finite complement clauses (see Joseph 1983); this feature is found most intensely in Greek, Macedonian, Bulgarian, Torlak Serbian, and Romani, but also in Albanian (especially Tosk) and Romanian, and to a lesser extent, the successor languages to Serbo‐Croatian; p. the pleonastic use of weak object pronominal forms together with full noun phrase direct or indirect objects (“object doubling”); this feature is found in Greek, Albanian, Romanian, Bulgarian, and Macedonian, dialectally in Serbian, and to a limited extent in Romani; q. the formation of the “teen” numerals as DIGIT‐“on”‐TEN; this is found in Albanian, South Slavic, Aromanian, Megleno‐Romanian, and Romanian; r. lexical parallels, including shared phraseology (e.g. a phrase that is literally “without (an)other” meaning “without doubt,” or “eat wood” meaning “suffer a beating”), and numerous shared loanwords, many of which are from Turkish. Some of these features are stated as synchronic typological characteristics, e.g. pleonastic use of weak object pronouns, while some are stated in historical terms, e.g. reduction of cases, while still others lend themselves to both sorts of framing, e.g. widespread use of finite complementation due to the replacement of infinitives. Both dimensions – the synchronic and the diachronic – are appropriate to consider in a discussion of the Balkan languages, since it is historical events (of contact and of reaction to that contact) that have led to the convergent typological state found in these languages. Without belaboring the point, it is important to note that these features are generally taken to be significant indications of contact‐induced convergence because, except as noted (e.g. with features involving the various forms of Romanian), they are not features inherited from a common protolanguage (e.g. Proto‐Indo‐European is typically reconstructed without a definite article and with synthetic (inflectional) analytic adjective formations, so (i, j, l) clearly could not be inheritances, and any features involving Turkish and Indo‐European languages similarly could not be due to genetic relatedness). Moreover, it is often the case that they are not found in varieties of the languages outside of the Balkans (e.g. other Romance languages use a preposed definite article and other Slavic languages generally lack an article, and other Romance languages have well‐developed infinitival usage, as do other Slavic languages). Further, occasional occurrence of some of these features in other closely or distantly related languages (e.g. a postposed definite article in northern Russian dialects and in Scandinavian languages, or object doubling in Spanish, or a “want”‐based future in English) does not vitiate the significant clustering of convergent features in the Balkans. To some extent, then, the geography here allows for what some (e.g. Campbell 1985, 1997: 330–1, 2006: 14) have called a “circumstantialist” argument for a sprachbund. It is also the case that much is known about the history of these languages, and comparisons can be made with earlier stages of Greek, for instance, where infinitival usage abounded, or Slavic, with a well‐developed case system, permitting a judgment as to the innovative convergence (and concomitant divergence from earlier structural patterns) and thus allowing for what some (e.g. Campbell 1985, 1997: 330–1, 2006: 14) have called a “historicist” argument for a sprachbund.

542  Brian D. Joseph

3  Causes of Convergence in the Balkans A key issue in the study of the Balkans is trying to determine what the causes are for the convergences noted. For the most part, scholars agree that language contact is at work, though it must be admitted that some of the developments may well be independent in each language, at least for individual features, even if not for all of the similarities between and among the various languages. Joseph (forthcoming; see also Friedman and Joseph 2021: ch.5) argues that the stressed schwa is not a contact‐induced feature but rather one that developed in each language on its own, and aspects of the emergence of the “want”‐based future, especially the reductions to a highly “abbreviated” form, may well have taken place on a language‐by‐language basis, given that full and reduced variants co‐existed (or continue to co‐exist) in each language for some time. But even if it is granted that contact is responsible, the question arises as to what kind of contact it was, and what contact‐related mechanisms were at work in the formation of the Balkan sprachbund. One can imagine several possibilities: (4)

a. substratum effect (i.e. first‐language speakers, shifting to a second language, carry over their habits and structures of the first language into the second, producing an altered form, via interference, of the second language); b. adstratum effects (i.e. structures from a second language are imported by speakers into their native (first) language, e.g. for reasons of prestige or communicative ease); c. pidginization (i.e. a simplified version of a target language is developed by speakers of several different languages in a situation of communicative necessity); d. speaker‐to‐speaker accommodation to (imperfect) skills of an interlocutor (i.e. a native speaker of the target language adjusts his/her speech to match the perceived level of ability in that language by a non‐native speaker; this may involve selection by both speakers of structures for the target language that are “comfortable” to both, that is, acceptable as a variant in the target language and matching some structural element in the other language); e. “enhancement via contact” (i.e. a feature which is latent or just emerging in a given language becomes strengthened or increases in frequency due to contact with speakers of another language in which that feature is found).

Several comments about these putative causes are in order. First and foremost, the same feature often has been explained by different scholars in different ways. For example, a substratum explanation has sometimes been proposed for the loss of the infinitive, but the chronology needed to make that work, involving a pre‐historic substrate language whose effects surface only in the medieval period, is difficult and argues against it; rather, either an adstratum account (so Sandfeld 1930, with Greek as the prestige language) or pidginization effects (Rozencvejg 1976), or, perhaps better, a mix of pidginization and accommodation drawing on language‐internal tendencies can be posited (thus rather like the enhancement scenario of (4e)  –  see Joseph 1983 for discussion of all of these possibilities). Second, these causes are not mutually exclusive, in that one might be right for one feature and another right for another feature. The postposed definite article, for instance, unlike the infinitive developments, could well be a real substratum effect, according to the rather compelling account offered by Hamp (1982). But other features are more amenable to other accounts, as the range of possibilities concerning the infinitive‐replacement shows.

Language Contact in the Balkans  543 What all these accounts have in common is that they involve, in one way or another, ­ ultilingualism. The question to be asked, then, is what sort of multilingualism is at issue: m casual and sporadic or intense and regular, unidirectional or mutual, intimate, or just what? As the earlier discussion makes clear, it seems that for the Balkans, for the most part (and maybe for sprachbunds in general), what is decisive is intense, intimate, sustained, and mutual multilingualism, involving what Friedman and Joseph (2017, 2021: ch. 8) call their “four‐M” model, with multilateral multidirectional mutual multilingualism. Speakers of different languages, living side by side for centuries (which, for the Balkans, corresponds to what is known historically about the co‐existence of several languages co‐territorially in multilingual villages, towns, and cities), and needing to communicate with one another on a variety of levels, necessarily were familiar with one another’s languages to some degree, and accommodated in their usage of their own language to the often imperfect (but possibly quite good) knowledge of that language on the part of speakers of other languages that they interacted with. Speakers of the target language, it can be posited, selected for structures that had ready analogs in the other language, in effect streamlining their own usage in the direction of that of others. Speakers of the other language, for their part, would often have produced structures in the target language that showed the effects of interference (substratum influence) from their own native language. This mutual accommodation on a base of native language interference would naturally lead to the sort of convergence results that characterize a sprachbund. Note that Thomason and Kaufman (1988) posit just such a social context as essential for the development of a sprachbund, namely with the relevant speech communities each maintaining their own linguistic identity in spite of the extensive and ­intimate contact and thus with some members of the groups of necessity being bi‐ or multilingual. Although it might seem that the evidence of overwhelming convergence alone confirms the hypothesis of language contact being involved in the formation of the Balkan sprachbund, there is direct evidence of the sort of contact that breeds a sprachbund, namely the intense and intimate bilingualism referred to herein throughout. The evidence in question is certain types of lexical borrowings, which requires a bit of explanation. Borrowing of lexical material in and of itself can occur with only casual or even very little contact between speakers; for the latter situation, for instance, the case of learned borrowings through the medium of written texts can be cited. But in some instances, borrowing occurs under circumstances of close and sustained contact in such a way that it is clear that the speakers of different languages must have been communicating with one another on a regular and everyday basis, and under conditions where there was some knowledge of the others’ language involved, in short an intimate contact situation with a degree of bilingualism. The particular borrowings in question are examples of what Friedman and Joseph (2014, 2021: ch.4) have called “ERIC” loans, an acronym for those loans that are “Essentially Rooted In Conversation.” They are loans that result from conversational interactions between speakers of different languages in bilingual situations. They are somewhat like the non‐ need/nonculturally based borrowings that Bloomfield (1933: 461) has called “intimate,” but they go beyond that to recognize those that derive from what we may characterize as “human‐oriented” interactions among speakers (as opposed to “object‐oriented,” as with commercial or work‐related interactions). They are especially revealing, since they necessarily involve real contact between and among speakers on an intense, regular, and sustained basis, in a more or less equal power situation.8 Moreover, the ERIC loans to be considered here involve items that are tightly tied to discourse and to the expressiveness that is part and parcel of human discourse. Clearly if such forms pass from language to language, there must have been discourse, i.e. conversational interactions, between speakers of these different languages.

544  Brian D. Joseph One large area of such discourse‐related borrowings involves various sorts of negation. There is no need for borrowing here at all, since the languages of the Balkans – as indeed surely all languages in general – had means for expressing negation. The incorporation of elements of negation from other languages, therefore, must represent the result of close contact among the speakers. Moreover, some of the forms in question have an expressive function that is intimately tied to conversational interaction and is not really found outside of that context, and one, moreover, is paralinguistic and thus could really only spread through visual contact.9 For example, Modern Greek and Macedonian have both borrowed the Turkish existential and emphatic negator yok ‘there is no …; no!’ in its emphatic function. Thus Greek has [yok] (spelled ) and Macedonian has jok, both with the meaning ‘no way; not in the least’.10 Interestingly, and significantly for the view advocated here, it is the more highly conversationally based function of Turkish yok that is borrowed, not the more denotational existential sense. Turkish, for its part, has an interjection ba with the meaning ‘oh!’,11 which, according to Redhouse (1984), is a borrowing from Greek ba (spelled ) ‘ah well’ (but also, as a negator, ‘unh unh; no way’). And, the widespread Balkan gesture of an upward head nod to signal negation, found at least among speakers of Albanian, Greek, Romanian, and Turkish, may well reflect a diffusion from Greek, given what is known about Ancient Greek gestures and the fact that the distribution especially in Italy coincides with geographic limits of Magna Graecia (Morris et  al. 1979); such an element of paralanguage could only spread through face‐to‐face interaction among speakers, that is, in an intimate conversationally based contact situation. Other discourse‐related borrowings include a large number of interjectional elements, presumably spread through face‐to‐face contact on a day‐to‐day basis. For instance, there is a form which can be glossed (roughly) as an “unceremonious term of address” and stems ultimately from Greek (where there are some 55 different variants across Greek dialects – see Joseph 1997), and which, in various of its shapes, is widespread in the Balkans, as indicated in (5): (5)

Turkish: Albanian: Romanian: Bulgarian: Macedonian: Serbian: Greek:

bre, bire, be ore, or, mor, more, moj, ori, mori, moré, mre, voré, bre bre, mă, mări more, mori, bre more, mori, bre more, mori, bre bre, vre, re, are, mare, marí, oré, voré, ori, mbre, pre, more etc. (this last being the source of practically all these forms)

Similarly, there are several parallel exhortative elements to be found across the Balkans, as given in (6), most probably from Turkish, possibly from (h)ay (interjection) + de (from de‐mek ‘to say’) though there are other possible sources within Turkish: (6)

Romanian: Serbian: Albanian: Greek:

haide/(2PL) haideţi/(1PL) haidem ‘c’mon; gw’an; let’s go’ hajde/hajdemo (1PL)/hajdete (2PL) hajde (SG)/hajdeni (PL) aide (spelled )

Note in this regard also Bulgarian and Macedonian ela, both borrowed from Greek éla ‘c’mon’ (the imperative of ‘come’).12 Continuing with interjections, one can also cite Albanian hopa and Greek opa! ‘oops’ (for something unexpected), ‘woo‐hoo!’ (expression of joy); Albanian pa pa pa and Greek pa pa pa ‘alas!’ (for disgust); and Albanian aman and Greek amán ‘oh my!’ (from Turkish aman).

Language Contact in the Balkans  545 Other highly expressive forms that are typically found in colloquial, and thus conversational, usage, also fit in here in that they too have diffused across the Balkans. In particular, one finds in these languages parallels in onomatopoeia (and the like). For example, for a dog’s noise, Albanian has ham‐ham, Romanian has ham, Greek has γav γav, and Turkish has hav hav, and for the noise for attracting a cat, Greek has ps ps ps, as also in Bulgarian and Romanian. There is of course the risk of attributing to contact here what might be thought of as universal, but since onomatopes (etc.) do vary across even related and contiguous languages (Spanish has [waw] for the bark of a dog while Portuguese has [kãw]), the Balkan similarities, especially when viewed against the backdrop of other parallels in expressive forms and other grammatical and lexical convergences, fit into a pattern worthy of the attention of contact‐minded linguists. In a similar vein, the expressive reduplication with m‐ that is found in Turkish and other more eastern languages,13 e.g. kitap‐mitap ‘books and such’, occurs in Greek, e.g. dzandzala mandzala ‘this and that’, literally ‘rags and such’ (cf. Levy 1980; Joseph 1984, 1995), Albanian, e.g. cingrë mingrë ‘trivia’, and Bulgarian (cf. Grannes 1978). Moreover, to elaborate somewhat on the parallel listed on p. 541 in (3r), there are many calques – phraseological loan‐translation parallels – in the Balkans in which native language material is substituted for elements in other‐language combinations. While there can be learnedisms that are calqued with no direct speaker contact, such as German Mitleid, which is a prefix‐plus‐noun combination based on Latin compassio (itself a calque on Greek sumpatheia ‘compassion, sympathy’), the colloquial and expressive nature of these Balkan calques suggests a social milieu for their creation that is different from that involving learnedisms. In particular, they point to face‐to‐face speaker interaction. Moreover, they offer direct evidence of bilingualism, since speakers must be familiar enough with the other language to be able to figure out equivalences in their own language to the other language’s pieces in the phrase or form or combination of elements being calqued. Thus, as noted in (3r), Greek has troγo ksilo for ‘I get a beating’, but it is literally ‘I‐eat wood’, with the choice of verb agreeing with the Turkish use of yemek ‘to eat’ in the expression kötek yemek ‘to get a beating’ but literally ‘a‐blow to‐eat’. Similarly, what is literally ‘to take [someone’s] eye’ means ‘to dazzle’ in ­several languages and ‘to cut [one’s] mind’ means ‘to decide’. Therefore, putting these two types of colloquial and expressive lexical evidence together, a strong case emerges for the conditions being present in the Balkans that were the essential ingredients for the convergence effects that characterize a sprachbund. They thus offer a further argument, along with the geography and the history of the Balkans, that the convergences listed in (3) are indeed indicative of a sprachbund.

4 Assessing the Sprachbund: Localized vs Broadly Realized Convergence As suggested in the enumeration of Balkanisms given on pp. 540–541 in (3), there are two general types of contact‐induced convergences to be recognized:14 those that occur on a widespread basis among the various languages and those that are highly local in nature. The loss of the infinitive and its replacement by finite forms would be an example of the former type, and the occurrence of prepositional marking of personal direct objects would be an example of the latter type. From the general approach taken here, with the emphasis on actual speaker‐to‐speaker contact as the source of the diffusion of features, it should be clear that all diffusion should be taken to be on a localized basis. And, as indicated on p. 540, there are actually dozens more such local convergences to be found in the Balkans. This observation leads to two questions: first, how one is to reconcile the local effects with the broadly realized Balkanisms, and second, whether, in the face of localized convergences,

546  Brian D. Joseph it makes any sense to think of the Balkan languages in the broad terms that the “sprachbund” designation requires. In other words, is a sprachbund a viable construct if all the relevant contact takes place locally? The answer is that local diffusion, if given enough time and the right sort of contact at the relevant “edges” of locales, can lead to spread across larger areas, a sort of “diffusionary chain reaction,” as it were. Thus a widely distributed feature, as with the infinitive, would have started small and spread widely from that point, and a feature with a more limited range is one that either has not had the chance yet to spread further or has been checked before spreading further. Indeed, since much remains to be known about the spread of any linguistic innovations – those that are the result entirely of language‐internal factors as well as those that are contact‐induced – it perhaps makes no sense to worry about features that remain localized in their distribution. What we really have then is clusters of convergent languages, where the convergence is on various features in various locales. A sprachbund – the Balkan sprachbund in this case – thus is really to be defined as a cluster of such clusters (see Hamp 1989 especially on this view).15 This view of the Balkans, and of sprachbunds in general, has several advantages and addresses a couple of issues typically raised as potential problems for the sprachbund as a viable construct. First, by looking at the overall picture on a feature‐by‐feature basis, the fact that some convergent features may not be as strongly realized as others is not a problem, since there is nothing that says that all features must be found to the same extent in all languages or even be found to any extent in all of them. Thus, the absence of a postposed definite article from Greek does not vitiate the importance of this feature in linking Albanian, Romanian, and Balkan Slavic. Second, as noted in section 3, different causes may underlie different Balkanisms; they need not be, and probably are not, uniform as to their source. And of the lists of Balkanisms that are generally offered, it is hard to see how all of them must involve contact. Third, the fact that there are differences among the languages even with respect to convergent features  –  for instance the relatively recent recrudescence of an infinitive in Tosk Albanian (the originally nominal purpose construction of the sort për të punuar ‘(in order) to work’ (literally ‘for (the‐act‐of‐)working’)) in the face of the loss of such constructs in the other languages – is not a problem, since just as features spread on a localized basis, so too can they go off in their own direction on a localized basis. What this really means is that different features have different histories, but that is as it should be since not all of the features are tied to one another such that a change in one would necessarily trigger a change in another. Moreover, the history is crucial to what the languages are synchronically. In a sense then, in the Balkans, we are dealing with the aftermath of a period of intense contact leading to convergence; the modern standard languages, inasmuch as they were generally formed on the basis of contact‐affected dialectal sources, show structural convergences as a relic of their histories; ongoing convergence continues, but on the local level rather than the “national language” level; the conditions that gave rise to the convergence, that created the sprachbund, are no longer present as far as the standard languages are concerned, though they do obtain in various multilingual locales still. Thus in the present just as in the past in the Balkans, the local dialects must be the main focus for the study of language contact, as they are, and have always been, where the action is (see Friedman and Joseph 2017, 2021: ch.8, for more discussion in this regard).

NOTES 1 Eric Hamp has argued, for instance (see Hamp 1994), that “Illyrian” may be a cover term used in ancient times in much the same way that “aboriginal language” is used by many in Australia today

Language Contact in the Balkans  547 or “Indian language” is used by many in the United States, in each case masking a considerably complex and diverse linguistic situation. 2 The state of what is known about these languages is given a careful treatment in Katičić (1976), and Friedman and Joseph (2021: ch.1) provides a good summary overview. 3 Although these four varieties are similar in many respects, there are some structural and lexical differences holding among them. Moreover, each is a national language: Bosnian of Bosnia and Hercegovina, Croatian of Croatia, Montenegrin of Montenegro (as of 2007), and Serbian of Serbia. Thus, treating them as separate languages is motivated as much by present‐day social and political realities as anything purely linguistic. 4 By some accounts, the Balkans may start at Vienna, for instance, while others look to the Danube River as the northern edge. 5 If Tagalog, Arabic, and such languages (including English even) are to be counted, then perhaps one might speak of an even broader designation of “languages in the Balkans,” of which “languages of the Balkans” would be a subset. 6 Despite distinguishing here among Aromanian, Megleno‐Romanian, and Romanian, quite rightly, as separate languages, I nonetheless occasionally, for the sake of convenience, refer simply to “Romanian” as a cover term for all three. 7 There is some controversy as to what the threshold is for recognizing a sprachbund. Thomason (2001: 99) opts for three as this lower limit on the grounds that it trivializes the notion to allow just two languages with structural features in common to determine what should properly be thought of as a special grouping, while Friedman and Joseph (2021: ch.3) argue that convergence is convergence and that therefore, assuming other criteria are met, a two‐member sprachbund should not be ruled out in principle. 8 The situation with Romani bilingualism admittedly does not involve equal power structures, since it was unidirectional: Romani speakers learned the other languages around them but speakers of those languages generally did not learn Romani. Still, assuming that what Romani speakers learned was already Balkanized varieties of these other languages, by a process of “reverse interference” (see Friedman and Joseph 2021: ch.3), whereby speaking another language can have an effect on one’s native language, Romani speakers could have assimilated their Romani to aspects of these other languages they came to speak. 9 See Joseph (2000, 2001, 2002a, 2002b) for more on parallels in the Balkans involving negation. 10 Turkish also has an emphatic negative, presumably related to yok, with the form yo. However, despite the similarity to the Albanian word for ‘no’, jo, this Turkish form is unlikely to be the source of the Albanian, since jo is found even in the Arbëresh Albanian of southern Italy, an Albanophone area that shows little or no influence from Turkish. I am indebted to Eric Hamp for clarification on this important point. 11 It is of course difficult to give precise definitions for interjections; the glosses here (and further on) are intended just to give a feel for the form’s use. 12 Albanian eja ‘come!’ probably has a different origin and is not connected to ela; see Joseph (2015) on the etymology of eja. 13 And elsewhere – see Southern (2005) on this particular expressive mechanism cross‐linguistically. 14 This assumes, of course, that the non‐contact‐induced convergences, especially those due to separate and independent developments in each language (as is probably the case with stressed schwa, as noted on p. 542), are properly excluded from consideration. 15 Just as a galaxy is made up of constellations and other groupings of stars and planets.

REFERENCES Asenova, Petja 1989 [2002]. Balkansko Ezikoznanie: Osnovni Problemi na Balkanskija Ezikov Săjuz [Balkan linguistics. Basic problems in the Balkan language union]. Sofia: Nauka i izkustvo (2nd edition 2002. Sofia: Faber).

Banfi, Emmanuele 1985. Linguistica balcanica. Bologna: Zanichelli. Beck, David. 2000. Grammatical convergence and the genesis of diversity in the Northwest coast Sprachbund. Anthropological Linguistics 42: 147–213.

548  Brian D. Joseph Bloomfield, Leonard 1933. Language. New York: Holt, Rinehart, and Winston. Campbell, Lyle 1985. Areal linguistics and its implications for historical linguistic theory. In Jacek Fisiak (ed.), Proceedings of the Sixth International Conference on Historical Linguistics. Amsterdam: Benjamins, pp. 25–56. Campbell, Lyle 1997. American Indian Languages: The Historical Linguistics of Native America. Oxford: Oxford University Press. Campbell, Lyle 2006. Areal linguistics: a closer scrutiny. In Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas: Convergence in Historical and Typological Perspective. New York: Palgrave Macmillan, pp. 1–31. Campbell, Lyle, Terrence Kaufman, and Thomas C. Smith‐Stark 1986. Meso‐America as a linguistic area. Language 62: 530–70. Demiraj, Shaban 1994 [2004]. Gjuhësi ballkanike [Balkan linguistics]. Skopje: Logos‐A. (Macedonian edition). 1994: Balkanska lingvistika [Balkan linguistics]. Skopje: Logos‐A (2nd edition. 2004: Tiranë: Akademia e Shkencave e Republikës së Shqipërisë (Instituti i Gjuhësisë dhe i Letërisë). Emeneau, Murray 1956. India as a linguistic area. Language 32: 3–16. Feuillet, Jack 1986. La linguistique balkanique (Cahiers Balkaniques 10). Paris: INALCO. Feuillet, Jack 2012. Grammaire comparée des langages balkaniques. Paris: Institut d’Études Slaves. Friedman, Victor A. 2003. Evidentiality in the Balkans. In Alexandra Y. Aikhenvald and Robert M. Dixon (eds.), Studies in Evidentiality. Amsterdam: Benjamins, pp. 189–218. Friedman, Victor A. and Brian D. Joseph 2014. Lessons from Judezmo about the Balkan Sprachbund and contact linguistics. International Journal of the Sociology of Language 226: 3–23. Friedman, Victor A. and Brian D. Joseph 2017. Reassessing Sprachbunds: a view from the Balkans. In Raymond Hickey (ed.), Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 55–87. Friedman, Victor A. and Brian D. Joseph 2021. The Balkan Languages. Cambridge: Cambridge University Press. Grannes, Alf 1978. Le redoublement turk à m‐ initial en bulgare. Balkansko Ezikoznanie 21: 37–50. Hamp, Eric P. 1982. The oldest Albanian syntagma. Balkansko ezikoznanie 25.1: 77–9. Hamp, Eric P. 1989. Yugoslavia: a crossroads of Sprachbünde. Zeitschrift für Balkanologie 25.1: 44–7.

Hamp, Eric P. 1994. Albanian. In Ronald E. Asher (ed.), The Encyclopedia of Language and Linguistics. Oxford: Pergamon Press, pp. 65–7. Joseph, Brian D. 1983. The Synchrony and Diachrony of the Balkan Infinitive: A Study in Areal, General, and Historical Linguistics. Cambridge: Cambridge University Press (reissued in paperback, 2009). Joseph, Brian D. 1984. Balkan expressive and affective phonology: the case of Greek ts/dz. In Kot Shangriladze and Erica Townsend (eds.), Papers for the Fifth Congress of Southeast European Studies (Belgrade, September 1984). Columbus, OH: Slavica Publishers (for the US National Committee of the AIESEE), pp. 227–37. Joseph, Brian D. 1995. Borrowing at the popular level: Balkan interjectional particles of Turkish and Greek origin. Septième Congres International d’Études du Sud Est Européen: Rapports. Athens: Greek National Committee for Southeast European Studies, pp. 507–20. Joseph, Brian D. 1997. Methodological issues in the history of the Balkan lexicon: the case of Greek vré/ré and its relatives. In Victor Friedman, M. Belyavski‐Frank, M. Pisaro, and D. Testen (eds.), Balkanistica: Studies Dedicated to the Memory of Zbigniew Golab 19 March 1923–24 March 1994, vol. 10. pp. 255–77. Joseph, Brian D. 2000. On the development of Modern Greek óxi “no.” In Chris Schaner‐ Wolles, John Rennison, and Friedrich Neubarth (eds.), Naturally! Linguistic Studies in Honour of Wolfgang Ulrich Dressler Presented on the Occasion of his 60th Birthday. Turin: Rosenberg and Sellier, pp. 207–14. Joseph, Brian D. 2001. Language contact and the development of negation in Greek and the Balkans. In Greek Linguistics ’99: Proceedings of the Fourth International Conference on Greek Linguistics, Nicosia, September 1999. Thessaloniki: University Studio Press, pp. 346–53. Joseph, Brian D. 2002a. Balkan insights into the syntax of *me: in Indo‐European. In Mark Southern (ed.), Indo‐European Perspectives (Journal of Indo‐European Studies Monograph Series 43). Washington, DC: Institute for the Study of Man, pp. 103–20. Joseph, Brian D. 2002b. Language contact and the development of negation in Greek – and how Balkan Slavic helps to illuminate the situation. In Irene Masing‐Delic and Mateja Matejic (eds.), A Festschrift for Leon Twarog: Working Papers in Slavic Studies, vol. 1. Columbus, OH:

Language Contact in the Balkans  549 The Ohio State University Department of Slavic and East European Languages and Literatures, pp. 131–9. Joseph, Brian D. 2015. Inheritance versus borrowing in Albanian etymology: The case of eja. In Bardhyl Demiraj (ed.), Sprache und Kultur der Albaner. Zeitliche und räumliche Dimensionen. Akten der 5. Deutsch‐albanischen kulturwissenschaftlichen Tagung (6.–9. Juni 2014, Buçimas/ Albanien) (Albanologische Forschungen 37). Wiesbaden: Harrassowitz, pp. 196–202. Joseph, Brian D., forthcoming. On the Need for History in Balkan Linguistics (Kenneth E. Naylor Memorial Lecture Series, vol. 10). Oxford, MS: Balkanistica. Katičić, Radoslav 1976. Ancient Languages of the Balkans, 2 vols. (Trends in Linguistics, State‐of‐ the‐Art Report 4, 5). The Hague: de Gruyter Mouton. Kopitar, J. 1829. Albanische, walachische und bulgarische Sprache. Jahrbücher der Literatur 46: 59–106. Levy, Harry 1980. An Anatolian language‐trait in Byzantios’ Babylonia and parallel traits on three continents. MGSA (Modern Greek Studies Association) Bulletin 12: 47–55. Masica, Colin 1976. South Asia as a Linguistic Area. Chicago: University of Chicago. Miklosich, Franz 1861. Die slavischen Elemente im Rumunischen. Denkschriften der kaiserlichen Akademie der Wissenschaften, philosophisch‐ historische Klasse 12: 1–70.

Morris, Desmond, Peter Collett, Peter Marsh, and Marie O’Shaugnessy 1979. Gestures: Their Origins and Distribution. New York: Stein and Day. Redhouse, James 1984. New Redhouse Turkish– English Dictionary. Istanbul: Redhouse Press. Rozencvejg, Viktor 1976. Linguistic Interference and Convergent Change (Janua Linguarum, Series Maior 99). The Hague: de Gruyter Mouton. Sandfeld, Kristian 1926 [1930]. Linguistique balkanique. Paris: Klincksieck. (First printed 1926: Balkanfilologien. Copenhagen: Luno.) Schaller, Helmut 1975. Die Balkansprachen: Eine Einführung in die Balkanphilologie. Heidelberg: Harrassowitz. Southern, Mark 2005. Contagious Couplings: Transmission of Expressives in Yiddish Echo Phrases. Westport, CT: Praeger. Steinke, Klaus and A. Vraciu 1999. Introducere în Lingvistica Balcanică. Iaşi: Editura universitătii “Al. I. Cuza.” Thomason, Sarah G. 2001. Language Contact. Washington, DC: Georgetown University Press. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Trubetzkoy Nikolay S. 1923. Vavilonskaja bašnja i smešenie jazykov [The Tower of Babel and the mixture of languages]. Evrazijskij vremennik 3: 107–24. Trubetzkoy, Nikolay S. 1928. Proposition 16. Actes du Premier congrès international de linguistes. Leiden: Sijthoff, p. 18.

28 Turkic Language Contacts LARS JOHANSON, ÉVA Á. CSATÓ, AND BIRSEL KARAKOÇ

Turkic has a vast area of distribution. It extends from the southwest with Turkey and her neighbors, to the southeast, to eastern Turkistan and farther into China. From here it stretches to the northeast, via South and North Siberia up to the Arctic Ocean, and finally back to the northwest, across West Siberia and East Europe. The area comprises a great number of languages. Regions in which Turkic is spoken include Anatolia, Azerbaijan, the Caucasus region, Iran, Iraq, Afghanistan, West and East Turkistan, South, North and West Siberia, and the Volga region. In the past, the Turkic‐speaking world also included enclaves in the Ponto‐ Caspian steppes, the Crimea, the Balkans, etc. Turkic offers particularly rich sources of data for the study of language contact. The continuous and massive displacements of Turkic‐speaking groups throughout their history have led to numerous new configurations of various kinds.

1  Intrafamily Contacts In one kind of areal contact situation, varieties of Turkic have encountered and influenced each other. This was the normal situation in the old tribal confederations with their mobile heterogeneous groups. The encounters led to the emergence of modified varieties. The population movements caused separation of linguistically close groups with the effect that related languages did not occur in clear geographic clusters. The interaction in a number of contact areas has led to new constellations involving convergence, innovation, mixture, leveling. Varieties with different backgrounds have developed common features. Several Turkic varieties have been used as koines, transregional codes for trade and intergroup communication, e.g. Azeri in Iran and the Caucasus region. Languages of the central part of the Turkic world have undergone a good deal of leveling, whereas those spoken in the periphery, e.g. Turkish, have preserved many older features. Languages such as Yakut, Salar, Yellow Uyghur, Khalaj, and Karaim have developed for centuries in relative isolation from their original close relatives, preserving old features and acquiring new ones in their respective environments.

2  Interfamily Contacts Turkic languages have for centuries been spoken in highly dynamic contact settings, entertaining numerous interfamily contacts with genetically and typologically different varieties. Areas of intense contact include Central Asia, Siberia, Transcaucasia, Anatolia, the Balkans,

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

552  Lars Johanson, Éva Á. Csató, and Birsel Karakoç the Volga‐Ural region, and northwest Europe. Complex situations of copying and language shift have led to intricate contact‐induced changes in linguistic subsystems. The contact languages include Indo‐European, especially Iranian, Slavic, recently West European languages, and non‐Indo European languages such as Mongolic, Uralic, Tungusic, Chinese, and Arabic. Because of the unique mobility of many Turkic‐speaking groups, contact‐driven developments have been especially important. The encounters have led to various contact‐induced processes of borrowing, or, to use a more adequate term: copying. The languages involved have undergone processes of change and shift in different dominance relations determined by various sociocultural conditions. There have been two types of code interaction: “take‐ over” and “carry‐over” copying (see Johanson 2008). Speakers have taken over copies from a foreign code into their native code, or they have carried over copies from their native codes into their own variety of a foreign code. Speakers of Turkic have thus taken over foreign lexical, morphological, phonological, and syntactic elements into their own varieties. Speakers of non‐Turkic languages, e.g. Iranian, Finno‐Ugric, Greek, Mongolic, Tungusic, Samoyedic, and Yeniseic, have shifted to Turkic and carried over native elements to their own varieties of Turkic, which has led to substrate effects of various kinds. Lexical elements and free function markers have been copied globally, as a whole, including their material shape (substance) and functions, i.e. properties of meaning, combinability, and frequency. They have also been copied selectively, as “loan translations” or “calques,” with respect to one or more semantic, combinational, and frequential properties, the material shapes being provided by indigenous morphological material. The same is true of affixes, i.e. bound derivational and inflectional elements. Structural patterns have been copied selectively as semantic‐combinational calques using indigenous morphemes. Phonological elements have been copied as elements occurring in loanwords, global lexical copies, or selectively, as elements copied onto indigenous units.

3  Structural and Social Factors The likelihood of a particular element being copied in Turkic language contacts has been determined in part by social factors, such as the prestige of the model language, and in part by structural factors of “attractiveness” (Johanson 2002: 2–3, 43–54). Attractive properties have been copied even in the absence of overwhelming social pressure. The presence of strong pressure has, however, ultimately led to the copying even of unattractive structures. If, in Turkic language contacts, a language has copied some element from another one, 1. the element might have been attractive, or 2. the social influence of the model language on the copying language might have been sufficient to overcome the unattractiveness of the element, or 3. both attractiveness and social influence might have been at work (see Comrie 2002: viii–ix). The contacts between Turkic and non‐Turkic languages show that, under appropriate social circumstances, in particular contact that is sufficiently intense and sufficiently long‐lasting, almost any feature from one language can ultimately be copied into another. Massive contact influence has sometimes caused considerable deviations from the original typological profile of the languages involved. It has been possible for languages to copy structures that appear to be typologically inconsistent with the rest of their structure. Turkic languages spoken under strong foreign impact and in relative isolation from the bulk of their relatives have abandoned old features and developed new ones. Languages such as Karaim and Gagauz, both spoken in Eastern Europe, have been strongly influenced by Slavic, displaying

Turkic Language Contacts  553 excessive copying of phonology, syntax, and lexicon (see e.g. Menz 1999; Csató 2000, 2016; Csató and Menz 2018). The influence of Turkic on its contact languages has been great. Many aspects of Turkic structure have turned out to be attractive, but the sociolinguistic questions concerning the nature of the migrations and political expansions have proven equally important. How did the speakers of the Turkic varieties enter the areas in question? Political expansion does not necessarily lead to linguistic expansion. The size of the relevant politically expanding incoming groups may have varied considerably. There were major, massive migratory movements and minor movements of a thin ruling layer. In both cases, code shifts took place. The question is: did local speakers shift to the incoming code, or did incoming groups shift to the local code? There is no need to postulate massive immigration as a precondition for code shift. Even small incoming elites have imposed their codes on comparatively large existing populations. In these cases we find Turkicized peoples of largely local origin without major demographic changes. A case in point seems to be the introduction of Azeri in the Transcaucasian area and Iran. A relatively small number of Turkic‐speakers seems to have moved in, displacing the existing elites and causing the replacement of existing codes. This kind of introduction of Turkic may also have taken place in other areas. The incoming Turkic‐ speaking groups mostly had an advanced political organization which contributed significantly to their dominance.

4  Examples of Contact Areas Some examples of major contact areas will be given below. It should be remembered that all languages belonging to the Russian sphere of influence show strong effects from Russian, since the second half of the nineteenth century at the latest. The influence is stronger in languages that were in contact with Russian relatively early, e.g. Tatar, Kazakh, and Yakut.

4.1  Central Asia Turkic and Iranian, genealogically unrelated and typologically different, have interacted for many centuries in Central Asia (Csató et al. 2016). The development dates back to early contacts of nomad groups in the Eurasian steppes. Long‐standing intense contacts between southeastern Turkic and eastern Persian, the forerunners of Uzbek and Tajik, have led to considerable influence in both directions and to close symbiotic bonds; see Csató et al. (2004), Johanson and Bulut (2006). Convergence processes resulted in numerous shared features in phonology, morphology, vocabulary, and syntax. Since the developments are highly complex, it is sometimes difficult to determine the direction of influence. It is often also difficult to pinpoint the developmental stages of the languages involved at the time of copying, i.e. to distinguish older and more recent changes. Large areas in Central Asia underwent increasing Turkicization. Turkic dialects were more or less Iranicized. They were first influenced by Soghdian and, after the Muslim conquest, New Persian, which took over the role of an interethnic lingua franca and was the medium through which Central Asian Turks became familiar with Islam and urban culture. Elements of Persian and Arabic origin were spread by merchants and religious teachers along the Silk Road. The Turkic varieties of eastern Turkistan, today’s Xinjiang, have been in contact with numerous languages. The oldest Turkic population of the area had close relations to speakers of Iranian. Strong substrate influences were exerted by speakers of Indo‐European shifting to Turkic. Speakers of Old Uyghur moving into the northeastern part of the Tarim basin also came into contact with Tokharian, a non‐Iranian Indo‐European language.

554  Lars Johanson, Éva Á. Csató, and Birsel Karakoç Uzbek has been significantly influenced by Persian. Its dialects exhibit various degrees of Iranicization. Uzbek and Tajik display many striking structural similarities. Certain shared features are due to Turkic influence on pre‐Tajik eastern Persian varieties. An increasing Uzbek influence on Tajik may be observed. Northern Tajik has even been described, albeit inadequately, as a Turkic language “in statu nascendi” (Doerfer 1967: 57). The development of New Persian in the direction of the Turkic type is obvious. Since the Middle Persian period, Persian had shared substantial typological characteristics with Turkic, developing into “the most atypical Iranian language” (Windfuhr 1990: 530). The Central Asian contact area also includes Mongolic influence on Kazakh, Kirghiz, etc., as well as Chinese and Tibetan influence on the Turkic languages of western China.

4.2 Siberia South Siberia is a melting‐pot of Turkic varieties characterized by contacts with Samoyedic, Yeniseic, Mongolic, and Russian. The historical and anthropological origins of the speaker groups are rather different. The varieties show various degrees of substrate and adstrate influence from Samoyedic and Yeniseic. Some groups speaking these languages have shifted to Turkic quite recently. Substrate effects of South Samoyedic, which belongs to the Uralic family, have played a major role. The Turkicization of the speakers of Yeniseic varieties, e.g. Ket and Kot, was still in progress in the nineteenth century. Ket is the last survivor of Yeniseic. South Siberian Turkic shows clear Mongolic influences, in particular strong Oirat impact from the fifteenth century on. Tuvan has been influenced by Middle Mongol, Oirat, and Khalkha. Western Buryat was once spoken in the area, but has now vanished. Altay Tuvan as spoken in China displays phenomena induced by contact with Chinese. Russian has exerted strong lexical and syntactic impact on all South Siberian varieties and displaced some of them. Yakut, or Sakha, spoken in North Siberia, deviates considerably from other Turkic languages, from which it has been isolated for many centuries. It displays some unique innovations partly due to Mongolic and Tungusic influence. There is an old Buryat Mongolic layer from the period when ancestors of Yakut speakers settled on the shore of Lake Baikal. An early impact may have been exerted by Yeniseic, a formerly widespread Paleoasiatic language. After their emigration to North Siberia, the Turkic language of the Yakut underwent strong substrate influence from Tungusic dialects. The closest neighbors are still the North Tungusic languages Evenki, in the northern and northwestern parts of Yakutia, and Even, previously called Lamut, in the northeastern parts, both probably with Paleoasiatic substrates. Contacts with the isolated language Yukagir have also been important. Dolgan, a Yakut dialect spoken on the Taimyr and considered a language in its own right, has both an Evenki and a Samoyedic (Nganasan) substrate. The complex problems of language contact and language shift in the area are, however, still partly unsolved.

4.3 Volga‐Kama The Volga‐Kama region has been a vital contact area for many centuries. The Turkic actors involved are Chuvash, Tatar, Bashkir, and their predecessors. The non‐Turkic actors are the Finno‐Ugric languages Mari, formerly called Cheremis, Mordvin, and Udmurt, formerly called Votyak, and their precursors, as well as Russian. The varieties show effects of long‐term areal contact processes due to complex combinations of “take‐over” and “carry‐over” processes. Though the relations of social dominance have varied through the centuries, the processes have led to the introduction of new linguistic patterns, typical Sprachbund phenomena.

Turkic Language Contacts  555 Probably as early as the fifth century, groups speaking Kipchak Turkic were present on the middle Volga, absorbing local Finno‐Ugric groups. Oghur Turkic influence came with the Volga Bulghars, who assimilated native groups of the region. Komi‐Zyrian features indicate that intensive contacts took place between Volga Bulghar and Permic in the tenth century. Oghur Turkic is commonly thought to have influenced the Finno‐Ugric and Russian varieties of the region, mainly in phonology (Róna‐Tas and Berta 2011). Oghur tribes came to dominate the Finnic groups on the left bank of the Volga, assimilating speakers of the predecessors of Meadow Mari and Udmurt. Chuvash, the only survivor of the Oghur Turkic type, displays substrate phenomena due to close contacts with Volga Finnic. The influence is strongest in Upper Chuvash, especially in the immediate Mari neighborhood. With the Mongol invasion in the thirteenth century, Kipchak‐speaking newcomers came to play a major role in the area. Tatar strongly influenced Mari, Udmurt, and Mordvin. No other Finno‐Ugric languages have been so strongly influenced by Turkic as Mari and Udmurt. On Mari influence upon Chuvash, see Agyagási (1998, 2019). The Turkic varieties of the area, for example, exhibit many loans of Middle Mongol origin, partly borrowed via Tatar. Certain Chuvash–Mongolic correspondences go back to early contacts of Oghur‐speaking and Mongolic‐speaking groups in South Siberia. The Russian impact increased rapidly from the middle of the sixteenth century on, i.e. after the fall of the Khanate of Kazan.

4.4  Transcaucasia and Iran Turkic and Iranian have interacted for many centuries in Transcaucasia and Iran (Csató et al. 2016). The groups that moved southwestwards from Central Asia to establish the future Oghuz branch of Turkic interacted closely with Persian‐speaking groups. The Seljuk groups who settled in Transcaucasia encountered speakers of other Iranian varieties, e.g. the Northwestern Iranian language Tati, closely related to Talysh, and Kurdish dialects. The Seljuk conquest of the eleventh century led to a massive Turkicization of the area, with many speakers of Iranian varieties shifting to Turkic. A good deal of the idiosyncratic features of Turkic of this area may thus be due to Iranian and other local substrates. Iranization is the most conspicuous feature of the Turkic varieties spoken in Iran; see, e.g. Kıral (2001), Bulut (2006). Direct contacts with spoken Persian, developed for many centuries in asymmetric settings, have left profound unidirectional influences at all linguistic levels in Azeri dialects: in South Oghuz (i.e. Kashkay and related varieties, and the transitional varieties between them), in Khorasan Turkic, and in the non‐Oghuz language Khalaj. Varieties of the adjacent border regions of Iraq and southeastern Anatolia share many of their features. The non‐Oghuz Turkic language Khalaj, spoken in central Iran, has been heavily influenced by Persian, Luri, etc., without losing its specific Turkic characteristics. The Iranicization of Kashkay, spoken in southern Iran, has become more dominant in the last decades.

4.5  The Caucasus The Caucasus offers rich materials for studying genetically diverse languages that have been in contact for millennia. Turkic languages are young languages in the Caucasus region. The ancestors of Kumyk and Karachay‐Balkar may have entered the area in the early Middle Ages. Karachay‐Balkar was an entrant from the steppes that was pushed into the mountainous regions in the thirteenth century and later on driven to still poorer locations in the highlands. Noghay arrived relatively late, after the end of the fourteenth century. Nationalist Turcologists, however, claim that Turkic has existed in the Caucasus for at least 5,500 years. In spite of its late arrival, Turkic played an important role in the region. Kumyk and Noghay

556  Lars Johanson, Éva Á. Csató, and Birsel Karakoç functioned as lingua francas in intergroup communications until Russian eventually took over this function. Karachay‐Balkar, Kumyk, and Noghay have been influenced by their Caucasian neighbors – Kumyk by the Northeast Caucasian languages of Daghestan, the southern dialects of Kumyk especially by Dargwa of the Nakho‐Daghestanian group, and the White Noghay dialect by Cherkes and Abazian of the Northwest Caucasian family, etc. This influence is mainly restricted to loanwords. The syntax of the Turkic varieties has widely remained untouched by foreign influence. In the 1930s, the Marr school of linguistics, founded by Nikolaj J. Marr (1864–1934), focused on Karachay‐Balkar, which was considered a cross‐breed of Turkic and Caucasian elements. This alleged status of Karachay‐Balkar is not supported by linguistic data. There is a certain degree of Caucasian influence, including substrate influence following code shift. But Caucasian languages have left relatively little imprint even upon this language. All languages indigenous to the region have been in intense contact with Russian.

4.6 Anatolia Anatolian Turkish has interacted with many languages: Indo‐European such as Greek, Kurmanji, Zaza, Armenian, Judeo‐Spanish; the Semitic languages Arabic and Syriac; and the Caucasian languages Cherkes, Georgian, and Laz. See Andrews (1989) and Bulut (2018) for details on the current language situation. The history of settlement and assimilation of Turkic‐speaking groups in Anatolia is long and complex. Oghuz‐speaking groups settled in the Byzantine territory before the Seljuk immigration. The large Seljuk immigration began in the eleventh century, and the Turkicization of the indigenous populations probably started toward the end of the Seljuk rule. The political power of the Ottoman state had a strong impact on the processes of Turkicization. Its history is full of minor or major population moves, e.g. immigration of new Turkic‐speakers, mostly from the east. The immigration increased with the extension of Russian rule in the neighborhood. Large masses of immigrants arrived after the annexation of the Crimea in 1783 and the definite subjugation of the Caucasian area in 1864. Owing to political developments, the presence of non‐Turkic languages is now rather reduced. The largest ones are the Iranian languages Kurmanji and Zazaki. Greek and Armenian are present almost only in Istanbul. On Greek as formerly spoken in Cappadocia, see Dawkins (1916). On traces of Greek in Trabzon, see Brendemoen (2002). Judeo‐Spanish, a Hispanic variety spoken by Jews of Spanish origin, is now vanishing. Arabic is spoken by Muslims on the borders to Syria and Iraq and by Christians in and around Mersin. Neo‐ Aramaic is still spoken by small groups in the eastern provinces, especially in Hakkâri. Caucasian varieties such as Laz, Georgian, Abkhas, Adyghe, and Cherkes are found in the northeast. The last speaker of Ubykh, of the Northwestern Caucasian group, died in 1992. Since direct contacts between the spoken languages are lacking, Persian impact is much weaker in Anatolia than in Iran. Ottoman Turkish was a typical member of the languages of the Islamic cultural sphere, but as a result of language reform, the share of Arabic‐Persian loans in modern Turkish has been drastically reduced. Anatolian Turkish as spoken in Cyprus has been in direct contact with Greek since the sixteenth century and with English since the end of the nineteenth century.

4.7  The Balkans The Balkans have for many centuries been a region of intense multilingualism, where Oghuz and Kipchak Turkic varieties established contact with South Slavic and Albanian. In the Ottoman period, vast areas of the Balkans were colonized from Anatolia, which led to the creation of West and East Rumelian Turkish dialects. Large non‐Turkic groups accepted

Turkic Language Contacts  557 Islam and were Turkicized. There is uncertainty about the numbers of Turkic‐speaking colonizers and the groups of autochthonous converts whose local languages served as substrates. Some scholars tend to classify the Turkish dialects as creoles, i.e. nativized pidgins. Turkish has exerted extensive influence on Romani dialects. Some Roma groups have their own varieties of Balkan Turkish. The dominance of Turkish in the Ottoman period gave the Balkan languages many common features. The Turkish impact was first dealt with by Franz von Miklosich in 1884. Spoken Turkish has been important for the formation of the so‐called Balkan Sprachbund of Slavic, Greek, Romance, and Albanian varieties sharing certain areal features. For an overview, see Friedman (2003). The autonomy of the Balkan peoples was followed by an abrupt decrease in Turkish influence. Turkish‐speaking masses left for Turkey, especially after World War I. Through the population exchange with Greece, about half a million Turks from Greece emigrated to Anatolia. Turkish has thus changed from a dominating language to a dominated language in the Balkans, but the linguistic contacts continue. West Rumelian Turkish is still spoken in Macedonia and Kosovo, East Rumelian Turkish mainly in Bulgaria. One language, Gagauz, has been subject to particularly strong Slavic impact, leading to drastic typological changes.

4.8  Western Ukraine and Lithuania A similar case of deep‐going typological changes in western Ukraine and Lithuania, is manifested by the varieties of the Kipchak language Karaim, whose precursors were transplanted from the Crimea several centuries ago. The largest settlements were established in the towns of Halich and Luck, today in western Ukraine, and Trakai (Polish Troki), today in present Lithuania. The only living variety is the Lithuanian one, spoken by some 20 mostly elderly persons. The small Karaim communities were multilingual and strongly dominated by the non‐Turkic languages of their areas. Nevertheless, they could maintain their Turkic vernacular in a long‐lasting asymmetrical contact situation on the periphery of the Turkic‐speaking world. The reason for this might be that Karaims are believers of the non‐Rabbanic Karaite Judaism, and thus readers of the Hebrew Bible. At the same time, they translated Hebrew biblical texts into Karaim. The native language thus played an important role in their religious practice. This function contributed to the survival of the language as long as the religion played a central role in their community life. Lithuania has another minority of Turkic origin, the Tatar minority. Like the Lithuanian Karaims, the Tatars moved into their present Lithuanian settlements about 600 years ago. They are Muslims using Arabic as their sacred language. The Tatar community rather early abandoned its Turkic vernacular. By contrast, Karaim survived, though subject to deep‐going typological changes. Its varieties were influenced by local Slavic and Baltic dialects and developed areal features, which led to great differences between them. They preserved the genuinely Turkic morphology, but their syntax underwent a basic typological change from SOV to SVO basic constituent order. The languages of the Bible translations mirror the structure of the original Hebrew texts (Olach 2013). Another northern European Turkic minority is the Tatar minority of Finland. Between 1870 and 1920, groups of Misher Tatars migrated to Finland from the Volga area and now constitute a small minority mostly settled in Helsinki and surroundings. They are speakers of Finnish but use Tatar as their community language.

4.9  Northwestern Europe The last century, in particularly from the early 1960s on, has seen the emergence of considerable Turkish immigrant communities in northwest European countries as well as in Australia. The initial reason for the migration from Turkey and also from former Yugoslavia was the

558  Lars Johanson, Éva Á. Csató, and Birsel Karakoç acute shortage of labor in the receiving industrial countries. Germany was the main receiving country. In the United Kingdom, the main bulk of the Turkish immigrant groups came from Cyprus. Although the active recruitment of immigrant workers was significantly restricted following the 1973 oil crisis, the migration continued in form of family reunification and family formation (Backus 2013). Today, the “guest workers” of the first generation and their descendants form well‐established communities in the respective countries. Linguistically, the speakers of Turkish live in unbalanced, asymmetrical contact situations, their primary language being dominated by the languages of the host societies, i.e. German, Dutch, Danish, Norwegian, Swedish, etc., which are used for group‐external communication. Code alternation appears to be a common discourse strategy in the speech of the second and third generations. The prevailing attitudes towards the heritage language and the policies of bilingual education and home‐language instruction vary a great deal in the countries concerned (see e.g. Yağmur 2016). It is still uncertain whether the Turkish varieties spoken in northwestern Europe will ultimately diverge from Turkish as spoken in Turkey. For a discussion see Rehbein, Herkenrath, and Karakoç (2009).

5  Written Turkic Languages A short overview of the development of written Turkic languages is given in Menges (1968: 165–73). The oldest documents of written Turkic, the Orkhon and Yenisey inscriptions, do not exhibit any loans except some Chinese and Iranian titles. The Old Uyghur literature of eastern Turkistan consists of Buddhist, Manichean, and Nestorian texts, almost exclusively translations. It displays loans from several non‐Turkic languages  –  Sanskrit, Tokharian, Soghdian, etc. – mostly pertaining to the religious, philosophical, political, and legal domains. It also documents interesting Turkic calques of Buddhist terminology. This literary culture came to an end with the Islamic conquest. In the Islamic successor languages of Old Uyghur, non‐Arabic terms in the areas of spiritual, social, and political life were not tolerated. Foreign words and numerous Turkic words were replaced by Arabic‐Persian ones. The Turkic literary languages that emerged in the late Middle Ages were subject to strong Persian influence from the very beginning. Chaghatay and Ottoman were thoroughly influenced by a prestigious Persian‐Arabic vocabulary that gave them a remarkable richness of expressive resources. Most Chaghatay authors were probably bilingual in Turkic and Persian. From the fifteenth century on, stylistically refined registers, overloaded with Persian‐Arabic elements, emerged in Ottoman elite literature. The abundance of Arabic‐Persian loans in Ottoman led to strong puristic efforts in the twentieth century. In spite of all differences, it is a fact that most written Turkic languages have been influenced by Persian and Arabic. A much more recent process is the strong Russian influence that the Turkic literary languages of the Russian political sphere have been subject to.

6  Types of Copying 6.1  Copied Lexicon Turkic languages have been exposed to a good deal of foreign lexical influence. Already Old Uyghur displays lexical copies from Chinese, Soghdian, and Tokharian. The vocabulary of modern languages, particularly Modern Uyghur, still mirrors these and other manifold contacts (Yakup 2005).

Turkic Language Contacts  559

6.2  Arabic‐Persian Loans Most Turkic languages possess words of Arabic and Persian origin that have partly superseded the native Turkic vocabulary. Persian lexical influence, which also includes Arabic words introduced via Persian, has been very strong in all Turkic languages of the Islamic world, covering various domains of Islamic culture and representing abstract concepts and concrete concepts pertaining to Oriental urban life. They represent all fields of traditional Islamic society. Many words are inherited from the Chaghatay literary tradition. Karakhanid was the first Turkic written language to contain Arabic‐Persian loans. Numerous loans are found in modern Azeri, South Oghuz, Khorasan Oghuz, Turkmen, Uzbek, Uyghur, Tatar, Bashkir, Kazakh, Noghay, Karakalpak, Kirghiz, Karaim, etc.

6.2.1  Examples of Persian Loans Azeri kor ‘blind’, ränk ‘color’, pul ‘money’ Uzbek gȯšt ‘meat’, låy ‘mud’, pȧrdȧ ‘curtain’, nån ‘bread’ Turkmen gül ‘flower’, šaːt ‘glad’ Tatar zur ‘big’, baḳča ‘garden’, atna ‘week’, šähär ‘city’, čaršaw ‘curtain’ Chuvash čaršav ‘curtain’ Kazakh apta ‘week’, künä ‘crime’, nan ‘bread’, bazar ‘market’ Karakalpak diydar ‘face’ Kirghiz künöː ‘sin’, šaːr ‘city’, baḳča ‘garden’ Noghay namart ‘treacherous’, nadan ‘ignoramus’, baḳša ‘garden’.

6.2.2  Arabic Loans Copied from Persian Azeri kef ‘pleasure’, dua prayer, ġädär ‘amount’ Uzbek mümkin ‘possible’, nihåyȧt ‘at last’, ḳuwwȧt ‘strength’ Tatar mäktäp ‘school’, taraf ‘side’, vaḳït ‘time’ Chuvash văχăt ‘time’ Kazakh aḳïl ‘intellect’, γïlïm ‘science’, maγïna ‘meaning’, waḳït ‘time’ Turkmen ïnθaːn ‘human being’, χat ‘letter’ Lithuanian Karaim jï̌ mat ‘congregation’, d́iń ‘religion’, kurban ‘sacrifice’, źerät́ ‘cemetery’ Noghay šaγat ‘witness’, ḳayrat ‘effort, zeal’, aḳïyḳat ‘truth’, ḳazna ‘treasure’. Ottoman Turkish displayed an overwhelming number of Arabic‐Persian loans, which ousted a considerable part of the native vocabulary. Even at the end of the Ottoman era, terms designating phenomena of the modern world, e.g. political and scientific terms, were mostly coined by means of Arabic devices. The modern Turkish lexicon still possesses a significant Arabic‐Persian component, though puristic language reform has weakened its dominance. Numerous old words have been abandoned and replaced by so‐called Öztürkçe (‘Pure Turkish’) neologisms. Azeri has preserved numerous words of Arabic‐Persian origin. The varieties of Iran also possess numerous copies from spoken Persian. There has not been any radical language reform comparable to the Turkish one. We thus find Azeri words such as lüγät ‘dictionary’ versus Turkish sözlük, müällim ‘teacher’ vs Turkish öğretmen, pul ‘money’ versus Turkish para. A good deal of the Turkish neologisms are not intelligible to native speakers of Azeri. South Oghuz and Khorasan Oghuz exhibit a strong Persian influence in their lexis. Turkmen and Uzbek copied words from Persian and inherited words from the old literary language Chaghatay. Long‐standing intensive Uzbek contacts with Iranian have resulted in numerous loanwords. Even female gender is expressed in some borrowed nouns, e.g. Uzbek šåirȧ ‘poetess’ vs šåir ‘poet’.

560  Lars Johanson, Éva Á. Csató, and Birsel Karakoç Modern Uyghur exhibits numerous Arabic‐Persian loans, introduced via urban varieties of Uzbek and through the Islamic literature, i.e. the common Central Asian heritage transmitted by the written language Chaghatay. Though the strong influence has now decreased, about one fifth of the vocabulary is still of Arabic‐Persian origin. In some Kipchak varieties, the amount of Arabic and Persian words is lower than in varieties whose speakers started a sedentary life by the time of their Islamization or even earlier. In Kazakh, Arabic‐Persian loans have entered via Tatar and Chaghatay. Karakalpak possesses additional loans introduced via Uzbek and Turkmen. Also Noghay has preserved many Arabic‐Persian words of the cultural vocabulary typical of the pre‐Soviet era. Even in Kirghiz, Arabic‐Persian lexical elements form a considerable part of the lexicon. They have been copied from Chaghatay and Uzbek, particularly into the southern dialects. The northern dialects, on which the standard language is based, are less influenced by the Islamic vocabulary. Middle Kipchak displays a great number of Persian words and Arabic words copied from Persian. Modern Kipchak Turkic languages, e.g. Tatar and Bashkir, have numerous loans of Arabic‐Persian origin. Most of the loans in Chuvash have entered via Tatar, though certain words were borrowed as early as the Volga Bulgar period (Scherner 1977). As a rule, Turkic languages outside the Islamic cultural sphere do not exhibit Arabic‐ Persian loans. However, some South Siberian languages offer a few exceptions, e.g. Khakas nan ‘bread’ or Altay Turkic urmat ‘reputation’, ultimately going back to Persian naːn and Arabic ḥ urmat, respectively.

6.3  Slavic Loans All languages in the Russian influence zone exhibit Russian loans. They mostly represent phenomena of modern life, aspects of Russian and European civilization, technical, scientific and administrative matters, and political and social concepts of the Soviet era. Many recent internationalisms have been borrowed via Russian. Russian loans mostly constitute a recent layer, introduced at the end of the nineteenth century at the earliest and during the Soviet era at the latest. In languages entertaining old contacts with Russia, words that have become archaisms in Russian may still be part of the active lexicon. Numerous loans are found in Yakut, Chuvash, Tatar, Bashkir, South Siberian Turkic, Kazakh, Kirghiz, Uzbek, Noghay, Uyghur (particularly outside Xinjiang), etc., e.g. Uzbek plan ‘plan’, stul ‘chair’, stakan ‘tumbler, glass’, studentka ‘female student’; Turkmen poθyolok ‘settlement’, ġaδyet ‘newspaper’, feːrma ‘farm’; Tatar par ‘steam’, kuχnya ‘kitchen’, vrač ‘medical doctor’; Chuvash χaśat ‘newspaper’, kĕneke ‘book’; Kazakh stol ‘table’, kerewet ‘bed’; Uyghur aptomobil ‘car’, kastum ‘costume, suit’. Russian impact on the modern vocabulary of Northern Azeri is strong, whereas it is almost absent in Southern Azeri, which has many loans from Persian. Thus Northern Azeri stol ‘table’ corresponds to Southern Azeri miz, aftobus ‘bus’ to otobus, kartof ‘potato’ to yer almasï, gäzet ‘newspaper’ to ruznamä, universitet ‘university’ to danïšgah, čaynïḳ ‘tea pot’ to čaydan, etc. Owing to the different political and cultural development over the last 600 years, the Northern Azeri and Turkish vocabularies differ from each other in many respects. Loans from Russian often correspond to Turkish loans from West European languages, e.g. Azeri zavod ‘factory’ versus Turkish fabrika, ġalstuk ‘necktie’ versus kravat, lampa ‘lamp’ versus lamba. In the post‐Soviet period there are tendencies to reduce Russian loans in favor of native or Arabic‐Persian words. But Russian is still the language of communication in professional domains, education, and science in many parts of the former Soviet Union. The native terminologies that are now being developed are thus often rather restricted to textbooks. In informal spoken registers, copies from Russian are still often used.

Turkic Language Contacts  561 Loans from other Slavic languages are numerous in dialects of the Balkans and some East European varieties. Numerous lexical items of Gagauz and Karaim are copied from Slavic. The lexicon of West Rumelian Turkish is heavily influenced by contacts with Slavic. Turkish used to be a rich lexical source for the Slavic contact languages, providing words relating to everyday life as well. This situation has changed radically. In Bulgaria, campaigns have been conducted against the use of Turkish words in local Turkish dialects and in favor of their replacement by Bulgarian words. The change in dominance relations is shown by the fact that colloquial West Rumelian Turkish is reborrowing loans in the shape they have in the languages of Macedonia and Kosovo that originally borrowed them, e.g. piper ‘pepper’ instead of biber, sapun ‘soap’ instead of sabun, pita ‘flat bread’ instead of pide, tašliya ‘stony’ instead of tašlï (Friedman 2006: 42).

6.4  Mongolic Loans Mongolic loans are found in many Turkic languages, Yakut, Tatar, Bashkir, Noghay, Kazakh, Kirghiz, South Siberian Turkic, Uyghur, etc. For an overview, see Schönig (2003); on loans in West Oghuz Turkic, see Schönig (2000). Languages spoken by nomadic groups have borrowed numerous Mongolic words as a result of close contacts especially in the Middle Ages. A number of languages have loans of Middle Mongolian origin. The literary language Chaghatay displays loans from the domains of warfare and administration (Kincses‐Nagy 2018). Most Yakut words of foreign origin are Mongolic loans. There is an old Buryat layer from the early period of settlement on the shore of Lake Baikal. Kirghiz possesses numerous old loans, e.g. sonun ‘remarkable’, dülöy ‘deaf’, belen ‘ready’. Owing to close contacts with other nomadic groups, Kazakh exhibits many words of Mongolic origin, e.g. olj̆a ‘booty’, ḳunan ‘colt in the third year’. Most of them date back to the eighteenth century, when Kazakh and western Mongol tribes fought in the steppes. The same is true for Noghay, which exhibits numerous Mongolic lexical loans that mostly originate from the Middle Mongolian period and are often related to animal breeding, kinds of animals, organs or diseases, e.g. bora ‘male camel’, dönen ‘a four‐year‐old male animal’, γunj̆ïn ‘a three‐year heifer’, kabïrγa ‘rib’, or to state administration and military organization (Menges 1957; Birtalan 1994). Languages of the Volga region also exhibit loans from Middle Mongolian, e.g. Tatar uram ‘street’, dala ‘steppe’. Chuvash copied its words of Mongolic origin from Tatar after the thirteenth century. Yellow Uyghur, spoken in western China, exhibits direct loans from neighboring Mongolic varieties. In South Siberian Turkic, Mongolic cultural loans often play a role comparable to that of Arabic‐Persian loans in Islamic Turkic languages. There is an older Mongolic layer common to all Sayan varieties. Even köz ‘eye’ has been replaced by ġaraḳ throughout the Sayan area. A West Mongolian layer goes back to the Oirat rule of the fifteenth to seventeenth centuries. There are also borrowings from Buryat, particularly in Tofan. Tuvan possesses loans from Mongolic in all word classes. It exhibits strong Khalkha influences dating back to the rule of the Altan Khans. Some recent words copied from Khalkha from 1930 to 1945 reflect modern spoken forms and are usually neologisms used as synonyms of Russian loans, e.g. χuviskaːl ‘revolution’ for revolyucya. Dukhan, spoken in northern Mongolia, has been subject to strong influence from Khalkha and Darkhad Mongolian during the last six decades. Many of its loans have abstract meaning, e.g. šašïn ‘religion’, j̆anj̆ïl ‘habit’, domoḳ ‘legend’, j̆ayan ‘destiny’, amdïral ‘life’ (Ragagnin 2008).

6.5  Chinese Loans On Chinese loans in Old Uyghur, see Menges (1968: 168–9). The modern Turkic languages spoken in China are strongly influenced by Chinese (Abish 2016; Abish and Csató 2010; Memtimin 2016). Older Chinese loans are found in all varieties of modern Uyghur (Jarring

562  Lars Johanson, Éva Á. Csató, and Birsel Karakoç 1964), but the lexical influence has become increasingly stronger. Most loans have entered through recent contacts with western Mandarin dialects of Xinjiang. Many of them denote items introduced by Chinese immigrants, e.g. j̆oza ‘table’, manta ‘stuffed bun’. Many Chinese neologisms denoting technological, political, bureaucratic, and military innovations have been copied. Not all are used in spoken registers, and most are lacking in Uyghur varieties outside Xinjiang. In the 1960s, the use of Chinese scientific terminology was obligatory. There is now a tendency to replace Chinese words by means of Turkic word formation devices and loan translations. The spoken varieties of Kazakh and Kirghiz spoken in Xinjiang exhibit a certain number of Chinese loans, but the written varieties are relatively little affected. Yellow Uyghur and Salar, spoken outside Xinjiang in western China, display various loans from neighboring Chinese varieties.

6.6  Uralic Loans Turkic languages of the Volga region, particularly Chuvash, possess many elements copied from neighboring Volga‐Finnic varieties. Mari loans are very common in Upper Chuvash, especially the Sundyr dialect, testifying to a Finnic substrate due to the assimilation of a local population, e.g. lĕpĕ ‘butterfly’, yantar ‘glass’, pürt ‘house’. Many of the copied words have now vanished in Mari. Mari also has a large number of Turkic loanwords. For a discussion of the initial contacts, see Róna‐Tas (1988). Tatar and Bashkir words of Finno‐Ugric origin mainly occur in the dialects. A layer of loanwords in Chuvash indicates old contacts with Samoyedic varieties in southwestern Siberia. The modern South Siberian languages exhibit South Samoyedic loans as substrate products. The loans primarily belong to the sphere of reindeer‐breeding, hunting, fishing, and botany (Helimski 1995).

6.7  Other Loans Words of Greek and Armenian origin are found in many Turkic languages of the western ́ esa ‘temple’, midraš sphere. Karaim has an old layer of Hebrew loanwords, for instance keń ‘religious school’, t́efiŋla ‘prayer’, γazzan ‘religious leader’, and pesaχ ‘Pesach’, and many copies from Lithuanian. Balkan Turkic varieties possess loans from Albanian and Rumanian. Yakut and other Siberian languages have loans from Tungusic, often belonging to the domains of husbandry and everyday life. Many Yakut elements are probably due to contact with Paleoasiatic languages. Words of Yeniseic origin are found in South Siberia. Salar and Yellow Uyghur display many Tibetan loans from neighboring languages. There are many words of unknown origin, e.g. in Chuvash. Kipchak varieties spoken in the Caucasus area have copied numerous words from the neighboring Caucasian languages and the Iranian language Ossetic. They pertain, in particular, to the sphere of everyday life, e.g. Karachay‐Balkar adaḳa ‘cock, roaster’, čubur ‘a short male garment’, čille ‘silk’, ġaḳḳï ‘egg’. On the other hand, Turkic words have been copied into various contact languages of the region and are often well adapted to the phonological systems of the recipient languages. Turkish, the most developed Turkic language, exhibits, in addition to the loans already mentioned, numerous lexical elements, e.g. technical, scientific and political terms, from Greek, Italian, and French, and after World War II also significantly more from English (Tietze 1990). For Turkish nautical terms of Italian and Greek origin, see Kahane, Kahane, and Tietze (1958).

Turkic Language Contacts  563

6.8  Copied Verbs Verbal stems are seldom copied as such, e.g. Kirghiz ḳara‐ ‘to look’, toḳto‐ ‘to stop’, Tuvan ägälä‐ ‘to start’. A more common way to accommodate copies to function as verbs is derivation of nominals, i.e. nouns and adjectives, by means of suffixes, e.g. Turkmen harč‐la‐ ‘to spend’ < harč ‘expenditure’. Verbs can also be formed analytically by means of auxiliary “light” verbs, e.g. Turkmen harč et‐ ‘to spend’. These compounds are lexicalized verbal phrases which form one syntactic constituent. The nominal element does not function as a free object, and it normally cannot be separated from the auxiliary verb by other elements than particles meaning ‘also’ or ‘even’. Light verbs meaning ‘to do’ include ät‐, äylä‐, ḳïl‐, ḳïn‐, yap‐, etc. The nominal is a petrified element, often an Arabic verbal noun, e.g. Azeri niyyet elä‐ ‘to intend’ < niyyet ‘intention’, or a Slavic infinitive, e.g. Karaim zvont’ et‐, Yakut svoni ġïn‐ ‘to phone’ < zvonit’, bïrasti ġïn‐ ‘to apologize’ < prostit’, Armeno‐Kipchak vïkupit et‐ ‘to ransom’ < vykupit’. This type is common as early as Old Uyghur, e.g. ḳšanti ḳïl‐ ‘to confess’. Other examples: Turkish memnun et‐ ‘to satisfy’, spor yap‐ ‘to do sport’, Kazakh ümit ḳïl‐ ‘to hope’, Middle Kipchak namaːz ḳïl‐ ‘to pray’, niyyet elä‐ ‘to intend’, Ottoman irsaːl et‐ ‘to send’, Karachay‐ Balkar razï et‐ ‘to satisfy’. Ottoman Turkic uses ät‐, äylä‐, and ḳïl‐ ‘to do’. Modern Turkish normally uses et‐, but has also introduced yap‐. Azeri displays elä‐, et‐, and ġïl‐. The verb ät‐ is used in the rest of Oghuz Turkic, in Kipchak Turkic, Kipchakoid South Siberian varieties, and Salar. It is not used in Sayan Turkic and Yakut. The verb äylä‐ < äδlä‐ is still used in the west, mainly in Southwest Turkic and Northern Kipchak Turkic, e.g. Azeri pušide elä‐ ‘to cover’, cf. Persian pušiːdan. The old auxiliary ḳïl‐is preserved in Southeast Turkic and Sayan Turkic. Oghuz and Kipchak ḳïl‐ is mostly confined to old formations of elevated style, e.g. Turkish namaz ḳïl‐ ‘to perform ritual worship’. The form ḳïn‐ is found in the Sayan Turkic language Tofan, ġïn‐ in North Siberian Yakut. The old verb tu‐ ‘to do’ is used in Chuvash, e.g. astu‐ ‘to remember’ < as ‘memory’. Noghay uses the light verb et‐ ‘to do’, e.g. älek et‐ ‘to destroy’, kïyanatlïḳ et‐ ‘to deceive, to betray’. Turkish varieties spoken in northwestern Europe have copied verbs by combining nominal forms with light verbs such as yap‐ ‘to do’, or to a lesser extent et‐ ‘to do’, e.g. vergessen yap‐ ‘to forget’, erinnern et‐ ‘to remember’ in Turkish as spoken in Germany (Pfaff 2000). Examples from Turkish spoken in Sweden include jaga yap‐ ‘to hunt’, släppa yap‐ ‘to release, to let go’, skada yap‐ ‘to damage’. In speech of some Turkish‐Swedish children, even inflected verb phrases are found in combination with yap‐, e.g. klättrade upp yaptı ‘X climbed up’ (Bohnacker and Karakoç, forthcoming). Intransitive verbs are formed with nominals, often participial forms, plus bol‐ ~ pol‐ ~ ol‐ ‘to be, to become’, e.g. Ottoman Turkic naːil ol‐ ‘to obtain’, zaːyi ol‐ ‘to be lost’, šaːd ol‐ ‘to rejoice’. Iranian languages use similar methods to accommodate borrowed Turkic verbs, combining a nominal form with a native auxiliary verb such as ‘to do’, e.g. Tajik ämr kärdän ‘to give an order’. It is also common to use nominal forms in the now unproductive Turkic marker ‐miš to copy Turkic verbs, e.g. Tajik tugulmiš kärdän ‘to be born’, cf. Uzbek tuγil‐.

6.9  Copied Function Words Turkic languages, both spoken and written, have shown a propensity for copying function words, e.g. conjunctions, postpositions, and discourse markers, from their contact languages. Languages of the Islamic sphere have copied from Persian, particularly all Persian‐influenced varieties of Iran; those of the Russian sphere of influence have also copied from Russian, e.g. no ‘but’, i ‘and’, ili ‘or’ (Johanson 1997): Uzbek gȯyå ‘as if’; Turkmen we ‘and’, χem ‘also’;

564  Lars Johanson, Éva Á. Csató, and Birsel Karakoç Tatar häm ‘and’, ämmä ‘but’, ägär ‘if’; Noghay em … em … ‘both … and …’; Turkish ve ‘and’, çünkü ‘for’, gerçi ‘though’, eğer ‘if’, no ‘but’; West Rumelian Turkic a ‘and, but’. Many borrowed junctors are complex, e.g. taː ki ‘until’. Some varieties in Iran have even copied Persian prepositions, e.g. Khalaj biː sän ‘without you’. The syntactic features of Slavic subordinative junctors are sometimes selectively copied (as “calques”), e.g. Gagauz ani ‘what, which’. Russian discourse particles are often used in spoken Noghay and Kazakh. Copying of temporal, purposive, causal and other conjunctions is connected with a reduced use of Turkic nonfinite constructions. The elements are often integrated into Turkic syntax in a way different from their behavior in the model languages. The free junctor ki ‘that’ (etc.) has a broad functional scope, serving as a general connector preceding relative and complement clauses. The Volga‐Finnic language Mari has copied postpositions and particles from Tatar and Chuvash, e.g. körä ‘in view of’, ‘because of’, the superlative particle en, and the interrogative particle mo.

6.10  Copied Affixes Copying of bound morphology is in general known to be relatively unattractive, but there are clear instances of non‐Turkic languages copying bound morphemes from Turkic, including both derivational morphemes and inflectional morphemes. This may be due to the agglutinative nature, the low level of fusion, of Turkic morphology, i.e. the frequent one‐on‐ one correspondence between grammatical categories and their exponents. Suffixes seem to be more or less pervious to copying depending on their position in the word. Turkic derivational morphemes such as the agentive nominalizer ‐či have been copied into many contact languages, e.g. Lezgian alver ‘trade, deal’, alverči ‘merchant’ < alber and alberči respectively in Kumyk; Abkhazian amalčï, Kabardian amalšï ‘enterprising, pushing’ < amalčï in Karachay‐Balkar; Adyghe amalsïz ‘hopeless, incurable, irreparable’ < Karachay‐ Balkar amalsïz. On the other hand, Balkan Turkic shows Slavic borrowings, e.g. markers of feminine nouns and diminutives. Turkic has copied comparative markers, e.g. Iranian Turkic ‐tar, whereas Tajik uses the corresponding Turkic suffix ‐råq. Kashkay has copied the suffix ‐(y)akï, a marker of familiarity in a variety of Lori spoken in the Kashkay confederation (Dolatkhah et al. 2016). In spite of the dominant suffixing morphology, the strong Iranian influence on Uzbek has led to the emergence of prefixes, e.g. nå‐toγri ‘untrue’ with nå‐ ‘non‐’ copied from Persian + Turkic toγri ‘right, true’. Inflectional morphemes, e.g. case and person‐ number markers, have been copied into varieties of Tajik and Anatolian Greek, respectively. The bound morphology of Mari has undergone strong Turkic impact, e.g. copying of word‐ formative and case suffixes. Copied Persian suffixes can be found in various Turkic languages. Turkmen, spoken in Central Asia, has several copied suffixes that do not only attach to Arabic or Persian stems but also to Turkic stems, e.g. ‐daːr and ‐hoːr in alġïdaːr ‘creditor’, bergidaːr ‘debtor, indebted’; süythoːr ‘the one who drinks a lot milk’, ġaːnhoːr ‘murderer’.

6.11  Adaptation of Loans Loanwords are subject to phonological adaptation (integration, assimilation) of different kinds in Turkic languages, i.e. the pronunciation is more or less accommodated to the indigenous sound systems. There are different styles or registers of pronunciation according the speakers’ proficiency in the model languages. Monolinguals have mostly tended to adapt the shape of loans to their native systems. Thus, older Russian loans show a high degree of adaptation, e.g. Yakut sïlabaːr ‘samovar’, bïragraːmma ‘program’; Kazakh kerewet ‘bed’ < Russian krovat’; Chuvash apat < Russian obed

Turkic Language Contacts  565 ‘dinner’. Loans belonging to the domains of modern life are less adapted. Items copied before the formation of the modern standard languages are often written as they were pronounced at the time of copying. The adaptation is sometimes reflected in old Arabic‐ and Roman‐based orthographies. Since the introduction of the Cyrillic script, Russian loans are written in their original graphic shape, which obscures possible adaptations. Though orthoepic norms recommend native Russian pronunciation, the loans may still be pronounced according to indigenous rules. Arabic‐Persian loans confront us with corresponding problems. The Arabic script often represents them in shapes indistinguishable from those of the model languages. Turkic languages differ with respect to the degrees of adaptation. Items copied directly from spoken Persian have generally accommodated closer to the Turkic phonological system, e.g. Kirghiz ubaḳtï ‘time’ < Arabic waqt, j̆oːp ‘answer’ < Arabic j̆awaːb (Johanson 1986), Noghay šaγat ‘witness’ < Arabic šaːhid. It is unknown to what extent lexical copies in the Old Uyghur language were adapted to Turkic phonology and phonotactics, or stand for marginal sound structures in learned pronunciation. Certain marginal sounds only occur in loans and are not common to all spoken styles. Long vowels appear in, at least conservative, pronunciation of Arabic‐Persian loans. The consonant šč, written щ, mostly pronounced as a long palatal fricative šː, occurs in more recent Russian loanwords. The affricate c (ts) occurs in loans from Russian and Chinese. The labiodental fricatives f and v are marginal in many languages. The glottal h occurs primarily in words of Arabic‐Persian origin. Initial h‐ is sometimes dropped, e.g. Noghay ȧr ‘each’ < har, Kirghiz apta ‘week’ < hafta. The consonant ž occurs primarily in words of Persian and French origin, e.g. Azeri žäst ‘gesture’ < geste. The glottal stop ʔ in words of Arabic‐Persian origin is realized in certain languages, e.g. Tatar mäsʔälä ‘question’, Uyghur täbiʔi ‘natural’. It may even occur in learned Turkish pronunciation, e.g. saʔat ‘hour’. Noninitial glottal stops also occur in Uyghur loans from Chinese, e.g. faŋʔän ‘scheme’. Also the pharyngeal ʕ is realized as a glottal stop. Both consonants may be lost, sometimes compensated by vowel length, e.g. Azeri teːsir ‘influence’, Uyghur mäsilä ‘question’. In certain languages, they are represented by the voiced fricative γ, e.g. Tatar γadät ‘habit’, šaγir ‘poet’, Noghay yamaγat ‘congregation’. Loans are more or less adapted to Turkic phonotactic rules. Though certain consonants, e.g. n and r, do not occur initially in native words, they may be realized as such in loans, e.g. Yakut naːda ‘necessary’. The initial position can be avoided by prothetic vowels, e.g. Turkish istasyon ‘station’, Turkmen ïrayoːn ‘district’, Kazakh ïras ‘true’ < Persian rast, Noghay orïs ‘Russian’. Nonpermissible consonant clusters are dissolved through consonant deletion or insertion of epenthetic or prothetic vowels, e.g. Turkmen uθθol ‘table’ < stol; Uyghur pikir ‘thought’ < fikr, čilen ‘member’ < člen’; Kazakh χalïḳ ‘people’ < χalq; Tatar dus < dost ‘friend’; Turkish isim ‘name’ < ism. In copied words, Noghay prevents the final consonant clusters by using epenthetic vowels, e.g. deris ‘lesson’ < ders, borïš ‘debt’ < borč (Karakoç 2005, 2013).

6.12  Phonological (Material) Copies The introduction of less adapted loanwords has often affected the phonological systems of Turkic languages. Two examples: Gagauz and Karaim have acquired palatalized front consonants due to Slavic influence. Chuvash displays palatalized consonants before and after front vowels. In Istanbul and Balkan dialects, k and g are palatalized before front vowels, e.g. g’äl ‘come’, k’itap ‘book’. Urban Uzbek, Iranian Turkic, West Rumelian Turkish, etc. possess vowels influenced by Iranian or Slavic, retracted variants of ö und ü, and absence of ï. The typical Turkic sound harmony structures have been disturbed in many dialects under Persian influence.

566  Lars Johanson, Éva Á. Csató, and Birsel Karakoç

6.13  Copies of Combinational Properties Copying of combinational properties has played an important role in Turkic contact situations. They have affected word structures and syntactic constructions, e.g. patterns of word order and clause combining. Copying of combinational and semantic properties has led to restructuring of morphosyntactic subsystems, e.g. aspect‐mood‐tense and case systems (Boeschoten and Johanson 2006). The contacts have sometimes been sufficiently intensive and long‐lasting to provide Turkic varieties with grammatical components strongly modeled on non‐Turkic patterns. Examples include Slavic influence on Karaim, Persian influence on Kashkay, etc. Vice versa, non‐Turkic languages have developed grammatical components patterned after Turkic models. One case in point is the Turkish influence on Greek dialects spoken in Central Anatolia. Copying of combinational and semantic properties will not be dealt with in length here. For details, see Johanson (2002, 2008). However, the following are some examples. Word order has been an important domain of contact influence. Shifts in sentential word order have, however, mostly affected the use of existing structures rather than leading to acquisition of new structures. Pragmatically marked word orders have often been treated as unmarked. In Karaim and Gagauz, strong foreign influence has led to the weakening of the basic verb‐final word order in favor of the verb‐object order. Whereas Turkic generally places relative clauses before their heads, Irano‐Turkic varieties prefer postposed relative clauses of the Persian type. Owing to Persian influence, some Turkic varieties spoken in Iran display prepositions. Karaim and some Balkan varieties display cases of genitive–head reversal, e.g. West Rumelian Turkish baba‐si Ali’nin ‘the father of Ali’ , Lithuanian Karaim orun‐lar‐ï karay‐lar‐nïn ‘the places of the Karaims’, but biź‐ń iń ata‐mïz ‘our father’. Morphosyntactic patterns have proven highly susceptible to contact‐induced changes. Under Slavic influence, the suffixed form of the Karaim postposition bïla ‘with’, ‐B2A2, has become a case. Karaim has copied the combinational properties of the Slavic instrumental case. Thus, for instance in copular clauses, the instrumental case is assigned to the predicative element, e.g. Ol vaχt‐ta e‐d’i üŕät́u̇ v́ću̇ ̌ ‐bä ‘At that time, he was a teacher’. Many Turkic languages have been strongly affected in the domain of complex sentences. The weakening or loss of nonfinite constructions, e.g. Turkish gel‐diğ‐in‐i duy‐du‐m ‘I heard that (s)he has come’, is a typical tendency of Turkic in contact with Indo‐European, e.g. Karaim tuy‐du‐m k’i k’el’‐d’i . For the less frequent use of nonfinite forms in Turkish as spoken in northwestern Europe see Rehbein, Herkenrath, and Karakoç 2009; Bohnacker and Karakoç (forthcoming). Conjunction‐marked postposed clauses with finite verb morphology have been introduced. In West Rumelian Turkish, this expanded usage is calqued from subordinate clauses in Macedonian and Albanian. Discourse motivations for calquing of relativizing and coordinating conjunctions is discussed in Matras (2006). Some non‐Turkic varieties, e.g. Tajik Persian, have acquired a head‐final clause subordination system with patterns copied from Turkic syntax. Turkic emulations of foreign hypotaxis are often subject to constraints, the constructions being integrated in a way different from those in the model languages. The junctors (relators, relativizers) used, e.g. ki ‘that’, do not always mark embedded clauses in the way Turkic subordinative devices do. The question of whether or not the calqued constructions represent genuine hypotaxis should be investigated further; see Johanson (1975).

Turkic Language Contacts  567

6.14  Copies of Frequential Properties Frequential patterns of features which the contact languages seem to have in common may be copied. They may cause increased or reduced use of anaphoric units, conjunctions, etc. Turkish spoken in Germany, for example, exhibits frequent use of the conjunction ve ‘and’ under the influence of German und. Lithuanian Karaim employs two possessive construć aχčam bart . The frequency of the locational construction has become higher under the influence of Russian, which only has the locational construction U menja (jest’) mnogo deneg ‘I have a lot of money’. The lack of a category in a contact language can cause the loss of the category in Turkic. Thus, for instance, peripheral Turkic languages such as Karaim and Turkic varieties of Iran have lost the typical Turkic evidential category. For the restricted use of evidential markers in Turkish as spoken in Germany, see Karakoç and Herkenrath (2019).

7 Conclusions Turkic languages present particularly rich sources of data for the study of language contact. The continuous massive displacements of Turkic‐speaking groups throughout their history have led to numerous encounters, to various contact‐induced processes of change and shift, and to new linguistic configurations under different sociocultural conditions. Though speakers of Turkic often entered new areas in the course of political expansion, this was far from always tantamount to linguistic expansion. There were major, massive migratory movements and minor movements of thin ruling layers. In some cases, local speaker groups shifted to the incoming codes; in other cases, incoming groups shifted to the local codes. Varying and intensive intrafamily contacts between highly mobile Turkic‐speaking groups led to reciprocal influence, to splits between closely related varieties, and to new constellations, in particular in the old heterogeneous tribal confederations. Turkic entertained close interfamily contacts with an astonishing number of genealogically and typologically different languages: Indo‐European such as Iranian and Slavic, non‐ Indo European such as Mongolic, Finno‐Ugric, and Tungusic. Major contact areas include Central Asia, Siberia, the Volga‐Kama region, Caucasus, Transcaucasia, Iran, Anatolia, the Balkans, and, recently, northwestern Europe. Certain contacts have been sufficiently intensive and long‐lasting to produce thorough linguistic changes. The grammatical components of some Turkic languages are almost totally modeled on non‐Turkic patterns; those of some non‐Turkic languages similarly patterned after Turkic. Factors determining linguistic copying have in part been social, “prestige” factors, and in part structural, “attractive” features. Under appropriate social circumstances, almost any features seem to have been copied, even those that appear to be typologically inconsistent with the rest of the structure of the copying language. Various types of copied features have been mentioned above. Turkic has copied parts of its lexicon, even function words, from Arabic‐Persian, Slavic, Mongolic, Chinese, Uralic, etc. Loanwords have been subject to phonological adaptation of different kinds. The introduction of less adapted loanwords has often affected the phonological systems. Copies of combinational properties have affected word structures and syntactic constructions, e.g. patterns of word order and clause combining. Copying of combinational and semantic

568  Lars Johanson, Éva Á. Csató, and Birsel Karakoç properties has resulted in restructuring of morphosyntactic subsystems, e.g. aspect‐mood‐ tense and case systems. Turkic language contacts have long been subject to detailed investigation, most intensively in the last decades. A good deal of research still remains to be done, in particular since the amplitude and the richness of the contact‐induced phenomena observed here appear to be of paradigmatic value for the study of language contact in general.

REFERENCES Abish, Aynur 2016. Modality in Kazakh as Spoken in China (Turcologica 107). Wiesbaden: Harrassowitz. Abish, Aynur and Éva Á. Csató 2010. Recent developments in Kazakh as spoken in the People’s Republic of China. Turkic Languages 15: 275–90. Agyagási, Klára 1998. On the characteristics of the Cheremiss linguistic interferences on Chuvash. In Lars Johanson (ed.), The Mainz Meeting: Proceedings of the Seventh International Conference on Turkish Linguistics, August 3–6, 1994 (Turcologica 32). Wiesbaden: Harrassowitz, pp. 667–82. Agyagási, Klára 2019. Chuvash Historical Phonetics (Turcologica 117). Wiesbaden: Harrassowitz. Andrews, Peter Alford (ed.) 1989. Ethnic Groups in the Republic of Turkey (Beihefte zum Tübinger Atlas des Vorderen Orients B, 60). Wiesbaden: L. Reichert. Backus, Ad 2013. Turkish as an immigrant language in Europe. In Tej K. Bhatia and William C. Ritchie (eds.), The Handbook of Bilingualism and Multilingualism. Oxford: Blackwell, pp. 770–90. Birtalan, Ágnes 1994. Further remarks on the Mongolian loanwords in Nogai. In Ingeborg Baldauf and Michael Friederich (eds.), Bamberger Zentralasienstudien (Islamkundliche Untersuchungen 185). Berlin: Klaus Schwarz Verlag, pp. 37–44. Bohnacker, Ute and Birsel Karakoç forthcoming. Subordination in children acquiring Turkish as a heritage language in Sweden. In Fatih Bayram and Deniz Tat (eds.), Studies in Turkish as a Heritage Language. Amsterdam: Benjamins. Boeschoten, Henrik and Lars Johanson (eds.) 2006. Turkic Languages in Contact (Turcologica 61). Wiesbaden: Harrassowitz. Brendemoen, Bernt 2002. The Turkish Dialects of Trabzon: Their Phonology and Historical Development (Turcologica 50). Wiesbaden: Harrassowitz.

Bulut, Christiane 2006. Syntactic traces of Turkic‐Iranian contiguity: an areal survey of language contact induced shift in patterns of relativization. In Lars Johanson and Christiane Bulut (eds.), Turkic‐Iranian Contact Areas: Historical and Linguistic Aspects (Turcologica 62). Wiesbaden: Harrassowitz, pp. 165–208. Bulut, Christiane 2018. Linguistic Minorities in Turkey and Turkic Minorities of the Periphery (Turcologica 111). Wiesbaden: Harrassowitz. Comrie, Bernard 2002. Introduction. In Lars Johanson, Structural Factors in Turkic Language Contacts. Richmond: Curzon, pp. i–xii. Csató, Éva Á. 2000. Some typological features of the viewpoint aspect and tense system in spoken north‐western Karaim. In Östen Dahl (ed.), Tense and Aspect in the Languages of Europe. Berlin: de Gruyter Mouton, pp. 671–99. Csató, Éva Á. 2016. 189. Karaim. In Peter O. Müller et al. (eds.), Word Formation. International Handbook of the Languages of Europe. (Handbooks of Linguistics and Communication Science, HSK). Berlin: de Gruyter, pp. 3379–88. Csató, Éva Á., Bo Isaksson, and Carina Jahani (eds.) 2004. Linguistic Convergence and Areal Diffusion: Case Studies from Iranian, Semitic and Turkic. London: RoutledgeCurzon. Csató, Éva Á., and Astrid Menz 2018. On the linguistic distances between Gagauz and Karaim. Turkic Languages 22: 43–62. Dawkins, Richard M. 1916. Modern Greek in Asia Minor: A Study of the Dialects of Sillí, Cappadocia and Phárasa with Grammar, Texts, Translations, and Glossary. Cambridge: Cambridge University Press. Doerfer, Gerhard 1967. Türkische Lehnwörter im Tadschikischen (Abhandlungen für die Kunde des Morgenlandes 37.3). Wiesbaden: Steiner. Dolatkhah, Sohrab, Éva Á. Csató, and Birsel Karakoç 2016. On the marker -(y)akï in the Kashkay variety spoken in the Jamehbuzurgi subtribe of the Amaleh tribe. In Éva Á. Csató,

Turkic Language Contacts  569 Lars Johanson, András Róna‐Tas, and Bo Utas (eds.), Turks and Iranians. Interactions in Language and History (Turcologica 105). Wiesbaden: Harrassowitz, pp. 283–95. Friedman, Viktor A. 2003. Turkish in Macedonia and Beyond: Studies in Contact, Typology and Other Phenomena in the Balkans and the Caucasus (Turcologica 52). Wiesbaden: Harrassowitz. Friedman, Viktor A. 2006. West Rumelian Turkish in Macedonia and adjacent areas. In Henrik Boeschoten and Lars Johanson (eds.), Turkic Languages in Contact (Turcologica 61). Wiesbaden: Harrassowitz, pp. 27–45. Helimski, Eugen 1995. Samoyedic loans in Turkic: check‐list of etymologies. In Barbara Kellner‐Heinkele and Marek Stachowski (eds.), Laut‐ und Wortgeschichte der Türksprachen (Turcologica 26). Wiesbaden: Harrassowitz, pp. 75–95. Jarring, Gunnar 1964. An Eastern Turkic‐English Dialect Dictionary. Lund: C. W. K. Gleerup. Johanson, Lars 1975. Some remarks on Turkic “hypotaxis.” Ural‐Altaische Jahrbücher 47: 104–18. Johanson, Lars 1986. Reproduktion, Widerstand und Anpassung: Zur lautlichen Iranisierung im Türkischen. In Rüdiger Schmitt and Prods Oktor Skjærvø (eds.), Studia Grammatica Iranica: Festschrift für Helmut Humbach. Munich: Kitzinger, pp. 185–201. Johanson, Lars 1997. Kopien russischer Konjunktionen in türkischen Sprachen. In Dieter Huber and Erika Worbs (eds.), Ars transferendi. Sprache, Übersetzung, Interkulturalität. Frankfurt‐am‐Main: Peter Lang, pp. 115–21. Johanson, Lars 2002. Structural Factors in Turkic Language Contacts. Richmond: Curzon. Johanson, Lars 2008. Remodeling grammar: copying, conventionalization, grammaticalization. In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages (Hamburg Studies on Multilingualism 7). Amsterdam: John Benjamins, pp. 61–79. Johanson, Lars and Christiane Bulut (eds.) 2006. Turkic‐Iranian Contact Areas: Historical and Linguistic Aspects (Turcologica 62). Wiesbaden: Harrassowitz. Kahane, Henry, Renée Kahane, and Andreas Tietze 1958. The Lingua Franca in the Levant: Turkish Nautical Terms of Italian and Greek Origin. Urbana: University of Illinois Press. Karakoç, Birsel 2005. Das finite Verbalsystem im Nogaischen (Turcologica 58). Wiesbaden: Harrassowitz.

Karakoç, Birsel 2013. Az konuşurlu bir Türk dili: Nogayca ve Nogay halkı üzerine [A small Turkic language: About Noghay and the Noghays]. Tehlikedeki Diller Dergisi 2. Journal of Endangered Languages 2: 273–301. Also published in Süer Eker and Ülkü Çelik Şavk (eds.) (2016) Tehlikedeki Türk Dilleri (Endangered Turkic Languages) 3. Ankara, Astana, pp. 65–80. Karakoç, Birsel and Annette Herkenrath 2019. Understanding retold stories: the marking of unwitnessed events in bilingual Turkish. Turkic Languages 23: 81–121. Kincses‐Nagy, Éva 2018. Mongolic Copies in Chaghatay (Turcologica 115). Wiesbaden: Harrassowitz. Kıral, Filiz 2001. Das gesprochene Aserbaidschanisch von Iran: Eine Studie zu den syntaktischen Einflüssen des Persischen (Turcologica 43). Wiesbaden: Harrassowitz. Matras, Yaron 2006. Layers of convergent syntax in Macedonian Turkish. In Hendrik Boeschoten and Lars Johanson (eds.), Turkic Languages in Contact (Turcologica 61). Wiesbaden: Harrassowitz, pp. 46–62. Memtimin, Aminem 2016. Language Contact in Modern Uyghur (Turcologica 108). Wiesbaden: Harrassowitz. Menges, Karl Heinrich 1957. Zum lexikalischen Bestand der aralo‐kaspischen Sprachen. Studia Altaica. Festschrift für Nikolaus Poppe zum 60. Geburtstag am 8. August 1957 Wiesbaden: Harrassowitz. Menges, Karl Heinrich 1968. The Turkic Languages and Peoples: An Introduction to Turkic Studies (Ural‐Altaische Bibliothek 15). Wiesbaden: Harrassowitz. Menz, Astrid 1999. Gagausische Syntax: Eine Studie zum kontaktinduzierten Sprachwandel (Turcologica 41). Wiesbaden: Harrassowitz. Miklosich, Franz von 1884–90. Die türkischen Elemente in den südost‐ und osteuropäischen Sprachen (Griechisch, Albanisch, Rumunisch, Bulgarisch, Serbisch, Kleinrussisch, Großrussisch, Polnisch), 2 vols. and 2 supplements (= Denkschriften der Kaiserlichen Akademie der Wissenschaften, Philosophisch‐Historische Classe). Vienna: Tempsky. Olach, Zsuzsanna 2013. A Halich Karaim Translation of Hebrew Biblical Texts (Turcologica 98). Wiesbaden: Harrassowitz. Pfaff, Carol 2000. Development and use of et‐ and yap‐ by Turkish/German bilingual children. In Aslı Göksel and Celia Kerslake (eds.) Studies on Turkish and Turkic Languages (Turcologica 46). Wiesbaden: Harrassowitz, pp. 365–73.

570  Lars Johanson, Éva Á. Csató, and Birsel Karakoç Ragagnin, Elisabetta 2008. Dukhan, a Turkic Variety of Northern Mongolia: Description and Analysis (Turcologica 76). Wiesbaden: Harrassowitz. Rehbein, Jochen, Annette Herkenrath, and Birsel Karakoç 2009. Turkish in Germany. On contact‐induced language change of an immigrant language in the multilingual landscape of Europe. Language Typology and Universals. Sprachtypologie und Universalienforschung. STUF. 62.3: Multilingualism and Universal Principles of Linguistic Change (Gisella Ferraresi and Esther Rinke, ed.), pp. 171–204. Róna‐Tas, András 1988. Turkic influence on the Uralic languages. In Denis Sinor (ed.), The Uralic Languages: Description, History and Foreign Influences (Handbook of Uralic Studies, Handbuch der Orientalistik 8). Leiden: Brill, pp. 742–80. Róna‐Tas, András and Árpád Berta 2011. West Old Turkic. Turkic Loanwords in Hungarian 1–2. (Turcologica 84). Wiesbaden: Harrassowitz. Scherner, Bernd 1977. Arabische und neupersische Lehnwörter im Tschuwaschischen: Versuch einer

Chronologie ihrer Lautveränderungen (Akademie der Wissenschaften und der Literatur. Veröffentlichungen der Orientalischen Kommission 29). Wiesbaden: Franz Steiner. Schönig, Claus 2000. Mongolische Lehnwörter in Westogusischen (Turcologica 47). Wiesbaden: Harrassowitz. Schönig, Claus 2003. Turko‐Mongolic relations. In Juha Janhunen (ed.), The Mongolic Languages. London: Routledge, pp. 403–19. Tietze, Andreas 1990. Die fremden Elemente im Osmanisch‐Türkeitürkischen. In György Hazai (ed.), Handbuch der türkischen Sprachwissenschaft 1. Wiesbaden: Harrassowitz, pp. 104–18. Windfuhr, Gernot 1990. Persian. In Bernard Comrie (ed.), The World’s Major Languages. New York: Oxford University Press, pp. 523–46. Yağmur, Kutlay 2016. Intergenerational Language Use and Acculturation of Turkish Speakers in Four Immigration Contexts. Frankfurt‐am‐Main: Peter Lang. Yakup, Abdurishid 2005. The Turfan Dialect of Uyghur (Turcologica 63). Wiesbaden: Harrassowitz.

29 Contact and Afroasiatic Languages ZYGMUNT FRAJZYNGIER AND ERIN SHAY

1  The Afroasiatic Phylum Afroasiatic languages (‘Afrasian’ in Diakonoff 1988 and Ehret 2015) are spoken by some 350 million people in North, West, Central, and East Africa, in the Middle East, and in scattered communities in Asia, Europe, the United States, and the Caucasus (Hetzron and Frajzyngier 2018; Frajzyngier and Shay, 2012; Frajzyngier, 2018). The phylum consists of Egyptian, Semitic, Berber, Chadic, Cushitic and Omotic families. The internal subgrouping within the family is subject to debate. Ehret (2015) proposes a three‐stage separation of the phylum, with Omotic separating first, followed by separation of a group called Northern Erythraic, then followed by the splitting of Northern Erythraic into Cushitic, Egyptian, Semitic, and Berber and Chadic. The wide geographical range of the languages belonging to the phylum implies numerous contacts with languages from other phyla, such as Indo‐European, Nilo‐ Saharan, Niger‐Congo, and Khoisan, as well as extensive contacts within the phylum and with the languages within the individual families. Some Afroasiatic languages, such as Egyptian, Akkadian, Phoenician, Hebrew, Arabic, and Ge’ez, have an old written tradition. Interaction through reading and writing has added further possibilities for language contact in addition to the usual cases of language contact through oral interaction. Three Semitic languages, Hebrew and Aramaic and Arabic, are languages of the Old Testament and the Koran, and as a result lexical borrowings from Hebrew and Arabic were transmitted to cultures that accepted Christianity or Islam. Thus, throughout Europe one can find proper names drawn from Hebrew via the Old and New Testament, and throughout Africa and Asia proper names drawn from Arabic via the Koran. The adaptation (‘borrowing’) of lexical items as a result of language contact is a common and expected phenomenon whenever two or more languages are spoken in the same area or whenever there exists extensive literary contact between languages. The change may be unidirectional, where only one language undergoes change, or bidirectional, where both languages in contact undergo changes. The presence of lexical borrowings is sufficient evidence of language contact, even though lexical items that are so transmitted may have originated in languages that are not in direct contact. The adaptation of lexical items from another language may contribute to changes in the phonological system and in phonotactics. In the discussion that follows, whenever we discuss instantiations of language contact, we take the existence of lexical adaptations for granted when other adaptations have taken place.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

572  Zygmunt Frajzyngier and Erin Shay The adaptation of lexical items (not free grammatical morphemes) does not necessarily imply that the grammatical properties of those items have been incorporated into the new language (Frajzyngier, Gurian, and Karpenko, this volume).

1.1  Time Depth and Migrations Ehret (2015) estimates that the time of the first split of Afroasiatic is between 22000– 15000 bce, and that the location of the split was in or near what is now Southeast Ethiopia. Over this stretch of time Berber and Chadic languages migrated from their place of origin, somewhere in East Africa, to their current locations in North, West, and Central Africa, while  some (or possibly all) Semitic languages migrated to their present locations in the Middle East. Some Semitic languages, the ones referred to now as Ethio‐Semitic, returned to the Horn of Africa.

1.2  Linguistic Environment in Space and Time Given the time depth of the Afroasiatic phylum, some language contacts are very old. To the north and east of the current location of the phylum there was contact between Old Egyptian and Indo‐European and contact between East Semitic and Persian, while in the south there was contact with all of the other phyla of Africa, cf. Niger‐Congo, Nilo‐Saharan, and Khoisan. The present chapter is organized as follows: we first discuss language contact among individual families of Afroasiatic and sometimes among individual languages, going from east to west, roughly corresponding to the consensus about the directionality of movement of Afroasiatic languages. For five families we present the most recent findings by other scholars with respect to language contact. Readers interested in deepening their knowledge of language contact in those areas are encouraged to consult references cited for each area. We then focus on language contact involving the Chadic family. These contacts most often do not involve the medium of writing, except for borrowings from Arabic. This part of the study consists of the results available in the literature as well as results presented for the first time. The discussion that follows focuses on two effects of (i) changes in the coding means used in various languages, and (ii) changes in the functions coded in those languages. Often, the two types of changes are intertwined.

1.3  Missing Prerequisites for the Study of Language Contact The literature on language contact in Egyptian, Berber, and Semitic languages is quite extensive. The literature on language contact in Cushitic, Omotic, and Chadic languages is by comparison more modest, although there has been a marked increase in the last 30 years. This state of the art corresponds to the time depth of studies of Egyptian, Berber, and Semitic as opposed to the much more recent studies of Cushitic, Omotic, and Chadic. For the whole Afroasiatic phylum, however, and for each of its families, there is a fundamental prerequisite that is missing, namely the reconstruction of the grammatical system of Proto‐Afroasiatic and the reconstruction of the grammatical systems of Proto‐ Berber, Proto‐Semitic, Proto‐Cushitic, Proto‐Omotic, and Proto‐Chadic. We take it for granted that the earliest Egyptian written records contain the forms and functions of the oldest grammatical system of Egyptian. The lack of reconstructions of other protolanguages prevents us from making confident statements as to what is a retention and what is an innovation within the grammatical system. This in turn prevents us from being confident with

Contact and Afroasiatic Languages  573 respect to assessing what is a result of language‐internal changes and what is a result of language contact. Here are some types of questions that will be hampered by the absence of reconstructions of grammatical systems (‘AA’ refers to Afroasiatic; the questions apply to the Afroasiatic phylum as well as to every Afroasiatic family other than Egyptian): What were the inherent syntactic properties of verbal roots in Afroasiatic? What grammatical relations were coded in Proto‐AA and how were they coded (Frajzyngier 2012b)? Can verbal extensions be reconstructed for Proto‐AA or are they innovations? Which functions (if any) with respect to tense, aspect, and modality can be reconstructed for Proto‐AA? Which pragmatic functions were coded in Proto‐AA and how were they coded? What types of relations between noun phrases were coded in Proto‐AA and how were they coded? (This question is pertinent to the discussion of so‐called construct states in Semitic languages.) What semantic relations between the verb and arguments were coded in Proto‐Afroasiatic? The rich system represented by the verbal derivational forms in Semitic languages, and the less elaborate systems in Egyptian and Berber, are not replicated in Chadic, Cushitic, and Omotic languages. On the other hand, Chadic languages have a rich system of verbal extensions coding a large variety of functions not found in other families. Given the absence of reconstructions of functions as stated above, the effects of language contact on the grammatical system described so far are, to a large degree, a product of accidental observations.

2 Language Contact in Old Egyptian, Late Egyptian and Coptic In a massive study, Kammerzell (2005) proposes that even before the emergence of the ­earliest written records, the valley of the Nile was inhabited by speakers of an Afroasiatic language and speakers of an Indo‐European language, and that Old Egyptian is a product of the fusion of the two languages. In support of his thesis Kammerzell uses the comparative method, essentially comparing lexical items of Egyptian with those of Indo‐European languages and typological argumentation. The comparative study demonstrates two sources of lexical items in pre‐Old Egyptian, one of Afroasiatic origin and the other of Indo‐European origin. Kammerzell (2005: 182) postulates that some of the features often associated with creoles can be found in Old Egyptian, e.g. transparent (morphologically complex) question word systems, no inversion in polar questions, morphologically complex reflexives, movement rules causing focused constituents to occur in sentence‐initial position, similar expressions of existence and possession, and adjectives as a subcategory of stative verbs. On the other hand, he states, there are several characteristics rather atypical of contact languages, e.g. existence of gender, passive forms, and tense/aspect/modality marked internally or/and by means of suffixes. The typological argumentation for the fusional nature of pre‐Old Egyptian, and consequently Old Egyptian, is not particularly convincing, as the number of features that allegedly characterize creole languages are shared by other Afroasiatic languages for which no creole origin has been postulated (Frajzyngier 2012a, 2012b, 2018). Over its 5,000 years of written history, Egyptian has undergone important internal changes and has been in contact with a variety of Semitic languages (Winand 2017) and Greek to the north and east, with Berber languages, and most likely with Cushitic languages to the south. Contact with Semitic and Greek has been studied over the years but there is very little known about the contact with Cushitic languages. There is evidence of lexical borrowings from Old Egyptian into Berber (Kossmann 2002). The latest stage of Egyptian, Coptic has undergone strong influence from Greek in both lexical and grammatical borrowings (see papers in

574  Zygmunt Frajzyngier and Erin Shay Grossman, Dils, Richter, and Schenkel 2017). The grammatical borrowings described include prepositions. Egyptian was also a source of lexical borrowings into Greek (Tovar 2017). Studies of the contact between Egyptian and Cushitic languages, which must have occurred at the earliest times, have yet to be made.

3  Language Contact in Semitic Semitic languages had contact with other languages through two modes: spoken, from the earliest split of the Afroasiatic family, and both written and spoken, in the case of Akkadian (East Semitic), Hebrew, Aramaic, Arabic, and some Ethio‐Semitic languages (West Semitic). For many of these languages contact resulted in bidirectional changes, whereby languages that have undergone changes become the source of borrowings, mainly lexical, from ­languages belonging to other families. Religious terminology in Christian and Muslim ­cultures all over the world reflects lexical borrowings from Arabic and Hebrew. The literature on language contact involving Arabic alone is very large and relates to all periods of time, to contact with different languages in different geographical areas, and to both lexicon and grammar (for the current state of the art, see Manfredi and Tosco (eds.) 2018). Lexical borrowings may contribute to changes in language structure through a chain reaction whereby the phonological system is affected first and then affects changes in the form of morphemes, which may lead to morpheme reduction and eventually to syntactic changes. The borrowing of grammatical morphemes, on the other hand, is motivated mainly by the introduction of a function that is not coded in the target language. In what follows, we discuss some of the most important changes in language forms and functions in Semitic languages and consider for which changes language contact was the only, or the contributing cause. For a recent study of language contact in Semitic see papers in Butts (ed. 2015). Most of these papers deal with lexical effects of language contact, but a few papers deal with morphology and syntax. Many studies of language contact in Semitic languages involve the effects of contact on the lexicon. Lexical effects of language contact are numerous and are well documented. Some of the major areas in which languages contact played a role in the grammatical systems of Semitic languages include: (i) Linear order of major constituents: in some languages with verb‐initial or verb‐final position, linear order is not a coding means for grammatical relations. In some languages, with verb medial position, linear order is sometimes a coding means for grammatical relations. (ii) Case marking: some languages have case marking and other languages do not. (iii) The category “state”: Some languages have the category state for nouns, whereby the head in the modifying construction has a shortened form when compared to the same noun in isolation. Other languages do not have this distinction, so the question is for which of those distinctions language contact contributed to a change.

3.1  Linear Order With respect to the linear order of predicates and arguments Semitic languages represent considerable variation: Akkadian, an East Semitic language, has the verb in clause‐final position. This contrasts with the verb in clause‐initial position in Biblical Hebrew, Classical Arabic, and Ge’ez, a Classical Ethiopian language. There is a consensus that Akkadian changed its word under from verb‐initial to verb‐final under the influence of Sumerian (von Sodden 1995: 227 §130; John Huehnergard p.c.). For a brief summary of the Sumerian word order, see Ebeling (2004). The modern Ethio‐Semitic languages Amharic and Tigrinya are verb‐final. The Classical Ethiopian language Ge’ez, which is retained as the liturgical language in the Coptic Church of Ethiopia, is verb‐initial. The common explanation of the word order in modern Ethio‐Semitic

Contact and Afroasiatic Languages  575 languages is that they have adopted the linear order of the Cushitic and Omotic languages all of which are verb‐final (Leslau 1945, 1952; Moreno 1948). Thomason (2001) demonstrates further typological changes linked with the change in word order in Ethio‐Semitic. The important fact about the changes in linear order in Akkadian and in Ethio‐Semitic languages is that in both cases the change was caused by language contact, and in both cases the change involved change from an order that was not a coding means for grammatical relations to a different order that is also not a coding means for grammatical relations (Frajzyngier 2011a). Ethiopian Semitic languages also display other characteristics that may be attributed to language contact with Cushitic languages. Appleyard (2015) considers the possibility that the development of the labialized velars kw, gw, qw, and h̬w in Ge’ez may have occurred through contact with the Cushitic language Agaw. Appleyard acknowledges that the process of labialization of velars may have happened independently. On the other hand, the development of a seven‐vowel system in Ge’ez as postulated by Appleyard may indeed be the effect of language contact. Appleyard also discusses a number of morphological features of Ethio‐Semitic languages, including loss of a gender distinction in plural pronouns, ­converbs, a means of chaining clauses, and the development of syntactically dependent verbal forms. All of these features are shared by Cushitic languages and occur, perhaps excluding converbs, in other languages of Africa. Modern Arabic varieties and Modern Hebrew are all predominantly SVO1. There have been claims with respect to Modern Hebrew that the change from the predominant VSO order to SVO was the result of the prolonged contact with West European languages (Schwartzwald 2011, as reported in Zeldes 2013). Frajzyngier (2011a) and Frajzyngier with Shay (2016), which treat linear order as a coding means, have demonstrated that the changes in the form and function of linear order can be triggered through language‐internal changes such as weakening or complete loss of case marking. In the case of Arabic and Hebrew we observe the change from a linear order that does not code grammatical relations to a linear order that does code grammatical relations. This change may have been language‐internally motivated as it has been noted in other languages from other families, e.g. English, Bulgarian (Frajzyngier 2011a). Language contact may have been a factor facilitating such a change. Zeldes (2013) argues against Modern Hebrew being a standard average European language, hence against the hypothesis that contemporary Hebrew is uniquely a product of language contact rather than a product of natural evolution from the earlier stages of the language with the usual adaptation of forms and functions from other languages.

4  Language Contact in Berber Berber languages are spoken in North and West Africa, namely in Morocco, Algeria, Mauritania, Mali, and Egypt (Siwi). Since the earliest time some Berber languages have been in contact with Egyptian (Kossmann 2002), Romance languages (Kossmann 2013b), and Chadic languages (Kossmann 2005), and starting in the mid eighth century with Arabic (Kossmann 2013b) and Northern Songhai languages (Niger‐Congo or Nilo‐Saharan ­language, depending on whose classification one chooses (Souag 2010, 2012)). The close contact between Berber languages and Arabic has existed for many centuries, and various varieties of Arabic and Berber languages have affected each other in all areas of grammatical systems including phonology, morphology, syntax, and lexicon (the most studied). Souag (2014) deals with contact between the Siwa language, spoken in Egypt, and Arabic. Kossmann (2013b) thoroughly describes and documents all kinds of borrowings from Arabic spoken in Maghreb into Northern Berber languages. There was a massive borrowing from all categories of Arabic vocabulary, not just those linked with Islam, into the Northern Berber varieties. Some Arabic words were borrowed with their Arabic morphology while in

576  Zygmunt Frajzyngier and Erin Shay others the Arabic morphology was suppressed. In some Berber languages with heavy Arabic borrowing, such as Ghomara, verbs of Arabic origin are conjugated according to Arabic patterns, and verbs of Berber origin are conjugated according to Berber patterns (Kossmann 2013b: 413). Arabic spoken in Maghreb has also borrowed a large number of Berber words. One of the most interesting aspects of the phonological effects of language contact was the use of Arabic phonemes as expressive markers in Berber words. The existence of similar parallel systems, as Kossmann calls them, is noted for other lexical categories. This phenomenon has also been observed by Souag (2014) for Siwi, which reports that the verbal morphology of Siwi has been simplified under the influence of Arabic. With respect to syntax Kossmann claims that the influence of Berber on Arabic spoken in Maghreb was stronger than the influence of Arabic on Berber. Berber did, however, undergo a change, namely the introduction of clausal coordination through borrowing of the Arabic conjunction w. Kossmann also notes the influence of Arabic relative clauses on the formation of relative clauses in those Berber languages that have been in contact with Arabic.

5  Language Contact in Cushitic and Omotic For the classification and typological characteristics of Cushitic and Omotic languages, see Mous (2012) and Amha (2012). As discussed on p. 572, Cushitic and Omotic languages share many typological characteristics that distinguish them from other Afroasiatic languages. Cushitic and Omotic have been in contact with each other (Appleyard 2006) as well as with Ethio‐Semitic languages (Moreno 1948; Appleyard 2015), Nilo‐Saharan languages, and Khoisan. They have also been contact with Bantu languages (Niger‐Congo). The mixed Ma’a/Mbugu language represents an extreme case of language contact as it has a largely Cushitic lexicon and a Bantu grammar (Mous 1995, 2013). Seen from the perspective of the Afroasiatic phylum, Cushitic and Omotic languages represent a puzzle because of their verb‐final linear order. Egyptian, Berber, Semitic (with exception of Akkadian as discussed earlier) and some Chadic languages are verb‐initial. In Proto‐Chadic, as proposed in Frajzyngier (1983), the verb was in clause‐initial position. Given that four families in Afroasiatic had the verb in clause‐initial position, the clause‐final position of the verb in Cushitic and Omotic languages may be considered an innovation. If it is an innovation, it has likely occurred as a product of language contact. And indeed, as documented in Dimmendaal (2012), many Nilo‐Saharan languages, including Eastern Sudanic languages, are verb‐final. The change from the presumed Proto‐Afroasiatic verb‐initial order to the verb‐final order would have been a change from one order that was not a coding means for grammatical relations to another order that also is not a coding means for grammatical relations. Another typological feature that distinguishes Cushitic and Omotic languages from other Afroasiatic families is the presence of the marked nominative case. Marked nominative case is a phenomenon whereby the nominal or pronominal agent is morphologically marked and the patient/undergoer is unmarked, i.e. it is not marked for the accusative case. The citation form of the noun is also unmarked (see König 2008a). König (2008b) proposes that the presence of the marked nominative case in East Africa is an areal feature. This conclusion is supported by facts discussed in Dimmendaal (2012).

6  Language Contact in Chadic Languages of the Chadic family, typologically the most diversified family of Afroasiatic, number around 150 languages spoken in Northern Nigeria, Northern Cameroon, Southern Niger, and Southern Chad. The Chadic family consists of three or four branches: West

Contact and Afroasiatic Languages  577 Chadic, Central Chadic, East Chadic, and Masa (Newman 1977) or three branches, West Chadic, Central Chadic, and East Chadic (Jungraithmayr and Ibriszimow 1994). For typological features of Chadic languages, see Frajzyngier and Shay (2012). Only about 50 of these languages have been described by reasonably complete reference grammars. There exists only one description per language, with the exception of Hausa (West Chadic), which has been described by dozens of grammars published in English, German, French, Russian, Italian, and Polish, and Ouldémé, which has two grammars by different authors. For majority of Chadic languages, we therefore lack different points of view and alternative analyses, a necessary component in scientific inquiry. Ehret (2018) postulates that Berber and Chadic once constituted a single branch of Afroasiatic. From the time of their separation (ca. 5000 bce, as per Ehret, 2018), Chadic languages have been in contact with Niger‐Congo languages and Western Nilo‐Saharan languages (Ehret 2006). Chadic languages display traces of old and recent language contact. There is significant literature on language contact involving Chadic languages as both recipients of lexical items and grammatical morphemes and as sources of the same. Several grammatical phenomena found in Chadic languages are also found in other families in Africa, indicating that they all share areal characteristics whose origin is often difficult to pin down (Cyffer and Ziegelmeyer 2009). For lexical evidence of old contact with Nilo‐Saharan languages, where some names of animals begin with forms that resemble prefixes in Nilo‐ Saharan languages, and for recent contact with Kanuri in particular, see Cyffer (2000, 2006), Ziegelmeyer (2009a–d), Schuh (2011), Allison (2017), and Frajzyngier and Ross (1991). For Berber borrowings in Hausa, see Kossmann (2005). In more recent times, numerous lexical borrowings from Arabic came directly through contact between speakers while some came indirectly through Fula, a West Atlantic language; Kanuri, a Nilo‐Saharan language (Allison 2007); and Hausa, the largest Chadic language. Chadic languages have also been in contact with each other. Given that lexical borrowings have been amply discussed in the existing literature, and that their existence is non‐controversial, the discussion below focuses on forms and functions that are either a product of language contact or whose presence has likely been facilitated by language contact. The additional reason for this focus is that grammatical changes are generally considered to be less frequent than lexical changes.

6.1  Phonology, Tone, Vowel Harmony Jungraithmayr (1991) notes that Tangale has nine vowels subject to A(dvanced) T(ongue) R(oot) vowel harmony, a unique system within Chadic languages. He attributes this system to the neighboring Adamawa (Niger‐Congo) languages. Vowel harmony rules of various types are found in other Chadic languages, e.g. Gidar (Central Chadic, Frajzyngier 2008), and Lele (East Chadic, Frajzyngier 2001). The motivation for vowel harmony in Lele and Gidar is most likely language internal, as it enables the listener to parse the utterance into words. The particular type of vowel harmony in Tangale may well be the product of ­language contact.

6.2  Free Grammatical Morphemes and Associated Functions Many Chadic languages have adopted functions coded by free grammatical morphemes in the source languages. The motivations for these adoptions appear to be functional. Languages borrow the formal means for functions that they do not encode themselves and that are salient enough to be perceived as serving a useful communicative need. Free grammatical morphemes are more easily perceived because they occur more frequently than other lexical items, as a consultation of any frequentative dictionary demonstrates. The ease of

578  Zygmunt Frajzyngier and Erin Shay recognition of a morpheme as a separate entity makes its functions easier to perceive. Free grammatical morphemes, like lexical items, are more easily inserted into the utterance in the target language, hence more easily adopted. One class of free grammatical morphemes is propositional connectors, i.e. markers that indicate the relationship of one proposition to another within discourse or within a sentence, if the language has such units as “sentences” (Frajzyngier and Katriel 2004). This class includes conjunctions and subordinators, markers of modality of the main and subordinate clause. These may include question and negation markers, markers of approval and disapproval, and epistemic and deontic modality markers. Here is a brief and incomplete list of functions coded by free grammatical morphemes adopted into Chadic languages. An anticipatory note: the number of free grammatical morphemes adopted through language contact in Chadic languages by far exceeds the number of affixes adopted through language contact.

6.2.1  Counter‐presuppositional Function One cannot reconstruct such a function for Proto‐Chadic as no Chadic language described so far has a dedicated marker of such a function. Many Chadic languages have borrowed the marker amma or ama ‘but’, which originally comes from Arabic. The direct sources from which the marker was borrowed may have been Fula (West Atlantic), Kanuri (Nilo‐Saharan), and Hausa (West Chadic) for many Chadic languages.

6.2.2  Clausal Coordination Although a few Chadic languages have a clausal coordinating conjunction, clausal coordination cannot be reconstructed for Proto‐Chadic. Chadic languages did not have a clausal coordinating conjunction corresponding to “and” in English, et in French, y in Spanish, and i in Russian or Polish (Frajzyngier 1995). Through contact with languages that have such a conjunction, e.g. French, English, or Arabic, some languages grammaticalized the nominal coordinating conjunction to serve also as a clausal coordinating conjunction. Lele (East Chadic) borrowed the exclamation ala originally from Arabic allah ‘God’. This marker was originally borrowed as a marker of surprise and can now function also as a marker of clausal coordination (Frajzyngier 2001, forthcoming).

6.2.3  Logical Conclusion Hausa has a form sai, which Bargery (1934: 884) characterizes as the conjunction ‘until’, the preposition ‘except’, which codes obligation, and several other modal functions. This particle has been widely borrowed in other Chadic languages, because this particular propositional connector cannot be reconstructed for Proto‐Chadic. Thus in Gidar, Central Chadic, the form séy is used as a conditional apodosis marker: (1) à‐ddə́ɗá‐ŋ    mé̃lìy    ví də́‐sə̀n ə́gɗà 3M‐replace‐pl    chief     for 3M‐know do zə̀gá ɓà ɓà séy   àn à‐mtə̀‐ká work NEG NEG except if 3M‐DIE‐PRF ‘One does not replace the chief just because he does not know how to do his work. One replaces the chief only when he dies.’ (Frajzyngier 2008: 296) Similarly in Wandala (Central Chadic): (2) sái then ‘I have to talk.’

bà FOC

yá 1SG

ndàhà speak

Contact and Afroasiatic Languages  579

6.3  Alternative Proposition Another modal marker that has been borrowed by many Chadic languages is the Hausa particle koo ‘or’, ‘either’, ‘whether’ (Bargery 1934: 616). This marker has been widely adopted by other Chadic languages (Ziegelmeyer 2009d). Reason. As described in Cyffer (2009), Hausa has adopted the Arabic lexeme sabab ‘reason, cause’ in the expression saboda ‘because of’. Intensifier. The Hausa or Fula intensifier daidai has been adopted by a number of Chadic languages, including Gidar: (3) à‐pə̀rmá‐w dáydáy ɓà 3M‐hear‐1SG even NEG ‘He does not even listen to me.’

6.4  Linear Orders: SV(O), VS(O), VOS Most contemporary Chadic languages from all three branches have the pragmatically neutral SV(O) order. The object (O) does not have to occur even with a transitive verb, and it is not clear that all languages have this category. Some Central Chadic languages are verb‐initial and have the neutral VS(O) order. Some VS languages have an alternative VOS order. Frajzyngier (1983) postulates that Proto‐Chadic had the verb in clause‐initial position and that the change to SVO was motivated by several language‐internal changes involving fronting of the noun phrase for topicalization and focus and perhaps other pragmatic functions. To those language‐internal motivations we would like to add language contact as another motivation. Most languages with which West Chadic, Central Chadic, and East Chadic languages are in contact are SVO. Kanuri is verb‐final. The change of linear order can happen within the lifetime of speakers if they are exposed to different word orders. Here is a not completely anecdotal piece of evidence: Hdi (Central Chadic, Frajzyngier with Shay 2002) is a predicate‐initial language, i.e. a language in which both the verb and the nominal predicate occur in clause‐initial position. Heusing (1999) elicited data from a Hdi speaker at the University of Maiduguri, where English is the language of instruction and where English and pidgin English are languages of communication among students. Most sentences Heusing elicited were of the SVO type. Frajzyngier had a similar experience with a Hdi speaker who spent about 20 years in Nigeria. Hence some speakers of Hdi have changed their syntactic pattern under the influence of SVO languages.

6.5 Negative Marker in Clause‐final Position – A Formal Feature Dryer (2009) observed an interesting peculiarity of Central African languages, cf. that the negative marker in these languages, unlike in most of the languages in other parts of the world, often occurs in clause‐final position. This formal feature is shared by Chadic and non‐Chadic languages alike, and for many years has been considered an areal feature. Given the rarity of the position of the negative marker in clause‐final position, and especially the fact that it does not occur in this position in other Afroasiatic languages, one may consider it to be an innovation in Chadic. But is it a product of language contact or did it emerge in Chadic as a result of language‐internal changes? One factor that would appear to favor the language‐internal processes is that in several Central Chadic languages, other modal markers also occur in clause‐final position.

580  Zygmunt Frajzyngier and Erin Shay Table 29.1  Pronominal subjects in Giziga. Person

Singular

Plural

First Second Third

yi ki a

yí … ‐am/‐kV kí … ‐am/‐kV a … ‐am/‐kV

6.6 Split Coding of Person and Number on the Verb: A Formal Feature Most Chadic languages code the person and number of the subject on one morpheme, which may precede or follow the verb. Gidar (Central Chadic, Frajzyngier 2008) and Giziga (Shay 2012, ms.) code person and number through separate morphemes (Table 29.1). The split coding of person and number occurs also in Gidar, but only in the third person, and in Mofu‐Gudur. Similar split coding of person and number markers occurs in Mundang, an Adamawa (Niger‐Congo) language. Mundang and Gidar are spoken in the same areas, e.g. in the town of Guider. Frajzyngier and Shay (2008) propose that the split coding of person and number was a language‐internal development within Chadic languages. Its presence in Mundang may well be the effect of language contact.

6.7 Pragmatically Dependent Tenses and Aspects: A Functional Feature Across African languages from several families there exists a double system of tense and aspectual markers, whereby two forms exist for the same tense or aspectual values (Jungraithmayr 1994). Frajzyngier (2004) demonstrates that the function of the two systems is to code the pragmatic status of the proposition. One set of tense and aspectual markers indicates that the proposition can be interpreted on its own and the other set indicates that the proposition must be interpreted in connection with some other proposition or in connection with some element in the speech environment. Such clauses are labeled as pragmatically dependent. Thus, comments on an element in focus, relative clauses, and content questions must be marked as pragmatically dependent. Polar questions and comments on topic are not marked as pragmatically dependent. In Chadic languages, a simple sentence may be marked as pragmatically dependent simply by the deployment of a dependent tense or aspect marker. This is shown in the following example, where the sentence (5) in Hausa (West Chadic) is interpreted as coding focus on the adverb of time jiyà ‘yesterday’ solely by the use of the aspectual system coding pragmatic dependency. Newman (2000), from which the example is taken, calls the dependent completive aspect “preterite”: (4) jiyà sun sana‐r̃ yesterday 3PL:COMPL know‐CAUS ‘Yesterday they informed us.’

dà ASSC

mu 1PL

(5) jiyà suka sana‐r̃ dà mu yesterday 3PL:PRET know‐CAUS ASSC 1PL ‘Yesterday they informed us.’ (Newman 2000: 572; glosses by Frajzyngier, using Newman’s grammatical labels)

Contact and Afroasiatic Languages  581 For a study of the two systems in Chadic languages, see Frajzyngier and Shay (2010). Given that the two tense/aspectual systems co‐exist in other languages of Africa one must consider it an areal feature. We have no clear methodology to establish in which family such a marking originated.

6.8 Coding of Nominal Plurality – A Formal and Functional Feature In Gidar (Central Chadic), the plurality of a noun is marked by the suffix ‐ɗi, with the phonologically predictable variant ‐ɗe. This suffix is also phonologically very similar to two nominal plural suffixes in the Diamare (North Cameroon) dialect of Fula ‐ɗe and ‐ɗi (Noye 1974: 197–9), which derive plurals of animals and humans. In Gidar this suffix is employed for coding the plurals of animals and inanimate objects. The evidence that this suffix was borrowed into Gidar from Fula, rather than the other way around, is provided by the fact that in the variety of Fula spoken in Mali, which is far from Cameroon, there are also the same plural suffixes ‐ɗe and ‐ɗi (Breedveld 1995: 296). This suffix in Gidar fully participates in the phonological processes, triggering vowel harmony and undergoing predictable phonological changes. The importance of the borrowing of this suffix is that there are no other productive nominal plural markers in Gidar. Moreover, the suffix is used only with some nouns. Thus body parts, whether those occurring in pairs or single, do not have plurals. Exceptions are nouns for “vagina” and “penis,” which do have plurals and which differ from other body parts in that they can occur freely in non‐possessive constructions. Kinship terms also do not have plural forms. These facts quite possibly indicate that the means to derive plural nouns are a relatively recent development.

6.9  Verbal Extensions – A Formal Feature Wolff and Gerhardt (1977) rightly noticed a similarity between some West Chadic languages and Benue‐Congo languages with respect to the presence of verbal extensions. Verbal extensions are suffixes to verbs that code a variety of functions. Since 1977 our knowledge of languages has increased and we can now provide a richer list of functions coded by verbal extensions. Verbal extensions in Hdi (Frajzyngier with Shay 2002) code the following functions: applicative, inverse, partial affectedness of the subject, point of view of the subject, goal, tentative, “do again,” associative, and directionality of movement, in, away, down, and up. This set of extensions is richer than the sets usually found in Bantu languages. Ehret (1995) proposes that at one time verbal extensions may have been much more widespread in Afroasiatic languages. If this assumption is correct, their survival in West and Central Chadic languages may have been reinforced by the presence of verbal extensions in neighboring languages.

6.10  Possessive Subject Suffixes aka ICP – A Formal Feature Newman’s widely accepted term ICP (intransitive copy pronoun) refers to morphemes suffixed to the intransitive verb. These morphemes are often identical to possessive suffixes. Their function is to represent the event from the point of view of the subject, a function similar to the reflexive pronouns in German and the short reflexive markers in Dutch, Polish, and Russian (Frajzyngier 2011b and references there). In some Chadic languages, including languages that have possessive subject pronouns, there exist other formal means to code the point of view of the subject. That is why we consider the possessive subject suffixes as a formal means for a function that is coded by a variety of formal means in Chadic. Wolff and

582  Zygmunt Frajzyngier and Erin Shay Gerhardt (1977) did observe that such pronouns occur in both West Chadic and Benue‐Congo languages. Such pronouns are considered to be an areal feature in West Africa (Cyffer and Ziegelmeyer 2009; Storch, Atindogbé, and Blench eds., 2011).

6.11 Additional Argument Marker (‐n Ending on the Verb): A Formal and Functional Feature Several Chadic languages have a suffix to the verb that some refer to as “causative” (Jaggar 2017) and some refer to as “benefactive” or “additional argument marker” (Frajzyngier 1989). In some languages, the function of this marker is to indicate that the proposition (not necessarily the clause) has more participants than one could deduce from the inherent properties of the verb. Here are some examples from Pero (West Chadic). The labels preceding each example refer to possible inference about the role of participants in the real world rather than to the grammatical function encoded: ‘Benefactive’: (6) bàtúurè n‐yè‐tù n‐wát‐tù múnù‐n aníni bèlòw white man SEQ‐call‐VENT SEQ‐came‐VENT give‐AAM anini two ’The white man called [him] and when he came he [the white man] gave him two anini.’ (7) n‐wálù‐n SEQ‐cook‐AAM ’and the gruel was made for her’

ɓwè gruel

The suffix is added to intransitive verbs to code presence of an object: glossed as AAM for “additional argument marker.” (8) pùrò‐kò‐n go out‐COMPL‐AAM ’He carried yam out.’

bírà OUT

kà ASSOC

cándù yam

‘Causative’: (9) ɓáy nì‐cí‐kò‐n wát‐nà dog lSG‐eat‐COMPL‐AAM come‐COMPL:VENT ’The dog that I fed came.’ (absence of the suffix ‐n would result in the nonsensical ‘the dog that I ate came’) The additional argument marker is not the expected third person object marker. Pero distinguishes between the third person masculine nì and the third person feminine ‐tò: (10) n‐tà‐lí‐ni 1SG‐FUT‐put‐3M ’I will hurt him.’

píɗà wound

An additional argument that the marker n is not an object marker is provided by the fact that it can co‐occur with object pronouns: (11) píl‐nò‐n buy‐1SG‐AAM ’Sell the goat for me.’

kà ASSOC

ɓíà‐i goat‐DEF

Contact and Afroasiatic Languages  583 Similar markers, albeit analyzed in different ways, have been recorded in other Chadic languages, e.g. Wandala (Frajzyngier 2012), Hausa (Jaggar 2017). The importance of this marker is that it is a prototypical inflectional marker as it is suffixed to a lexical, in this case verbal, form and that it indicates some relationship between constituents of a proposition. It appears, however, that this marker may well be a product of language contact. Breedveld (1995: 174) describes a causative marker in Fula, ‐n, which is also suffixed to the verb and whose functions are similar to that of the additional argument marker in Chadic. It is a possible that the similarity is purely accidental. There are three factors that argue against this possibility: (i) the phonological form of the marker; (ii) the fact that in all languages involved the marker is suffixed to the verb rather than to some other constituent; and (iii) that the marker carries the same function across languages. The possibility that Fula borrowed this marker from a Chadic language is ruled out by two factors: Fula is a vehicular language in Northern Cameroon but not in Nigeria, where Pero is spoken. Pero is spoken by around 20,000 people. The variety of Fula reported by Breedveld (1995) is spoken in Mali, an area where there are no Chadic languages.

6.12  Serial Verb Constructions – A Formal Feature Serial verb constructions are formal means frequently found in Chadic languages from West, Central, and East Chadic branches. Serial verb constructions in Chadic function in several domains. One of these is the coding of movement predication. Serial verb constructions are used here to code the manner of movement, which may be directionality towards a goal, directionality away from the source, and the spatial orientations up, down, in, and out. Here are a few examples coding directionality and manner of movement: (12) wùr sú 3M runs ’He runs to school.’

sé go‐away

də̄m go‐to

mákárántá school

(13) wùr kə̄ tóñ n‐jíñ jì 3m HABIT stay PREP‐Jing come ’He comes from Jing.’ (Mupun, West Chadic) (Frajzyngier 1993) Serial verb constructions are also used to code comparative function, where the verb corresponding “to surpass” precedes the target of comparison: (14) pàk mó máa mò dém ás met some 3PL also 3PL like dog surpass ’Some of them like dogs more than people’ (Mupun, West Chadic) (Frajzyngier 1993)

ngúrùm people

Serial verb constructions are not found in Semitic, Berber, Egyptian, Cushitic or Omotic languages. They are, however, found frequently in Niger‐Congo languages. Serial verb constructions or traces thereof can be found in all branches of Chadic. In West Chadic languages they are a product of language contact with Niger‐Congo languages, where serial verb constructions can be found in several families. Compare a serial verb construction coding manner and directionality in Yoruba (Kwa, Niger‐Congo): (15) Àjàó rìn lo Ajao walked go ’Ajao walked home.’ (George 1975: 82)

(sí) (to)

ilé home

584  Zygmunt Frajzyngier and Erin Shay And a comparative construction in Yoruba (Kwa): (16) Ayò Ayo

ní has

ogbòn cleverness

jù surpass

mí me

lo go

’Ayo is cleverer than I am.’ (Stahlke 1970: 64) Language contact provided Chadic languages with the formal means to code functions that were coded by other means, mainly inflectional, that were lost. It is possible that serial verb constructions emerged independently in Chadic languages. Their emergence may have been reinforced by the presence of similar formal means in neighboring languages.

6.13  Logophoricity – A Functional Feature Logophoricity, in its original sense as proposed by Hagège (1974), is the coding of the third person subject of the complement clause of the verbs of saying as being coreferential with the subject of the matrix clause. This is a function coded in West and East Chadic languages but not coded in other Afroasiatic families. The phenomenon was first observed in Adamawa (Cloarec‐Heiss 1969), Kwa (Clements 1975), Benue‐Congo, Gokana (Hyman and Comrie 1981), all Niger‐Congo languages and in Chadic (Frajzyngier 1985a, 1985b). In many languages logophoricity is coded by a distinct set of pronouns. In Lele (East Chadic) logophoricity is coded by the position of the pronoun before the verb (coreference) as opposed to the position of the pronoun after the verb (disjoint reference). Disjoint reference: (17) yàá‐dú ná bòy‐dú say‐3F COMP break‐3F ‘She1 said that she2 broke her1/2 hoe.’

kójò hoe

kò‐tó GEN‐3F

Coreference with the subject of the matrix clause: (18) yàá‐dú ná dú bòy kójò say‐3F COMP 3F break hoe ‘She1 said that she1 broke her1/2 hoe.’ (Frajzyngier 2001)

kò‐tó GEN‐3F

This coding corresponds to the use of the pronouns in Mupun, where the default third person pronouns code disjoint reference and a special set of pronouns codes coreference (Frajzyngier 1993). The presence of the logophoric function in Chadic must be considered to be the adaptation of the function without the adaptation of the formal means, as none of the formal means deployed to code logophoricity in Chadic is present in the neighboring languages.

6.14 Focus on Object Through a Preverbal Position – A Formal Feature This section is devoted to just one structural innovation in just one language, Mina (Central Chadic), where focus on the object is coded through the position of the object before the verb and the use of the locative preposition n: (19) á n kə̀ɗə́m 3SG PREP calabash ‘She took her calabash.’

ngə̀n 3SG

ɓə̀t take

Contact and Afroasiatic Languages  585 (20) tíl á ndə̀   zə́ bíŋ depart 3SG   go    EE    room mì bíŋ à n 3SG PREP mouth room ‘He went to the room and closed the door.’

dzáŋ close

á 3SG

dzáŋ close

ká POS

Compare the non‐focused object, which occurs after the verb: (21) bìɮáv və̀l‐á‐k God give‐GO‐1SG ‘God gave me that child …’

mbə̀ child

táŋ DED

Compare also the use of the preposition n to code the locative complement when the preposition is not inherently locative: (22) hà táŋ tə́wə̀r á 2SG DED suffer PRED ‘You suffer a [a lot] among them.’

nə̀ PREP

fálà among

tə̀tàŋ 3PL

This particular formal means of coding the object through the position before the verb, S n O V, rather than just clause‐initial position has not been observed in other Central Chadic languages. It is, however, a feature that has been described for Western Nilo‐Saharan languages (see Heath 1998 on Koyra‐Chini) and for other varieties of Songhai, including Frajzyngier’s own observation of Zarma spoken in Niger. The problem here is that Mina is not spoken anywhere in the vicinity of Songhay. Could it be the case that Mina has preserved a trace of old language contact?

6.15  Conclusions about Chadic Chadic languages differ from other Afroasiatic languages with respect to some typological features and at the same time share a number of features with Niger‐Congo and Nilo‐Saharan languages. This fact reflects the current geographical situation whereby Chadic languages are in direct contact with one or more languages from the Niger‐Congo or Nilo‐Saharan families. One West Atlantic Niger‐Congo language, Fula, is a vehicular language in Northern Cameroon and was a language through which Islam came to Nigeria. Another, Kanuri, a Western‐Nilo‐Saharan language is a language through which Islam came into some communities in Northern Cameroon. The introduction of Islam into Nigeria and Cameroon was a source of numerous lexical borrowings from Arabic. Many Chadic languages have also undergone the influence of Hausa, the most widely spoken Chadic language, which has become the vehicular language in Northern Nigeria. Language contact has brought or contributed to major structural and functional innovations in Chadic languages, including: 1. The change from the original verb‐initial order to the current SV order in many languages from all branches of Chadic 2. The introduction of serial verb construction as a coding means 3. The emergence of verbal extensions as a coding means (reinforced by language contact) 4. The emergence of two tense and aspectual systems to code the pragmatic status of clauses 5. The emergence of logophoricity as a function coded by the grammatical system 6. The emergence of clausal coordination as a function coded in the grammatical system, and 7. The emergence of the counter‐presuppositional function.

586  Zygmunt Frajzyngier and Erin Shay

7  Implications for the Theory of Language Contact Two of the issues dealt with in the field of “contact linguistics” (Winford 2003) are what determines the borrowing of lexical items and what determines the spread of grammatical elements. With respect to the borrowing of lexical items, language contact in Afroasiatic languages attests to the same conditions that have been found in other areas of language contact: languages borrow terms for semantic notions for which they do not have a lexical item; languages borrow lexical items for objects and concepts that the speakers did not have before; languages borrow lexical items that for social reasons appear to bestow more prestige on the user; and languages borrow lexical items because the speakers want to demonstrate new religious or political beliefs and affiliations. With respect to the spread of grammatical elements, the studies so far have not been very careful in separating the formal means of coding from the functions coded. The study of the spread of grammatical features within the Afroasiatic phylum demonstrates the existence of the following processes: 1. The adaptation of a form that does not code any new function and that replaces another form that did not carry a function. This is the case in modern Ethio‐Semitic languages that have adopted verb‐final order based on the patterns of Cushitic and Omotic languages. This is also the case of the Akkadian language adopting the verb‐final order on the pattern of Sumerian. 2. The borrowing of formal means without any associated function and the use of this means to code a variety of functions. This is the case of the spread of serial verb constructions, which are used in Chadic for a variety of functions. 3. The borrowing of a formal means to code a function that was already coded by a different formal means. This is demonstrated by the spread of the coding of negation through the use of clause‐final negative markers. 4. The borrowing of a formal means and of the associated function. This is demonstrated by the borrowing of the nominal plural marker in Gidar (Central Chadic). Contemporary Gidar does not have an alternative plural marker. There are only weak indicators that Gidar may have had another nominal plural marker. The adaptation of a formal means along its function is also illustrated in Berber, where the clausal coordinated conjunction w was borrowed from Arabic. 5. The adaptation of a function without borrowing formal means to code it. In Chadic languages this is the case of the coding of clausal coordination, which did not exist as a function until there was contact with languages that did have the function. This is also the case with the coding of logophoricity in Mupun and Angas (West Chadic) and Kera (East Chadic). The study has also demonstrated the existence of similar changes occurring in parallel in different languages, motivated by similar processes and resulting in similar constructions. This is the case with the emergence of verbal extensions in Chadic. We know that some verbal extensions emerged relatively recently as a result of suffixation of what used to be prepositions or independent verbs (Frajzyngier 1987). Verbal extensions are also present in other African languages. There is no evidence that the presence of the same formal characteristics is the result of language contact. Studies of language contact have for a long time focused on the sociolinguistic conditions that allow language contact to occur, the chief among them being multilingualism and, more specifically, bilingualism. However, for a number of linguistic functions discussed in this study there is no attested bilingualism. Multilingualism can trigger linguistic changes in the

Contact and Afroasiatic Languages  587 languages involved, as certain forms and functions in multilingual situations can become salient enough to be picked up by speakers who have no mastery, or only limited knowledge, of the languages involved. Raymond Hickey (p.c.) has raised an important question: how much knowledge of a donor language is necessary for contact‐induced change to occur? We cannot answer this question in a systematic way. Some observations, however, indicate that knowledge of the donor language is not required at all for the borrowing of a form or function. It is enough to observe a salient feature in the donor language: if this feature is not present in the language of the observer, such a feature may very well be borrowed. This is the case with the borrowing of the originally Arabic counter‐presupposition marker amma ‘but’ into a variety of Chadic languages whose speakers do not know Arabic or Hausa or Fula through which the word was transmitted. This is also the case with the surprise exclamation ala originally Arabic Alla ‘God’, into Lele, whose speakers do not know Arabic. The borrowing of exclamations into a variety of languages lends further support to the notion that knowledge of other languages is not an absolutely necessary prerequisite for borrowing.

Acknowledgments The work on the Chadic part of this study was supported by the Innovative Seed Grant from the University of Colorado for the project “Linguistic evidence of Chadic migrations,” with Zygmunt Frajzyngier, Erin Shay, Marielle Butters, and Megan S. Schwabauer as participants. We are grateful to specialists in various areas for their help with various queries: Sean Allison (Trinity Western University, Canada) for his work on Makary Kotoko; David Appleyard (SOAS, University of London) for references on Semitic‐Cushitic contact; Christopher Ehret (UCLA), for sharing his recent work on several issues related to languages of Africa; John Huehnergard (University of Texas at Austin), for providing information and references on Akkadian‐Sumerian contact; Maarten Kossmann (University of Leiden), for generously sharing his work on Berber; Maarten Mous (University of Leiden), for sharing his work on Cushitic; and Georg Ziegelmeyer (University of Vienna), for sharing his work on language contact. We are also most grateful to Raymond Hickey for the insightful comments and questions. None of these colleagues is responsible for any errors of fact or interpretation in the present study.

Abbreviations ASSOC = associative, BEN = benefactive, COMP = complementizer, COMPL = completive, DED = deduced (reference), DEF = definite, DEM = demonstrative, EE = end‐of‐event, F = feminine, FOC = focus, FUT = future, GO = goal, HABIT = habitual, M = masculine, NEG = negative, POS = point of view of the subject, PRED = predicator, PREP = preposition, PRF = perfective, SEQ = sequential, SG = singular, VENT = ventive.

NOTE 1 Frajzyngier did observe use of the verb‐initial order in complements of verbs of saying in a Northern Moroccan variety of Arabic. If this was not an accidental use it may indicate that the verb‐initial order in Arabic left its traces in complement clauses.

588  Zygmunt Frajzyngier and Erin Shay

REFERENCES Allison, Sean 2007. Linguistic evidence for the Islamization of the Makary Kotoko by the Kanuri. In Henry Tourneux (ed.), Topics in Chadic Linguistics III. Historical Studies. Papers from the 3rd Biennial International Colloquium on the Chadic Languages, Villejuif, November 24–25, 2005. Cologne: Rüdiger Köppe, pp. 9–25. Allison, Sean 2017. Borrowings but no diffusion: a case of language contact in the Lake Chad basin. Journal of Language Contact 10: 395–421. Amha, Azeb 2012. Omotic. In Zygmunt Frajzyngier and Erin Shay (eds.), The Afroasiatic Languages. Cambridge: Cambridge University Press, pp. 423–504. Appleyard, David 2006. Agaw and Omotic: a case of language contact? In Siegbert Uhlig. (ed.), Proceedings of the 15th International Conference of Ethiopian Studies. Wiesbaden: Harrassowitz, pp. 697–703. Appleyard, David 2015. Ethiopian Semitic and Cushitic. Ancient contact features in Ge’ez and Amharic. In Aaron Michael Butts (ed.), Semitic Languages in Contact. Leiden: Brill, pp. 16–32. Bargery, G. P. 1934. A Hausa‐English Dictionary and English‐Hausa Vocabulary. Oxford: Oxford University Press. Breedveld, J. O. 1995. Form and meaning in Fulfulde. Leiden: Research School CNWS. Butts, Aaron Michael (ed.) 2015. Semitic Languages in Contact. Leiden: Brill. Clements, George N. 1975. The logophoric pronoun in Ewe: its role in discourse. Journal of West African Languages 10: 141–77. Cloarec‐Heiss, France 1969. Les modalités personnelles dans quelques langues oubanguiennes. Discours direct‐Discours indirect. Paris: SELAF. Cyffer, Norbert 2000. Areale Merkmale im TAM‐ System und der Syntax der Saharanischen Sprachen. In Rainer Voßen, Angelika Mietzner, and Antje Meißner (eds.), Mehr als nur Worte …. Afrikanistische Beiträge zum 65. Geburtstag von Franz Rottland. Cologne: Rüdiger Köppe, pp. 159–82. Cyffer, Norbert 2006. Kanuri and its neighbours: when Saharan and Chadic languages meet. In Paul Newman and Larry M. Hyman (eds.) 2006. West African Linguistics: Papers in Honor of Russell G. Schuh. Studies in African Linguistics, Supplement 11. Cleveland, OH: Ohio State University, pp. 33–55. Cyffer, Norbert 2009. Cause and reason in Kanuri – the impact of areality on linguistic

change. In Norbert Cyffer and Georg Ziegelmeyer (eds.), When Languages Meet. Language Contact and Change in West Africa. Cologne: Rüdiger Köppe, pp. 7–22. Cyffer, Norbert and Georg Ziegelmeyer (eds.) 2009. When Languages Meet. Language Contact and Change in West Africa. Cologne: Rüdiger Köppe. Cyffer, Norbert, Erwin Ebermann, and Georg Ziegelmeyer 2009. Negation Patterns in West African Languages and Beyond. Amsterdam: Benjamins. Diakonoff, Igor M. 1988. Afrasian Languages. Moscow: Nauka. Dimmendaal, Gerrit 2012. Marked nominative systems in Eastern Sudanic and their historical origin. Afrikanistik Online. Available at www. researchgate. net/ publication/274954995_Marked_nominative_ systems_in_Eastern_Sudanic_and_their_ historical_origin Dryer, Matthew S. 2009. Verb‐object‐negative order in central Africa. In Norbert Cyffer, Erwin Ebermann, and Georg Ziegelmeyer (eds.), Negation Patterns in West African Languages and Beyond. Amsterdam: Benjamins, pp. 307–62. Ebeling, Jarle 2004. Sumerian word order alterations. Available at http://etcsl.orinst. ox.ac.uk/edition2/pdf/sumwordorderalt.pdf Ehret, Christopher 1995. Reconstructing Proto‐ Afroasiatic (Proto‐Afrasian). Vowels, Tone, Consonants, and Vocabulary. University of California Publications in Linguistics 126. Ehret, Christopher 2006. The Nilo‐Saharan background of Chadic. In Paul Newman and Larry M. Hyman (eds.), West African Linguistics: Papers in Honor of Russell G. Schuh. Studies in African Linguistics, Supplement 11. Cleveland, OH: Ohio State University, pp. 56–66. Ehret, Christopher 2015. Africa from 48,000 to 9500 bce. In David Christian (ed.), The Cambridge World History, vol. 1: Introducing World History (to 10,000 bce). Cambridge: Cambridge University Press, pp. 362–93. Ehret, Christopher 2018. Berber peoples in the Sahara and North Africa: Linguistic historical proposals. In Maria Gatto, David Mattingly, Nick Ray, and Martin Sterry (eds.), Burial, Migration and Identity in the Ancient Sahara and

Contact and Afroasiatic Languages  589 beyond. Cambridge: Cambridge University Press, pp. 464–94. Frajzyngier, Zygmunt 1983. Marking syntactic relations in Proto‐Chadic. In Ekkehard Wolff and Hilke Meyer‐Bahlburg (eds.), Studies in Chadic and Afro‐Asiatic Linguistics. Hamburg: Buske, pp. 115–38. Frajzyngier, Zygmunt 1985a. Borrowed logophoricity? Precis from the 15th Conference on African Linguistics, Studies in African Linguistics, Supplement 9. (Russell G. Schuh, ed.) pp. 114–18. Frajzyngier, Zygmunt 1985b. Logophoric systems in Chadic. Journal of African Languages and Linguistics 7: 23–37. Frajzyngier, Zygmunt 1987. Ventive and centrifugal in Chadic. Afrika und Übersee 70.1: 31–47. Frajzyngier, Zygmunt 1989. A Grammar of Pero. Berlin: Reimer. Frajzyngier, Zygmunt 1993. A Grammar of Mupun. Berlin: Reimer. Frajzyngier, Zygmunt 2001. A Grammar of Lele (Stanford Monographs in African Linguistics). Stanford: CSLI. Frajzyngier, Zygmunt 2004. Tense and aspect as coding means. Journal of West African Languages 30.2: 53–67. Special issue, edited by Bernard Comrie and Ekkehard Wolff. Frajzyngier, Zygmunt 2008. A Grammar of Gidar. Frankfurt‐am‐Main: Peter Lang. Frajzyngier, Zygmunt 2011a. Les fonctions de l’ordre linéaire des constituants. Bulletin de la Société de Linguistique de Paris, 107.1: 7–37. Frajzyngier, Zygmunt 2011b. Possessive subject pronouns in Wandala: point of view of the subject. In Anne Storch, Gratien G. Atindogbé, and Roger M. Blench (eds.), Copy Pronouns. Case Studies from African Languages. Cologne: Rüdiger Köppe, pp. 47–64. Frajzyngier, Zygmunt 2012a. Typological outline of Afroasiatic languages. In Zygmunt Frajzyngier and Erin Shay (eds.), The Afroasiatic Languages. Cambridge: Cambridge University Press, pp. 505–624. Frajzyngier, Zygmunt 2012b. Coding relations between the verb and noun phrases in Afroasiatic: a sketch of typological explanations. In Matthias Brenzinger and Anne‐Maria Fehn (eds.), Proceedings of the 6th World Congress of African Linguistics, Cologne, 17–21 August 2009. Cologne: Koeppe, pp. 57–68. Frajzyngier, Zygmunt 2018. Afroasiatic languages. In Mark Aronoff (ed.), Oxford Research Encyclopedia of Linguistics. New York:

Oxford University Press. doi:10.1093/ acrefore/9780199384655.013.15 Frajzyngier, Zygmunt forthcoming. Borrowed grammaticalization and grammaticalizing on the borrowing. In Rainer Vossen and Christa Koenig (eds.), Language Contact in Africa. SUGIA (Sprache und Geschichte im Afrika). Cologne: Koeppe. Frajzyngier, Zygmunt, and Tamar Katriel 1991. Functions of propositional relators. In Allan Kaye (ed.), Semitic Studies in Honor of Wolf Leslau. Wiesbaden: Harrassowitz, pp. 449–61. Frajzyngier, Zygmunt and Wendy Ross 1991. Methodological issues in applying linguistics to the study of pre‐history. In Stanisław Piłaszewicz and Eugeniusz Rzewuski (eds.), Unwritten Testimonies of the African Past, Orientalia Varsoviensia 2: 21–44. Frajzyngier Zygmunt, with Erin Shay 2002. A Grammar of Hdi. Berlin: de Gruyter Mouton. Frajzyngier, Zygmunt and Erin Shay (eds.) 2012. The Afroasiatic Languages. Cambridge: Cambridge University Press. Frajzyngier, Zygmunt and Erin Shay 2008. Language‐internal versus contact‐induced change: the split coding of person and number: a Stefan Elders question. Journal of Language Contact THEMA 2 (2008). Language Contact and the Dynamics of Language. Available at www. jlc‐journal.org/ Frajzyngier, Zygmunt, and Erin Shay 2010. Tense and aspect in Chadic: a typological overview. Folia Orientalia 47: 278–315. Frajzyngier, Zygmunt with Erin Shay 2016. The Role of Functions in Syntax: A Unified Approach to Language Theory, Description, and Typology. Amsterdam: JBenjamins. George, Isaac 1975. Typology of verb serialization. Journal of West African Languages 10.1: 78–97. Grossman, Eitan, Peter Dils, Tonio Sebastian Richter, and Wolfgang Schenkel (eds.) 2017. Greek Influence on Egyptian‐Coptic: Contact‐Induced Change in an Ancient African Language. Berlin: Database and Dictionary of Greek Loanwords in Coptic (DDGLC). Working Papers 1. Hagège, Claude 1974. Les pronoms logophoriques. Bulletin de Société de Linguistique de Paris 69:1: 287–310. Heath, Jeffrey 1998. A Grammar of Koyra Chiini, the Songhay of Timbuktu. Berlin: de Gruyter Mouton. Hetzron, Robert and Zygmunt Frajzyngier 2018. Afroasiatic. In Bernard Comrie (ed.), The World’s Major Languages. London: Routledge, pp. 562–67.

590  Zygmunt Frajzyngier and Erin Shay Heusing, Gerald 1999. Aspects of the Morphology‐ Syntax Interface in Four Nigerian Languages (Beiträge zur Afrikanistik 11). Münster: LIT Verlag. Hyman, Larry M. and Bernard Comrie 1981. Logophoric reference in Gokana. Journal of African Languages and Linguistics 3: 19–37. Jaggar, Philip J. 2017. The Morphological‐to‐ Analytic Causative Continuum in Hausa. New Insights and Analyses in a Typological Perspective. Wiesbaden: Harrassowitz. Jungraithmayr, Herrmann 1980. Kontakte zwischen Adamawa‐Ubangi‐ und Tschad‐ Sprachen: Zur Übertragung grammatischer Systeme. Zeitschrift der Deutschen Morgenländischen Gesellschaft 130.1: 70–85. Jungraithmayr, Herrmann 1991. A Dictionary of the Tangale Language. Berlin: Reimer. Jungraithmayr, Herrmann 1994. “Zweite Tempora” in afrikanischen Sprachen – Ägyptisch‐Tschadische Gemeinsemkeiten. In Manfred Bietak, Johanna Holaubek, Hans Mukarowsky and Helmut Satzinger (eds.), Zwischen den beiden Ewigkeiten: Festschrift Gertrud Thausing. Vienna: Institut für Ägyptologie der Universität Wien, pp. 101–22. Jungraithmayr, Herrmann and Dymitr Ibriszimow 1994. Chadic Lexical Roots. Berlin: Reimer. Kammerzell, Frank 2005. Old Egyptian and pre‐Old Egyptian. Tracing linguistic diversity in Archaic Egypt and the creation of the Egyptian language. Thesaurus Linguae Aegyptiae 3, 165–247. Kossmann, Maarten G. 2002. Deux emprunts à l’égytien ancien en berbère. In K. Nait‐Zerrad (ed.), Articles de linguistique berbère: mémorial Werner Vycichl. Paris: L’Harmattan, pp. 245–52. Kossmann, Maarten G. 2005. Berber Loanwords in Hausa. Cologne: Rüdiger Köppe. Kossmann, Maarten G. 2007. The borrowing of aspects as lexical tone: y‐initial Tuareg verbs in Tasawaq (Northern Songhay), Studies in African Linguistics 36.2: 151–66. Kossmann, Maarten G. 2008. The collective in Berber and language contact. In Mena Lafkioui and Vermondo Brugnatelli (eds.), Berber in Contact. Linguistic and Sociolinguistic Perspectives. Cologne: Rüdiger Köppe, pp. 53–62. Kossmann, Maarten G. 2013a. Borrowing. In Jonathan Owens (ed.), The Oxford Handbook of Arabic Linguistics. Oxford: Oxford University Press, pp. 349–68. Kossmann, Maarten G. 2013b. The Arabic Influence on Northern Berber. Leiden: Brill.

Kossmann, Maarten G. 2015. Contact‐induced change. In Matthias Baerman (ed.). The Oxford Handbook of Inflection. Oxford: Oxford University Press, 251–71. König, Christa 2008a. Case in Africa. Oxford: Oxford University Press. König, Christa 2008b. The marked‐nominative languages of East Africa. In Bernd Heine and Derek Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, pp. 251–71. Leslau, Wolf 1945. The influence of Cushitic on the Semitic languages of Ethiopia: a problem of substratum. Word 1: 59–82. Leslau, Wolf 1952. The influence of Sidamo on the Ethiopic languages of Gurage. Language 28: 63–81. Löhr, Doris 1998. Sprachkontakte bei den Malgwa (Gamergu) in Nordostnigeria. In Ines Fiedler, Catherine Griefenow‐Mewis, and Brigitte Reineke (eds.), Afrikanische Sprachen im Brennpunkt der Forschung. Cologne: Rüdiger Köppe, pp. 251–69. Manfredi, Stefano and Mauro Tosco (eds.) 2018. Arabic in Contact. Amsterdam: Benjamins. Moreno, Martino Mario 1948. L’azione del cuscito sul sistema morfologico delle lingue semitiche dell’Ethiopia [The effect of Cushitic on the morphological system of the Semitic languages of Ethiopia]. Rassegna di Studi Etiopici 7: 121–30. Mous, Maarten P. 1995. Ma’a or Mbugu. In Peter Bakker and Maarten P. Mous (eds.), Mixed Languages: 15 Case Studies in Language Intertwining, Amsterdam: Benjamins, pp. 175–200. Mous, Maarten P. 2013. Mixed Ma’a/Mbugu. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), Atlas and Survey of Pidgin and Creole Languages, vol. 3: Contact Languages based on Languages from Africa, Asia, Australia, and the Americas. Oxford: Oxford University Press, pp. 42–9. Mous, Maarten P. 2012. Cushitic. In Zygmunt Frajzyngier, and Erin Shay (eds.), The Afroasiatic Languages. Cambridge: Cambridge University Press, pp. 342–422. Mous Maarten P. 2013. Mixed Ma’á /Mbugu. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds), Atlas and Survey of Pidgin and Creole Languages, vol. 3: Contact Languages based on Languages from Africa, Asia, Australia and the Americas. Oxford: Oxford University Press, pp. 42–9. Newman, Paul 1977. Chadic classification and reconstructions. Afroasiatic Linguistics 5.1: 1–42.

Contact and Afroasiatic Languages  591 Newman, Paul 2000. The Hausa Language: An Encyclopedic Reference Grammar. New Haven: Yale University Press. Noye, Dominique 1974. Cours de foulfouldé. Maroua: Mission Catholique, and Paris: Paul Geuthner. Schuh, Russell G. 2011. Grammatical influences of Kanuri on Chadic languages of Yobe State. In Doris Löhr, Eva Rothmaler, and Georg Ziegelmeyer (eds.), Kanuri, Borno and Beyond. Current Studies on the Lake Chad Region. Cologne: Rüdiger Köppe, pp. 137–54. Schwarzwald, Ora R. 2011. Modern Hebrew. In Stefan Weninger, Geoffrey Khan, Michael P. Streck, and Janet C. E. Watson (eds.), Semitic Languages: An International Handbook (Handbooks of Linguistics and Communication Science 36). Berlin: de Gruyter Mouton, pp. 523–36. Shay, Erin 2012. Ms. A grammar of Giziga. Souag, Lameen 2010. Grammatical contact in the Sahara: Arabic, Berber, and Songhay in Tabelbala and Siwa (School of Oriental and African Studies, University of London Ph.D. dissertation). Souag, Lameen 2012. The subclassification of Songhay and its historical implications. Journal of African Languages and Linguistics 33.2: 181–213. Souag, Lameen 2014. Berber and Arabic in Siwa (Egypt): A Study in Linguistic Contact. Cologne: Rüdiger Köppe. Stahlke, Herbert 1970. Serial verbs. Studies in African Linguistics 1.1: 60–99. Storch, Anne, Gratien G. Atindogbé, and Roger M. Blench (eds.) 2011. Copy Pronouns. Case Studies from African Languages. Cologne: Rüdiger Köppe. Thomason, Sally 2001. Contact‐induced typological change. In Martin Haspelmath, Ekkehard Koenig, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals, Sprachtypologie und sprachliche Universalien: An International Handbook. Berlin: Gruyter, pp. 1640–8. Torallas Tovar, Sofía 2017. The reverse case: Egyptian borrowing in Greek. In Eithan Grossman, Peter Dils, Tonio Sebastian Richter, and Wolfgang Schenkel (eds.), Greek Influence

on Egyptian‐Coptic: Contact‐Induced Change in an Ancient African Language. Berlin: Database and Dictionary of Greek Loanwords in Coptic (DDGLC). Working Papers 1, pp. 97–113. von Sodden, Wolfram 1995. Grundriss der akkadischen Grammatik, 3rd edition. Rome: Pontificium Institutum Biblicum. Winand, Jean 2017. Identifying Semitic loanwords in Late Egyptian. In Eithan Grossman, Peter Dils, Tonio Sebastian Richter, and Wolfgang Schenkel (eds.), Greek Influence on Egyptian‐Coptic: Contact‐Induced Change in an Ancient African Language. Berlin: Database and Dictionary of Greek Loanwords in Coptic (DDGLC). Working Papers 1, pp. 481–511. Winford, Donald 2003. An Introduction to Contact Linguistics. Oxford: Blackwell. Wolff, H. Ekkehard and Ludwig Gerhardt 1977. Interferenzen zwischen Benue‐Kongo und Tschad‐Sprachen. Zeitschrift der Deutschen Morgenländischen Gesellschaft, Suppl. 3: 1518–43. Zeldes, Amir 2013. Is Modern Hebrew standard average European? The view from European. Linguistic Typology 17.3: 439–70. Ziegelmeyer, Georg 2009a. Between Hausa and Kanuri – on the linguistic influence of Hausa and Kanuri on Bade and Ngizim. In Eva Rothmaler (ed.), Topics in Chadic Linguistics V. Comparative and Descriptive Studies. Cologne: Rüdiger Köppe, pp. 173–85. Ziegelmeyer, Georg 2009b. Negation of non‐ indicative mood in Hausa, Fulfulde and Kanuri. In Norbert Cyffer, Erwin Ebermann, and Georg Ziegelmeyer 2009. Negation Patterns in West African Languages and Beyond. Amsterdam: Benjamins pp. 7–20. Ziegelmeyer, Georg 2009c. Areal features in northern Nigeria – towards a linguistic area. In, Petr Zima (ed.), The Verb and Related Areal Features in West Africa – Continuity and Discontinuity within and across Sprachbund Frontiers. Munich: LINCOM Europa, pp. 269–306. Ziegelmeyer, Georg 2009d. The Hausa particle “koo” – a widely spread formative in northern Nigeria. In Norbert Cyffer and Georg Ziegelmeyer (eds.) 2009. When Languages Meet. Language Contact and Change in West Africa. Cologne: Rüdiger Köppe, pp. 65–91.

30 Contact and North American Languages MARIANNE MITHUN

Languages indigenous to North America offer excellent opportunities for investigating effects of contact, because of their genealogical diversity and areas of intense interaction among their speakers. Around 275 distinct languages are known to have been spoken north of Mexico at the time of first contacts with Europeans. They are not a monolithic group: they fall into over 55 distinct families. Yet against this backdrop of diversity, waves of typological similarities suggest pervasive, long‐standing multilingualism. North America shows many of the kinds of contact effects common elsewhere in the world. Some can be seen between pairs of languages, but others pervade linguistic areas, particularly on the Northwest Coast, in California, in the Southeast, as well as sub‐areas, such as Northwest California, the Clear Lake Region of California, and the Pueblo Southwest (Bright 1973; Sherzer 1973; Haas 1976; Thompson and Kinkade 1990; Silverstein 1996; Campbell 1997; Mithun 1999, 2017; Beck 2000; Jany 2007, Golla 2011). In some cases all domains of language have been affected, but in others only some, and not always those that might be predicted.

1  Detecting Contact Without Philology For most languages of the Americas, there are no written records comparable to those for major languages of Europe. Many communities did not encounter Europeans until the late eighteenth or nineteenth century. It is thus not generally possible to trace the effects of ancient contact philologically. Alternative strategies must often be explored. The clearest evidence of contact is of course loanwords. Some North American languages show the same kinds of lexical loans as languages elsewhere. The word háyu ‘dog’, for example, appears in neighboring but genetically unrelated languages of Northern California: in the Pomoan languages; in Bodega Miwok, Lake Miwok, and Southern Sierra Miwok but not the related Central Miwok or Northern Sierra Miwok; in Hill Patwin but not its sister Wintu; in Maidu but not its sister Nisenan; in the Western dialect of Wappo but not the Southern (Napa) dialect. Many American languages contain loans from the European languages of colonists: French in the Northeast, French and Spanish in the Southeast, Spanish in the Southwest and California, and Russian in Alaska (Mithun 1999: 311–13). But the contact suggested by loanwords need not have been direct. Spanish loanwords in many California languages, for example, were not adopted directly from Spanish speakers but rather through the intermediary of other California languages.

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

594  Marianne Mithun Among the words copied across indigenous languages are items once thought to be unborrowable. Pronouns, particularly full paradigms, have sometimes been cited as indicators of deep genetic relationship. Yet Yuki and Wappo, two California languages, borrowed first and second person pronouns from the neighboring but unrelated Pomoan languages (Mithun 2008). (Third persons are unmarked or demonstratives.) There are even cases of borrowed bound pronouns. In Alsea, on the Oregon Coast, subject enclitics are attached to the first element of the clause. The full set of enclitics perfectly matches that reconstructed for the unrelated Proto‐Salishan, immediately to the north (Kinkade 1978). Perhaps surprisingly, many languages show relatively little copied vocabulary despite longstanding multilingualism. Multilinguals sometimes take special pains to keep their languages distinct, often with a focus on vocabulary. They may, however, be less conscious of abstract structures, and the result can be parallel grammatical distinctions and constructions. As will be seen, such effects are pervasive in areas of North America. Establishing contact as the source of structural similarities, however, can be challenging. Chance can be a significant factor: many languages show basic verb‐final (SOV) clause structure, for example, simply because the alternatives are limited. One important strategy for detecting contact‐induced grammatical change is the comparison of structures in related languages spoken outside of the geographical area. Features shared by a language with its neighbors but not its relatives are more likely to be a result of contact. This situation can be illustrated with consonant inventories in the Algic family. The family consists of the Yurok and Wiyot languages on the California Coast, and the nearly 30 Algonquian languages distributed across the continent from Alberta, Montana, and Wyoming to the Atlantic. The inventories of the two California languages differ strikingly from those of their Algonquian relatives. (1) Algic consonant inventories a. Yurok: 27 (Blevins 2003) p, t, č, k, kw, p’, t’, č’, k’, k’w, s, ł, š, x, m, n, l, r, w, y, w, γ ’w, ’l, ’r ’y, ’γ, ʔ, h b. Wiyot: 25 (Teeter and Nichols 1993) p, t, c, č, k, kw, ph, th, ch, čh, kh, khw, s, ł, š, h, b, d, g, l, r, m, n, w, y c. Proto‐Algonquian: 13 (Bloomfield 1946) p, t, č, s, š, h, m, n, θ, l, w, y The Yurok and Wiyot inventories resemble those of their Northern California neighbors, Chimariko and the Pacific Athabaskan languages (Hupa, Tolowa, Mattole, and Eel River dialects). Chimariko contains 33 distinctive consonants with plain, ejective, and aspirated obstruents, and front and back apicals: p, t, ṭ, c, č, k, q, p’, t’, ṭ’, c’, č’, k’, q’, ʔ, ph, th, ṭh, ch, čh, kh, qh, s, š, x, x̣, h, m, n, l, r, y, w (Jany 2007). The Pacific Athabaskan languages contain in addition a voiceless lateral and labio‐velars: b, d, c, č, ky, g, q, th, ch, čh, čhw, khy, kh, t’, c’, č’, tł’, ky’, k’, q’, s, ł, š, xw, x, m, n, ŋ, l, y, w, w̥ , ’n, ’ŋ, ’l, ʔ, h (Golla 1996). There is consensus that there are no genealogical links between the Algic languages and their Chimariko and Athabaskan neighbors. But Northern California is known as an area of longstanding multilingualism. Communities have always been small, and intermarriage the norm. The consonant inventories reflect this history. Where languages in contact are related, it can be difficult to distinguish common structural inheritances from copied ones. Similar structures could be the result of common inheritance or drift, developments that were triggered by an inherited feature but occurred after the languages had separated. In some cases a clear understanding of relationships within the family and documented histories of locations and contacts can help. The Iroquoian languages present such puzzles (Mithun 2013a, 2013b; Lukaniec and Chafe 2016). The family first split into two main branches, Southern Iroquoian and Northern Iroquoian. Then one

Contact and North American Languages  595 group, the Tuscarora, separated from the Northern branch. They were first encountered by Europeans in what is now North Carolina in the sixteenth century. In 1713 they migrated north to settle near their Five Nations relatives in New York State. Their language thus apparently developed independently over several millennia, but was then in close contact with related languages until the late nineteenth century, when people in all communities began to learn English. All of the Iroquoian languages are highly polysynthetic, with elaborate but similar morphological structures. Tuscarora contains some copied words, noun and verb stems, roots, and affixes. Fortunately, certain phonological changes in Tuscarora make it possible to distinguish many of the copies from cognates. Intriguingly, some Tuscarora words contain clearly copied roots but native affixes. What we cannot know is whether the bound morphemes were copied directly, since the morphological structures were so close, or came into the language inside of whole words, which were later analyzed once a large enough set of morphologically related copied words had been accumulated. Not surprisingly, there are also replicated semantics, pragmatics, and extension of grammatical functions. A Proto‐Northern Iroquoian verb root *‐iyo, for example, meant ‘be large, great, grand’. In the modern Five Nations languages it now means ‘be nice, good’. This use also occurs on occasion in modern Tuscarora. In Proto‐Northern‐Iroquoian, there were just two basic genders, masculine (for male persons) and feminine‐neuter (for others), as well as a generic category ‘one, people’. At a certain point, speakers of Five Nations languages began using the generic as a sign of respect for some women. In Cayuga and Seneca, this became the basic feminine. This usage was adopted by the Tuscarora, though earlier forms survive in traditional personal names (Mithun 2014).

2  Structural Copying: Sound Patterns Like the Algic example just seen, much of what we know about contact effects is necessarily based on circumstantial evidence. Such evidence can be strong, however. Further examples of replicated phonological distinctions and processes occur elsewhere on the continent. The Uto‐Aztecan languages are spoken over an area from Idaho south to El Salvador, and the California coast east to Oklahoma. In Southern California, languages of the Cupan subbranch branch of the family, Luiseno, Cahuilla, and Cupeño, were in close contact with Yuman languages. Hinton (1991) shows that the Cupan phonological inventories were strongly affected by this contact, with added distinctions between labio‐velar and labio‐ uvular stops kʷ/qʷ, back and front sibilants s/ṣ, liquids r/l, and distinctive xʷ, ñ, and lʸ. In Northern California, the Utian languages generally have a single series of stops, but one, Lake Miwok, has developed distinctive series of plain, ejective, aspirated, and voiced stops as a result of contact with speakers of neighboring Pomoan languages and Patwin (Wintuan family) (Callaghan 1964), as well as a velar/uvular k/q contrast. Phonological processes have also spread. Somewhat unusually typologically, languages from three different families on the Northwest Coast all lack nasals: Twana and Lushootseed (Salishan), Quileute (Chimakuan), and Makah and Nitinat (Wakashan) (Thompson and Thompson 1972; Kinkade 1985). But not all languages in any of the families show the shift. Kinkade points out that “in virtually every littoral language of the Northwest from the 46th to the 50th parallel nasals were sometimes pronounced without full closure of the velum” (1985: 478). Early phonetic transcriptions of languages in the area show alternations between nasal and oral voiced stops. “These sounds then settled out as voiced stops in six languages, but reverted to pure nasals in the rest” (1985: 479). Another widespread pattern is in the West, where a number of languages show sound symbolism, whereby certain sounds or sound shifts are associated with specific meanings, often size or intensity (Nichols 1971; Mithun 1999: 31–3).

596  Marianne Mithun

3  Structural Copying: Morphology It has often been noted that morphological structure is highly resistant to the influence of contact. Morphological similarities were once even proposed as better indicators of deep genetic relationship than the traditional comparative method. In his attempts to group North American families into larger superstocks, Sapir was adamant that morphology outlasts cognates: “so long as such direct historical testimony as we have gives us no really convincing examples of profound morphological influence by diffusion, we shall do well not to put too much reliance in diffusion theories” (1921: 206). The principle seems reasonable. The internal structure of words is generally less accessible to the consciousness of speakers, and, one would expect, less easily manipulated by bilingual speakers seeking to bring structures from one of their languages into the other. Yet numerous morphological parallelisms appear in neighboring but genetically unrelated American languages. It might seem difficult to imagine how such structures could be transferred under contact: they involve abstract, largely unconscious patterns without the words or morphemes that carry them. But various routes could lead to this result. It is likely that most of these routes begin with what is sometimes termed frequential copying, the replication by bilinguals of frequencies of expression of particular distinctions or translations of turns of phrases from one language into another Johanson (2002, 2008, 2013). We know that grammatical structures most often arise via grammaticalization processes: lexical items gradually lose their concrete meanings and phonetic substance, ultimately evolving into grammatical affixes. Transfer could occur at various points in such developments. A language might contain co‐existing cognate constructions at various stages of grammaticalization, one form retaining its original concrete, lexical meaning, the other having developed a more abstract grammatical meaning. By analogy, bilinguals might extend the use of the lexical counterpart in their other language to a similar grammatical function. Alternatively, a construction might be translated word‐for‐word by bilinguals into their other language, perhaps also replicating its frequency of use. One or more of the components might later undergo processes of grammaticalization in each language individually, again resulting in parallel morphological constructions. Finally, a compound construction and/or its frequency, might be copied by bilinguals from one language to another. At this point the component roots, stems, or words would still retain concrete lexical meanings. One component could then undergo functional abstraction and phonetic erosion in place, either in each language in isolation or under continuing contact. Again the result could be parallel morphological structure. Contact‐induced grammaticalization has been discussed by a number of authors, among them Matisoff (1991), Heine (1994, 2009, 2011), Haase and Nau (1996), Bisang (1996), Kuteva (1988, 2000, 2008), Heine and Kuteva (2003, 2005, 2006), Aikhenvald and Dixon (2007), Matras and Sakel (2007), Ramat and Roma (2007), Ramat (2008), Narrog and Heine (2011), Gast and van der Auwera (2012), Heine and Nomachi (2013), Robbeets and Cuyckens (2013), and Mithun (2018). All of these processes can be seen in North America. In the sections which follow, three are discussed which are particularly characteristic of the continent.

3.1  Lexical Suffixes All of the Wakashan languages of the Northwest Coast contain large inventories of suffixes with meanings typical of roots in other languages, meanings that often seem more concrete and specific than those usually associated with affixes cross‐linguistically. Some of the suffixes have meanings expressed in other languages by noun roots, such as ‐sac ‘bag’, ‐sii ‘family’, ‐’aqs ‘woman’, ‐qʔičh ‘year’, and ‐ʔin ‘costume’. Some are expressed in other languages by verb roots, such as ‐ḥa ‘buy’, ‐naqa ‘use as bait’, ‐taqa ‘blame’, ‐ht ‘exit the woods’,

Contact and North American Languages  597 ‐’ałak ‘love’, and ‐’i’s ‘copulate’. Some are expressed by adjectives, such as ‐ap’ii ‘coiled’, ‐siikw ‘complete’, ‐aaʔa  ‘destined for’, ‐ʔał ‘aware of’, and ‐isim ‘principal’. Some have adverbial meanings, many indicating locations or directions, such as ‐su:ʔis ‘far out at sea’, ‐aˑt ‘out of the woods’, ‐’a:či ‘in a bay’, ‐spuł ‘between the legs’, ‐it ‘in the body’, ‐yin ‘at the bow of a boat’, and ‐saq  ‘under the covers’ (Stonham 2005). Some examples of their uses are below. (2) Ahousaht Nuuchahnulth (Nakayama 2003: 323) susw’isčaʔa. sus‐w’isč‐’aʔa’. swim‐move.up.bank‐on.rock ‘They swam ashore onto the rocks.’ (3) Ahousaht Nuuchahnulth: (Nakayama 2003: 378) ḥaaw’iła łučḥaa qwaaaa ḥa:w’iła łuč‐ḥaˑ qwa:‐ a: woman‐buy also young.man ‘On the other hand, a young man proposes ʔucači ḥaw’iłʔi ḥakwanakʔi. ʔu‐ca‐či ḥaw’iłʔi ḥakwa‐naˑk‐ʔi. it‐go.to‐MOM chief‐DEF daughter‐having‐DEF by going to the chief who has a daughter.’ One might wonder whether these are indeed suffixes. Formally, there is no question about their status. The languages are uniquely suffixing, and these morphemes never occur at the beginning of a word. They always follow a stem. Functionally they can differ subtly from stems as well. The languages contain many suffix:stem counterparts with similar meanings. Stonham lists both a suffix ‐sac ‘bag’ and an unrelated word niisaakw ‘bag’; a suffix ‐sii ‘family’ and a word ʔuštaqimł ‘family’; a suffix ‐’aqs ‘woman’ and a word łuucsma ‘woman’; a suffix ‐’aap ‘buy’ and a root maakuk ‘buy’; a suffix ‐taqa ‘blame’ and a root wišk ‘blame’. Despite their concrete and specific translations, many of these suffixes do have broader meanings than their stem counterparts. The suffix ‐sac is translated by Stonham variously as ‘vessel, dish, box, container’. The suffix ‐saq ‘under the covers’ is also ‘under one’s clothing’, ‘in a shelter’, and ‘inside’. Still, the lexical suffixes found throughout the Wakashan family are typologically unusual for their concrete meanings and large inventories, numbering in the hundreds. What is even more surprising is that large inventories of similar suffixes appear in the neighboring Chimakuan and Salishan families as well. A sample of Quileute (Chimakuan) suffixes includes ‐takil‐ ‘footprint’, ‐wi:yi’‐ ‘wall’, ‐spe:‐‘fire’, ‐t’ło:‐ ‘dirt’, ‐sida ‘water’, ‐x̣a‐ ‘eat’, ‐k ‘go to a definite place’, t’s‐ ‘make’, ‐t’co’‐ ‘have inside’, ‐kits’ ‘dance, kick’, ‐qo‐ ‘make use of’, and ‐sqa‐ ‘carry’ (Andrade 1933: 262). Lushootsheed (Salishan) suffixes include ‐us ‘face, head, upper part’, ‐ač ‘hair of head, crest or topknot of bird, hackles of dog’, ‐ič ‘cover, surface, on top of, over, string, cord, spine’, ‐ulč ‘container, belly’, ‐iʔł ‘baby, child’, ‐ikw ‘shirt’, ‐lc ‘round thing, money, curved objects’, ‐dup ‘ground, floor’, ‐iqad ‘incline, slope, bank, hill’, ‐ucid ‘body of water’, ‐ac ‘tree, bush’, and ‐ic’aʔ ‘clothe, wear, support from shoulder’ (Bates, Hess, and Hilbert 1994). In these languages as well there is no question about the formal status of the suffixes. They never serve as the foundation of a word on their own, but must always follow a stem. They also tend to show more general and diffuse meanings than their root counterparts. The cross‐linguistic rarity of affixes like these suggests that the parallelism here is the result of contact. But the suffixes themselves have not been borrowed. How could such a deeply embedded morphological structure be transferred? The pattern may not have been copied in its current state. The functions of the suffix constructions are strikingly close to

598  Marianne Mithun those of compounds, in many cases noun incorporation. They are used to create lexical items, such as the Nuuchahnulth qawaš‐sac ‘salmonberry dish’ (‘salmonberry‐container’) and łuč‐ḥaa ‘propose’ (‘woman‐buy’). They are also used to convey background information not worthy of the attention given to separate words. A likely origin of the lexical suffix constructions is in compounds. Compounding is common cross‐linguistically and can be reconstructed for Proto‐Salishan. A propensity for compounding could easily be spread by bilinguals. As single words, compounds have just one primary stress. Over time, roots that occurred as unstressed members of a substantial number of compounds could undergo further phonological reduction, resulting in the lexical suffixes of today. The large inventories would be explained by the fact that the components of compounds are drawn from the full inventory of stems. Their relatively concrete meanings are unsurprising, since they became bound while they still had their earlier stem meanings. The fact that they do not designate syntactic arguments or specify particular syntactic roles follows from an origin as members of compounds. (More detailed discussion can be found in Mithun 1997, 1998.)

3.2  Manner and Direction Another abstract morphological structure shows a wide distribution in the West, particularly in modern Oregon, California, and Nevada. It appears in languages of numerous distinct families, crossing boundaries between even the deepest hypothesized superstocks. Central Pomo, a Northern California language seen earlier, contains a set of prefixes that occur pervasively throughout the verbal lexicon: (4) Some Central Pomo verbs t̯’é:č’ ‘stick together, be alongside of each other’ da‐t̯’é:č’ ‘push on something that sticks in your hand’ ’‐t̯’é:č’ ‘stick on with fingers, as chewing gum under a table’ ma‐t̯’é:č’ ‘step on a nail or something that sticks in your foot’ ča‐t̯’é:č’ ‘sit on a thorn, put a patch on pants’ h‐t̯’é:č’ ‘stick up a pole, pitchfork, shovel, etc. in the ground’ m‐t̯’é:č’ ‘catch fire’ ph‐t̯’é:č’ ‘hammer a nail into the wall, nail something on’ pha‐t̯’é:č’ ‘something floating downriver gets stuck on the bank’ s‐t̯’é:č’ ‘while one is drinking, something gets into the mouth that does not belong, such as a bug or dirt’ ša‐t̯’é:č’ ‘stick a support, as a box, next to something long, like fence posts stored upright for use’ Such morphemes are sometimes called ‘instrumental prefixes’, because they suggest a means or manner of motion, but there is no explicit specification of the role of an entity beyond general involvement. They are not specifically nominal or verbal: often translations like either ‘with the foot’ or ‘by stepping’ would be appropriate. They can also co‐occur with nouns specifying an instrument: (5) Central Pomo prefixed verbs with nouns čhbá:č’. Qhabéwi h č h‐bá:‐č’ q abé=wi rock=with massive.object‐split‐sml.pfv ‘He cracked it open with a rock.’

Contact and North American Languages  599 The prefixes can be reconstructed for Proto‐Pomoan. Prefixes with similar functions occur in a number of other families and isolates in the area: Yuman, Karuk, Yana, Palaihnihan, and Washo. All of these were included at one time or another in proposals for a larger Hokan stock by Sapir and others. But the prefixes are absent from other languages and isolates grouped as Hokan: Shasta, Esselin, and Salinan. They also occur in California languages not grouped as Hokan. They appear throughout the Chumashan family, in Wappo, and Yuki. They occur in some isolates and families hypothesized by Sapir to be part of a larger Penutian stock in California, Oregon, and Idaho: Maiduan, Klamath, Takelma, and Sahaptian (Sahaptin and Nez Perce). But they are absent from other families and isolates grouped as Penutian: Wintuan, Utian (Miwok‐Costanoan), and Yokutsan. They even occur in a number of Uto‐ Aztecan languages of the area: Kawaiisu, Tümpisa (Panamint) Shoshone, etc. The languages vary substantially in their inventories and the productivity of prefixes, and none of the forms themselves are cognate across genetic lines. Central Pomo also contains a set of directional suffixes: (6) Central Pomo suffixes čá‐w ‘run’ (one) čá‐:la‐w ‘run down’ čá‐:qač’ ‘run up (as up a hill)’ čá‐č’ ‘run away’ čá‐way ‘run against hither, as when a whirlwind came up to you’ čá‐:’w‐an ‘run around here and there’ čá‐mli‐w ‘run around it (a tree, rock, house, pole etc.)’ čá‐mač’ ‘run northward’ čá‐:q’ ‘run by, over (along on the level), southward’ čá‐m ‘run over, on, across (as bridge)’ The suffixes can appear in verbs containing the means/manner prefixes described in the beginning of the current section. (7) Central Pomo prefix–suffix combinations da‐dí‐:la‐w ‘push something over a cliff’ ma‐dí‐:la‐w ‘kick something over a cliff’ ‘slowly glide into a swimming pool’ pha‐dí‐:la‐w ‘jump down, over a cliff, into the water’ ph‐dí‐:la‐w ča‐dí‐:la‐w ‘chase (dog) downhill’ ba‐dí‐:la‐w ‘walk downhill singing’ ’‐dí‐:la‐w ‘carry something downhill in hands’ š‐dí‐:la‐w ‘carry something downhill by the handle’ Like the prefixes, these directional suffixes are pervasive throughout the verbal lexicon and can be reconstructed for Proto‐Pomoan. Suffixes with similar meanings occur in other languages once proposed as part of Hokan: Karuk, Shasta, Palaihnihan (Atsugewi, Achumawi), and Yana. There is, however, no mention of them in other languages grouped as Hokan, even some that contain means/manner prefixes, such as languages of the Yuman family and Washo. They appear in some isolates and families proposed as part of Penutian: Maidun, Klamath, and Sahaptian (Sahaptin, Nez Perce). But other languages and families grouped as Penutian lack them, including some that contain the prefixes: Wintun, Utian, Yokuts, Takelma, Coos, Siuslaw, and Alsea. Again, the forms are not shared across genetic lines. There is thus a widespread morphological structure, appearing with varying degrees of robustness over a geographical area that extends through California, Oregon, and regions

600  Marianne Mithun inland. The prefixes occur in over a dozen genealogically distinct units, and the suffixes in seven. Their functions are strikingly similar across the languages, but the forms differ. An obvious explanation would be contact. But how could such abstract structure, below the level of consciousness of words, be transferred without the forms themselves? It seems unlikely that bilingual speakers would spontaneously create affixes in one of their languages by analogy with affixes in the other. And it is unlikely that the affixes rode into the language on borrowed words, because there is so little copied lexicon. Here again, the structures need not have been transferred in their modern state. It is more likely that the precursors to these structures were copied. The most obvious precursors are particular compounding patterns. Prefixes at an early stage of development can be seen in languages near the periphery of the area. All of the Uto‐Aztecan languages show extensive compounding: Noun–Noun, Verb–Verb, and Noun–Verb compounds. Some languages of the Numic branch of the family show prefixed verbs as well. (8) Kawaiisu (Zigmond, Booth, and Munro 1991) ‘hand‐wash’ Noun root moʔo‐ ‘hand’ moʔo‐zɨgɨ moʔo‐paʔa ‘hand‐stir’ ‘stir by hand’ ma‐gavi‐ ‘manually‐cut’ Prefix ma‐ ‘manually’ ma‐guri‐ manually‐circle’ ‘stir by hand’

‘wash one’s hands’ ‘break off’

A typical feature of nouns in compounds is the lack of a specific syntactic relationship to the verb. As can be seen on above, the noun root in Kawaiisu noun–verb compounds can represent entities in a variety of semantic roles, including that of an instrument. Kawaiisu also contains directional suffixes. (9) Kawaiisu directional suffixes (Zigmond et al. 1991) ʔga‐kwee‐ ‘go in’ ʔga‐ ‘enter’ yaa‐kwee‐ ‘take one’ yaa‐ ‘carry one’ ‘take several’ huʔma‐ ‘carry several’ huʔma‐kwee‐

ʔga‐ki‐ yaa‐ki‐ huʔma‐ki‐

‘come in’ ‘bring one’ ‘bring several’

Kawaiisu still has a verb root ‐kwee ‘go’ and a verb root ‐ki ‘come’. The directional suffix constructions appear to be descended from verb–verb compounds. The morphological structure so prevalent in the area today appears to have developed from compounding patterns. Bilinguals could easily spread a tendency to form noun–verb or verb–verb compounds with an initial member indicating a means or manner of motion, and a tendency to form verb–verb compounds with a second member specifying direction. Over time, frequently‐occurring unstressed elements of such compounds could be reduced to affixes like those seen throughout the area (Mithun 2007a).

3.3  Morphological Form: Clitic Structures The North Wakashan language Kwak’wala shows a somewhat unusual morphological structure. Case is marked on demonstratives which precede the noun phrase. What is surprising is that the clitics are attached phonologically not to the following noun phrase in their scope, but to the preceding word, whatever its function. (10) Kwak’wala nominative and accusative case (Boas 1911a: 557) dó:x’waλél=e Dzá:wadalalisa=x̲‐a élkwa. doqw‐’aλela=e Dzawadalalisa=x̲‐a elkw see‐suddenly=nom.proper  name=acc.com‐dem blood ‘Dzawadalalis [nominative] saw the blood [accusative].’

Contact and North American Languages  601 (11) Kwak’wala nominative and oblique case (Boas 1911a: 533) ’né:xyso:’la=e Q’ámtalała=s Q’áne:qe:’lakw ’ne:xy‐so:‐’la=e Q’ámtalała=s Q’áne:qe:’lakw tell‐pass=hrs=nom.proper name=obl name ‘It is said, Q’amtalał [nominative] was told by Q’aneqe’lakw [oblique].’ The full clitic structure cannot be reconstructed for Proto‐Wakashan. In the two other North Wakashan languages, Heiltsuk and Haisla, it occurs only before obliques. (12) Heiltsuk (Rath 1981: 1.85) hi=s Dáduqvlá wísmá‐x̣i w’ác’iá‐x̣ dem=obl watch man‐dem dog‐dem ‘The man watched a dog with [oblique] binoculars.’

dúğváyúa‐x̣i. binocular‐dem

Case clitics do not occur at all in the South Wakashan languages. Even more remarkably, this unusual clitic structure appears in languages in the unrelated Tsimshianic languages, immediately to the north of the North Wakashan languages. The systems are not identical: while the Kwak’wala enclitics show a nominative/accusative pattern, the Tsimshianic enclitics show an ergative/absolutive pattern, which distinguishes common from proper nouns. Tsimshianic clitics are discussed in detail in Stebbins (2003). (13) Sm’algyax (Coast Tsimshian) (Boas 1911b cited in Mulder 1994: 204) Da gwaant=ga’ wii gyisiyaask. then blow=com.absent.abs great northwind ‘Then the great northwind blew.’ (14) Sm’algyax (Boas 1911b cited in Mulder 1994: 81) Dm dzakda=sga ġibaw=ga FUT kill=com.absent.erg wolf=com.absent.abs ‘The wolf will kill the dog.’

haas‐ga dog‐dem

There is no doubt about the phonological bond between the clitics (called “connectives” in the Tsimshianic literature) and the preceding word. Dunn observes: In hesitating and pausing, speakers always tie the connective to the preceding word, that is, they always pause after a connective. They never continue a sentence (after a pause) by starting with a connective; they may repeat the last word before the pause but never just the connective. (Dunn 1979: 131–2)

The cross‐linguistic rarity of this structure suggests that the similarity is unlikely to be due to chance. The transfer of such a pattern of bound morphology seems at first unlikely. But again, the structure need not have been transferred in its modern form. A number of languages in North America show a recurring rhetorical structure by which speakers manage the flow of information. Typically speakers introduce no more than one major new item of information at a time in a prosodic phrase (Chafe 1987, 1994; Pawley 2000). This information might set the scene, introduce a new participant, present a new event, etc. Particularly in predicate‐initial languages like those of the Wakashan and Tsimshianic families, a prosodic phrase may consist of an initial verb that provides an outline of an event, followed by a demonstrative that functions cataphorically to promise further information to come, such as more precise identification of a participant. Such a structure can be seen in the South Wakashan language Nuuchahnulth.

602  Marianne Mithun (15) Ahousaht Nuuchanulth (Nakayama 2003: 586) histaqšiitquu his‐taq‐ši‐it‐qu: there‐come.from‐mom‐past‐cond.3

ʔaḥ, ʔaḥ, this

hitaqił. hita‐aqił. there.MOM‐in.a.sound ‘He could have started from here – the Puget Sound area.’ A recurring rhetorical pattern of this type, where a demonstrative is grouped prosodically with the preceding material, could set the stage for morphological fusion. It would be easy to transfer such a pattern through contact: the initial predicates and demonstratives already existed in both languages. The social circumstances for transfer were in place as well. There was intense contact among Wakashan and Tsimshianic‐speaking peoples, including intermarriage and extensive multilingualism, which extended into recent times (Codere 1990: 360; Halpin and Seguin 1990: 275–6; Hamoi‐Torok 1990: 306; Hilton 1990: 314–17). Ceremonies involving elaborate oratory were also shared.

4  Patterns of Core Argument Structure On the basis of a survey of 174 genetically and areally diverse languages, Nichols (1992) proposes that core argument patterns such as nominative/accusative and ergative/absolutive have “high genetic stability” and are potentially capable of revealing genetic relations more ancient than those recoverable through the comparative method: “Dominant alignment is genetically stable and not greatly susceptible to areal spread” (1992: 166). The proposal makes sense. Grammatical relations are typically coded by morphology, one of the most tightly integrated, systematic domains of grammatical structure, less accessible to the consciousness of speakers than independent words. Yet clusters of the core argument patterns identified by Nichols as the rarest cross‐linguistically appear in several geographical areas of North America, often cutting across genealogical lines.

4.1  Semantically Based Systems The Athabaskan‐Eyak‐Tlingit languages are distributed over a large area from the Southwest through Alaska. All of the nearly 40 Athabaskan languages identify core arguments by pronominal prefixes in their verbs. Subject prefixes occur immediately before the classifier+stem complex. (Basic third person subjects are zero.) Object prefixes occur further from the stem, potentially separated from it by various modal, aspectual, and adverbial prefixes. (16) Navajo pronominal prefixes (Faltz 1998: 112–13, 156) a. ha‐n‐sh‐łteeh up.out‐2sg.obj‐1sg.sbj‐cl.handle.animate.obj ‘I’m carrying you up.’ b. ha‐sh‐ni‐łteeh up.out‐1sg.obj‐2sg.sbj‐cl.handle.animate.obj ‘You’re carrying me up.’ c. ha‐ni‐łteeh up.out‐2sg.obj‐cl.handle.animate.obj ‘(He/she) is carrying you up.’ d. ha‐ni‐d‐eesh‐tééł up.out‐2sg.obj‐fut‐1sg.sbj‐cl.handle.animate.obj.fut ‘I’ll carry you up.’

Contact and North American Languages  603 (17) Navajo subject sh‐ [š‐] ‘I’ (Young, Morgan, and Midgette 1992) yi‐sh‐háád ‘I shook it.’ (a rattle) ’adah ’ii‐sh‐ááh ‘I went down, descended.’ ni‐sh‐chon ‘I stink.’ ’ádadii‐sh‐nih ‘I got hurt.’

230 664 82 456

The Athabaskan languages are related as a group to the Eyak language of Alaska. Eyak subject prefixes, which are cognate with those in Athabaskan languages, also occur immediately before the classifier+stem complex. (18) Eyak pronominal subject x‐ ‘I’ (Krauss 1982) ‘I will show you.’ ich’ qanuh qu’‐x̲‐tah ‘I am going to boat back.’ q’e’ qu’‐x̲‐daqe: ‘I already shoveled them out.’ Datli: a’q’ six̲uhłkłinu: ‘I am alive.’ gala‐x̲‐tah ‘I will die.’ qu’‐x̲‐sinh

42 75 75 75 119

Objects are represented by pronominal clitics before the full verb. (19) Eyak object clitic sik’ah ‘me’ (Krauss 1982) sik’ah q’e’ sdile’kł ‘He released me.’ de:dal six̣a’ k’usałyahł? ‘What’s this interfering with me?’

43 99

The Athabaskan‐Eyak group is related in turn to the Tlingit language of Alaska. Tlingit also contains a set of pronominal prefixes in the verb immediately before the classifier+stem complex, cognate with the subject prefixes in Athabaskan and Eyak. Again basic third persons are zero. (20) Tlingit pronominal prefix x̲‐ ‘I’ (Story and Naish 1973) ‘I’m writing.’ kax̲ashxéet ‘I’m writing a letter.’ x′úx′ kax̲shaxéet x̲waaják ‘I killed it.’ kúnax̲ oox̲dzikàa ‘I’m really lazy.’ yan sh kax̲wjix’ákw ‘I’m sitting very comfortably, just the way I want to.’

374 374 377 122 53

A set of pronominal clitics precede the verb. (21) Tlingit pronominal clitic x̲at ‘me’ (Story and Naish 1973) ‘He saw me.’ x̲at woositèen ‘Somebody hit me.’ x̲at woodoowagwál ‘He poked me in the face.’ x̲at yawsiták ‘They paid me.’ x̲at woodoodzikéi ‘Don’t you recognize me?’ tléil agé x̲at yayeetèen?

384 366 155 146 169

The Tlingit pronominals differ in a fundamental way from those in the Athabaskan languages and Eyak, however. While the Athabaska‐Eyak pronominals show a clear nominative/ accusative pattern, those in Tlingit show an agent/patient pattern. The Tlingit prefixes, like x̲‐ in (20), represent participants who typically instigate and are in control of situations: grammatical agents. The clitics, like x̲at in (21), represent those who are not in control but are significantly affected: grammatical patients. Some patients, like those in (21), would be categorized as direct objects in English or Athabaskan languages. Others, like those in (22), would be categorized as subjects.

604  Marianne Mithun (22) Tlingit clitic x̲at ‘I’ (Story and Naish 1973) ‘I’m cold.’ x̲at seiwa.át’ ‘I’m lonesome.’ x̲at woolitèesh ‘I’m going to die.’ x̲at googanáa ‘I was paralyzed, so shocked I couldn’t act.’ x̲at woodi.éik kúnax̲ x̲at yanéekw ‘I’m real sick.’ ‘I failed completely.’ x̲at kawdikèi ‘I’m growing old.’ yaa x̲at nadashán ‘I need new shoes.’ yées téel èetee‐nax x̲at ya tèe kut x̲at woodzigèet ‘I was lost.’ x̲at k’eiwawáash ‘I yawned.’ ‘I sobbed.’ x̲at oowakích ‘I frequently have hiccups.’ x̲at yadút’kw ‘I woke up.’ x̲at woodzigít ‘I fell flat on my back.’ kindayígin x̲at woodzigèet

366 127 366 145 190 85 141 139 129 251 201 108 241 247

Though a semantic basis underlies the Tlingit system, speakers do not make online decisions about agency, control, or affectedness as they speak. The pronominal set associated with each verb is lexicalized. One of the patterns represents an innovation. Since the Athabaskan‐Eyak languages are related as a group to Tlingit, the innovation could have occurred in either branch of the family. (Recent work by Vajda (2008a) indicates that Athabaskan‐Eyak‐Tlingit may be related to the Yeneseic languages of Siberia, but as reconstructed by Vajda (2008b), their common ancestor had not yet developed a full system of either type, so the Yeneseic languages provide no help here.) Suggestive evidence of the direction of shift can be found in a neighbor. Immediately to the south of the Tlingit are the Haida, who speak an unrelated language. Modern Haida territory was occupied until around 1700 by the Tlingit (De Laguna 1990: 203). De Laguna reports that there was intense Tlingit–Haida contact and intermarriage, and that “the Tlingit are known to have absorbed increments of Haidas and Tsimshians” (1990: 213). The two languages are quite different typologically. Haida pronouns are independent words or clitics rather than prefixes, and they show no similarity in form to those of Tlingit. They do, however, follow an agent/patient pattern. (23) Haida agent/patient system: 1SG Agent hl and Patient dii (Enrico 2003) hl sral‐gan ‘I fixed it.’ 491 Joe hl qing‐gan ‘I saw Joe.’ 51 ’laa hl st’ida‐gan ‘I warned him.’ 433 hl ’iij‐angqasaa‐ang ‘I am going to go.’ 565 dii ’la gu’laa‐gang ‘He likes me.’ 79 dii‐gingaan ’la qeenggaa ‘He looks like me.’ 84 dii hlrwaaga‐ang ‘I am afraid.’ 87 dii rahgal‐gang ‘I am tired of it.’ 82 dii gudang‐gang ‘I want to.’ 71 dii q’ud‐ang‐gan ‘I wasn’t hungry.’ 41 ’laa‐gingaan dii qeenggaa ‘I look like him.’ 84 The evidence strongly suggests that the Tlingit system developed under Haida influence. There is no philological record of the transfer, but a likely scenario can be imagined. It is not uncommon cross‐linguistically for third persons not to be mentioned overtly in every clause, so long as reference is clear. Such a propensity can even be borrowed (Myers‐Scotton 2002: 210). The absence of overt third person reference can set the stage for the reanalysis of

Contact and North American Languages  605 nominative/accusative systems as agent/patient systems and vice versa. Transitive clauses with a single overt object argument could be reinterpreted as intransitive clauses with a single patient argument, or the reverse. (24) (subject) (tr) (It/something)

verb scared

object me

patient I

(intr) verb was/am scared

Such a development could happen spontaneously in a language. It could also be stimulated by contact, as bilinguals strive to reconcile their two grammatical systems. The Tlingit–Haida parallel is not an isolated case. Clusters of agent/patient systems appear in several other areas in North America. The Wappo and Yuki languages of California mentioned earlier are distantly related to each other, but no further relationships have been identified. The first and second person singular pronouns in the two are nearly identical in form (borrowed from Pomoan). Third person pronouns, used only for emphasis, developed recently in each language from demonstratives. The Wappo pronouns show a nominative/ accusative pattern. The Yuki pronouns show an agent/patient pattern, one which matches that of their Pomoan neighbors down to the finest detail (Mithun 1991, 2008). Agent/patient systems also appear in the Southeast, Great Plains, and Northeast, in all languages of the Siouan‐Catawba, Caddoan, and Iroquoian families, as well as in all languages of the Muskogean family and isolates Chitimacha, Tunica, Natchez, and Atakapa. Together these languages cover an area from Canada to the Gulf of Mexico, and from the Atlantic across the Great Plains. Such systems also appear in the Pueblo Southwest, in languages of the Kiowa–Tanoan family, and dialects of the Keresan language. In some languages the pronominals are prefixes, in some suffixes, and in some both. The affixes also show no similarities in form across family boundaries. Nichols found agent/patient patterns rare cross‐linguistically, occurring in just 13.5% of her sample (1992: 101). This rarity, combined with the pervasiveness of agent/patient systems in North America, suggests contact effects. The most likely mechanism of transfer is not unusual: a reanalysis of clause structure by bilinguals seeking to reconcile the argument categories of their two languages.

4.2  Hierarchical Systems The rarest pattern found by Nichols is that termed hierarchical, occurring in just 5% the languages she examined. The Wakashan languages are indigenous to the Northwest Coast. Core arguments are identified by pronominal enclitics to the predicate, which is basically clause‐initial. (25) Ahousaht Nuuchahnulth clitic =s (Nakayama 2003; George Louie, speaker) wałšiʔa = s ‘I went home’ 195 wikʔaq = s suutił wiiqhap ‘I will not harm you’ 383 naʔaa = s ‘I understood’ 169 ʔuuyimłckwi=s ‘I was born’ 163 ʕiiḥši=s ‘I cried’ 166 iič’imʔa=s ‘I am old’ 451 wałsaap’at=s ‘They sent me home’ 167 n’aacsaat=s q wayac’iikʔi ‘The wolf was watching me’ 383 Clitic choice is not affected by transitivity, so this is not an ergative/absolutive system: the clitic =s ‘I’ appears in both ‘I went home’ and ‘I will not harm you’. This is not an agent/ patient system: the same clitic appears in ‘I was born’. It is not active/stative: the same clitic

606  Marianne Mithun appears in ‘I am old’. But it is not nominative/accusative either: the same clitic represents both subjects (‘I will not harm you’) and objects (‘They sent me home’). It is a hierarchical system. Only one argument is represented pronominally in a verb. The choice of argument depends on person, according to the hierarchy 1, 2 > 3. If a first or second person acts on a third (1>3, 2>3), that first or second person is represented. If a third person acts on a first or second (3>1, 3>2), again the first or second person has priority (‘The wolf saw me’). One might wonder how speakers could distinguish ‘I found him’ from ‘He found me’. Nuuchahnulth has a suffix ‐’at, somewhat comparable to a passive in other languages. Agents may or may not be mentioned lexically in ‐’at clauses. (26) Nuuchahnulth ‐’at (Nakayama 1997: 168, 170) a. ha:ʕanʔanits ha:ḥan‐’at‐it=s invite‐pass‐past=1sg ‘I was invited.’ b. ’ičiʔat  a mamałn’i ’i‐či‐ʔat‐ a: mamałn’i shoot‐mom‐pass‐also Whiteman ‘He was shot at by White men again.’ The ‐’at construction is used extensively for ensuring that a continuing discourse topic is the core argument of the clause. It also functions to maintain the hierarchy. If a first or second person acts on a third (‘I found her’) the ‐’at construction cannot be used. If a third person acts on a first or second, the ‐’at construction must be used (‘I was found’). When a clause involves only first and second persons, just the agent is represented by a clitic. The other participant may be identified in a separate word. (27) Local relations (1>2, 2>1) (Nakayama 2003: 383) wikʔaq  s suutił wiiqʕap wik‐ʔaq  =s sut‐čił wi:q‐ʕap not‐fut‐1sg you‐doing.to unpleasant‐do ‘I will not harm you.’ Special transitive pronominal enclitics representing combinations of first and second persons are used in imperatives. The hierarchical system cannot be reconstructed for Proto‐Wakashan. To the south of Nuuchahnulth are the two other South Wakashan languages: Nitinaht (Ditidaht) and Makah. Both also show the 1, 2 > 3 hierarchy. Both maintain it by means of constructions cognate with the Nuuchahnulth ‐’at construction. The hierarchical system has not penetrated their grammars quite as thoroughly, however. Any time first and second persons act on each other, transitive clitics are used. The hierarchical system has been extended even less deeply in the three North Wakashan languages. Immediately to the north of Nuuchahnulth is Kwak’wala. In this language subjects are identified by enclitics, and objects by verbal suffixes. There is one gap in the pronominal paradigm: there are no first person object forms. In place of an inherited object form, a word based on the verb ‘come’ is used, or an oblique construction. North of Kwak’wala are Heiltsuk and Haisla. These two languages show no trace of a hierarchy. Full sets of subject and object pronominals exist and are used in all combinations. To the south of the Wakashan family is the Chimakuan family, consisting of Chemakum and Quileute. Documentation of Chemakum is sparse, but the Quileute system is clear.

Contact and North American Languages  607 Arguments are identified by pronominal subject enclitics and object suffixes, but not all subject > object combinations occur. There is a hierarchy, 2 > 3, also maintained through passivization, but the forms of the Quileute pronominals and passive suffixes are completely different from those of the Wakashan languages. West of the Wakashan and Chimakuan languages are the 23 Salishan languages. The northernmost Salishan languages Bella Coola, Comox, and Sechelt show no restrictions whatsoever on argument combinations. Immediately to the south along the coast, Squamish, Halkomelen, and the Saanich dialect of Northern Straits show a limited hierarchical system: 2 > 3. South of them, the Sooke and Lummi dialects of Northern Straits, and the Klallam language, show a more extensive hierarchical system, equivalent to those of their South Wakashan neighbors to the west, Nitinaht and Makah: 1, 2 > 3. None of the Salishan languages further south (Lushootseed, Twana, Quinault, Lower Chehalis, Upper Chehalis), nor those of the Interior, shows hierarchies at all. The differences in the extent to which the hierarchical systems have penetrated the grammars of the different languages and dialects cut across genetic lines. The patterns show clear areal grouping, however, with the most extensive system, that of Nuuchahnulth, at the geographical core. But how could such abstract structures be transferred without the morphemes that carry them? The systems need not have been transferred in their modern forms. It is more likely that what was transferred was a precursor to the systems: a recurring stylistic choice. In languages with a grammatical subject category, certain kinds of participants tend to be preferred over others for this role. Animates tend to be preferred over inanimates, humans over nonhumans, first and second persons over third, agents over patients, given referents over new, and identifiable (definite) over indefinite. Such characteristics do not always coincide in a single participant: the speaker (first person) may not be a semantic agent, for example. Speakers of one of the Northwest Coast languages, perhaps Nuuchahnulth, may have tended to prioritize person under such circumstances, often passivizing clauses with third person agents acting on first or second person patients. This stylistic tendency could easily be transferred by bilinguals from one language to another. The structural equivalences already existed in all of the languages: first, second, and third person pronominals, and passive constructions. What would have been transferred was the frequency of the structures. Recurring choices could become routinized and ultimately obligatory. (The systems are further described in Mithun 2007b.) Hierarchical patterns are found elsewhere in North America as well. An intriguing cluster is in Northern California. There the mechanisms used to maintain the hierarchies vary, drawn from various resources originally present in different languages, but the resulting systems have begun to converge. Chimariko, an isolate, shows a strong hierarchical system. Verbs contain pronominal affixes with an agent/patient base. Most verbs appear with prefixes, but one set appear with suffixes. First, second, and third persons are distinguished, and singular and plural number. In addition, different pronominal affixes distinguish first person singular agents and patients, and also first person plural agents and patients. Only one argument is represented in any verb. In transitive verbs, the choice depends on a 1, 2 > 3 hierarchy: speech‐act participants have priority over others. Verbs with meanings like ‘I found him’ and ‘He found me’ both contain only a first person pronominal affix, but the difference is clear from the form of the first person prefix. ‘I found him’ contains just a first person agent prefix; ‘He found me’ contains just a first person patient prefix. When both arguments are speech‐act participants (‘I found you’, ‘You found me’), only the agent is represented. A second argument may be identified with an independent emphatic pronoun. Special transitive forms are used in imperatives. To the southeast of Chimariko is Yana. In Yana, core arguments are identified by pronominal suffixes on verbs. This system shows a nominative/accusative basis. The same

608  Marianne Mithun pronominal forms are used to represent subjects of intransitives and transitives, semantic agents and patients (‘I pound it up’, ‘I am shaking with fear’), and those involved in events and states (‘I killed him’, ‘I am ugly’). The transitive pronominal suffixes are now fused complexes, but earlier internal structures can be detected. As in many pronominal affix paradigms, there is no overt marker for third persons. When a first or second person acts on a third (1>3, 2>3), the form is the same as for intransitives (1, 2). The third person object is simply not mentioned. When a third person acts on a first or second (3>1, 3>2), an additional element ‐wa‐ appears, and the stem shows ablaut. The source of this ‐wa‐ element is a passive marker. Passive formation involves the suffix ‐wa(ʔa) plus ablaut. Pronominal suffixes representing combinations of first and second persons are fossilized, but they also contain recognizable elements. All include a remnant of the passive suffix ‐wa(ʔa). Various other markers have been added over time, apparently to clarify reference in potentially ambiguous situations. An element ‐ki‐ in forms involving first person plurals comes from a verbal suffix ‘hither’. An element ‐wii‐ in combinations with second person plural subjects matches a noun plural. An element ‐m‐ in ‘I>you.all’ and ‘we>you.all’ is a second person pronominal apparently reinforcing the second person. There is a third hierarchical system in the area. To the west of Chimariko, on the Coast, is Yurok, an Algic language clearly unrelated genetically to either Chimariko or Yana. In Yurok, core arguments in indicative verbs are identified by pronominal suffixes. The suffixes generally show nominative/accusative patterning: transitivity, semantic role, and aspect make no difference. In transitive constructions, however, both arguments are not always represented overtly. In certain combinations involving third person patients, the third persons are not represented at all. In certain other combinations, there is obligatory passivization by means of the passive suffix ‐ey or ‐oy: nekcenoy ‘he/she meets us’ is literally ‘we are met’ (meet‐ PASSIVE). Yurok thus shows some of the strategies at work in neighboring languages to ensure a person hierarchy, but they have not been extended through the full grammar. Both participants are still represented in the combinations 1SG>2SG, 1SG>3SG, 1SG>2PL, 1SG>3PL, 2SG>1SG, 2>3SG, 3SG>1SG, 1PL>2SG, 1PL>3SG, 2PL>1SG, and 3PL>1SG. There is a slight priority given to second persons: third person agents are never expressed in the presence of second persons. There is also a fourth hierarchical system in the area. The isolate Karuk is spoken to the north of Chimariko and immediately to the east of Yurok. Here arguments are identified by pronominal prefixes on verbs. The system has a nominative/accusative base. First person subject and object suffixes have different forms: ní‐mmah ‘I see him’, ná‐mmah ‘He sees me.’ But here, too, only one argument is expressed in a verb. As in the other languages, first and second persons are always chosen over third. Third persons are simply unmentioned. The difference in form between first person subjects and objects keeps roles clear for first persons. When a second person pronominal prefix represents an object, an inverse suffix ‐ap is added to the verb (Macaulay 1992). Interestingly, second person plurals are chosen over all other participants, resulting in the hierarchy 2PL > 1 > 2SG > 3. Speakers of other languages in the area, such as those of the Pomoan family, use second person plural forms for respect, particularly to elders and in‐laws. Northern California thus provides another example of shared abstract structure not transferred with substance. Chimariko, Yana, Yurok, and Karuk all show person hierarchies in their pronominal affixes on verbs. The forms of their pronominal affixes are different. Some are even prefixes while others are suffixes. The bases for the pronominal systems are different: Chimariko shows an agent/patient base, while Yana, Yurok, and Karuk show a nominative/accusative base. The person hierarchies are slightly different, and they have penetrated the pronominal paradigms to differing extents. Importantly, the hierarchies are insured by different mechanisms: differences in the forms of first person agent and patient or subject and object markers, obligatory passivization, a directional marker ‘hither’, and an

Contact and North American Languages  609 old inverse marker. The systems differ in their pathways and endpoints of development, but the similarities are striking. Given the overall rarity of hierarchical systems cross‐linguistically, and the long‐standing, intense multilingualism in the area, there is every indication that the  similarities are due to language contact. As in the Northwest, it is likely that the modern  structures were not transferred directly as abstract grammatical systems. Rather, what may have been transferred were their precursors, certain recurring patterns of ­expression, which subsequently crystallized in each language. (Further details are contained in Mithun 2012.)

5 Conclusion The Americas provide rich examples of effects of language contact, far beyond those mentioned here. In most cases, the details of these effects are not documented by a philological record comparable to those for certain languages of Europe, but comparisons with modern languages can be revealing. In particular, many of the languages offer insights into the potential role of contact in stimulating the development of structural parallelisms, even in the absence of borrowed substance. Considering the structures in diachronic perspective can often open the way to understanding the mechanisms by which such parallelisms can arise. Abstract patterns need not be transferred as such. Their development can have been set in motion by contact at an earlier stage, characterized by the transfer of particular patterns of expression and frequencies of stylistic choices.

Abbreviations ABS = absolutive, CL = classifier, COM = common, COND = conditional, DEM = demonstrative, ERG = ergative, FUT = future, HRS = hearsay, MOM = momentaneous, NOM = nominative, OBJ = object, OBL = oblique, PASS = passive, PFV = perfective, SML = semelfactive, SBJ = subject.

REFERENCES Aikhenvald, Alexandra 2007. Grammars in contact: a cross‐linguistic perspective. In Alexandra Aikhenvald and Robert M. W. Dixon (eds.), Grammars in Contact: A Cross‐ Linguistic Typology. Oxford: Oxford University Press, pp. 1–37. Andrade, Manuel J. 1933. Quileute. In Franz Boas (ed.), Handbook of American Indian Languages, vol. 3. New York: Columbia University, pp. 151–292. Bates, Dawn, Thom Hess, and Vi Hilbert 1994. Lushootseed Dictionary. Seattle: University of Washington Press. Beck, David 2000. Grammatical convergence and the genesis of diversity in the Northwest Coast Sprachbund. Anthropological Linguistics 42.2: 147–213.

Bisang, Walter 1996. Areal typology and grammaticalization: processes of grammaticalization based on nouns and verbs in East and Mainland South East Asian languages. Studies in Language 20.3: 519–97. Blevins, Juliette 2003. The phonology of Yurok glottalized sonorants: segmental fission under syllabification. International Journal of American Linguistics 69: 371–96. Bloomfield, Leonard 1946. Algonquian. In Harry Hoijer et al. (eds.), Linguistic Structures of Native America (Viking Fund Publications in Anthropology 6), facsimile reprint 1963. New York: Johnson, pp. 85–129. Boas, Franz 1911a. Kwakiutl. Handbook of American Indian Languages (Bureau of American

610  Marianne Mithun Ethnology Bulletin 40.1). Washington DC: Smithsonian Institution, pp. 423–558. Boas, Franz 1911b. Tsimshian. Handbook of American Indian Languages (Bureau of American Ethnology Bulletin 40.1). Washington DC: Smithsonian Institution, pp. 283–422. Bright, William 1973. North American Indian language contact. In Thomas Sebeok (ed.), Linguistics in North America (Current Trends in Linguistics 10). The Hague: de Gruyter Mouton, pp. 713–26. Campbell, Lyle 1997. American Indian Languages: The Historical Linguistics of Native America. Oxford: Oxford University Press. Callaghan, Catherine 1964. Phonemic borrowing in Lake Miwok. In William Bright (ed.), Studies in Californian Linguistics (University of California Publications in Linguistics 34), pp. 46–53. Chafe, Wallace 1987. Cognitive constraints on information flow. In Russell Tomlin (ed.), Coherence and Grounding in Discourse. Amsterdam: Benjamins, pp. 21–51. Chafe, Wallace 1994. Discourse, Consciousness and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: University of Chicago Press. Codere, Helen 1990. Kwakiutl: traditional culture. In Wayne Suttles (ed.), Handbook of North American Indians, vol. 7: Northwest Coast. Washington DC: Smithsonian Institution, pp. 59–77. De Laguna, Frederica 1990. Tlingit. In Wayne Suttles (ed.), Handbook of North American Indians, vol. 7: Northwest Coast. Washington DC: Smithsonian Institution, pp. 203–28. Dunn, John 1979. Tsimshian connectives. International Journal of American Linguistics 45: 131–40. Enrico, John 2003. Haida Syntax. Lincoln: University of Nebraska Press. Faltz, Leonard 1998. The Navajo Verb. Albuquerque: University of New Mexico. Gast, Volker and Johan van der Auwera 2012. What is “contact‐induced grammaticalization”? Evidence from Mesoamerican languages. In Bjorn Wiemer and Bernhard Hansen (eds.), Grammatical Replication and Grammatical Borrowing in Language Contact. Berlin: de Gruyter Mouton, pp. 381–426. Golla, Victor 1996. Hupa. In Ives Goddard (ed.), Handbook of American Indians, vol. 17: Languages. Washington, DC: Smithsonian Institution, pp. 364–89. Golla, Victor 2011. California Indian Languages. Berkeley: University of California Press. Haas, Mary 1976. The Northern California linguistic area. In Margaret Langdon and Shirley Silver (eds.), Hokan Studies. The Hague:

de Gruyter Mouton, pp. 347–59. (Reprinted 1978 in Mary R. Haas, Language, Culture, and History: Essays by Mary R. Haas. Stanford, CA: Stanford University Press, pp. 353–69.) Haase, Martin and Nicole Nau 1996. Sprachkontakt und Grammatikalisierung. Special issue of Sprachtypologie und Universalienforschung 49.1. Bremen: Akademie Verlag. Halpin, Marjorie and Margaret Seguin 1990. Tsimshian peoples: Southern Tsimshian, Coast Tsimshian, Nisgha, and Gitksan. In Wayne Suttles (ed.), Handbook of North American Indians, vol. 7: Northwest Coast. Washington DC: Smithsonian Institution, pp. 267–84. Hamor‐Torok, Charles 1990. Haisla. In Wayne Suttles (ed.), Handbook of North American Indians, vol. 7: Northwest Coast. Washington DC: Smithsonian Institution, pp. 306–11. Heine, Bernd 1994. Areal influence on grammaticalization. In Martin Putz (ed.), Language Contact and Language Conflict,. Amsterdam: Bejnamins, pp. 56–68. Heine, Bernd 2009. Identifying instances of contact‐induced grammatical replication. In Samuel Gyasi Obend (ed.), Topics in Descriptive and African Linguistics: Essays in Honor of Distinguished Professor Paul Newman. Munich: Lincom Europa, pp. 29–56. Heine, Bernd 2011. Areas of grammaticalization and geographical typology. In Osamu Hieda, Christa König, and Hirosi Nakagawa (eds.), Geographical Typology and Linguistic Areas, with Special Focus on Africa (Tokyo University of Foreign Studies 2). Amsterdam: John Benjamins, pp. 41–66. Heine, Bernd and Tania Kuteva 2003. On contact‐induced grammaticalization. Studies in Language 27.3: 529–75. Heine, Bernd and Tania Kuteva 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva 2006. The Changing Languages of Europe. Oxford: Oxford University Press. Heine, Bernd and Motoki Nomachi. 2013 Isomorphic processes: grammaticalization and copying of grammatical elements. In Martine Robbeets and Hubert Cuyckens (eds.), Shared Grammaticalization. Amsterdam: John Benjamins, pp. 67–100. Hilton, Susanne 1990. Haihais, Bella Bella, and Oowekeeno. In Wayne Suttles (ed.), Handbook of North American Indians, vol. 7: Northwest Coast. Washington DC: Smithsonian Institution, pp. 312–22.

Contact and North American Languages  611 Hinton, Leanne 1991. Takic and Yuman: a study in phonological convergence. International Journal of American Linguistics 57.2: 133–57. Jany, Carmen 2007. Chimariko in cross‐linguistic perspective (University of California, Santa Barbara doctoral dissertation). Johanson, Lars 2002. Contact‐induced change in a code‐copying framework. In Mari Jones and Edith Esch (eds.), Language Change: The Interplay of Internal, External and Extra‐Linguistic Factors. Berlin: de Gruyter Mouton, pp. 285–313. Johanson, Lars 2008. Remodeling grammar: copying, conventionalization, grammaticalization. In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages. Amsterdam: Benjamins, pp. 61–80. Johanson, Lars 2013. Isomorphic processes: grammaticalization and copying of grammatical elements. In Martine Robbeets and Hubert Cuyckens (eds.), Shared Grammaticalization. Amsterdam: Benjamins, pp. 101–9. Kinkade, M. Dale 1978. Alsea pronouns. Paper presented at the 1978 annual meeting of the American Anthropological Association, Los Angeles. Kinkade, M. Dale 1985. More on nasal loss on the Northwest Coast. International Journal of American Linguistics 51.4: 478–80. Krauss, Michael 1982. In Honor of Eyak: The Art of Anna Nelson Harry. Fairbanks: Alaska Native Language Center. Kuteva, Tania 1998. Large linguistic areas in grammaticalization: auxiliation in Europe. Language Sciences 20.3: 289–311. Kuteva, Tania 2000. Areal grammaticalization: the case of the Bantu‐Nilotic borderland. Folia Linguistica 34.3–4: 267–83. Kuteva, Tania 2008. On the frills of grammaticalization. In María José López‐ Couso and Elena Seoane (eds.), Rethinking Grammaticalization: New Perspectives. Amsterdam: Benjamins, pp. 189–218. Lukaniec, Megan and Wallace Chafe 2016. Huron/Wendat interactions with the Seneca language. In Andrea Berrz‐Kroeker, Diane Hintz, and Carmen Jany (eds.), Language Contact and Change in the Americas. Amsterdam: Benjamins, pp. 189–218. Macaulay, Monica 1992. Inverse marking in Karuk: the function of the suffix ‐ap. International Journal of American Linguistics 58: 182–201. Matisoff, James 1991. Areal and universal dimensions of grammatization in Lahu. In Elizabeth Traugott and Bernd Heine (eds.),

Approaches to Grammaticalization, vol. 2. Amsterdam: Benjamins, pp. 383–453. Matras, Yaron and Jeanette Sakel 2007. Grammatical Borrowing in Cross‐Linguistic Perspective. Berlin: de Gruyter Mouton. Mithun, Marianne 1991. Active/agentive case marking and its motivations. Language 67: 510–46. Mithun, Marianne 1997. Lexical affixes and morphological typology. In Joan Bybee, John Haiman, and Sandra A. Thompson (eds.), Essays on Language Function and Language Type. Amsterdam: Benjamins, pp. 357–72. Mithun, Marianne 1998. The sequencing of grammaticization effects. In Monika Schmidt, Jennifer Austin, and Dieter Stein (eds.), Historical Linguistics 1997. Amsterdam: John Benjamins, pp. 291–314. Mithun, Marianne 1999 [2001]. The Languages of Native North America. Cambridge: Cambridge University Press. Mithun, Marianne 2007a. Grammar, contact, and time. Journal of Language Contact (e‐journal) THEMA 1: 133–55. Available at www.jlc‐ journal.org Mithun, Marianne 2007b. Integrating approaches to diversity: argument structure on the Northwest Coast. In Yoshiko Matsumoto, David Oshima, Orrin Robinson, and Peter Sells (eds.), Diversity in Language. Stanford: CSLI (Center for the Study of Language and Information), pp. 9–36. Mithun, Marianne 2008. The emergence of agentive systems in core argument marking. In Mark Donohue and Søren Wichmann (eds.), The Typology of Semantic Alignment. Oxford: Oxford University Press, pp. 297–333. Mithun, Marianne 2012. Core argument patterns and deep genetic relations: hierarchical systems in Northern California. In Bernard Comrie (ed.), Typology of Argument Structure and Grammatical Relations. Amsterdam: Benjamins, pp. 257–94. Mithun, Marianne 2013a. Challenges and benefits of contact among relatives: morphological copying. Journal of Language Contact 6: 243–70. Mithun, Marianne 2013b. Wendat among the Iroquois. In Louis‐Jacques Dorais and Jonathan Lainey (eds.), Continuités/Continuities. Proceedings of the First Wendat and Wyandot Studies Conference. Wendake: Éditions Hannenorak, pp. 79–89. Mithun, Marianne 2014. Gender and culture. In Greville Corbett (ed.), The Expression of Gender. The Expression of Cognitive Categories Volume 6. Berlin: Mouton de Gruyter, pp. 131–160.

612  Marianne Mithun Mithun, Marianne 2017. Native North American languages. In Raymond Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 878–933. Mithun, Marianne 2018. Shaping typology through grammaticalization: North America. In Heiko Narrog and Bernd Heine (eds.), Grammaticalization from a Typological Perspective. Oxford: Oxford University Press, pp. 309–36. Mulder, Jean 1994. Ergativity in Coast Tsimshian (Sm’algyax) (University of California Publications in Linguistics 124). Berkeley: University of California Press. Myers‐Scotton, Carol 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Nakayama, Toshihide 1997. Discourse‐pragmatic synamism in Nuu‐chahnulth (Nootka) Morphosyntax (University of California, Santa Barbara doctoral dissertation). Nakayama, Toshihide (ed.) 2003. George Louie’s Nuu‐chah‐nulth (Ahousaht) Texts with Grammatical Analysis (Endangered Languages of the Pacific Rim Publications Series A2‐028). Kyoto: Nakanishi Printing Co. Narrog, Heiko and Bernd Heine 2011. The Oxford Handbook of Grammaticalization. Oxford: Oxford University Press. Nichols, Johanna 1971. Diminutive consonant symbolism in Western North America. Language 47: 826–48. Nichols, Johanna 1992. Linguistic Diversity in Space and Time. Chicago: University of Chicago Press. Pawley, Andrew 2000. The one‐clause‐at‐a‐time hypothesis. In Heidi Riggenbach (ed.), Perspectives on Fluency. Ann Arbor: University of Michigan Press, pp. 163–99. Ramat, Anna Giacalone 2008. Areal convergence in grammaticalization processes. In María José López‐Couso and Elena Seoane (eds.), Rethinking Grammaticalization: New Perspectives. Amsterdam: John Benjamins, pp. 129–68. Ramat, Paolo and Elisa Roma (eds.) 2007. Europe and the Mediterranean as Linguistic Areas: Convergences from a Historical and Typological Perspective. Amsterdam: Benjamins. Rath, John C. 1981. A Practical Heiltsuk‐English Dictionary with a Grammatical Introduction (Mercury series, Canadian Ethnology Service Paper 75). Ottawa: National Museums of Canada. Robbeets, Martine and Hubert Cuyckens 2013. Towards a typology of shared grammaticalization. In Martine Robbeets and Hubert Cuyckens (eds.), Shared

Grammaticalization. Amsterdam: John Benjamins, pp. 1–20. Sapir, Edward 1921. Language. New York: Harcourt, Brace and World. Sherzer, Joel 1973. Areal linguistics in North America. In Thomas Sebeok (ed.), Current Trends in Linguistics, vol. 10: Linguistics in North America. The Hague: de Gruyter Mouton, pp. 749–95. (Reprinted 1976 in Thomas Sebeok (ed.), Native Languages of the Americas, vol. 1. New York: Plenum, pp. 121–73.) Silverstein, Michael 1996. Dynamics of linguistic contact. In Ives Goddard (ed.), Handbook of North American Indians vol. 17: Languages. Washington, DC: Smithsonian Institution, pp. 117–36. Stebbins, Tonya 2003. On the status of intermediate form classes: words, clitics, and affixes in Sm’algyax (Coast Tsimshian). Linguistic Typology 7: 383–416. Stonham, John 2005. A Concise Dictionary of the Nuuchahnulth Language of Vancouver Island (Native American Studies 17). Lewiston, NY: Edward Mellen Press. Story, Gillian and Constance Naish 1973. Tlingit Verb Dictionary. Fairbanks: Alaska Native Language Center. Teeter, Karl and John Nichols 1993. Wiyot Handbook (Algonquian and Iroquoian Linguistics Memoirs 10, 11). Winnipeg: Algonquian and Iroquoian Linguistics. Thompson, Laurence and M. Terry Thompson 1972. Language universals, nasals, and the Northwest coast. In M. Estellie Smith (ed.), Studies in Linguistics in Honor of George L. Trager. The Hague: de Gruyter Mouton, pp. 441–56. Thompson, Laurence and M. Dale Kinkade 1990. Languages. In Wayne Suttles (ed.), Handbook of North American Indians 7: Northwest Coast. Washington, DC: Smithsonian Institution, pp. 30–51. Vajda, Edward 2008a. A Siberian link with the Na‐Dene languages. Dene‐Yeneseic Symposium, Ms., University of Alaska, Fairbanks. Vajda, Edward 2008b. Losing semantic alignment: from Proto‐Yeniseic to Modern Ket. In Mark Donohoe and Søren Wichmann (eds.), The Typology of Semantic Alignment. Oxford: Oxford University Press, pp. 140–61. Young, Robert, William Morgan, and Sally Midgette 1992. Analytical Lexicon of Navajo. Albuquerque: University of New Mexico. Zigmond, Maurice, Curtis Booth, and Pamela Munro 1991. Kawaiisu: A Grammar and Dictionary with Texts (University of California Publications in Linguistics 119). Berkeley: University of California Press.

31 Contact and Mayan Languages DANNY LAW

1 Introduction The effects of contact in Mayan languages run the gamut from run‐of‐the‐mill lexical ­borrowing to the direct borrowing of inflectional morphemes, copying of grammatical categories, and diffusion of phonological innovations. At least one Mayan language has been argued to be a bilingual mixed language which emerged in a context of community‐wide bilingualism, and there are preliminary descriptions of contemporary patterns of code‐ switching, new dialect formation, and other ongoing situations of language contact. Some Mayan languages today are showing effects of attrition, as the vast majority of their speakers have shifted to Spanish. Other Mayan languages are regionally prominent and are the language of everyday communication for hundreds of thousands of individuals, but still operate in a fundamentally multilingual environment in which speakers of indigenous languages confront continual social and political marginalization, with language being one site of this oppression. In this diverse and dynamic setting, Mayan language contacts involve influence from colonial languages, particularly Spanish, as well as other indigenous non‐Mayan languages, including Nahuatl and Mixe‐Zoquean languages. There is also a striking amount of evidence for centuries‐long contact between different Mayan languages. This means that studying the history of Mayan languages unavoidably involves a deep engagement with the methodological difficulties of distinguishing similarities due to contact from similarities that languages may share because of inheritance from a common ancestor. Contact between related Mayan languages also seems, at this stage, to have had different linguistic consequences than contact with unrelated languages, which poses questions about the impact that factors such as relatedness or preexisting linguistic similarity more generally may have on the mechanisms and outcomes of language contact.

2  Background and Historical Context The Mayan language family is relatively small, with some 32 languages spoken today in Guatemala, Belize, and parts of Mexico. Two of those languages (Itza’ and Mocho’) are highly endangered, with only a small number of very elderly speakers. Several Mayan languages, including K’iche’, Yucatec Maya (known locally simply as ‘Maya’), Q’eqchi’, Mam, Kaqchikel, Tsotsil, and Tseltal, are robust by comparison, with half a million to a million speakers or more and many communities in which the languages continue to be the

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

614  Danny Law dominant language of community interaction and the first language of children. Q’eqchi’ is reported to actually be gaining speakers, as there has been a fairly recent expansion of Q’eqchi’ speakers into new territories. The colonial language of the region is Spanish, and where Mayan languages are being lost, speakers are shifting to Spanish. Mayan languages are fairly linguistically diverse. Time depth estimates for the family range from about 4,200 years (Kaufman 1976, derived from glottochronology) to around 6,000 to 6,500 years (Atkinson 2006: 6.18, Bayesian phylogenetics). Most of the languages have dialectal variation and some dialectal variants are not mutually intelligible. Mam, for example, is composed of something approaching a dialect chain, with northern and southern varieties not mutually intelligible. The Ixil variety of Cotzal is also quite different from Nebaj and Chajul Ixil varieties, but it is common for speakers of the Cotzal variety to have some familiarity with the Nebaj variety. In addition, at least two mutually intelligible close dialectal variants have been labeled, for social and political reasons, to be separate languages: Achi, which is mutually intelligible with K’iche’, and Chalchiteko, which is mutually intelligible with Awakateko. In addition to contemporary languages, we have attested data for two now‐extinct languages, Ch’olti’ and Chicomuseltec. There are also some 20,000 pre‐Columbian hieroglyphic texts which mostly record a single fairly geographically and temporally uniform language, referred to as Classic Mayan (Law and Stuart 2017), which belongs to the Ch’olan subgroup of the family. There are also colonial era texts, written, in many cases by Spanish priests or missionaries, using an adaptation of the roman script that record sixteenth‐ and seventeenth‐ century varieties of several different Mayan languages, including particularly K’iche’, Yucatec Maya, Mam, Q’eqchi’, Kaqchikel, and Tseltal. The cultural history of the Maya area often distinguishes between the Highland Maya and the Lowland Maya. This division is roughly based on topography. The Lowland Maya traditionally occupy a mostly flat, tropical region comprising the Yucatán peninsula, parts of Chiapas and Tabasco, and the Lowland jungles of the Petén in Guatemala. The Highland Maya live traditionally in the mountainous and volcanic zone in southern Guatemala. In ancient and modern times, the Maya have participated in far‐reaching networks of trade and exchange. In ancient time, Maya goods reached to the American Southwest and at least down to Costa Rica, and at least some of that far‐reaching trade also involved movements of Mayan‐speaking peoples. Aside from far flung recent migrant communities, the only Mayan language spoken outside of the contiguous Maya region is Teenek (Huastec), which is spoken in northern Veracruz and San Luis Potosí in Mexico, far north of the rest of the Mayan languages. By around 600 bce, major urban population centers had begun to emerge in the Maya region, both in the southern Highlands and coastal piedmont of modern‐day Guatemala, and in the northern Lowlands. Cities like El Mirador, Tikal, and Kaminajuyu’ all had massive populations, even compared to contemporary cities in the Mediterranean, the Middle East, Asia and Africa. It is during this period that we have the first evidence of writing in the Maya area, in select Highland sites and in the Lowlands, though only a couple dozen texts from this period have been found. There is archeological evidence of a large demographic collapse in the first centuries ce, followed by another spike in both population and elite architectural and artistic production from around 300 to around 900 ce, which is commonly referred to as the Maya Classic Period. In the Lowlands, Classic civilization made extensive use of hieroglyphic inscriptions on permanent objects, such as stone and ceramic, and we accordingly have access to thousands of Lowland Classic Mayan hieroglyphic texts from this period. The texts cover many themes and paint a picture of a highly dynamic and multifaceted political system of semi‐independent polities whose alliances shifted over the centuries primarily between two major powers in the region centered at Calakmul and Tikal (Martin and Grube 2008). Another demographic collapse and dramatic political upheaval struck the Maya Lowlands at the end of the Classic Period. Classic Period urban centers were

Contact and Mayan Languages  615 largely abandoned and political power migrated north to the Yucatán centers of Chichen Itza and, later, Mayapan. Mayapan itself fell to civil war in 1441, mere decades before the arrival of Europeans. In the Highlands, several major urban centers and kingdoms consolidated power in the centuries after the Classic Period, including the K’iche’, the Mam, and the Kaqchikel, all of whom were prominent political entities at the time of the first Spanish military incursion in what is now Guatemala in the early sixteenth century. European contact had generally devastating effects in terms of demographics, both because of military incursions and because of disease throughout the Maya area, and, indeed, the rest of the Americas, where somewhere between 75 and 90% of the population is thought to have died in the decades following the initial European contact (Livi Bacci 2008). The linguistic consequences of this precipitous loss of population are not clear, but it seems likely that such a dramatic change would have affected the social and political networks and patterns of linguistic interaction in the region in substantive ways. In the following sections, I will attempt to characterize the nature of language contact effects in Mayan languages. In many cases, it is not clear exactly when these patterns first emerged. In some cases, we can see ongoing areal innovations that may well have been initiated by upheavals following European contact. In other cases, colonial and hieroglyphic linguistic evidence confirms that interactions preceded European contact and can be understood in the context of the complex and multilingual pre‐Columbian Maya society.

3  Mayan Language Contact Areas One of the most striking aspects of the contact history of Mayan languages is the pervasiveness of contact effects from linguistic interaction between Mayan languages. The Maya area can be roughly divided into three general areas of linguistic interaction. The first, which we might call the K’iche’an area, was investigated in early work by Campbell (1977), who focused particularly on lexical and phonological innovations and showed that isoglosses for a palatalization rule /k(’)/ > kj (’)/_V[‐round]{q(‘), x} which is found in some dialects of almost every K’iche’an language and is not found in all dialects of any K’iche’an language. Campbell argues that this palatalization was most likely introduced to K’iche’an languages from Mam, which has the same rule (Campbell 1977: 117) and Barrett (2002: 311–12) discusses its development within the Mamean subgroup. Campbell also notes a more restricted areal sharing of the unconditioned change of ts > s in the K’iche’an languages Poqomam, Poqomchi’, and Q’eqchi’. Because this region consists mostly of Mayan languages from the same subgroup, K’iche’an, the study of areal innovations that are not ongoing is difficult and little work beyond Campbell’s pioneering study has been done on this. The K’iche’an area also corresponds roughly to the former domain of the ancient K’iche’ kingdom, so it is ­possible that contact in the region, which is ongoing, has continued since prior to the Spanish invasion. A second area, which has also received little scholarly attention, has been called the Huehuetenango Sprachbund. Barrett (2002) is the only source to discuss this area in any detail. He identifies several areally shared innovations centered in the Guatemalan department of Huehuetenango and involving, primarily, languages in the Mamean and Q’anjob’alan subgroups of Mayan. Innovations include a novel phonemic contrast between retroflex and non‐retroflex affricates and fricatives, fixed VSO (rather than the proto‐Mayan VOS) word order, and the spread of a system of noun classifiers which is innovative and restricted to the languages of this area. Barrett also describes more complex areal patterns involving inclusive/exclusive/dual distinctions in first person plural and conditions for nominative alignment in certain morphosyntactic contexts in these normally strongly

616  Danny Law ergative languages. Where it can be inferred, the direction of borrowing for these features is mixed, with some features, like retroflex consonants, most likely originating in Mamean, while others, like noun classifiers, possibly originating in Q’anjob’alan. While the research on contact effects in the Huehuetenango area has been only preliminary, there is great potential for investigating areally shared versus inherited similarities and there is some indication that further substantial contact effects between languages in this region will yet be identified. The historical impetus for the formation of this contact zone is unclear, though forced relocations of Mamean speaking peoples to Q’anjob’alan areas in the colonial period may have played a role, as well as trade. Currently, the best studied of the contact areas in the Maya region is the Maya Lowlands (Justeson et al. 1985; Law 2014). Justeson et al. (1985) distinguished between the Lowland language area, which consisted of Ch’olan and Yukatekan languages, and the Greater Lowland language area, which added Q’anjob’alan and Tseltalan languages to that. Law (2014) cast this contrast as being between a core and periphery of the Lowland area and noted that some of the areal features excluded the core Eastern Ch’olan subgroup. He also provides a more detailed discussion of the many areally shared features of the Lowland linguistic area, which include numerous areally shared phonological innovations, morphosyntactic innovations like the innovation of aspect‐based split ergativity, the development of obligatory systems of numeral classifiers, and the innovative postverbal placement of a set of pronominal enclitics. Law (2009) describes several different patterns of areally shared innovations related to person marking. These include areally shared grammatical distinctions, such as the introduction of an inclusive/exclusive distinction for first person plural and neutralization of number distinctions in some persons. The core Lowland languages, (Eastern) Ch’olan and Yukatekan also appear to have shared the phonological forms of almost the entire paradigm of person markers. Finally, as Justeson et al. (1985) point out, the Lowland language area also includes a great deal of shared Lowland vocabulary. In many cases, the source language of the borrowing cannot be firmly established, but only languages in the Lowland or Greater Lowland area have those words. Directionality of the influence in the Maya Lowlands is problematic for more than just the lexical borrowings. Virtually all of the areal innovations that have been described have slightly different distributions and few of the innovations have elements that would uniquely identify the source language. Another question that the cross‐linguistically remarkable features of the Lowland Maya language area raise is, “what might have led to these languages sharing linguistic material at what appear to be higher than average levels?” One potential motor driving the spread of so many linguistic features through the region would be the clearly dominant role of Ch’olan languages historically, and specifically the prestigious language recorded in hieroglyphic inscriptions. Some of the most widespread areal features are, in fact, present in the language of the hieroglyphs, and might therefore have been spread by that language. However, Law (2017a) has argued that for most of the features identified as areally shared in the Maya Lowlands, the language of the glyphs retains the earlier inherited form and does not have the areal innovation. And for a large number of areally shared features, the hieroglyphic corpus does not give the necessary contexts to determine if the language had those particular areal features or not. This argues against assuming that Classic Maya elites were the primary motivator driving Lowland linguistic diffusion.

4  Mayan Mixed Languages and Dialect Mixing The long history of multilingualism among Mayan languages is logically likely to have led to both dialect and language mixing. Relatively little work has been done to describe the emergence of new dialects or bilingual code‐switching in Mayan communities. One exception

Contact and Mayan Languages  617 to this is DeChicchis’ (1989) study of new dialect formation in newly formed Q’eqchi’‐ speaking communities in the Petén and Belize. In the communities he investigated, speakers of several Highland Q’eqchi’ dialects all moved into the same community. The variety of the language that emerged was a classic koiné variety, drawing on the least marked features of each of the contributing varieties to develop a neutral variety. Hofling (2006) has argued for dialect or close‐language mixing in colonial times in the history of Northern Lacandon. Based primarily on evidence from person marking, he concludes that, “At the time the missions were established, refugee groups, likely including Itzaj, Mopan, Kowoj, and Kehache Maya, fled to the Lakantun forest mixing with other refugees, and the Northern Lakantun are their descendants” (Hofling 2006: 395). He also concludes that the Southern Lacandon variety does not have the same mixture and was primarily a mixture of more closely related varieties of Yucatec Maya. A similar situation has been argued to apply in the case of a variety of K’iche’ spoken in Santa Maria Cauque, in which a large group speaking the closely related language Kaqchikel shifted to K’iche’, but retained high levels of Kaqchikel lexicon. This led Stevenson (1990) to label this variety a mixed language, à la Media Lengua (Thomason and Kaufman 1988). However, little direct data has been published in support of this analysis and is not clear from what has been published how dominant identifiably Kaqchikel terms actually are in the variety, since Kaqchikel and K’iche’ more generally are close sister languages that already share a large amount of the lexicon because of common inheritance. It may be simpler to understand the Cauque K’iche’ as a variety of K’iche’ with moderate levels of lexical borrowing from Kaqchikel, rather than as a mixed language. Further research is needed here. Another proposed mixed language, this one involving a mixture of much more distantly related Mayan languages, is the Mayan language Tojol‐ab’al. Law (2011, 2017b) first proposed that Tojol‐ab’al was a mixed language based on an investigation of the etymologies of a variety of morphological and morphosyntactic features of Tojol‐ab’al as well as a comparison of lexical similarity between Tojol‐ab’al and other Mayan languages. The investigation showed that the grammatical morphology and lexicon of Tojol‐ab’al was of mixed origin, with a large portion of virtually every grammatical subsystem deriving from the Mayan language Chuj, and large portions of those same subsystems deriving from the Mayan language Tseltal. Gómez Cruz (2017) expanded this investigation substantially, looking at many more morphological features, and a broader range of syntactic features, all of which showed the same pattern observed by Law. In contrast to other well‐known mixed languages (e.g. Michif, or Media Lengua, Bakker, this volume), which have whole subsystems (verb phrases, or lexicon) that are taken in their entirety from one donor language or another, Tojol‐ab’al is not tidily compartmentalized in terms of etymology. Law (2017a) speculates that this thorough mixing of etymological sources of the grammar of Tojol‐ab’al might be understood as deriving from community‐wide norms of code‐switching involving closely related languages, what Muysken (2000) calls “congruent lexicalization.”

5  Mayan Languages in Mesoamerica The long history of contact involving Mayan languages is in no way restricted to contact among related languages. Contact with Spanish has been pervasive since the colonial period and has increased as development and modernization increasingly affect more rural Mayan communities, and that agenda is almost exclusively carried out in Spanish. Spanish influence on Mayan is noticeably different in quality from contact described above with other Mayan languages. Influence has primarily been lexical, though, as noted by Brody (1987), the borrowing of certain function words, or “discourse utterance modifiers” (Matras 2007) from Spanish is pervasive in Mayan languages. Minor effects of Spanish contact have also been

618  Danny Law argued to be present in the syntax of some Mayan languages for basic word order (Quizar 1994) the periphrastic “to go” future (Montgomery‐Anderson 2010) or, possibly, in the emergence of definite articles across the family. In addition, most Mayan languages have adopted Spanish numbers and a base 10 counting system for larger numbers. Thus, while Spanish continues to displace Mayan languages in an increasing range of social domains of use, its influence on the grammatical systems of Mayan languages is comparatively minimal. Contact between Mayan and non‐Mayan languages also involves neighboring indigenous Mesoamerican languages, particularly the Mixe‐Zoquean language family. Mayan languages belong to the oft‐cited “Mesoamerican linguistic area” (Campbell et al. 1986). They have all of the features highlighted by Campbell et al. as characteristic of the area, including relational nouns, a vigesimal numeral system (where inherited Mayan numbers are still in use), nonverb‐final basic word order, possessum‐possessor word order and 13 particular semantic calques that are pervasive in Mesoamerica. In is unclear what historical movements and relations led to this areal pattern of features, though there is unambiguous evidence of millennia of extensive networks of trade and the sharing of other elements of culture in the region. It also worth noting that, while they ultimately only highlight five defining features of the Mesoamerican linguistic area, the survey that Campbell et al. made of other linguistic features speaks to extensively shared linguistic and discursive features that are very likely to have been shared across languages and families via contact even if they do not rise to the level of a defining characteristic of the Mesoamerican linguistic area. Finer‐grained studies of patterns of language contact in Mesoamerica geared towards understanding the history of linguistic interaction, rather than linguistic areas, are a particularly promising area for future research. Kaufman and Justeson (2009: 222) assert that, in Mesoamerica generally, lexical borrowing is infrequent, though they do not specify what set of lexical borrowings they use to support this assertion, nor what that proposed infrequency is relative to. However, it is clear that Mayan languages have at least some identifiable lexical borrowing from several different Mesoamerican languages and language families. Highland Mayan languages have fairly numerous lexical borrowings from Nahuatl. Words like Ixil tenam ‘town’ from Nahuatl tinam(itl) ‘town’, or Tz’utujil masaat from Nahuatl mazatl are widespread. These borrowings are typically assumed to be relatively recent, borrowed either during the centuries immediately before the Spanish conquest, when Nahuatl speaking Aztecs were a prominent economic and political presence in the region, or during the Spanish Colonial era, when Nahuatl was sometimes used by Spanish interests as a regional indigenous lingua franca. Other contributing languages include Totonac and, particularly, Mixe‐Zoque. Evidence of apparently fairly recent intensive contact (within the last millennia) between a Zoquean language and the Mayan language Ch’ol has been noted by Zavala (2002). Perhaps one of the clearest and most striking cases is the borrowing of a Zoque relativizer ‐pä as ‐b’ä in Ch’ol. In Mayan languages, relative clauses typically follow relativized nouns. They may be marked with an overt relative clause initial relativizer or through simple juxtaposition. Ch’ol has adopted the Zoque order of elements in which the relative clause precedes the relativized noun. In addition, as with Zoque, the Ch’ol relativizer is a second position clitic, encliticizing to the first element of the relative clause. Another likely Mixe‐Zoquean feature borrowed into Ch’ol is the ergative (or “set A”) third person prefix i‐. Contact effects in the other direction, from Ch’ol (or another Mayan language) and Zoque can also be identified. Zavala (2002) has argued that innovations in directional verbs shared by many Mayan languages and Zoque can be shown to have originated in Mayan. Mixe‐Zoquean loanwords in Mayan are also distinctive because they are widespread and some are reconstructable to proto‐Mayan, suggesting that contact between Mixe‐Zoquean and Mayan languages has been ongoing since then. Kaufman (2003) reconstructs four such Mixe‐Zoquean loanwords for proto‐Mayan: pM *tzima(’) ‘dipper gourd, jícara’, pM *kakaw

Contact and Mayan Languages  619 ‘cocoa’ (though cf. Dakin and Wichmann 2000), pM *poom ‘incense’, and pM *’am ‘spider’. It has also been claimed that Mixe‐Zoquean and Mayan (and sometimes Totonacan) are, in fact all genetically related (Mora‐Marin 2016; McQuown 1956). However, given the clear evidence of continued contact between Mixe‐Zoquean and Mayan from the very earliest stages of the Mayan family’s development, it becomes extremely difficult to determine if those ­similarities that do exist are a faint signal of genetic relatedness, or similarities due to extensive contact. That ambiguity in how to interpret similarity between languages is perhaps even more problematic in the case of Teenek (Huastec), a Mayan language that is currently spoken some 1,000 kilometers from the rest of the Mayan languages, in northern Veracruz and San Luis Potosí. A close sister language to Teenek, Chicomuseltec, was spoken until the mid nineteenth century in Chicomuselo, in Chiapas, contiguous with the rest of the Mayan family. Together, these two languages form the Huastecan branch of Mayan. There is disagreement amongst specialists about how Huastecan relates to the rest of the family. The most widely accepted view, promoted by Kaufman (1976) is that Huastecan was the first language to separate from proto‐Mayan, and that speakers of proto‐Huastecan migrated from the proto‐Mayan homelands in what is present‐day Highland Guatemala to the region in which Teenek is spoken, the Huasteca, around 4,200 years ago. According to this model, speakers of Chicomuseltec, which is too similar to have been separated from Teenek for millennia, would have migrated back to the Maya homeland and their present location nearly 3,000 years later during the postclassic period (around 1100 ce). Another model (Roberston 1992), places Huastecan as a sister to the Ch’olan‐Tseltalan subgroup, separating from other languages much later than is proposed by Kaufman. Under this model, rather than Chicomuseltec speakers back‐ migrating to the Maya area after millennia away, they remained in the Huastecan homeland in Chiapas, and the Teenek would not have arrived in their present‐day northern location in the Huasteca until the postclassic era. At its heart, these conflicting models rest on different interpretations of similarity between Teenek and Mayan languages on the one hand, and, on the other hand, Teenek and non‐Mayan, particularly Oto‐Manguean, languages spoken around the Huasteca. While the empirical foundation of Kaufman’s model is largely unpublished, the key evidence, at least as far as can be inferred from publications (Kaufman and Justeson 2007, 2009) is the presence of two proposed lexical borrowings in Kaufman’s unpublished reconstruction of proto‐Oto‐ Pamean (*mu’ ‘cross‐sex sibling‐in‐law’, and *tzoa‐tz’ ‘bat’) adopted from proto‐Mayan (or perhaps, pre‐proto‐Huastecan) *mu’ and *so’tz’ respectively. Because these are forms that are both claimed to be reconstructable to earlier stages of Huastecan and earlier stages of Oto‐ Pamean, Kaufman argues that they require that speakers of Huastecan were present in the region when proto‐Oto‐Pamean was spoken, at least 3,000 years ago. Kaufman and Justeson (2008) also cite evidence of lexical borrowing from a Huastecan language into two now‐ extinct languages from the northern Huasteca, Yemé, and Pajalat. One of these lexical items they argue to be relevant for dating the migration of Huastecans to the present‐day location is the Yemé word ‘how many?’ The proto‐Mayan term for ‘how many?’ is fairly straightforwardly reconstructible as *jar, but in proto‐Huastecan, pM *r became a palatal glide *y and both Teenek and Chicomuseltec have jay for this term. The argument in this case is that the Yemé form is more plausibly an adaptation of /jar/ than /jay/, and so must have been borrowed prior to the shift of proto‐Mayan *r to proto‐Huastecan *y. Thus, the argument in support of Kaufman’s model is based primarily on evidence of lexical borrowing in reconstructed proto‐languages. The argument for placing Huastectan at a shallower node, as a sister to the Ch’olan‐ Tseltalan languages (Robertson 1992; Robertson and Houston 2003), is based primarily on shared morphological and phonological innovations. Campbell (1988: 221), for example, notes that Teenek and Ch’olan‐Tseltalan languages share the shift of pM *r>y, *q>k, *tj>t and

620  Danny Law *k>ch. Robertson (1992) also points out several morphological innovations that Teenek shares with Ch’olan‐Tseltalan languages, including the passive of derived transitives ‐at (restricted in function from a more general passive marker in proto‐Mayan) and the innovative passive for root transitives ‐ey, as well as marking of plural persons with a separate plural enclitic, rather than suppletive plural person markers, and a suffix of the form ‐Vl used on intransitive verbs in the imperfective aspect. The shared morphological and phonological innovations seem too numerous to allow chance parallel development as an explanation. Under the assumption that phonological features and bound grammatical morphology were “unborrowable,” Robertson therefore concluded that Huastecan similarities with Ch’olan‐Tseltalan languages were evidence of a Huastecan/Ch’olan‐Tseltalan subgroup. Another possibility, suggested by Law (2011, 2017b) is that the features that Huastecan shares with Ch’olan‐Tseltalan are in fact borrowable and are shared between these languages by contact. This conclusion follows on a study of contact effects in the Maya Lowlands (Law 2014) where high levels of areally shared phonological and morphological patterns and matter were shown to have circulated, including the vast majority of the features that Robertson highlighted as shared between Huastecan and Ch’olan‐Tseltan. This can be understood as an argument in favor of Kaufman’s analysis: if we attribute these similarities to contact, there is no reason to think that Huastecan languages form a subgroup with any other Mayan language, and they may plausibly have been the first branch to separate from proto‐Mayan. However, the observation of areally shared similarities between Huastecan and Lowland Mayan languages complicates the other element of Kaufman’s model – that Huastecan languages have been spoken in the Huasteca for thousands of years. The areal innovations identified by Law (2014) for the Maya Lowlands cannot all be firmly dated, but most must have taken place during the Maya Classic Period (200–900 ce) or later. This means that whatever scenario is entertained for the history of Huastecan languages, it must include a mechanism for the areal sharing of phonological and morphological innovations between the Maya Lowlands and the Huastecan languages. One possibility is that the Huastecan languages were spoken for millennia in roughly the area where Chicomuseltec was spoken, contiguous to languages involved in the Lowland sphere of linguistic interaction. The Teenek, under this model, would have only migrated to their present‐day location, relatively late, during the postclassic period, after much of the substantial linguistic exchange documented for the area had already taken place.

6 Conclusions While this brief overview of contact involving Mayan languages cannot do justice to the complexity of thousands of years of multilingualism, it showcases several themes of broader interest to theorists of language contact. In these brief concluding remarks, I will only highlight three of these. First, Mayan languages provide a rich context for investigating questions of borrowability and language mixing. The range of effects of contact that have been identified to date for Mayan languages include dramatic instances of the areal sharing of virtually every kind of linguistic feature, from syntactic patterns, to phonological processes, from lexical items to inflectional paradigms. While it is fairly well‐accepted today that, given the right conditions, any kind of linguistic feature can be copied from one language to another, the fact that the outcomes of Mayan language contact skew so strongly toward contact effects that are far less frequent in the literature for other parts of the world invites further investigation into how the conditions, both linguistic and socio‐historical, of Mayan language contact may be similar or different from other parts of the world and what further refinement that can suggest with regard to what can and cannot happen in different situations of language contact.

Contact and Mayan Languages  621 Related to this interest in the limits and possibilities of outcomes of language contact is the role that similarity due to relatedness has in the effects of contact. As argued by Law (2013), preexisting linguistic similarity between source and recipient languages, due to genetic relatedness, has arguably had a substantial role in shaping the outcomes of language contact in the Maya area. The borrowing of a paradigm of person markers, for example, was doubtless facilitated by the fact that both Yukatekan and Eastern Ch’olan languages (recipient and donor) had virtually identical morphosyntactic slots for person markers, and two paradigms of person markers (ergative and absolutive) with virtually identical functional distributions. In addition to its possible relevance to the processes and mechanisms of contact‐induced change, the fact that many languages in contact in the Maya area are related makes Mayan language contact of interest methodologically because of the challenges introduced when a researcher must attempt to disentangle similarity due to contact from similarity due to common inheritance. In this case, a holistic approach is necessary in order to evaluate the plausibility of different hypothetical scenarios. This is particularly visible in the case of Teenek/Huastec, where different assumptions about what can be borrowed, and what threshold should be used for allowing chance similarity as an explanation for shared linguistic features, result in dramatically different historical accounts. Finally, Law (2013, 2014) has also argued that genetic relatedness has been drawn on as a resource for defining and expressing group identity among the Maya, and that these dynamics of group identity have played a role in influencing the outcomes of contact. The particular outcomes of language contact in the Maya region have clearly been influenced by local ideologies of language and identity. Specifically, there is evidence that the European ideology that equates language to a large ethnic or national group identity was not prevalent among the Maya (and perhaps Mesoamericans more generally) prior to the Spanish conquest. And it is still not a key factor in many Maya communities, where kinship and place are more salient indicators of group identity than the language that one speaks. Law (2014) specifically argues that the lack of focus on linguistic systems as markers of identity, in a sense, provided no ideological obstacles to the transfer of specific linguistic elements across language boundaries. This provides an interesting contrast to other regions, such as the Amazon, in which it has been argued that hyper salience of language for group identity (because of the practice of linguistic exogamy) has been a strong deterrent to the transfer of certain types of linguistic features (Aikhenvald 1996; Epps 2007). More work remains to be done to understand how social factors such as ideologies of language and identity constrain or facilitate particular outcomes of language contact among the Maya, in Mesoamerica, and beyond.

REFERENCES Aikhenvald, Alexandra Y. 1996. Areal diffusion in North‐West Amazonia: The case of Tariana. Anthropological Linguistics 38: 73–116. Atkinson, Quentin D. 2006. From species to language: a phylogenetic approach to human prehistory (University of Auckland doctoral dissertation). Epps, Patience 2007. The Vaupés melting pot: Tukanoan influence on Hup. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Grammars in Contact: A Cross‐linguistic Typology (= Explorations in Linguistic Typol.ogy, 4). Oxford: Oxford University Press, pp. 267–89.

Barrett, Rusty 2002. The Huehuetenango Sprachbund and Mayan Language Standardization in Guatemala. In Mary Andronis, Erin Debenport, Anne Pycha, and Keiko Yoshimura (eds.), Proceedings of the 38th Chicago Linguistics Society: The Panels. Chicago: Chicago Linguistics Society, 309–18. Brody, Jill 1987. Particles borrowed from Spanish as discourse markers in Mayan languages. Anthropological Linguistics 29: 507–21.

622  Danny Law Campbell, Lyle 1977. Quichean Linguistic Prehistory. Berkeley: University of California Press. Campbell, Lyle 1988. The Linguistics of Southeast Chiapas, Mexico (Papers of the New World Archaeological Foundation, 50). Provo: New World Archaeological Foundation. Campbell, Lyle, Terrence Kaufman, and Thomas C. Smith‐Stark 1986. Mesoamerica as a linguistic area. Language 62: 530–70. Dakin, Karen and Søren Wichmann 2000. Cacao and chocolate: a Uto‐Aztecan perspective. Ancient Mesoamerica 11: 55–75. DeChicchis, Joseph 1989. Q’eqchi’ variation in Guatemala and Belize (University of Pennsylvania doctoral dissertation). Gómez Cruz, José 2017. Estructuras morfosintácticas del tojol‐ab’al en perspectiva comparativa: el caso de una lengua maya mixta [Morphosyntactic structures of Tojol‐ab’al from a comparative perspective: the case of a Mayan mixed language] (Centro de Investigación y Estudios Superiores en Antropología Social (CIESAS) Mexico doctoral dissertation). Hofling, Charles A. 2006. A sketch of the history of the verbal complex in Yukatekan Mayan languages. International Journal of American Linguistics 72.3: 367–96. Justeson, John S., William M. Norman, Lyle Campbell, and Terrence Kaufman 1985. The Foreign Impact on Lowland Mayan Language and Script. New Orleans: Middle American Research Institute, Tulane University. Kaufman, Terrence 1976. Archaeological and linguistic correlations in Mayaland and associated areas of Meso‐America. World Archaeology 8: 101–18. Kaufman, Terrence with John Justeson 2003. A Preliminary Mayan Etymological Dictionary. Foundation for the Advancement of Mesoamerican Studies. Available at www. famsi.org/reports/01050/index.html Kaufman, Terrence and John Justeson 2007. The history of the word for cacao in ancient Mesoamerica. Ancient Mesoamerica 18: 193–237. Kaufman, Terrence and John Justeson 2008. The Epi‐Olmec language and its neighbors. In Philip Arnold III and Christopher Pool (eds.), Classic‐period Cultural Currents in Southern and Central Veracruz. Cambridge, MA: Harvard University Press, pp. 55–83. Kaufman, Terrence and John Justeson 2009. Historical linguistics and pre‐Columbian Mesoamerica. Ancient Mesoamerica 20: 221–31.

Law, Danny 2009. Pronominal borrowing among the Maya. Diachronica 26: 214–52. Law, Danny 2011. Linguistic inheritance, social difference, and the last two thousand years of contact among lowland Mayan languages (University of Texas at Austin doctoral dissertation). Law, Danny 2013. Inherited similarity and contact‐induced change in Mayan languages. Journal of Language Contact 6: 271–99. Law, Danny 2014. Language Contact, Inherited Similarity and Social Difference: The Story of Linguistic Interaction in the Maya Lowlands (Current Issues in Linguistic Theory Series). Amsterdam: Benjamins. Law, Danny 2017a. Language mixing and the case of Tojol‐b’al. Diachronica 34: 40–78. Law, Danny 2017b. Language contact with(in) Mayan. In Judith Aissen, Nora England, and Roberto Zavala Maldonado (eds.), The Mayan Languages. London: Routledge, pp. 112–27. Law, Danny and David Stuart 2017. Classic Mayan: an overview of language in ancient hieroglyphic script. In Judith Aissen, Nora England, and Roberto Zavala Maldonado (eds.), The Mayan Languages. London: Routledge, 128–72. Livi Bacci, Massimo 2008. Conquest: The Destruction of the American Indios (Carl Ipsen, trans). Oxford: Polity. Martin, Simon and Nikolai Grube 2008. Chronicle of the Maya Kings and Queens, 2nd edition. New York: Thames & Hudson. Matras, Yaron 2007 The borrowability of structural categories. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross‐linguistic Perspective. Berlin: de Gruyter Mouton, pp. 31–73. McQuown, Norman A. 1956. Evidence for a synthetic trend in Totonacan. Language 32: 78–80. Montgomery‐Anderson, Brad 2010. Grammaticalization through language contact: the periphrastic passive in Chontal Mayan. Journal of Language Contact 3: 84–100 Mora‐Marin, David 2016. Testing the proto‐ Mayan‐Mije‐Sokean hypothesis. International Journal of American Linguistics 82.2: 125–80. Muysken, Pieter 2000. Bilingual Speech: A Typology of Code‐mixing. Cambridge: Cambridge University Press. Quizar, Robin 1994. Split ergativity and word order in Ch’orti’. International Journal of American Linguistics 60.2: 120–38.

Contact and Mayan Languages  623 Robertson, John 1992. The History of Tense/Aspect/ Mood/Voice in the Mayan Verbal Complex. Austin: University of Texas Press. Robertson, John and Stephen Houston 2003. El problema del Wasteko: una perspectiva lingüística y arqueológica [The Wastek problem: a linguistic and archaeological perspective]. In Juan Pedro Laporte, Bárbara Arroyo, Héctor Escobedo, and Héctor Mejía XVI (eds.), Simposio de Investigaciones Arqueológicas en Guatemala. Guatemala: Ministerio de Cultura y Deportes, pp. 723–33.

Stevenson, Paul 1990. Santa Maria Cauque: A Case of Cakchiquel–Quiche Language Mixing. Grammar Series. Guatemala: Summer Institute of Linguistics Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Zavala, Roberto 2002. Calcos sintácticos en algunos complejos verbales Mayas y Mixe‐ Zoques [Syntactic calques in some Mayan and Mixe‐Zoquean verbal complexes]. Pueblos y Fronteras 4: 169–87.

32 Contact and South American Languages LYLE CAMPBELL, THIAGO CHACON, AND JOHN ELLIOTT 1 Introduction1 The sheer size of South America (henceforth SA), the large number of languages, and its diverse geography, ecology, and cultures make language contact in SA a “vast and almost intractable” problem (Muysken 2012: 235; see also Epps and Michael 2017: 934). The range of language contact phenomena and multilingualism is also daunting. Our goal here is to provide an overview of language contact involving indigenous languages in SA, emphasizing cases that may have broader implications.

1.1  Classification and Diversity SA is home to ca. 107 language families – 53 language families and 54 language isolates (language families which have only a single member). Nine of the 53 families are relatively large, while 44 are small with six or fewer languages. The exact number of languages and language families in South America is just not known. For example, there are literally hundreds of unclassified languages in SA, most long extinct with little or no attestation (cf. Campbell 2012a: 114–31). There are also scores of uncontacted indigenous groups across the continent, especially in Brazil, where FUNAI (Fundação Nacional do Índio) lists about 107 isolated Indians groups throughout the Amazon.2 For many of these uncontacted groups, the language they speak is unknown. With about 406 language families in the world, SA’s 107 language families constitute 26% of the world’s linguistic diversity, calculated in terms of language families (Campbell 2018). Moreover, SA languages are typologically very diverse, as well (Campbell 2012b). This great linguistic diversity with its concomitant cultural diversity developed on the last major land mass to be colonized by humans, and in a relatively short period of time. Recent archeological and genetic research argues for a peopling of SA between 15,000 and 13,500 years before the present, probably stemming from one major migration from Central America (Rothhammer and Dillehay 2009: 546).

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

626  Lyle Campbell, Thiago Chacon, and John Elliott

2  Loanwords and Linguistic Acculturation There has been no wholesale study of loanwords in SA indigenous languages, though many loanwords are identified in studies of particular languages. Some languages are known to have borrowed many words from other indigenous languages, especially languages in contact with Quechua (Muysken 2012: 239). Amuesha underwent extensive borrowing from Quechua, which resulted also in some phonological change in the language (Adelaar 2006; Wise 1976). Esmeralda borrowed heavily from Barbacoan languages (Adelaar 2012: 583). Jivaroan has a number of apparent loanwords from Tupí‐Guaranían (Adelaar and Muysken 2004: 434). Viegas Barros (2001: 28–40) identified numerous loans and probable loans in Lule and Vilela from various neighboring languages. Zamponi (2003: 11) identified loans from Cariban languages in Maipure (a long extinct Arawakan language). Carvalho (2018) identified loans from Guaicuruan languages into Terena (Arawakan). Numerous other cases could be cited. Haspelmath and Tadmor (2009) has chapters on loans in particular SA languages, Epps (2009) on loans in Hup (Nadahup), Golluscio (2009) for Mapudungun, Gómez Rendón and Adelaar (2009) for Imbabura Quichua, Renault‐Lescure (2009) for Kali’na (Cariban), and Vidal and Nercesian (2009) for Wichí (Matacoan). Most of the loans identified in these works are not from other indigenous languages; for example, for Imbabura Quichua 95.3% are from Spanish. Muysken (2012: 251–2) speaks of some loanwords in “a process of chain borrowing,” borrowed from a prestige language into other languages and then passed on further to still others. A well‐known case is Quechua atawal ypa ‘chicken’, widely spread “through 35 pre‐Andine Amazonian languages” and well beyond (Adelaar and Muysken 2004: 500–1; cf. Nordenskiöld 1922). Aymara pataka ‘hundred’ was borrowed into Araucanian, and then into Allentiac, Tehuelche, and Gününa Yajich, and Quechua and Aymara numerals above ‘three’ or ‘four ’ are widespread in pre‐Andine languages (Muysken 2012: 252). Some Wanderwörter have been reported, words found widely diffused but where the original donor language is usually unknown (Haynie et al. 2014; Epps 2012).3 Most of the Wanderwörter in SA involve items of cultural significance. Among them are widespread terms for ‘spider monkey’ (originally from Arawakan?), ‘iguana’ (ultimately from Cariban?), ‘gourd dipper’ and ‘beans’ (Tupí‐Guaranían originally?), ‘coca’ (originated in Boran or Witotoan?), ‘maize’, and ‘signal drum’ (with connections to ‘canoe’, ‘bench’, ‘laurel tree’, ‘shaman/curer’). On the other hand, numerous languages of lowland SA have been reported to avoid lexical borrowing. To mention some cases, this is true of most languages in the Chaco (Campbell and Grondona 2012b: 653, 657; Campbell and Grondona 2012a); relatively little lexical borrowing occurs in languages of the Upper Xingu Linguistic Area, though some loans exist (Seki 1999, 2011; cf. Epps and Michael 2017: 946); and in languages of the Southern Guiana area there are relatively few loanwords (Carlin 2007). Epps and Michael (2017) believe that emphasis on language as an emblem of identity in these regions is what is behind restricted lexical borrowing, due to the desire to keep the languages distinct from one another. In the Vaupés region, known for being particularly resistant to direct borrowing, there are nevertheless reports of some lexical borrowings. For instance, Kubeo (Tukanoan) borrowed many words from Baniwa (Arawakan) (Chacon 2013); Aikhenvald (2002) mentions borrowings, particularly of verbs, from Tukano into Tariana. Rose et al. (2017) report borrowings in Tanimuka‐Retuarã (Tukanoan) from different Arawakan sources. Chacon (2017) analyzes several cases of lexical and grammatical borrowings from Arawakan languages into Proto‐Tukanoan and its daughter languages throughout the long history of contact between these families.

Contact and South American Languages  627 It is not, however, the case that the lexicon in languages intolerant of loanwords is not affected in language contact. The lexical effects of contact are seen, for example, in calquing and development of similarities in semantic domains (Epps and Michael 2017: 940). These languages accommodate acculturation by developing new words, without borrowing. Silva (1962: 57) and Floyd (2013: 275–91) show systematic calquing among languages in contact in the Vaupés area, particularly in the domains of place names, ethnonyms, and cosmology. For new words, languages of the Chaco rely on a derivational suffix meaning ‘similar to’, compounding, and sometimes onomatopoeia. For example, in Nivaclé (Matacoan) the suffix ‐tax ‘similar to’ derives new words for introduced concepts, as in tašinš‐tax ‘goat’ (< tašinša ‘gray brocket deer’ + ‐tax ‘similar to’). Some new concepts are brought in by metaphorical extension of the meaning of existing words. For example Nivaclé siwȏklȏk ‘spider’ was extended to mean also ‘bicycle’, and tukus ‘ant’ to mean ‘soldier’. Other examples created by metaphorical extensions include ‘flour’ < ‘dust’; ‘lemon’ < ‘sour;’ ‘needle’ < ‘thorn’; and ‘wristwatch’ < ‘bracelet’. Compounding and lexical items formed from phrases gave ‘glass’ < ‘sight goes through it’, ‘guitar’ < ‘horse’s tail’, ‘hair of horse’s tail’; ‘radio’ < ‘it sings’; ‘scissors’ < ‘knife’s legs’, etc. New words based on onomatopoeia include, for example, ‘motorcycle’: Nivaclé k’ututut and Chorote pohpoh (Campbell and Grondona 2012a). Some of the languages extensively use nominalization for verbs to develop new nouns, for example Enxet sẽlpextetamo ‘soldier’ is literally from ‘those who tie us up’, from underlying seg‐el‐pextet‐amo 1pl.inv‐ dist‐tie‐nmlz (pl = plural, inv = inverse, dist = distributive, nmlz = nominalization). These numerous SA cases with restrictions against lexical borrowing but with fairly abundant structural borrowing (see pp. 627–33) go against the claim that structural borrowing is only possible if connected with lexical borrowing (e.g. Labov 2007).

3  Linguistic Areas of South America A linguistic area (also called Sprachbund; henceforth LA) is a geographical area where languages share structural features due to language contact. Several LAs have been proposed in SA; some are more securely defined than others. We survey them briefly.

3.1  Proposals of Macro Linguistic Areas in SA David Payne (1990) pointed out some widely shared traits among South American languages that he believes indicate either diffusion or an undocumented deep genetic relationship. They include: a negative morpheme of the approximate shape ma, a causative affix of the approximate shape mV, a causative verbal prefix (a back vowel), a directional verb suffix (often pV or Vp), an auxiliary ‘to have’, ‘to do’, or ‘to be’, usually containing ka, often ­coinciding in the same language with the verb ‘to say, to work’ and often with a valency‐ changing verbal affix of the same or similar shape. Other proposals have suggested macro subdivisions within SA. Doris Payne (1990a, 1990b) and Harriet Klein (1992: 33–4) considered the possibility that all lowland SA languages constitute a LA. An east‐west, or Highland‐Lowland, division among SA languages has been contemplated since Lafone Quevedo (1896), who based his division on suffixed pronominal markers (his “Quichua” type languages) vs prefixed ones (his “Guaraní” type). Payne (1990: 213–40) and Birchall (2014) suggest a division between Western and Eastern Amazonian languages. In Payne’s and Birchall’s proposals, their Eastern area covers Cariban, “Macro‐Jêan,” and Tupían languages, where Birchall’s also includes the Chaco‐Planalto area.4 Their Western contact areas also coincide, except Birchall also includes Southern Cone and Andean

628  Lyle Campbell, Thiago Chacon, and John Elliott languages in his Western South American Linguistic Area (WSALA). Payne’s (1990: 213–40) western group includes languages from Pano‐Takanan, Arawakan, Tukanoan, Saliban (Sálivan), Zaparoan, Yaguan, Witotoan, and Cahuapanan families. It is characterized by highly verbal polysynthetic languages, verbal directionals (that may have tense‐aspect‐modality functions), noun classification systems (missing in Pano‐Tacanan and some Arawakan ­languages), and verb‐initial and postpositional word orders (found in some Arawakan and Zaparoan languages, and in Taushiro and Yagua). Payne’s eastern group includes languages belonging to the “Je‐Bororo” (Bororoan), Tupían, Cariban, and “Makúan” (Nadahup) families. They share a more isolating morphology and minimal (or no) verbal directionals, and they lack noun classification. Birchall (2014: 203) found that ergative alignment, which had been considered an Amazonian feature by some, connects only with the southern Amazon region and his Eastern South American Linguistic Area (ESALA), not with northern or ­western Amazonia. The inclusive‐exclusive distinction, claimed as an Andean feature, is not limited to that area, but is a feature of Birchall’s WSALA. Many Cariban and Tupían languages also have a first person plural inclusive‐exclusive distinction, despite belonging to ESALA. Birchall found several other features that had been associated with particular SA LAs turn out to reflect either WSALA or ESALA (Birchall 2014: 205–6). Birchall’s east‐west division is seconded by Epps and Michael (2017: 952–3), who report that phonological similarities in SA also cluster in ways consistent with Birchall’s WSALA vs ESALA.

3.2  Amazonia as a Linguistic Area? Amazonia is the lowland region drained by the Amazon and Orinoco Rivers. Its status as a LA is much talked about but is uncertain. Derbyshire and Pullum (1986: 16, 20) found typological tendencies across Amazonia, but they caution that the amount of information available was not enough for certainty. The language families included are: Arawakan, Arawan, Cariban, Chapacuran, Jêan, Pano‐Takanan, Puinave, Kakua‐Nukak, Tucanoan, and Tupían (Derbyshire and Pullum 1991: 3). Derbyshire and Pullum mentioned the tendency of these languages to have O‐before‐S orders (VOS, OVS, OSV); Derbyshire (1987: 313) noted that in many of the OVS and OSV languages the word order tends to be flexible, making it difficult to decide which word order is basic. Other widely shared Amazonian traits include: verb agreement with both subject and object (plus null realization of subject and object nominals); nominalizations for relative clauses and other subordinate clauses; nominal modifiers following their heads (Noun‐Adjective, Genitive‐Noun, Noun‐postposition orders); no agentive passive construction; lack of indirect speech forms (reliance on direct speech constructions); absence of coordinating conjunctions (juxtaposition to express coordination); extensive use of right‐dislocated paratactic constructions; extensive use of particles that are phrasal subconstituents syntactically and phonologically but are sentence operators or modifiers semantically; tendency toward ergative subject marking (but see p. 628); complex morphology; noun classifier systems (Derbyshire 1986: 560–1; Derbyshire and Pullum 1986: 16–19; Derbyshire and Payne 1990). Constenla Umaña (1991: 135) questioned some of the traits listed for Amazonia. He believed that Amazonia is not a single LA, but rather that some of its languages should be assigned to neighboring LAs instead. He believed the following were common also in LAs beyond Amazonia: absence of a passive construction; agreement of transitive verb with subject and object, and the correlated null realization of full noun phrases in cases of “given” information; and the predominance of the orders Noun‐Adjective, Genitive‐Noun, and postpositions. He noted also that the use of nominalizations for subordinate clauses and the absence of direct‐indirect speech indicators have broader distributions. Epps and Michael (2017: 952) also challenge some of these traits (see also Klein 1992: 33–4).

Contact and South American Languages  629 Aikhenvald (2012: 68–71) revised the claim of Amazonia as a LA by introducing the term “language region”, in contrast to the more rigorously defined notion of “linguistic area,” to characterize Amazonia as a region of recurrent but sporadically attested shared features, such as antipassives and complex classifier systems.

3.3  Northern Amazonia 3.3.1  The Vaupés (Vaupés‐Içana) Linguistic Area More has been written about language contact in the Vaupés region than anywhere else in SA. It involves dozens of languages belonging to the Eastern Tukanoan, Arawakan, Nadahup, and Kakua‐Nukak language families, with Nheengatú also in the area (see pp. 634–5). The area is famous for its widespread multilingualism, associated with linguistic exogamy (see pp. 638–9). Language contact among the languages of the area has resulted in a high degree of morpheme‐to‐morpheme and word‐to‐word intertranslatability. As Aikhenvald repeats in several p ­ ublications, Tariana (the sole Arawakan language in the area) has undergone extensive changes in its grammar under Tukanoan influence, for example change in word order, development of nominal cases, reduction of the set of multiple locative markers to a single one), elaborate changes in the nominal classifier system, change of the two evidential distinctions to five, verb compounding, switch reference, new complementizer structures, and phonological changes (cf. Epps and Michael 2017: 940). There has also been diffusion among the Eastern Tukanoan languages (Chacon 2013; Gómez‐Imbert 1993). The Nadahup languages and Kakua have also changed markedly due to influence from Tukanoan (Epps 2005, 2007, 2008; Epps and Michael 2017). On the other hand, Arawakan languages have influenced some Eastern Tukanoan languages, for example with the development of aspirated stops in Kotiria (Wanano), development of an alienable vs inalienable distinction, possessive proclitics in Kotiria, Kubeo, Tatuyo, and some other languages, use of shape classifiers with non‐human animates in Kubeo and (to a lesser extent) in Kotiria (Gómez‐Imbert 1996; Stenzel and Gómez‐Imbert 2009; Stenzel 2013). The influence of Tukanoan on Tariana, as well as on Kakua and Nadahup languages, seems to reflect more recent language contact in the region, where Arawakan languages were in a more dominant position in the more distant past (Chacon 2017; Epps 2017).

3.3.2  The Caquetá‐Putumayo Linguistic Area A Caquetá‐Putumayo LA has been proposed for southern Colombia and northern Peru, though not well studied yet. Languages of this LA are the Boran and Witotoan families, Resígaro (Arawakan), and Andoke (language isolate). Speakers of these languages refer to themselves together as “People of the Center” and are widely multilingual. The features shared in the area include: first person plural inclusive/exclusive contrast, second‐position tense‐aspect‐mood clitics, loss of object cross‐referencing suffixes, restructuring of verbal morphology (Seifart 2011: 14; Aikhenvald 2001: 189). Chang and Michael (2014) argue that these languages have undergone significant convergence in their phonological inventories (Epps and Michael 2017: 944). Resígaro added new phonemes (/ɸ/, /dʒ/, /ʔ/), syllable structure restrictions, and a two‐tone contrast due to influence from other languages of this LA (Aikhenvald 2001; Seifart 2012; Chang and Michael 2014).

3.3.3  The Venezuelan‐Antillean Area Constenla Umaña (1991) proposed a Venezuelan‐Antillean LA that includes several Arawakan languages (e.g. Taino, Island Carib, Caquetio, Lokono) and various Cariban languages (e.g. Cumanagoto [Cumana], Chaima [Cumana], Tamanaco, Cariña), Otomacoan,

630  Lyle Campbell, Thiago Chacon, and John Elliott Guamo, Yaruro, and Warao). The traits common to the languages of the area are: exclusively VO order (absence of SOV), absence of voicing opposition in obstruents, Numeral‐Noun order, Noun‐Genitive order, and prepositions (Constenla Umaña 1991: 125–6). Constenla Umaña (1991: 136) believed that this area could be extended to the south to include the western part of the Amazonia, where Arawakan languages with VO order predominate (Constenla Umaña 1991: 125–6; Campbell 2012b: 307; Epps and Michael 2017: 948–9).

3.3.4  The Colombian‐Central American Area Constenla Umaña (1991: 103) also proposed what he called the Colombian‐Central American LA, composed primarily of Chibchan languages, but also including Lencan, Jicaquean, Misumalpan, Chocoan, and Betoi, with the following shared traits: voicing opposition in stops and fricatives, exclusive SOV order, postpositions, mostly Genitive‐Noun order, Noun‐ Adjective order, Noun‐Numeral order, clause‐initial question words, suffixation or postposed particle for negative, absence of gender opposition in pronouns and inflection, absence of accusative case marking in most of the languages, absence of possessed/non‐possessed and alienable/inalienable possession oppositions, and the presence of lexical compounds rather than independent roots (Constenla Umaña 1991: 126–9; Campbell 1997: 347).5

3.3.5  The Northeastern Amazonia: Southern Guiana Linguistic Area The Northeast Amazon region is located within a wide geographical area encompassing eastern and southeastern Venezuela, Guyana, Suriname, French Guiana, and the Brazilian states of Roraima, Amapá, and northeastern Pará. It is an area predominantly occupied by speakers of Cariban languages, but also of a number of Arawakan languages and languages of the Tupí‐Guaranían branch of Tupían, plus Yanomaman, Warao, Arutani, Sapé, and Mako. In addition, five official European languages (English, Portuguese, Spanish, French, and Dutch) are in use in the region, and numerous creoles developed or were brought into the region during its colonial history, for example Sranantongo, Haitian Creole, Guyanese English Creole, Kheuól, etc. There is intense interaction with structural diffusion in languages of Southern Guiana, involving several Cariban, Arawakan, and Sálivan languages (Carlin 2007, 2011). Carlin (2007) found grammatical diffusion from Cariban languages (principally Trio and Waiwai) into Mawayana (Arawakan), seen in comparison with Wapishana, a sister of Mawayana: creation of a first person exclusive distinction (borrowing of a Waiwai form), nominal tense marking, marking on nouns or verbs to express “pity” or “recognition of unfortunate circumstance,” frustrative marker on verbs, “similative” marker on nominals, etc. (Epps and Michael 2017: 948–9).

3.4  Southern Amazonia 3.4.1  Guaporé‐Mamoré Linguistic Area The Guaporé‐Mamoré linguistic area is linguistically very diverse; it includes parts of Rondônia state in Brazil and the Amazonian region of Bolivia, with over 50 languages from 17 different families (Arawakan, Chapacuran, Jabutían, Nambikwaran, Pano‐Takanan, Tupían, 11 language isolates, 12 unclassified languages, and 1 pidgin). It has many shared structural features together with shared cultural traits, with considerable morphological borrowing in the Brazilian part of the zone. Some of the shared traits include grammatical gender, evidentiality, classifiers, verbal classification, subordination by nominalization, switch reference, many prefixes, preference for verbal number (alteration of the verb by ­suppletion, reduplication, or affixation to express the number of the subject or object),

Contact and South American Languages  631 accompanied by a general lack of nominal number, verbal cross‐reference systems with ­similar morpheme positions, and complex systems of directional morphemes (cf. Braga et al. 2011: 224–9; Crevels and van der Voort 2008: 167; Epps and Michael 2017: 947–8). Diffusion in the Guaporé‐Mamoré LA also involves cases of direct borrowing of morphological forms, especially in the nominal classifier systems (van der Voort 2005: 397; Crevels and van der Voort 2008: 167) – a parallel to which is known from the Caquetá‐Putumayo LA (Epps and Michael 2017: 947–8).

3.4.2  A Tocantins‐Mearim Interfluvium Linguistic Area? Another possible LA is in the land between the Tocantins and Mearim Rivers in northeastern Brazil, with several Tupí‐Guaranían and Jêan languages. The Tupí‐Guaranían languages include Guajajára and Tembé (Branch IV, Tenetehára), Guajá and Urubú‐Ka’apór (Branch VIII), memory of Anambé and Amanajé (Branch V), and evidence of Ararandewára (Branch V), Turiwára (branch IV), Tupinambá, and Língua Geral Amazônica (Branch III). Jêan is represented by three dialects of Timbira (Braga et al. 2011: 224). Shared features include change in a grammatical particle, change in gerund, loss of first person plural inclusive‐exclusive contrast, and argumentative case (Cabral et al. 2007).

3.4.3  The Upper Xingu Linguistic Area The Upper Xingu is another possible LA. The Xingu is an area of intense interaction among several different ethnic groups, involving especially ritual and trade (Chacon 2020; Epps and Michael 2017: 946). There are more than a dozen languages in the area, belonging to the Cariban, Arawakan, Jêan, and Tupían families, and Trumai (language isolate). Several of these groups arrived since the sixteenth century due to colonial pressures. Seki (1999, 2011) considered the Xingu an “incipient” LA. The features she thought were diffused in the region include: loss of a masculine‐feminine gender distinction in the Arawakan languages of the region, diffusion of /ɨ/ into Xingu Arawakan languages from their Cariban or Tupí‐Guaranían neighbors (Chang and Michael 2014), p > h shift in Cariban and Tupí‐Guaranían, change to CV syllable structure in Cariban, diffusion of /ts/ into the Xingu Cariban languages from Arawakan, and diffusion of nasal vowels into the Xingu Arawakan and Cariban languages from their Tupí‐Guaranían neighbors. The ­diffusion among the Xingu languages appears to be multidirectional (cf. Epps and Michael 2017: 947).

3.5  Highlands (West) 3.5.1  The Andean Area The central highland Andean region is considered a LA, though it has not been investigated carefully. It is dominated by Quechua and Aymara, though several other languages are also associated with the area, several of them long extinct: Atacameño (Kunza), Callahuaya, Candoshian, Chacha, Cholonan (Cholón and Hibito), Culle (Culli), Esmeralda, Mochica (Yunga, Muchik), Puquina, Quingnam, Sechura‐Catacaoan (Sechura, Tallán), Uru‐Chipayan, and numerous other now extinct languages about which we have little information (Adelaar 2008, 2012). Constenla Umaña (1991: 123–4) postulated a broad Andean area which also included his Ecuadorian‐Colombian subarea, containing the languages of highland Colombia, Ecuador, Peru, and Bolivia. He also believed that some languages spoken in the region east of the Andes could also be incorporated into the Andean LA. Büttner (1983: 179) included Quechua, Aymara, Callahuaya, and Chipaya in the LA he defined based on phonological traits.

632  Lyle Campbell, Thiago Chacon, and John Elliott Quechuan and Aymaran share a large number of structural, phonological, and lexical similarities, constituting “one of the most intriguing and intense cases of language contact to be found in the entire world” (Adelaar 2012: 575). “Often treated as a product of long‐ term convergence, the similarities between the Quechuan and Aymaran families can best be understood as the result of an intense period of social and cultural intertwinement” (Adelaar 2012: 575). Some of the traits that Adelaar (2008) lists as having been considered representative of languages of the Andes include: complex number system; “agglutinative structure with an exclusive or near exclusive reliance on suffixes for all morphological and morphosyntactic purposes” (Adelaar 2008: 25); “constituent order is relatively free in Andean languages, although there seems to be a preference for the order in which subject/actor and object precede the verb (SOV)” (Adelaar 2008: 26); in many of the languages, “subordinate clauses are strictly verb‐final” (Adelaar 2008: 26); in most of the languages, in noun phrases “modifiers must precede the modified” (Adelaar 2008: 26); few vowels, no tone contrasts, and no “nasality spread” (Adelaar 2008: 26); many have a contrastive uvular stop (Adelaar 2008: 27); “absence of Amazonian‐type classifier systems” (Adelaar 2008: 28); “gender distinctions are not expressed morphologically” (Adelaar 2008: 28); “case marking on noun phrases expressed by means of suffixes or postpositions is common” (Adelaar 2008: 29); “accusative case marking is found in several central Andean languages” (Adelaar 2008: 29); “the stative‐ active distinction, which is attested in eastern lowland languages … and in languages of the Gran Chaco … has not been found in the Andes” (Adelaar 2008: 30); “verbal morphology is extremely rich and varied” (Adelaar 2008: 30); first person plural inclusive‐exclusive contrast (Adelaar 2008: 31); etc. Several of these, however, as seen in other LAs, are not exclusive to languages of the Andean region. As Adelaar (2008: 23) mentions, it “has been a common practice among linguists working on South American languages to make an intuitive distinction between ‘Amazonian’ and ‘Andean’ languages on the assumption that there would be two different language types corresponding to these labels.” Nevertheless, as several have pointed out (e.g. Campbell 2013; Campbell and Grondona 2012b; Dixon and Aikhenvald 1999: 9; Epps and Michael 2017: 952), the separation of the Andean and Amazonian LAs is not supported; rather the evidence speaks for the west‐east split (pp. 627–8), where the western area corresponds roughly to the Andes, Southern Cone, and the Western Amazonian region, matching Birchall’s (2014) WSALA (cf. Payne Doris 1990b), while the other large linguistic area to the east consists of the remainder of the continent.

3.5.2  The Ecuadorian‐Colombian Subarea The Ecuadorian‐Colombian subarea of the Andean LA, as Constenla Umaña (1991) defined it, includes Paez; Guambiano; Cuaiquer, Cayapa, Colorado (Barbacoan); Kamsá; Cofán; Esmeralda; and Ecuadoran Quichua. Most of these languages share the traits: high‐mid opposition in front vowels, absence of glottalized consonants, glottal stop, absence of uvular stops, voiceless labial fricative, rounding opposition in non‐front vowels, lack of person inflection in nouns, prefixes to express tenses or aspects (Constenla Umaña 1991: 123–5).

3.6  Southern Regions 3.6.1  The Southern Cone Area? Klein (1992: 35) noted several traits common to languages of the Southern Cone (e.g. Mapudungun, Guaicuruan, and Chonan), including: semantic positions signaled morphologically by means of “many devices to situate the visual location of the noun subject or object

Contact and South American Languages  633 relative to the speaker; tense, aspect, and number are expressed as part of the morphology of location, direction, and motion” (Klein 1992: 25); more back consonants than front ones; SVO basic word order (cf. Campbell 1997: 351). This possible LA has not received much attention.

3.6.2  A “Fuegian” (Tierra‐del‐Fuego and Patagonia) Linguistic Area? The possibility has been raised, in different guises, that the languages of Tierra del Fuego and possibly also of Patagonia form a LA, but this remains inconclusive (Adelaar and Muysken 2004: 578–82). Enclitizicization, suffixation, compounding, and reduplication are widely shared among these languages. The languages of Tierra del Fuego mostly have OV word order (where the position of the subject varies) (Adelaar and Muysken 2004: 579). The Tierra de Fuego languages are: Chonan, Kawesqaran, Chono (isolate), and Yahgan (isolate).

3.7  The Gran Chaco Linguistic Area? The Gran Chaco is the extensive lowland plain of central SA, stretching across northern Argentina, Paraguay, southeastern Bolivia, and southern Brazil. It is also a culture area (Métraux 1946; Murdock 1951) characterized by cultural traits shared across ethnic boundaries. More than 20 languages from six language families are known from the Chaco: Guaicuruan, Guachí, Payaguá, Matacoan, Enlhet‐Enenlhet (Mascoyan), Lule‐Vilelan, Zamucoan, some Tupí‐Guaranían languages, and Charrúan. A fair amount has been written about shared traits among these languages and the possibility of a Chaco LA. A number of traits are shared among these languages but several of them are too widespread to help in defining a LA. These traits include: SVO word order, grammatical gender, possessive classifiers, complex set of demonstratives, active‐stative verb alignment, lack of verbal tense and presence of nominal tense, extensive set of directional verbal affixes, etc. (Campbell 2006, 2013, 2017a; Campbell and Grondona 2012b). Most shared structural traits in languages of the region are found also in languages beyond the Chaco, and some that were thought possibly to be areal traits are actually characteristic of only a few of the languages within the region. Few are true of a majority of Chaco languages; none is unique to the area and, as mentioned, some are commonplace (e.g. SVO word order). Active‐stative verb alignment is shared by most of the languages of the area, but not exclusively so. The Tupí‐Guaranían languages of the Chaco share active‐ stative verb alignment but since Tupí‐Guaranían extends far beyond the Chaco, a LA based on this feature or on Tupí‐Guaranían languages that have it would hardly reflect a “Chaco” LA per se; Tupí‐Guaranían also shares many traits with languages of the putative Amazonian LA (pp. 628–9). This overlapping of shared traits between Chaco languages and languages beyond the Chaco has complicated attempts to define a Chaco LA. (Campbell 2013, 2017a.)

3.8  Linguistic Areas and Rivers? Van Gijn et al. (2017) asked whether river systems played a role in the formation of linguistic areas in SA. They conclude that river systems appear not to contribute significantly to patterns of diffusion in South, based on their examination of Noun-Phrase features. This is a surprising result, since several of the linguistic areas most talked about in SA involve rivers, the Vaupés area, the Guaporé-Mamoré area, the Upper Xingu area, and others, where patterns of diffusion and language contact do seem to rely on rivers. Could it be the case that an investigation of other shared structural traits, not just noun-phrase traits, might render a different sort of conclusion? This question definitely should be investigated.

634  Lyle Campbell, Thiago Chacon, and John Elliott

4 Contact Languages, Lingua Francas, Mixed Languages, and Pidgins and Creoles Here we survey the contact languages involving indigenous languages of SA.

4.1  Lingua Francas Epps and Michael (2017: 954) believe cases of lingua francas are “rare or even unattested in … Amazonian contact zones.” There are, nevertheless, several cases of lingua francas in Amazonia as well as elsewhere in SA, most of them attested since colonial times. The use of indigenous lingua francas appears to have been common in the pre‐colonial Andes, but less common in the lowlands, although more research on this topic is needed. Some argue that a confluence of economic and sociolinguistic factors made lingua francas unlikely or at least unstable in some regions of lowland SA prior to changes brought about by colonization (Ball 2011: 95). We list known examples of SA lingua francas here.

4.1.1  Lenguas Generales / Línguas Gerais In Latin America, Lenguas Generales (Línguas Gerais in Portuguese‐controlled areas) were established in the sixteenth century to facilitate religious and political administration in the colonial period. These were languages of groups with strong influence over larger territories that were used as lingua francas among speakers of different languages. The status of these “general languages” was strengthened by King Philip II’s favorable attitude towards these languages; in 1580 he ordered university chairs of the general languages to be established. The major “lenguas generales” (“línguas gerais”) of South America included Aymara, Chibcha (Muisca), Guaraní, Puquina, Quechua, and Tupinambá (cf. Estenssoro 2015; Itier 2011).

4.1.2  Other Lingua Francas Chiquitano was promoted as a lingua franca by Jesuit missionaries, imposed on the groups reduced to and united in their mission probably because it was the language of the most numerous indigenous group there (cf. Santana et al. 2009: 101). Kamayurá (Tupí‐Guaranían branch of Tupían) serves as a kind of lingua franca in the Upper Xingu region (https://pib.socioambiental.org/en/Popular:Trumai). Makurap (Tuparaían branch of Tupían): at the time of the rubber tappers, Makurap speakers were the first and main interlocutors of White rubber bosses and other Whites. Makurap became a lingua franca in the Rio Branco region used among several other indigenous groups (https://pib.socioambiental.org/en/Popular:Tupari). Portuguese has taken the place as the lingua franca formerly occupied by Makurap in the Middle Guaporé region (https://pib.socioambiental.org/en/Povo:Wajuru). Kubeo functions as the lingua franca for the northwest Vaupés region. Its role is expanding (Chacon 2013). Língua Geral do Sul (Língua Geral Paulista, Língua Geral of São Paulo, Tupí Austral) was originally the Tupí language of São Vicente and the upper Tiete River, different from Tupinambá, less well‐known than Língua Geral Amazȏnica (Nheengatú) (pp. 634–5). Língua Geral do Sul was the dominant language in São Paulo until it was displaced by Portuguese in the eighteenth century (Rodrigues 1986: 102). Língua Geral do Sul was spoken by those from São Paulo who in the seventeenth century explored the states of Minas Gerais, Groias, and Mato Grosso, and southern Brazil. As the language of these settlers and adventurers, it penetrated far into the interior. Later, government policy promoted Portuguese and Língua Geral do Sul became extinct.

Contact and South American Languages  635 Nheengatú (Língua Geral Amazȏnica, “Língua Boa,” Língua Brasílica, Lingua Geral do Norte, called Yeral in Venezuela) is a simplified version of Tupinambá that developed as a lingua franca for interethnic communication in northern Brazil. Tupinambá (Tupían), spoken broadly in Maranhão and Pará along the northern Brazilian coast, came to be used widely among the Portuguese colonial population during the seventeenth century and gave rise to Língua Geral (Língua Geral Amazȏ nica) or Nheengatú, the language spoken in the missions and by the colonizers who pushed into the Amazonian interior. It became the lingua franca for a large region of the upper Amazon, used by various indigenous groups, spreading throughout the Amazon basin, reaching Peru, Colombia, and Venezuela. It is still spoken extensively along the Rio Negro and in pockets elsewhere in the Amazon region, but it waned with the decline of rubber extraction, with Tukano taking over its lingua franca function throughout the northwest Vaupés region. Today’s Nheengatú differs from both Tupinambá and eighteenth‐century Língua Geral. It simplified the demonstratives from a system that contrasted visible/invisible and three degrees of proximity (like ‘this’/‘that’/‘that yonder’) to just two, ‘this’ and ‘that’. The personal pronouns system was reduced from various plural forms for “inclusive” and “exclusive” to only one. Its five moods, “indicative, imperative, gerund, circumstantial, and subjunctive,” collapsed into a single one corresponding to the old “indicative”; it lost its system of six cases. It developed subordinate clause structures that are more like Portuguese (Bessa Freire 1983; Moore et al. 1994; Rodrigues 1986: 99–109; Taylor 1985). Tonocoté is long extinct and essentially nothing is known of the language itself, though it was reported as in use as a lingua franca to communicate with Lules and Abipones in the Chaco (Caraman 1975: 192, 209). Tukano is the lingua franca for the southeastern Vaupés region, used by speakers of many languages there. Its role is expanding. (https://pib.socioambiental.org/pt/ Povo:Tukano.) Waiwai (Cariban) has become a lingua franca, the language predominantly spoken in the indigenous General Assemblies in its region (https://pib.socioambiental.org/en/ Public:Waiwai).

4.2  Mixed Languages A mixed language is one which has two source languages for different components of its grammar and so has no single ancestor. There are very few true mixed languages, though SA has well‐known examples. It is believed that they can arise only in very limited social circumstances, typically correlating with how newly formed or reorganized ethnic groups see their identity.

4.2.1  Media Lengua and Related Varieties Media Lengua is a mixed language in Salcedo (Cotopaxi province, Ecuador); the name is Spanish meaning ‘half language’. Muysken (1980: 75) defines Media Lengua as “a form of Quechua with a vocabulary almost completely derived from Spanish, but which to a large extent preserves the syntactic and semantic structures of Quechua,” where “all Quechua words, including all core vocabulary, have been replaced” by Spanish (Muysken 1994b: 203). Muysken speculates that Media Lengua probably originated with acculturated Quechua who did not identify completely with either rural Quechua culture or Spanish culture, and Media Lengua served the role of ethnic self‐identification (Muysken 1981, 1994b). It began around 1967 when many young men started working in construction in a nearby town and began learning Spanish. Muysken (1997: 376) claims that this occurred “because acculturated Indians could not completely identify with the traditional Quechua culture or the urban Spanish culture.”

636  Lyle Campbell, Thiago Chacon, and John Elliott The mixture is seen in the following example, where relexification is reflected in the lexical roots, all of Spanish origin, in boldface, with the grammatical morphology from Quechua, in italics: (1)

Unu fabur‐t pidi‐nga‐bu bini‐xu‐ni one favor‐acc ask‐nmlz‐ben come‐prog‐1sg ‘I come to ask a favor.’ (Muysken 1997: 365) (acc = accusative, nmlz = nominalization, ben = benefactive, prog = progressive)

The lexical roots, all from Spanish, are: /fabur/ < Sp. favor ‘favor’; /pidi/ < Sp. pedi(r) ‘to ask’; /bini/ < Sp. veni(r) ‘to come’. The Media Lengua of Salcedo may no longer be spoken. There are other “media lenguas,” Catalangu, Saraguro Media Lengua (Muysken 1980), and Imbabura (Pijal) Media Lengua (Gómez Rendón 2005, 2008). Catalangu is also a mixed language but is said to be “much closer to Spanish than [Salcedo] Media Lengua” (Muysken 1980: 78; see also Lipski 2017).

4.2.2  Callahuaya (Machaj‐Juyai, Kallawaya) Callahuaya is a mixed language or jargon based predominantly on lexical items from Puquina (long extinct) with Quechua morphology. It is used only for curing ceremonies by male curers from Charazani and a few villages in the provinces of Muñecas and Bautista Saavedra, Department of La Paz, Bolivia. These curers travel widely to practice their profession. They also speak Quechua, Aymara, and Spanish. Stark (1972) found that 70% of Callahuaya words in the Swadesh 200–word list are from Puquina, 14% from Quechua, 14% from Aymara, and 2% from Uru‐Chipaya; Muysken (1994a) reports that some Callahuaya words are also from Takana. The grammatical morphology, however, is almost wholly from Quechua, for examples: ‘accusative’ ‐ta (Callahuaya usi‐ta, Quechua wasi‐ta ‘house’), ‘imperative’ ‐y (Callahuaya tahra‐y, Quechua lyank’a‐y ‘work!’), ‘plural’ -kuna (Callahuaya simi‐kuna, Quechua ñan‐kuna ‘roads’). The possessive pronominal suffixes are identical in the two languages (‐y ‘my’, ‐yki ‘your’, ‐n ‘his/her/its’, ‐nčis ‘our [inclusive]’, ‐yku ‘our [exclusive]’). The locative suffixes are identical (‐man ‘to’, ‐ manta ‘from’, ‐pi ‘in’, ‐wan ‘with’, ‐rayku ‘because of’). The verb morphology is also from Quechua (e.g., ‐a ‘causative’, ‐na ‘reciprocal’, ‐ku ‘reflexive’, ‐mu ‘directional hither’, ‐rqa ‘past’, ‐sqa ‘narrative past’) (see Büttner 1983: 23; Muysken 1994b; Oblitas Poblete 1968; Stark 1972).

4.2.3 Kokama? Kokama (aka Kokama‐Cocamilla, Kukama‐Kukamiria, Kokáma/Omágua) is endangered, spoken in the Peruvian Amazon (Loreto and Ucayali departments) by about 1,500 people. The Kokama lexicon is mostly Tupinambá, but contains also words of Portuguese, Spanish, Arawakan, and Panoan origin, with a substantial number from Quechua. In Cabral’s (1995) analysis, Kokama began when a group of Tupinambá speakers migrated in the late fifteenth century from the Atlantic coast to the upper Amazon and came in contact with speakers of other languages. Cabral argues that the large number of lexical elements from Tupinambá, coupled with an almost complete absence of shared morphological and grammatical features, and a number of phonological changes untypical of Tupían languages, indicate that Tupinambá was learned imperfectly as a second language by other groups, meaning that Kokama is a mixed language (see also Braga et al. 2011:223; Cabral 1995, 2007; Vallejos Yopán 2010; Vallejos Yopán 2005). Michael (2014) argues that Kokama and Omagua were not the product of colonial‐era language contact, but were rather the outcome of language contact in

Contact and South American Languages  637 the pre‐Columbian period. He concludes that Proto‐Omagua‐Kokama, the parent language from which Omagua and Kokama derive, was a pre‐Columbian contact language.

4.3  Indigenous Pidgins and Creoles in South America 4.3.1  Ndjuka‐Trio Pidgin (Carib Pidgin, Ndjuka‐Amerindian Pidgin) This pidgin was used by speakers of Wayana and Trio (Cariban) of southeastern Suriname in their dealings with the Ndyuka‐speaking Maroon tribes of Suriname and French Guiana, formed by escaped slaves (De Goeje 1906, 1908, 1946; Huttar 1982; Huttar and Velantie 1997). Nimuendaju (1926: 112–13, 124, 140–3) reported it also in Brazilian territory, used there by the Palikur speakers. Its syntax is essentially Trio, not Ndyuka; its lexicon is from both Ndyuka (Sranan) and Trio, with mainly Ndyuka pronouns, verbs, and adverbs, and with nouns being equally from the two sources, Trio and Ndyuka (Huttar and Velantie 1997: 117).

4.3.2  Carib Pidgin‐Arawak Mixed Language Taylor and Hoff (1980) argued that a mixed language involving Carib Pidgin and Arawak is the ancestor of the Island Carib men’s language, basically an Arawakan language with a special men’s jargon based on Cariban lexical items. Most of what is known about this language is from Raymond Breton’s (1665, 1667) works on the Island Carib of Dominica and St. Vincent (where the language is now extinct), forms he designated as “language of men” and “language of women.” “A few remnants of the male register” are also preserved in Garífuna of Central America, whose speakers are descendants of Island Caribs who were deported from St. Vincent in 1797, though most to the men’s language is now lost (Hoff 1994: 161). The Cariban elements of the male jargon in Garífuna are now limited to lexical items and one postposition, while the grammatical morphemes of both male and female styles are all of Arawakan origin. Hoff (1994) also holds that the Cariban elements in the men’s jargon are from the Carib Pidgin and not directly from the Mainland Carib language itself. This is based on the observation that, in both the Carib Pidgin of Cayenne (French Guiana) and the Island Carib men’s jargon, the markers of transitive and intransitive verbal subclasses are derived historically from frozen personal pronominal prefixes. He speculates that in the wars against the Arawakan inhabitants of the Lesser Antilles, able‐bodied native men were killed off and Cariban men took their place, resulting in a mixed society consisting of Iñeri [Arawakan]‐ speaking women and children and Mainland Carib‐speaking men, who used Pidgin Carib to bridge the language gap between them. Children born of these unions learned their mothers’ Arawakan language; the men’s Carib failed to be imposed as a community language, but Pidgin Carib continued to be used because these people continued to identify themselves ethnically as Caribs and maintained political and trading relations with Mainland Caribs. The Pidgin Carib that was retained for these functions became the men’s jargon.

4.3.3  The Akuntsú‐Kanoe “pidgin” The so called “Akuntsú‐Kanoe” pidgin (Crevels and van der Voort 2008: 156) is a contact language that arose among the surviving speakers of Akuntsú (Tupían) and the Kanoe (isolate) after the Akuntsú were massacred with only six survivors. The two groups were placed in the same indigenous territory on the Omeré River (Aragon 2014: 1–11); the Akuntsú seek communication more often with the Kanoe and for that purpose they use the grammar of their native language with a mixture of Kanoe and Akuntsú vocabulary (Aragon, personal communication 2018).

638  Lyle Campbell, Thiago Chacon, and John Elliott

5  Socio‐cultural‐political Factors in SA Language Contact Contact‐induced linguistic changes are the consequences of interaction among people. It is sometimes said that there are differences in the sort of language contact in some regions of SA from that characteristic of elsewhere in the world, for example, less lexical borrowing, less code‐switching, and different patterns of language shift (long‐term language maintenance), often involving extensive multilingualism (cf. Epps and Michael 2017: 936–7). This may be relatively true, though only to a degree. Consequently, the differences in language contact in regions of SA from cases known from elsewhere need more careful investigation. In particular, the role of trade and subsistence patterns in language contact in SA needs more attention (see Epps 2017; Epps and Michael 2017: 950). For example, it is repeatedly claimed that there is a marked absence of code‐switching, in particular in the upper Amazon region (cf. Aikhenvald 2002: 95; Epps and Michael 2017: 937, 954; Epps and Stenzel 2013: 36). However, some cases have been reported recently of contemporary speakers of multiple Amazonian languages practicing code‐switching between indigenous languages. For example, Silva (2020) reported code‐switching among speakers of several languages in a Desano‐Siriano community in Colombia; for another case, see Paula (2001). However, neither restricted lexical borrowing nor limited code‐switching seems limited to only particular regions of SA. Preferences for neologisms formed from native language resources over lexical borrowing have long been reported for numerous North American languages (cf. Boas 1911: 50), and it is claimed that languages that have complex verbal morphology (as many SA languages do, and as do others all over the world) tend to have limits on the adoption of loanwords (Bakker and van der Voort 2017: 421). Consequently, potential differences between language contact in regions of SA and those known elsewhere in the world need investigation, but need to be approached with caution. There are, of course, numerous SA cases of languages with complex verb morphology that have also undergone considerable lexical borrowing, for example, Amuesha, Aymara, Cholón, Kamsá, and Quechua, among others (cf. Section 2; Muysken 2012). Nevertheless, work on language contact in SA has shown how social, political, and economic factors affect the diffusion of linguistic traits among languages. Here we focus on one factor, linguistic exogamy.

5.1  Marriage Patterns, Multilingualism, Linguistic Exogamy Some of the most intensive forms of language contact in SA occur in communities with a tradition of linguistic exogamy, where marriage is between speakers of different languages. The extensive multilingualism and linguistic exogamous marriage patterns of the Vaupés region are famous, not matched elsewhere in the world. Another unusual case is that of Misión La Paz, Argentina (MLP); a comparison of the two cases is instructive. In MLP three indigenous languages, Chorote, Nivaclé, and Wichí (members of the Matacoan family, not closely related), are spoken, as well as Spanish. Most interactions in MLP are multilingual – each participant in a conversation typically speaks his or her own language, and the other participants in the conversation each speak their own particular language in return. People communicate regularly with speakers of different languages, but commonly not in the same language as the one addressed to them. This sort of non‐reciprocal language use where each speaker speaks his or her own language but understands the other’s language in return has been called “dual‐lingualism” (Lincoln 1979). MLP also practices linguistic exogamy. Each spouse speaks his/her own language and is addressed in/understands the other spouse’s language in return – a spouse does not accommodate by speaking

Contact and South American Languages  639 the other spouse’s language; each maintains and uses his or her own. People “identify” with a single language and speak it with all others. (Campbell and Grondona 2010.) The situations in the Vaupés and Misión La Paz are similar in that spouses marry someone who speaks a different language, and they avoid borrowings from other languages. Both involve small populations with low population density and the languages are considered of equal status (Jackson 1983: 164, 175). However, the two differ in fundamental ways. In the Vaupés area, one identifies with and is loyal to the father’s language. Wives were expected to learn and use the language of the husband’s longhouse, and people learn and use other languages eagerly. In MLP the language a person identifies with can be either the father’s or the mother’s and it is a personal choice. People typically use only their own language, not shifting to the language others speak, resulting in multiple languages used in single conversations. Men exchange women in the Vaupés; in MLP, women select the men they marry. Residence is patrilocal in the Vaupés; Vaupés patrilocality with residential exogamy leads somewhat automatically to linguistic exogamy. In MLP, the man usually lives with the wife’s family initially until the first child is born, and then the couple typically goes to live with the husband’s family. Women are community outsiders in the Vaupés (who come to the longhouse because of patrilocal residence rules), not the case in MLP. Linguistic exogamy in the Vaupés is a major principle of social organization, residence patterns, kinship, and regional interactions; failure to conform is considered incest. In MLP there is linguistic exogamy, but the practice is not talked about as a necessary rule. The dual‐lingualism sometimes practiced in the Vaupés is sporadic and short‐term; in MLP, dual‐lingualism is the predominant kind of verbal interaction and is institutionalized. Interestingly, these examples of intensive long‐term contact and multilingualism do not conform to some assumptions about convergence in language contact. Linguists agree with Leonard Bloomfield’s (1933: 476) declaration, “when two speech communities are in continuous communication, linguistic convergence is expected, and any degree of divergence requires an explanation.” The expectation is that in intensive language contact, languages become more similar to one another and should not undergo changes that make them less similar. For example, in famous cases from India, different languages in extensive long‐term contact became so structurally similar that rather exactly matching morpheme‐by‐morpheme translations became possible (Gumperz and Wilson 1971; Nadkarni 1975). However, ­examples of linguistic exogamy in SA give us examples of the opposite phenomenon: languages tenaciously maintaining distinctions or actively diverging in situations of ­ intensive contact. For example, contrary to expectations, the three indigenous languages of MLP, though in intensive daily contact with one another, are diverging structurally in several traits, not converging as expected. They have undergone several structural changes in both morphosyntax and phonology that make the languages more different from one another (see Campbell and Grondona 2010; Campbell 2017b; Campbell et al. 2019). It would be expected that the intensive multilingual interactions of MLP would influence these languages not to change but to keep traits that the other languages spoken there have. On the other hand, in the highly multilingual Vaupés region, languages have converged in semantic and structural features which operate below the level of speakers’ awareness, although speakers may actively avoid code‐switching and lexical borrowing, borrowing of grammatical structures occurs (Epps 2012: 224). We see in these two contact situations a range of results. Although both the Vaupés and MLP have intensive language contact as a result of linguistic exogamy, the linguistic outcomes of contact differ in the two. It is mistaken to insist that languages in intensive contact must necessarily change only in the direction of more convergence, as we see in the MLP instances of structural divergence and in the explicit avoidance of lexical borrowing in both the Vaupés and MLP (and other Chaco languages).

640  Lyle Campbell, Thiago Chacon, and John Elliott

6 Language Endangerment, Language Shift, and Extinction There are 322 indigenous languages still spoken in SA, all endangered to one degree or another, according to the Catalogue of Endangered Languages (www.endangeredlanguages. com). It is uncertain how many languages there were in South America at the time of first European contact, though the number was certainly large. Loukotka’s (1968) 1,492 “languages” is the number often cited, although many of these “language” names are known to be errors, alternate names for groups known by other names or merely names of rivers, clans, etc. Whatever the actual figure, it reflects the catastrophic loss of languages in SA. Many other SA languages will soon become extinct. For example, currently 38 severely endangered SA languages have ten or fewer speakers (www.endangeredlanguages.com). The causes of language endangerment and in many cases of extinction in SA vary from region to region, and have varied over time, although some recurring factors are clear. European diseases decimated many indigenous populations. Estimates of pre‐European contact population sizes vary widely; a figure of 90% decline in the indigenous populations soon after 1520 is often repeated (cf. Denevan 1992). Economic greed is a major cause of language endangerment in SA. The rubber boom (1879–1912) had devastating consequences across the Amazon basin, where Indians were pressed into forced labor and large numbers died. Resource exploitation has a long record of murder and massacres of indigenous groups by loggers, miners, ranchers, and exploiters of oil, gas, and other mineral resources, resulting in displacement and genocide of whole groups. Drug trafficking and armed conflicts have impacted several groups, leading also to population displacement in several regions. The causes also include other political, social, and economic factors. Colonial policies fostered shift to Spanish or Portuguese. Bilingualism in more influential indigenous languages – such as Quechua, Tukano, and Nheengatú – fostered shift to these larger languages. The issues of discrimination, repression, stigmatization, negative status of the indigenous languages, official language policies, and lack of institutional support drive language endangerment. Intermarriage with Europeans has contributed to loss of indigenous languages and assimilation of these populations. In some instances, intermarriage among indigenous groups has also contributed to endangerment of their languages.

7 Conclusions In this survey, we have seen many different consequences of the various kinds of language contact among SA languages, as might well be expected in such a large politically, culturally, and linguistically diverse region. We saw numerous cases where lexical borrowing is vigorously resisted and yet we also encounter some Wanderwörter, borrowing chains, and other cases of borrowing, which call out for explanation. Many languages with restrictions against lexical borrowing have nevertheless undergone considerable structural borrowing – showing that the claim that structural borrowing is only possible if connected with lexical borrowing must be reconsidered. Some of the recently postulated linguistic areas raise questions about how linguistic areas are defined in general, given the number of traits in these areas that also overlap with other proposed linguistic areas. Of particular interest are the two major contact zones in SA for which Birchall (2014) has argued, WSALA and ESALA, and the relationships between the two. In spite of the belief that conditions in SA make the presence of lingua francas unlikely, several have been documented and described. While there are few mixed languages in the world, SA has several (Media Lengua, Callahuaya, and Kokama), providing rich grounds for

Contact and South American Languages  641 investigating general properties of mixed languages. While most pidgins and creoles in the world involve European colonial languages in their formation, SA has several pidgins or creoles that are not based on European languages. We have seen that claims of the absence of code‐switching are overstated, although we still need an explanation for its scarceness in SA. Cases of linguistic exogamy are well‐known in SA but mostly not found elsewhere in the world. As we learn more of the social and cultural patterns of different regions of SA, and as we discover more about the important role personal and group identity plays in determining kinds of language contact and their outcomes in SA, we will advance understanding of language contact in general.

NOTES 1 The abbreviations used in this chapter are: ACC accusative; BEN benefactive; LA linguistic area; MLP Misión La Paz (Salta Province, Argentina); NMLZ nominalization; PL plural; PROG progressive; SA South America; SG singular; Sp. Spanish. 2 www.funai.gov.br/index.php/nossas‐acoes/povos‐indigenas‐isolados‐e‐de‐recente‐contato? limitstart=0#. 3 Epps and Michael (2017: 940) see a connection between the presence of the Wanderwörter and the “remarkably low levels of lexical borrowing despite the intense, long‐term contact among them.” The combination of low lexical borrowing and substantial grammatical diffusion is, they say, “undoubtedly linked to speakers’ conscious efforts to avoid language mixing … which fosters the convergence of grammatical structures and categories below speakers’ ‘limits of awareness.’” They sense “the impression that if a lexical item is to be borrowed, it is likely to travel widely, possibly because it loses an association with a particular language in the process and thus becomes ‘fair game’” (cf. Muysken 2012: 252). 4 It should be noticed that there have been inconclusive proposals of distant genetic relationship among three major linguistic families grouped within a purported eastern South America LA, namely: Tupían, Cariban and Macro‐Jêan (Rodrigues 1985, 2000). 5 A larger Orinoco‐Amazon LA that would include parts of the Orinoco and northern Amazon watersheds was proposed by Migliazza (1985[1982]: 20). He viewed the Northern Amazon Culture Area as also constituting a linguistic area, with the languages: Yanomaman family; Salivan: Piaroa; Arawakan: Baniwa (Karutiana‐Baniwa), Wapixana, Bare, Mandahuaca, Warekena, Baniva (Yavitero); Cariban: Panare, Yabarana (Mapoyo‐Yavarana), Mapoyo, Yekuana, Pemon, Kapong, Makuxi, Waiwai, Waimiri, Hixkaryana, Warikyana; language isolates: Joti; Arutani (aka Uruak, Awake); Sapé (Kaliana); and Maku. Some traits that Migliazza noted as shared are: a pattern of discourse redundancy, ergative alignment (except in a few Arawakan languages), OV (SOV or OVS) order (except in a few Arawakan languages), lack of active‐passive distinction, and relative clauses formed by apposition and nominalization. Diffusion from west to east of nasalization, aspiration, and glottalization has also been suggested (Migliazza 1985[1982]: 20, 118). As Epps and Michael (2017: 951) point out, Amazonianists today would now see several of these features as having a broader Amazonian distribution; they note that the features that Derbyshire (1987: 311) tentatively proposed as defining an Amazonian linguistic area include most of these. That being the case, it seems best to abandon the idea of an Orinoco‐Amazon LA.

REFERENCES Adelaar, Willem F. H. 2006. The Quechua impact in Amuesha, an Arawak language of the Peruvian Amazon. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Grammars in

Contact: A Cross‐Linguistic Typology. Oxford: Oxford University Press, pp. 290–312. Adelaar, Willem F. H. 2008. Towards a typological profile of the Andean languages.

642  Lyle Campbell, Thiago Chacon, and John Elliott Studies in Slavic and General Linguistics 33: 23–33. Adelaar, Willem F. H. 2012. Languages of the middle‐Andes in areal‐typlogical perspective. In Lyle Campbell and Verónica Grondona (eds.), The Indigenous Languages of South America: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 575–624. Adelaar, Willem F. H. and Pieter C. Muysken 2004. The Languages of the Andes. Cambridge: Cambridge University Press. Aikhenvald, Alexandra Y. 1999 The Arawak languge family. In Robert M. W. Dixon and Alexandra Y. Aikhenvald (eds.), The Amazonian Languages. Cambridge: Cambridge University Press, pp. 65–106. Aikhenvald, Alexandra Y. 2000. Classifiers: A Typology of Noun Categorization Devices. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. 2001. Areal diffusion, genetic inheritance, and problems of subgrouping: a North Arawak case study. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press, pp. 167–94. Aikhenvald, Alexandra Y. 2002. Language Contact in Amazonia. Oxford: Oxford University Press. Aikhenvald, Alexandra. 2012. The Languages of the Amazon. Oxford: Oxford University Press. Aragon, Carolina 2014. A grammar of Akuntsú, a Tupian language (University of Hawai‘i at Mā noa Ph.D. dissertation). Bakker, Peter and Hein van der Voort 2017. Polysynthesis and language contact. In Michael Fortescue, Marianne Mithun, and Nicholas Evans (eds.), The Oxford Handbook of Polysynthesis. Oxford: Oxford University Press, pp. 408–27. Ball, Christopher 2011. Pragmatic multilingualism in the Upper Xingu speech community. In Bruna Franchetto (ed.), Alto Xingu: uma sociedade multilíngue [Upper Xingu: a multilingual society]. Rio de Janeiro: Museo do Índio – FUNAI, pp. 87–113. Bessa Freire, José 1983. Da “fala boa” ao português na Amazônia brasileira [From the “good speech” to Portuguese in the Brazilian Amazon]. Amerindia 8: 39–83. Birchall, Joshua 2014. Argument marking patterns in South American languages (Radboud University Ph.D. dissertation). Bloomfield, Leonard 1933. Language. New York: Holt, Rinehart and Winston. Boas, Franz 1911. Introduction. Handbook of American Indian Languages (Bureau of

American Ethnology, Bulletin 40, part 1.5–83). Washington, DC: Government Printing Office. Braga, Alzerinda, Ana Suelly Arruda Câmara Cabral, Aryon Dall’Igna Rodrigues, and Betty Mindlin 2011. Línguas entrelaçadas: uma situação sui generis de línguas em contato [Intertwined languages: a sui generis situation of languages contact]. PAPIA – Revista Brasileira de Estudos do Contato Linguístico [PAPIA – Brazilian Journal of Language Contact Studies] 21: 221–30. Breton, Raymond. 1665. Dictionnaire caraïbe‐ françois. Auxerre: Gilles Bouquet. (Facsimile edition 1892, ed. by Jules Platzmann. Leipzig: Teubner.) Breton, Raymond. 1667. Grammaire caraïbe. (New edition 1877, Grammaire caraibe suivie du catechisme caraibe, ed. by Lucien Adam and Claude‐André LeClerc. Paris: Maisonneuve.) [Reprinted 1968, Nendeln, Liechtenstein: Kraus Reprint.]) Büttner, Thomas T. 1983. Las lenguas de los Andes centrales: estudios sobre la clasificación genética, areal y tipológica [The languages of the Central Andes: studies on the genetic, areal, and typological classification]. Madrid: Ediciones Cultura Hispánica del Instituto de Co‐operación Iberoamericana, Magerit. Cabral, Ana Suelly Arruda Câmara 1995. Contact‐induced language change in the Western Amazon: the non‐genetic origin of the Kokama language (University of Pittsburgh Ph.D. dissertation). Cabral, Ana Suelly Arruda Câmara 2007. New observations on the structure of Kokáma/ Omágwa. In W. Leo Wetzels (ed.), Language Endangerment and Endangered Languages: Linguistics and Anthropological Studies with Special Emphasis on the Languages and Cultures of the Andean‐Amazonian Border Area. Leiden: CNWS Publications, pp. 365–79. Cabral, Ana Suelly Arruda Câmara, Ana B. Corrêa da Silva, Marina M. S. Magalhães, and Maria R. S. Julião 2007. Linguistic diffusion in the Tocantins‐Mearim area. Línguas e Culturas Tupí [Tupian languages and cultures] 1: 357–74. Campbell, Lyle 1997. American Indian Languages: The Historical Linguistics of Native America. Oxford: Oxford University Press. Campbell, Lyle 2006. Areal linguistics: a closer scrutiny. In Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas: Convergence in Historical and Typological Perspective by. Basingstoke, Hampshire: Palgrave Macmillan, pp. 1–31.

Contact and South American Languages  643 Campbell, Lyle 2012a. The classification of South American indigenous languages. In Lyle Campbell and Verónica Grondona (eds.), The Indigenous Languages of South America: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 59–166. Campbell, Lyle 2012b. Typological characteristics of South American indigenous languages. In Lyle Campbell and Verónica Grondona (eds.), The Indigenous Languages of South America: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 259–330. Campbell, Lyle 2013. Language contact and linguistic change in the Chaco. In Ana Suelly Arruda Câmara Cabral and Jorge Domingues Lopes (eds.), Encontro internacional arqueologia e linguística histórica das línguas indígenas sulamericanas [International meeting for archeology and historical linguistics of the South American indigenous languages]. Special issue of Revista Brasileira de Linguística Antropológica [Brazilian journal of linguistic anthropology] 5: 259–591. Campbell, Lyle 2017a. Why is it so hard to define a linguistic area? In Raymond Hickey (ed.), Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 19–39. Campbell, Lyle 2017b. Language contact and language documentation: whence and whither? In Hannah Foster, Michael Everdell, Katie Bradford, Lorena Orjuela, Frances Cooley, Hammal Al Bulushi, and Ambrocio Gutiérrez Lorenzo Proceedings of SALSA 25, Texas Linguistics Form 60. Available at http:// salsa.ling.utexas.edu/proceedings/2017/ Campbell.pdf Campbell, Lyle 2018. How many language families are there in the world? Studia Diachronica et Philologica in Honorem Joakin Gorrotxategi: Vasconica et Aquitanica [Diachronic and philological studies in honor of Joaquín Gorrochategi: Basque and Acquitanian). Special issue of ASJU (Anuario del Seminario de Filología Vasca “Julio de Urquijo” [Yearbook of the “Julio de Urquijo” seminar of Basque philology] 52. Gasteiz: UPV/EHU (Universidad del País Vasco), pp. 133–52. Campbell, Lyle, Luis Díaz, and Fernando Ángel 2019. Gramática nivaclé. Salt Lake City: University of Utah Press. Campbell, Lyle and Verónica Grondona 2007. Internal reconstruction in Chulupí (Nivaclé). Diachronica 24: 1–29. Campbell, Lyle and Verónica Grondona 2010. Who speaks what to whom: Multilingualism

and language choice in Misión La Paz – a unique case. Language in Society 39: 1–30. Campbell, Lyle and Verónica Grondona 2012a. Linguistic acculturation in Nivaclé (Chulupí) and Chorote. International Journal of American Linguistics 78: 335–67. Campbell, Lyle and Verónica Grondona 2012b. Languages of the Chaco and Southern Cone. In Lyle Campbell and Verónica Grondona (eds.), The Indigenous Languages of South America: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 625–68. Caraman, Philip 1975. The Lost Paradise: The Jesuit Republic in South America. New York: Seabury Press. Carlin, Eithne 2007. Feeling the need: the borrowing of Cariban functional categories into Mawayana (Arawak). In Alexandra Aikhenvald and Robert M. W. Dixon (eds.), Grammars in Contact: A Cross‐Linguistic Perspective. Oxford: Oxford University Press, pp. 313–32. Carlin, Eithne 2011. Nested identities in the southern Guyana‐Suriname corner. In Alf Hornborg and Jonathan D. Hill (eds.), Ethnicity in Ancient Amazonia. Boulder, CO: University of Colorado Press, pp. 225–36. Carvalho, Fernando O. de 2018. Arawakan‐ Guaicuruan language contact in the South American Chaco. International Journal of American Linguistics 84: 243–63. Chacon, Thiago 2013. Kubeo: Linguistic and cultural interactions in the Upper Rio Negro. In Patience Epps and Kristine Stenzel (eds.), Upper Rio Negro: Cultural and Linguistic Interaction in Northwestern Amazonia. Rio de Janeiro: Museu do Índio‐FUNAI, pp. 403–40. Chacon, Thiago 2014. A revised proposal of Proto‐Tukanoan consonants and Tukanoan family classification. International Journal of American Linguistics 80: 275–322. Chacon, Thiago 2017. Arawakan and Tukanoan contacts in Northwest Amazonia prehistory. PAPIA: Revista Brasileira de Estudos Crioulos e Similares [Brazilian journal of creole and similar studies] 27: 237–65. Chacon, Thiago 2020. Migration, trade and language contact in Latin America. In Salikoko Mufwene and Ana María Escobar (eds.), The Cambridge Handbook of Language Contact. Cambridge: Cambridge University Press. Chang, Will and Lev Michael 2014. A relaxed admixture model of language contact. Language Dynamics and Change 4: 1–26. Constenla Umaña, Adolfo 1991. Las lenguas del área intermedia: introducción a su estudio areal.

644  Lyle Campbell, Thiago Chacon, and John Elliott San José: Editorial de la Universidad de Costa Rica. Crevels, Mily and Hein van der Voort. 2008. The Guaporé‐Mamoré region as a linguistic area. In Pieter Muysken (ed.), From Linguistic Areas to Areal Linguistics. Amsterdam: John Benjamins, pp. 151–79. de Goeje, Claudius Henricus 1906. Handelsdialekt, tussen Aucaners en Indianen gebruikelijk (Handelsdialekt, gebrauchlich zwischen Aucanern und Indianern) [Trade dialect, common between Aucaners and Indians (Handelsdialekt, gebrauchlich zwischen Aucanern und Indianern [Trade dialect, common between Aucaners and Indians])]. Bijdrage tot de ethnographie der surinaamsche Indianen [Contribution to the ethnography of the Surinamese Indians]. Supplement to Internationales Archiv fur Ethnographie [International Archive for Ethnography] 17: 109–11. de Goeje, Claudius Henricus 1908. De “handelstaal” der Joeka’s met de Trio’s en de Ojana’s. Appendix to Verslag van Toemoekhoemak‐ expeditie [The “trade language” of the Yukas with the Trios and the Ojanas. Appendix to Report of the Toemoekhoemak expedition]. Leiden: Brill, pp. 204–19. de Goeje, Claudius Henricus 1946. Études linguistiques Caraïbes, vol. 2. Verhandelingen der Koninklijke Nederlandse Akademie van Wetenschappen te Amsterdam, Afdeeling Letterkunde [Transactions of the Royal Netherlands Academy of Arts and Sciences in Amsterdam, Department of Literature] n.s. 40.2: 101–2. Denevan, William M. 1992. The Native Population of the Americas. Madison, WI: University of Wisconsin Press. Derbyshire, Desmond C. 1986. Comparative survey of morphology and syntax in Brazilian Arawakan. In Desmond C. Derbyshire and Geoffrey K. Pullum (eds.), Handbook of Amazonian Languages, vol. 1. Berlin: de Gruyter Mouton, pp. 469–566. Derbyshire, Desmond C. 1987. Morphosyntactic areal characteristics of Amazonian languages. International Journal of American Linguistics 53: 311–26. Derbyshire, Desmond C. and Doris L, Payne 1990. Noun classification systems of Amazonian languages. In Doris L. Payne (ed.), Amazonian Linguistics: Studies in Lowland South American Languages. Austin: University of Texas Press, pp. 243–71.

Derbyshire, Desmond and Geoffrey Pullum 1986. Introduction. In Desmond Derbyshire and Geoffrey Pullum (eds.), Handbook of Amazonian Languages, vol. 1. Berlin: de Gruyter Mouton, pp. 1–28. Derbyshire, Desmond and Geoffrey Pullum 1991. Introduction. In Desmond C. Derbyshire and Geoffrey K. Pullum (eds.), Handbook of Amazonian Languages, vol. 3. Berlin: de Gruyter Mouton, pp. 3–18. Dixon, Robert M. W. and Alexandra Y. Aikhenvald 1999. Introduction. In Robert M. W. Dixon and Alexandra Y. Aikhenvald (eds.), The Amazonian. Cambridge: Cambridge University Press, pp. 1–21. Epps, Patience 2005. Areal diffusion and the development of evidentiality: evidence from Hup. Language 29: 617–50. Epps, Patience 2007. The Vaupés melting pot: Tukanoan influence on Hup. In Alexandra Aikhenvald and Robert M. W. Dixon (eds.), Grammars in Contact: A Cross‐Linguistic Typology. Oxford: Oxford University Press, pp. 267–89. Epps, Patience 2008. Grammatical borrowing in Hup. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing: A Cross‐linguistic Survey. Berlin: de Gruyter Mouton, pp. 551–65. Epps, Patience 2009. Loanwords in Hup, a Nadahup language of Amazonia. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter Mouton, pp. 992–1014. Epps, Patience 2012. On form and function in language contact: a case study from the Amazonian Vaupés region. In Isabelle Léglise and Claudine Chamoreau (eds.), Dynamics of Contact‐Induced Language Change. Berlin: de Gruyter Mouton, pp. 195–230. Epps, Patience 2017. Subsistence pattern and contact‐driven language change. Language Dynamics and Change 7: 47–101. Epps, Patience and Lev Michael 2017. The areal linguistics of Amazonia. In Raymond Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 934–63. Epps, Patience and Kristine Stenzel 2013. Introduction. In Patience Epps and Kristine Stenzel (eds.), Upper Rio Negro: Cultural and Linguistic Interaction in Northwest Amazonia. Rio de Janeiro: Museu Nacional / Museu do Índio‐FUNAI, pp. 13–52. Estenssoro, Juan Carlos 2015. Las vías indígenas de la occidentalización: lenguas

Contact and South American Languages  645 generales y lenguas maternas en el ámbito colonial americano (1492–1650) [The indigenous ways for Westernization: general languages and mother tongues in the American colonial context (1492–1650)]. Mélanges de la Casa de Velázquez 45: 15–36. Floyd, Simeon 2013. Semantic transparency and cultural calquing in the Northwest Amazon. In Patience Epps and Kristine Stenzel (eds.), Upper Rio Negro: Cultural and Linguistic Interaction in Northwestern Amazonia. Rio de Janiero: Museu do Indio, pp. 271–308. Golluscio, Lucía 2009. Loanwords in Mapudungun, a language of Chile and Argentina. In Martin Haspelmath and Uri Tadmoor (eds.), Loanword in the World Languages: A Comparative Handbook. Berlin: de Gruyter Mouton, pp. 1035–71. Gómez Rendón, Jorge A. 2005. La Media Lengua de Imbabura [The Media Lengua of Imbabura]. In Hella Olbertz and Pieter Muysken (eds.), Encuentros y conflictos: bilingüísmo y contacto de lenguas en el mundo andino [Encounters and conflicts: bilingualism and language contact in the Andean world]. Frankfurt‐am‐Main: Vervuert, pp. 39–58. Gómez Rendón, Jorge A. 2008. Mestizaje lingüístico en los Andes: génesis y estructura de una lengua mixta [Linguistic mixing in the Andes: genesis and structure of a mixed language]. Quito: Abya‐Yala. Gómez Rendón, Jorge A. and Willem Adelaar 2009. Loanwords in Imbabura Quichua. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter Mouton, pp. 944–67. Gómez‐Imbert, Elsa 1993. Problemas en torno a la comparación de las lenguas Tukano orientales [Problems concerning the comparison of the Eastern Tukano languages]. In María Luisa Rodríguez de Montes (ed.), Estado actual de la clasificación de las lenguas indígenas de Colombia [Current status of the classification of the indigenous languages of Colombia. Bogotá: Instituto Caro y Cuervo, pp. 235–67. Gómez‐Imbert, Elsa 1996. When animals become “rounded” and “feminine”: conceptual categories in a multilingual setting. In John J. Gumperz and Stephen C. Levinson (eds.), Rethinking Linguistic Relativity. Cambridge: Cambridge University Press, pp. 438–69.

Gumperz, John J. and Robert Wilson 1971. Convergence and creolization: a case from the Indo‐Aryan/Dravidian border in India. In Dell Hymes (ed.), Pidginization and Creolization of Languages. Cambridge: Cambridge University Press, pp. 151–67. Haspelmath, Martin and Uri Tadmor (eds.) 2009. Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter Mouton. Haynie, Hannah, Claire Bowern, Patience Epps, Jane Hill, and Patrick McConvell 2014. Wanderwörter in languages of the Americas and Australia. Ampersand 1: 1–18. Hoff, Berend J. 1994. Island Carib, an Arawakan language which incorporated a lexical register of Cariban, used to address men. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: Fifteen Case Studies in Language Intertwining. Amsterdam: Institute for Functional Research into Language and Language Use, pp. 161–8. Huttar, George L. 1982. A Creole‐Amerindian Pidgin of Suriname. (Society for Caribbean Linguistics Occasional Paper 15.) St. Augustine, Trinidad: School of Education, University of the West Indies. Huttar, George L. and Frank J. Velantie 1997. Ndyuka‐Trio Pidgin. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 99–124. Itier, César 2011. What was the lengua general of Colonial Peru? In Paul Heggarty and Adrian J. Pearce (eds.), History and Language in the Andes. New York: Palgrave Macmillan, pp. 63–85. Jackson, Jean 1983. The Fish People: Linguistic Exogamy and Tukanoan Identity in Northwest Amazonia. Cambridge: Cambridge University Press. Klein, Harriet E. Manelis 1992. South American languages. In William Bright (ed.), International Encyclopedia of Linguistics, vol. 4. New York: Oxford University Press, pp. 31–5. Klein, Harriet E. Manelis 1993. Aculturación lingüística de los toba. [Linguistic acculturation of the Toba] Primeras Jornadas de Lingüística Aborigen 1: 185–200. Labov, William 2007. Transmission and diffusion. Language 83: 344–87. Lafone Quevedo, Samuel A. 1896. Grupo mataco‐mataguayo del Chaco: dialecto Vejoz [Mataco‐Mataguayo group of the Chaco, Vejoz dialect]. Boletín del Instituto Geográfico Argentino [Bulletin of the Argentine Geographical Institute] 17: 121–76.

646  Lyle Campbell, Thiago Chacon, and John Elliott Lafone Quevedo, Samuel A. 1915. Lenguas guaicuras y mataguayas [Guaicuran and Mataguayan (Matacoan) languages]. Boletín del Instituto Geográfico Argentino [Bulletin of the Argentine Geographical Institute] 31: 11–25. Lincoln, Peter C. 1979. Dual‐lingualism: passive bilingualism in action. Te Reo 22: 65–72. Lipski, John M. 2017. Ecuadoran Media Lengua: more than a “half” language? International Journal of American Linguistics 83: 233–62. Loukotka, Č estmír 1968. Classification of South American Indian Languages. Los Angeles: University of California, Latin American Center. Métraux, Alfred 1946. Ethnography of the Chaco. In Julian H. Steward (ed.), Handbook of South American Indians, vol. 1: The Marginal Tribes (Smithsonian Institution Bureau of American Ethnology, Bulletin 143). Washington, DC: Government Printing Office, pp. 197–370. Michael, Lev 2014. On the pre‐columbian origin of Proto‐Omagua‐Kokama. Journal of Language Contact 7: 309–44. Migliazza, Ernest C. 1982. Languages of the Orinoco‐Amazon region: current status. Antropologica 53: 95–162. (Reprinted 1985 in Harriet E. Manelis Klein and Louisa R. Stark (eds.), South American Indian Languages: Retrospect and Prospect. Austin: University of Texas Press, pp. 17–139.) Moore, Denny, Sidney Facundes, and Nadia Pires 1994. Nheengatu (língua geral amazonica): its history and the effects of language contact. In Margaret Langdon (ed.), Proceedings of the Meeting of the Society for the Study of the Indigenous Languages of the Americas, July 2–4, 1993, and the Hokan‐Penutian Workshop, July 3, 1993 (SCOIL Report no. 8.) Berkeley: Department of Linguistics, University of California, pp. 93–118. Murdock, George Peter 1951. South American culture areas. Southwestern Journal of Anthropology 7: 415–36. Muysken, Pieter 1980. Sources for the study of Amerindian contact vernaculars in Ecuador. In Pieter Muysken and Norval Smith (eds.), Amsterdam Creole Studies 3, (Publikaties van het Instituut voor Algemene Taalwetenschap [Publications of the Institute for General Linguistics] 31). Amsterdam: Universiteit van Amsterdam, pp. 66–82. Muysken, Pieter 1981. Halfway between Quechua and Spanish: the case for relexification. In Arnold Highfield and Albert Valdman (eds.), Historicity and Variation in

Creole Studies. Ann Arbor: Karoma Publishers, pp. 52–78. Muysken, Pieter 1994a. Callahuaya. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: Fifteen Case Studies in Language Intertwining. Amsterdam: Institute for Functional Research into Language and Language Use, pp. 207–11. Muysken, Pieter 1994b. Media Lengua. In Peter Bakker and Maarten Mous (eds.), Mixed Languages: Fifteen Case Studies in Language Intertwining. Amsterdam: Institute for Functional Research into Language and Language Use, pp. 201–5. Muysken, Pieter 1997. Media Lengua. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: John Benjamins, pp. 365–426. Muysken, Pieter 2012. Contacts between indigenous languages in South America. In Lyle Campbell and Verónica Grondona (eds.), The Indigenous Languages of South America: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 235–58. Nadkarni, Mangesh V. 1975. Bilingualism and syntactic change in Konkani. Language 51: 672–83. Nimuendaju, Curt 1926. Die Palikur‐Indianer und ihre Nachbarn [The Palikur Indians and their neighbors] Göteborgs Kungliga Vetenskaps‐och Vitterhets‐samhälles Handlingar [Acts of the Gothenburg Royal Society of Science and Letters] 31.2. Gothenburg: Elanders. Nordenskiöld, Erland 1922. Deductions Suggested by the Geographical Distribution of some Post‐ Columbian Words Used by the Indians of South America. (Comparative Ethnographical Studies 5). Gothenburg: Elanders. Oblitas Poblete, Enrique 1968. El idioma secreto de los incas [The secret language of the Incas]. La Paz: Editorial “Los Amigos del Libro.” Paula, Luiz Gouvêa de 2001. Mudanças de código em eventos de fala na língua tapirapé durante interações entre crianças [Code‐switching in Tapirapé speech events in interactions among children] (Universidade Federal de Goiás master’s dissertation). Payne, David L. 1990. Some widespread grammatical forms in South American languages. In Doris L. Payne (ed.), Amazonian Linguistics: Studies in Lowland South American Languages. Austin: University of Texas Press, pp. 75–87. Payne, Doris 1990a. Introduction. In Doris Payne (ed.), Amazonian Linguistics: Studies in Lowland

Contact and South American Languages  647 South American Languages. Austin: University of Texas Press, pp. 1–10. Payne, Doris L. 1990b. Morphological characteristics of lowland South American languages. In Doris L. Payne (ed.), Amazonian Linguistics: Studies in Lowland South American Languages. Austin: University of Texas Press, pp. 243–71. Renault‐Lescure, Odile 2009. Loanwords in Kali’na, a Cariban language of French Guiana. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: Comparative Handbook. Berlin: de Gruyter de Gruyter Mouton, pp. 968–91. Rodrigues, Aryon 1986. Línguas brasileiras: para o conhecimento das línguas indígenas [Brazilian languages: for knowledge of the indigenous languages]. São Paulo: Edições Loyola. Rodrigues, Aryon Dall’Igna. 1985. Evidence for Tupi‐Carib relationships. In Harriet E. Manelis Klein and Louisa R. Stark (eds.), South American Indian Languages: Retrospect and Prospect,. Austin: University of Texas Press, pp. 371–404. Rodrigues, Aryon Dall’Igna. 2000 “Ge‐Pano‐ Carib” X “Jê‐Tupí‐Karib”; sobre relaciones lingüísticas prehistóricas en Sudamérica. In L. Miranda Esquerre (ed.), Actas I Congreso de Lenguas Indígenas de Sudamérica, Vol. 1. Lima: Universidad Ricardo Palma, Facultad de Lenguas Modernas, pp. 95–104. Rodrigues, Douglas A. 2014. Proteção e assistência à saúde de povos indígenas isolados e de recente contato no Brasil [Protection and health care of isolated and recently contacted indigenous peoples in Brazil]. Organización del Tratado de Cooperación Amazónica [Organization of the Amazon Cooperation Treaty]. São Paulo. Available at www.academia.edu/12865129/ Prote%C3%A7%C3%A3o_e_ Assist%C3%AAncia_%C3%A0_Sa%C3%BAde_ de_Povos_Ind%C3%ADgenas_Isolados_e_de_ Recente_Contato_no_Brasil. Rose, Françoise, Thiago Costa Chacon, Magdalena Lemus, and Natalia Eraso 2017. A new look into Arawak‐Tukanoan contact: the Yukuna‐ Tanimuka bidirectional hypothesis. Paper presented at the SSILA annual meeting, Austin, Texas. Rothhammer, Francisco and Tom D. Dillehay 2009. The late Pleistocene colonization of South America: an interdisciplinary perspective. Annals of Human Genetics 73: 540–9. Santana, Áurea Cavalcante and Ema Marta Dunck‐Cintra 2009. Estudos da língua chiquitano do Brasil: trajetória e perspectivas

[Studies of the Chiquitano language of Brazil: trajectory and perspectives]. Polifonia [Polyphony]15.17: 91–109. Seifart, Frank 2011. Bora loans in Resígaro: massive morphological and little lexical borrowing in a moribund Arawakan language. Cadernos de Etnolingüística [Ethnolinguistic notebooks] (Série Monografias [Monograph series] 2). Seifart, Frank 2012. The principle of morphosyntactic subsystem integrity in language contact: evidence from morphological borrowing in Resígaro (Arawakan). Diachronica 29: 471–504. Seki, Lucy 1999. The Upper Xingu as an incipient linguistic area. In Robert M. W. Dixon and Alexandra Aikhenvald (eds.), The Amazonian Languages. Cambridge: Cambridge University Press, pp. 417–30. Seki, Lucy 2011. Alto Xingu: uma sociedade multilíngue? [Upper Xingu: a multilingual society?] In Bruna Franchetto (ed.), Alto Xingu: uma sociedade multilíngue [Upper Xingu: a multilingual society]. Rio de Janeiro: Museu do Índio/Funai, pp. 57–86. Silva, Alcionílio Bruzzi Alves da 1962. A civilização indígena do uaupés [The indigenous civilization of the uaupés]. São Paulo: Centro de Pesquisas de Iauareté, Missão Salesiana do Rio Negro. Silva, Wilson 2020. Multilingual interactions and code‐mixing in northwest Amazonia. International Journal of American Linguistics, 86: 133–54. Stark, Louisa R. 1972. Machaj‐Juyai: secret language of the Callahuayas. Papers in Andean Linguistics 2: 199–227. Stenzel, Kristine 2013. Contact and innovation in Vaupés possession‐marking strategies. In Patience Epps and Kristine Stenzel (eds.), Upper Rio Negro: Cultural and Linguistic Interaction in Northwestern Amazonia. Rio de Janeiro: Museo do Índio, pp. 353–400. Stenzel, Kristine and Elsa Gómez‐Imbert 2009. Contato linguístico e mudança linguística no noroeste amazônico: o caso do Kotiria (Wanano) [Language contact and language change in the northwest Amazon: the case of Kotiria]. Revista da ABRALIN [Journal of the ABRALIN] 8: 1–100. Taylor, Douglas R. and Berend J. Hoff 1980. The linguistic repertory of the Island Carib in the seventeenth century: the men’s language – a Carib pidgin? International Journal of American Linguistics 46: 301–12. Taylor, Gerald 1985. Apontamentos sobre o nheengatu falado no Rio Negro, Brasil [Notes

648  Lyle Campbell, Thiago Chacon, and John Elliott on the Nheengatu spoken on Rio Negro, Brazil]. Amerindia 10: 5–23. Vallejos, Rosa 2005. Entre flexión y derivación: examinando algunos morfemas en Kokama‐ Cocamilla [Between inflection and derivation: examining some morphemes in Kokama‐ Cocamilla]. Memorias del Congreso de Idiomas Indígenas de Latinoamérica‐II, 27–29 October [Memoirs of the Congress of Indigenous Languages of Latin America‐II, October 27–29]. Austin: University of Texas at Austin. Vallejos Yopán, Rosa 2010. A grammar of Kokama‐ Kokamilla (University of Oregon Ph.D. dissertation). Van Gijn, Rik, Harald Hammarström, Simon van de Kerke, Olga Krasnoukhova, and Pieter Muysken 2017. Linguistic areas, linguistic convergence and river systems in South America. The Cambridge Handbook of Areal Linguistics, ed. by Raymond Hickey, 964–996. Cambridge: Cambridge University Press. Van der Voort, Hein 2005. Kwaza in a comparative perspective. International Journal of American Linguistics 71: 365–412.

Vidal, Alejandra and Verónica Nercesian 2009. Loanwords in Wichí, a Mataco‐Mataguayan language of Argentina. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: Comparative Handbook. Berlin: de Gruyter Mouton, pp. 1015–34. Viegas Barros, José Pedro 2001. Evidencias del parentesco de las lenguas Lule y Vilela [Evidence of the relatedness of the Lule and Vilela languages]. Santa Fe, Argentina: Gobierno de Santa Fe. Viegas Barros, José Pedro 2005. Voces en el viento: raíces lingüísticas de la Patagonia [Voices in the wind: linguistic roots of Patagonia]. Buenos Aires: Mondragón Ediciones. Wise, Mary Ruth 1976. Apuntes sobre la influencia inca entre los amuesha: factor que oscurece la clasificación de su idioma [Notes on the Inca influence among the Amuesha: factor that obscures the classification of their language]. Revista del Museo Nacional [Journal of the National Museum] 42: 355–66. Zamponi, Raoul 2003. Maipure. Munich: LINGCOM Europa.

33 Contact among African Languages KLAUS BEYER 1  Introduction The African continent is currently home to 2,1431 indigenous languages which account for approximately 30% of all languages worldwide. Most modern African states count at least several dozens of national languages in various degrees of vitality. In Nigeria, for instance, 519 languages are currently spoken and even very small (in geographical terms) states like, e.g. The Gambia (11) or Lesotho (5) still display considerable linguistic variety (Simons and Fenning 2018). Therefore, contact among African languages seems merely inevitable and is in fact ubiquitous. The continent can thus be considered a natural research area for language‐ contact questions and related fields. Notwithstanding this proneness to the contact issue, the subject only came to the full awareness of Africanists (linguists specializing in African languages) a few decades ago. This fact is related to developments of linguistics in general and to the special circumstances of Africa’s colonial past. In the late nineteenth and the beginning of the twentieth century, research on African languages was predominantly occupied with description and classification tasks to serve colonial demands (Pugach 2018: 18, 19) which were also guided by European concepts of nation states and concomitant questions of ethno‐linguistic power relations. In fact those who delimited language boundaries were European linguists, giving names to the entities created thus and describing them in terms of Latin‐based concepts of grammar. They thus defined languages and their speakers from a purely Eurocentric perspective. Moreover, in the racist and evolutionary thinking of this time morphological complexity of languages was considered to reflect the cultural development of their speakers. Following a scenario first outlined by Speke (1863), it was believed that sub‐Saharan African languages were originally only of the isolating type and that the development of more complex morphology was due to “superior” people migrating from the Near East to Africa. Especially Meinhof (1912) put forward the so‐called “Hamitic theory” stating that the intrusion of speakers of inflectional Semitic languages yielded the slightly “higher” developed agglutinating type of Bantu languages2 through the mixing with the isolating type, widely spread across western and southern Africa. The reigning linguistic paradigm of that time was diachronically oriented and divergence was the governing principle in classification and reconstruction techniques. Tendencies of language convergences, albeit observed and commented on, were merely neglected as

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

650  Klaus Beyer unfruitful complications for the tasks at hand. Subsequently, the upcoming generative paradigm with its postulate of a coherent language system residing in the language faculty of a competent speaker also dominated African linguistics. Proponents of this view regarded any contact related distortion of the ideal “pure” linguistic system as “corrupted speech” and a mere obstacle to grammar writing. Complications such as code‐switching, heavy borrowing, or completely mixed languages counted as irregularities triggered by individual behavior or special social contexts which were not treated within a coherent linguistic theory.3 From about the early 1980s onward the dominant linguistic paradigm changed again and more researchers became interested in general dynamics and variability of language systems and speakers’ multilingual repertoires also began to attract attention. The postulate of a “completely homogeneous speech‐community” was, again,4 called into question and research on the impact of convergence on language classifications and areal typology gained ground. Moreover, functional approaches to language (Dik 1978, 1997a, 1997b) that put language use and pragmatics center stage as well as the ongoing development in variational linguistics set the scene for a wider appraisal of language‐contact issues also in the study of African languages. By now, it is widely accepted among both Africanists and general linguists alike that most languages of the world are shaped to a very large extent by contact‐induced features that defy essentialist views of closed language systems. Accordingly, language history and grammar description can no longer be modeled exclusively along the genetic metaphor of diverging dendrograms and on the assumption of ideal speaker‐listeners in the Chomskyan sense. As a consequence of that insight, a huge amount of literature considers the impact of language contact and data from all over the world and substantiates the claim that contact‐ driven language change is rather the norm then the exception. Research on African languages does not fall short of this new movement, all the more so as Africanists are convinced that the multiple variations of the language‐contact theme observable in Africa help to develop the subject further and may, eventually, add to the formulation of an integrated theory of language contact.

1.1  Early Works on Language Contact in Africa In the nineteenth and early twentieth centuries the contact issue in connection with Africa was mainly discussed in terms of the “Hamitic theory” (see p. 649). Meinhof (1910, 1912, 1932) who applied the comparative method to Bantu languages and tried to reconstruct Proto‐Bantu was convinced that the first boost for the Proto‐Bantu family and the Bantu expansion was the external input from Semitic languages. In showing close genetic relations between the Bantu languages and most language families of the Western Sudan (nowadays called the Niger‐Congo phylum), Westermann (1927) had already all the elements for the deconstruction of the Hamitic theory at hand. Finally, it was the comparative work of Greenberg (1963) that completely rejected this theory. Still, however, after World War II, Africanists were by and large occupied with classification and reconstruction tasks employing only the principles of divergence in the treatment of comparative data. Heine’s overview of African lingua francas (1968) was an exception and a first serious step in considering linguistic contact relations in Africa. His description of 38 African languages of wider communication, including pidgins and/or creoles, led him to reflect on the origins and possible classification of this kind of so‐called mixed languages. In a later study of seven pidgin languages from the Bantu area, Heine (1973) concluded that they are not mixed languages and that any model of pidginization based on language mixing must be rejected (1973: 50). In his view, contact phenomena only played a role inasmuch as formerly unrelated language communities were suddenly forced to communicate in specialized domains (1973: 67, 68), and therefore used reduced forms of otherwise genetically defined languages.

Contact among African Languages  651 The French “Laboratory of Languages and Civilizations of Oral Tradition” (LACITO) i­nitiated a series on “Contact of Languages and Contact of Cultures” (Contact de langues et contact de cultures) in the late 1970s5 mainly intended as a reflection on practical problems resulting from language contact and multilingualism for e.g. education and nation‐building. Concentrating on a contact zone in the geographical center of Africa (Chad, Cameroon, and Central African Republic), a number of authors looked at linguistic demography and language repartition among multilingual speakers, analyzed lexical borrowings and neologisms derived from French, and added detailed studies of the concomitant sociolinguistic backgrounds. The series therefore can be considered as the first attempt to unravel the multiple levels and different aspects of language contact in Africa. Further important contributions came from Myers‐Scotton who worked on multilingualism and code‐switching in African contexts from the 1970s onwards (Scotton 1971; Myers‐Scotton 1983, 1993a, 1993b, 2002). Her descriptions and analyses of data on the mixing of languages from eastern Africa influenced the development of contact linguistics and contributed to the ever rising interest in questions of contact and convergence, not only in Africa but worldwide (see also Section 2).

1.2  Africa as a Contact Zone, Recent Findings For African language research, as well as for other subfields of linguistics, the seminal Language Contact, Creolization and Genetic Linguistics (Thomason and Kaufman 1988) marks the starting point for a growing body of research on contact‐induced language change. Scholars by and large began to question the hitherto widely accepted language change and development models mainly based on the principle of divergence and started looking at convergence phenomena more closely. By the same token, a growing number of Africanists began to rethink classification models and the concomitant sociolinguistic scenarios for individual languages as well as for whole subgroups and families. The exhaustive survey on literature on language‐contact research all over Africa presented by Güldemann (2018: 454–72) also illustrates the growing awareness of contact issues in Africa related linguistics. The sheer number of contact related publications literally exploded from the beginning of the 1990s. The research history on Songhay may serve as a case in point. Spoken in Mali and adjacent countries alongside the river Niger, it was classified throughout the decades as either Mande, Voltaic, Saharan or an isolate by various authors. The arguments for its inclusion into these different families depended on whether one concentrated more on the syntax, the morphology or the lexicon and which of its closely related varieties served as the main source for comparison. Greenberg (1963), mainly basing his attempt on lexical resemblances with Zerma and the Gao‐dialect, included Songhay into his Nilo‐Saharan phylum, closely related to the Saharan branch. Notwithstanding critics renouncing the poor evidence for this (Lacroix 1971; Nicolaï 1990), the connection with Nilo‐Saharan was still held up for some time (Bender 1996; Ehret 2001; and with some reservations, Kossmann 2005; Souag 2012). It is only quite recently that Songhay was finally excluded from Nilo‐Saharan and counted as a language family in its own right Dimmendaal (2011: 326, 407, 408). It was the then world‐leading specialist on Songhay, Robert Nicolaï, who first attempted a different explanation to account for the language’s multiple relations (Nicolaï 1990, 2003) thereby also criticizing the ubiquity of the language tree model in all former attempts. For him, the proto‐language from which all modern Songhay varieties derive was a mixed language consisting of a grammatical system mainly based on Mande (independent family)6 and a lexicon derived from an unknown source of Afro‐Asiatic origin. At a later stage, a pidgin based on Mande syntax and a Touareg (Afro‐Asiatic) lexifier language developed as a means of communication between Sub‐Saharan and Berber traders at the southern end of an ancient trans‐Saharan trade route. This pidgin creolized with the rise of the

652  Klaus Beyer Songhay‐Empire from the fifteenth century onwards and all western dialects of contemporary Songhay derive from it. Although his hypothesis may suffer from some flaws and inconsistencies (Kossmann 2005) it is a much better attempt at unifying all the different linguistic relations than keeping it in the generally debatable Nilo‐Saharan phylum while dismissing all the other connections. Moreover, while Bender’s and Ehret’s comparative arguments – both stick to the genealogical connection with Nilo‐Saharan – are also far from convincing (Nicolaï 2003), Nicolaï’s hypothesis takes into account social and historical evidence (Nicolaï 2009) and thus advances the epistemology of historical linguistics in Africa considerably. What since stands out from Nicolaï’s work is his refusal to accept mono‐ genetic origin as the standard representation of language history. The Songhay example also shows that there is no need for outside counter‐opinions7 to spur the rethinking of some views on African language relationships. By the beginning of the 1990s, many important scholars had already begun to review family tree models on their own initiative and responsibility. Even Westermann (1935) and Greenberg (1959, 1983) commented on areal distributions of some linguistic features that cut across their own genetically defined language families.8 Based on these earlier suspicions some of the features formerly regarded as prime elements for an eventual reconstruction of Proto‐Niger‐Congo (PNC) were reanalyzed and convergence as an alternative or at least supplementary scenario for historical processes of language development was taken into account (cf. Dimmendaal 2001; Childs 2010). Likewise, the Khoisan family formerly presented as a tentative genealogical unit by scholars as Greenberg (1963) and Köhler (1975) was reconsidered and is nowadays understood as an areal agglomeration of small language groups that defies broad‐scale genetic reconstructions (Güldemann and Voßen 2000). Only the well‐established Afro‐Asiatic family seems to have stood the test of time and is still considered a proven genealogic unit (but see Dimmendaal 2018: 36 for some minor uncertainties). Despite these insights and the growing awareness of the areal distribution of linguistic elements and convergence as an equally important forming principle of contemporary African languages, contact linguistic research in Africa at the beginning of the new millennium was still in the “hunting and gathering stage” (Haspelmath 2004: 221). The emerging field of contact linguistics propelled mostly large‐scale areal overviews and critical reviews of the genetically based language families. For instance, Heine and Nurse (2008) bring together data on language convergence from all geographic regions and picture Africa as composed of different linguistic areas thus defying classical genealogical views. In this book, Güldemann (2008) presented the “Macro‐Sudan belt” as a vast convergence area sandwiched between the Sahel in the north and the Congo basin in the south and spanning the continent from the Atlantic Ocean in the west to the Ethiopian plateau in the east. Based on the analysis of six cross‐linguistically highly marked features (logophoricity, labial velars, Advanced Tongue Root (ATR) vowel harmony, S‐(Aux)‐O‐V‐X, V‐O‐Neg, labial flaps) that are frequently present in the area’s languages but absent or rare outside, a geographical cohesion cutting across major language families and subfamilies was demonstrated (Güldemann 2008: 154–70). The author refrained from developing a historical scenario for the linguistic findings but from the specific clustering within the Macro‐Sudan it is assumed that the features did not spread simultaneously from one source only. The only non‐linguistic arguments for the existence of the convergence area are the geographical boundaries in the north (Sahara) and south (the Congo basin rainforest) that constantly reinforced population movements along the west‐east axis (Güldemann 2008: 180, 181). So, approximately a decade ago, the “prevailing orthodoxy” (Williamson 1989: 8) of Greenberg’s African language classification has been overcome on the basis of a general appraisal of contact phenomena in African languages. What is still largely missing are detailed accounts of the socio‐historical processes leading to linguistic convergence phenomena and the resulting contact‐induced language change. Most scholars kept working on

Contact among African Languages  653 what Muysken (2008: 5) calls the meso‐ and macro‐level of language contact thereby neglecting micro‐level processes that can only be observed in communicative events between speakers in their social frameworks. One counter example to this general tendency is the insightful Repertoires and Choices in African Languages by Lüpke and Storch (2013). In this book, the authors highlight historical and contemporary cultural practices in Africa (e.g. exogamy, child fostering, economic interdependencies, mobility) that all show how multilingual repertoires of speakers arise and finally defy any essentialist views of traditional languages. The authors criticize that the growing awareness of the misconceptions of “fixed languages” and “fixed linguistic identities” in disciplines like sociolinguistics, variation studies, and the ethnography of speaking has not been taken up by mainstream descriptive and documentary linguistics, especially in the Africanist tradition (Lüpke and Storch 2013: 2). While such a disciplinary bias towards an (over)simplification in terms of coherence has been, and is still, observable in many works on African languages, the radical deconstruction of the linguistic object does not seem to be the solution either. Many scholars (including the present author) still see the need for description and analysis of languages as rule‐governed objects, if only to understand how contact‐induced processes of change work on the cognitive level for individual speakers. The fact that in daily multilingual practices speakers with highly heterogeneous linguistic backgrounds and repertoires nonetheless communicate successfully which each other solicits a theoretical approach that integrates the different forces and various layers at play in such multilingual encounters. However, the above mentioned authors also have no remedy to offer but the proposition “…to set language free from the artificial, its existence as artifact, and acknowledge speaking languages as being something tantamount to an artful and tangible expression of the Self, in different contexts” (Lüpke and Storch 2013: 346). Thus, many questions concerning different aspects of contact between speakers of African languages rest unsolved – especially on how social contexts shape linguistic repertoires and why an individual innovation eventually spreads to a wider community.

2 Contributions from Africa to a General Framework of Language Contact The above briefly sketched history of language‐contact research shows how developments in general linguistics and in sociolinguistics began to loom large in African contexts throughout the years. There are, however, also contributions from African language research to the further development of the general framework of language contact. Such contributions are rarely tidal events that wash away whole paradigms and establish new schools of thought but mostly add data on various contact issues that often lead to reconsidering tenets drawn from general northern contexts. A case in point is the work of Carol Myers‐Scotton, who added substantial insights to the issue of code‐switching (CS) in analyzing data from fieldwork in east and southern Africa. Already in 1971 she put forward a first analysis of data on “language switchings,” as it was called in this study (Scotton 1971: 114). Therein she explained choices between English and Swahili among educated respondents in terms of costs and rewards related to the speaker’s individual communicative situation against the backdrop of multilingual Kampala (Scotton 1971: 121, 122). Subsequently, she developed the “cost and rewards” approach further into the larger framework of the “Markedness model” which describes the underlying social motivations for CS based on a large corpus of naturally occurring communicative data from Africa (Myers‐Scotton 1993a). Furthermore, the author also analyzed the structural

654  Klaus Beyer constraints on CS which eventually led to the “Matrix Language Frame” model (Myers‐ Scotton 1993b). This model added substantially to the theoretical framework in which CS is understood even today. In more recent years, she enlarged her vision and included other types of language‐contact phenomena (e.g. child bilingualism, Creole structuring, mixed languages) in her analysis. The resulting “4‐M model” can be understood as a psycholinguistic approach to language production that explains various phenomena of bilingual speech on the micro level of the individual speaker (Myers‐Scotton and Jake 2000, 2016). Apart from such ground‐breaking work, the majority of research on language contact in Africa kept on focusing on descriptions of linguistic features affected by convergence, their areal distribution and on their impact on language classification and representations of genealogical relations. A recurrent theme in these strands of research was the detection of features that can or cannot be transferred through contact between languages (as opposed to inheritance). A closer look at linguistic features that were foremost held to be diagnostic for genealogical relations revealed that they may also be prone to areal distribution. For ­instance, highly marked diagnostic features like vowel harmony and labiovelars, noun‐ class morphology and serial verb constructions are often also present in non‐genetically related languages within a vicinity (Dimmendaal 2001: 365–87).9 The detection of probabilities for the borrowing of elements from all linguistic subsystems became a major interest, mainly for methodological reasons. For instance, the method of mass comparison (Swadesh 1955) relied on the assumption that lexical material from the “core” or “basic” vocabulary is less prone to borrowing then any sort of cultural vocabulary. That this is an oversimplification has since been shown by many studies of lexical borrowings in African contexts. In fact, it seems that, given the right sociolinguistic background, lexical borrowings are possible in any semantic domain. A recent example from the Northern Gurunsi language Pana (Gur, Niger‐Congo) shows how speakers of this language adopted basic vocabulary like “animal,” “blood,” “smell,” “root,” and “mountain” from speakers of the non‐related neighboring Manding (Mande, independent family) languages with whom they are in a long‐lasting and multilayered social relationship (Beyer 2006: 14). The fact that the word “fire,” which is equally part of basic vocabulary, is transferred in the other direction, namely from Pana to Manding and Northern Samo (Mande, independent family) seems strange at first sight. One possible explanation is to attribute it to Pana women who frequently marry into neighboring villages and somehow reverse the general pattern of linguistic interference (Beyer and Schreiber, forthcoming). An important methodological step was to distinguish between the “borrowing” of linguistic material and interference phenomena where elements of a model language are traceable in replica languages only on a structural level. Thus, in an import extension of the theory of grammaticalization, Heine and Kuteva (2003) demonstrate how cases of contact‐ induced grammaticalization lead to the areal diffusion of linguistic structures and thus help us to understand the formation of linguistic areas (Heine and Kuteva 2003: 530). Notwithstanding the examples of borrowings from core vocabulary, there are by now accepted empirically based probabilities for the general borrowability of linguistic material. Most scholars nowadays agree that bound morphology is less easily borrowed than free morphemes or that agglutinated elements are easier incorporated as loan elements then inflectional morphology (Tadmor et al. 2010). While many other scales also rate the relative borrowability of, e.g. words, word classes or semantic meanings (Haspelmath 2008: 48), it also seems true that linguistic constraints on the borrowability of a given item may always be overridden by a specific social context. As Thomason (2008: 42) phrases it “… when social factors and linguistic factors might be expected to produce opposite results in a language‐contact situation, the social factors will be the primary determinants of the ­ linguistic outcome.”

Contact among African Languages  655 In short, empirical data from Africa (and  –  of course  –  from many other multilingual contact contexts in the world) have led to scholars ascertaining some probabilities for the non‐genealogical transfer of linguistic material in a language‐contact situation. The most prominent factors seem to be: a. The type of contact in terms of social relations between speakers b. The intensity of contact between speech communities c. The typological fit between two languages Whether these factors correlate in a deterministic manner or at least display some robust statistical tendencies as to their connectedness is still a large and widely debated question. African contexts provide a natural testing ground for these questions and that is what many Africanists interested in the contact issue are currently doing. However, given that context information on historical speech‐group events and earlier sociolinguistic scenarios (i.e. factors (a) and (b)) are often very scarce and difficult to come by, the appraisal of non‐ linguistic factors is not always easy. One possibility is the study of contemporary sociolinguistic behavior and extrapolation into the past. That the different transfer types just mentioned can be observed in one contact context and that it is also possible to add some highly probable socio‐historical hypothesis can been demonstrated with the example of the Western Nilotic language Luo (Eastern Sudanic, Nilo‐Saharan) spoken in western Kenya and adjacent areas of Tanzania and Uganda alongside the eastern banks of Lake Victoria. As Dimmendaal (2011: 193–6) points out, Luo has developed a set of preverbal tense markers absent in all genealogically related varieties. Typologically comparable systems, albeit based on different morphological elements, are only used among some neighboring Bantu (Benue‐Congo, Niger‐Congo) languages (referred to as “Suba” by Rottland and Okombo 1992: 273). Luo, then, has created its set of new preverbal tense markers replicating the Bantu‐model but using its “own” linguistic material, in this case from independent adverbs of time. What is more, Luo also developed prefixed number marking on nouns, a feature usually alien to Nilotic languages. Owing to an already formally reduced number marking system (stemming from older long‐term contact with Central‐Sudanic and Ubangian languages), Luo borrowed Bantu class 1 and class 2 markers (cf. mɪ‐hɪa, wa‐hɪa “child, children”) and also developed additional number prefixes from its own sources (Dimmendaal 2011: 195). Such a calquing process in addition to morphological borrowing (cl1{mɪ} and cl2{wa} are based on Bantu class‐morphemes *m/ba) and far‐reaching borrowing of cultural vocabulary (Reh 2000: 290, 291) was possible through intense language contact resulting in a massive primary language shift from Bantu towards Luo. Apparently, Luo infiltration of the contact zone was completed around 1850. From that time onwards, assimilation of erstwhile Suba speakers mainly worked along trade and intermarriage lines. With the advent of colonialism, Luo was established as a prestige language of upward mobility which further accelerated the language shift from Suba to Luo. Finally, long‐lasting bilingualism among the Suba led to a complete language shift (Rottland and Okombo 1992: 276, 277; Kembo‐Sure 2000). In the Luo/Suba case, many of these bilingual speakers have been Bantu women who married into Luo households and thus gradually changed the language. For Reh (2000: 589) this process already started in the sixteenth century and has parallels with the Songhay scenario developed by Nicolaï (1990; see p. 651). She pictures the present Luo variety as the result of massive intermarriage of Luo men with non‐Luo women that led to multilingual households. From these households a vehicular variety of the husband’s language emerged that eventually developed into the modern Luo language (Reh 2000: 591). While Reh’s scenario seems to reach far further back than that of Rottland and Okombo, both have in common that the heavy influence of bilingual speakers finally changed the language in many important parts.

656  Klaus Beyer The Luo case thus illuminates the general sociolinguistic context for the replication of concrete linguistic material and more abstract forms of structural borrowing: intense (long‐ lasting) contact and the existence of numerous bilinguals. Cases like this with more or less clear social contexts and correlated linguistic outcomes have been documented by and large in Africa. Some extreme cases like, e.g. Ma’a / Mbugu, a kind of consciously made‐up mixed language with Bantu and Cushitic ingredients (Mous 2001, 2003), but also the cited Songhay and Luo cases only seem to be the tip of the iceberg. They all broaden our empirical basis and help to refine our models of language contact even on a global scale. But there are many questions still open and far from clear. For instance, we do not really know how fast and through which channels contact‐induced changes are processed and become stable in bi‐ or multilingual communities. So, nearly 20 years after the most urgent need for “sociolinguistic micro‐studies of speech communities in contact” (Heine and Kuteva 2001: 409) was formulated, empirical micro‐studies are still comparatively rare.

3 Micro‐studies of Contact in African Multilingual Environments In this Section I present two case studies that were designed to fill the gap in empirical data on the process of linguistic norm building against the backdrop of non‐standardized languages in multilingual contexts. The two research projects were designed to capture contact‐ induced changes concentrating on the language behavior of individual speakers. The first one plays out in a rural context of long‐standing multilingualism in a somewhat remote contact area named Souroudougou. The second project looks at the highly dynamic situation of multilingual professional groups in the capital city of the Adamawa region in northern Cameroon, Ngaoundéré. Although empirical data collection has already been completed for Souroudougou, the analysis of the very rich data base is still continuing. The Ngaoundéré project only started in 2017 and is still in the data collecting phase with just preliminary analyses done. Notwithstanding this, some insights can already be drawn from the two ­projects that further illuminate the complicated and multilayered issue of language contact in Africa.

3.1  Language Contact in Souroudougou The overarching goal of the research project10 was the presentation of a fine‐grained documentation and analysis of the relevant processes of language change in a complex multilingual situation characterized by long‐standing contact of various ethno‐linguistic groups embedded in predominantly oral societies of a non‐Western type. More precisely, Beyer and Schreiber (2013, 2017, forthcoming) set out to study diachronic and synchronic outcomes of contact‐induced language change and to correlate them with both, linguistic and extra‐ linguistic factors. That is, they aim at an application of variational sociolinguistics in a rural African setting and are concentrating on contact‐induced elements of language change in two neighboring speech communities. The contact zone east of the Upper Sourou River (hence Souroudougou = Sourou land) in northwest Burkina Faso and adjacent areas in Mali has always been an area of retreat for smaller ethno‐linguistic groups withdrawing from the dominant West African power centers.11 The northern fringe is dominated by Dogon languages (independent family) and the east is occupied by Northern Samo‐speaking groups. In the center, the endangered Gurunsi language Pana is spoken and the southern and western parts are dominated by Marka‐ Dafing. Linguistically, apart from the Marka‐Dafing and Northern Samo which belong to

Contact among African Languages  657 different branches of the independent Mande family, the ethno‐linguistic groups are not genetically related. Additionally, a few Fulani (Atlantic, Niger‐Congo) speakers, the long‐ standing vehicular language Jula (Mande, independent family), and French are part of the linguistic landscape of Souroudougou. All ethno‐linguistic groups have been in place for at least 300 years, but autochthony is disputed. They are organized in village alliances that cut across linguistic and

Map 33.1  The Souroudougou Research area.

658  Klaus Beyer blood‐relational boundaries and were originally set up for defensive purposes (Beyer 2010; Beyer and Schreiber 2013). The long‐standing history of active exchange and regular communication between the different groups resulted in a complex situation of multilingualism that left its effects on all languages of the area. For instance, as already mentioned, lexical borrowings, even from basic vocabulary, betray the long‐lasting contact relationship between the ethno‐linguistic groups in Souroudougou (Beyer 2006). Further outcomes of this prolonged language contact are visible in all linguistic subsystems (e.g. noun‐class morphology, phrase‐final marking, negative constructions) (Beyer and Schreiber 2013; Beyer, forthcoming). Against the backdrop of this complex historical and linguistic scenario we concentrate on contact‐induced variation in Northern Samo and Pana and describe how individual speakers’ social backgrounds shape linguistic behavior. A case in point is the phonological variation between labialized and non‐labialized consonants. The contact area lies on the northern fringe of the Macro‐Sudan‐Belt (see Section 1.2), which is defined, amongst other features, by the presence of labiovelar consonants (Güldemann 2008: 156–8). Such labiovelars have also been reconstructed for Pana’s genealogical ancestors Proto‐Gurunsi and Proto‐Central Gur (Manessy 1969, 1979). In Pana, reflexes of the reconstructed *kp/*gb are the labialized variants [kw/gw] and labialization extends to other word‐initial obstruents. This labialization is ubiquitous and, at first sight, just in free variation there. Northern Samo and Marka‐Dafing are also northern fringe languages of the Macro‐Sudan‐Belt (MSB) and labialization of velars is described as variation in their synchronic phonological systems (Diallo 1988; Schreiber 2008). All other contemporary contact languages in the Souroudougou area do not display this feature. A closer look at speakers in the two adjacent villages, Pini (Northern Samo) and Donon (Pana) (see Map 33.1) reveals that labialization there is not just in free and unregulated variation but that it correlates with exposure to contact languages and social status. Schreiber (for details see Beyer and Schreiber, forthcoming) describes the overall picture for Pini as follows: labialization of velars (and other consonants) is a marginal variation in Northern Samo as spoken in Pini with the majority of speakers not using this feature at all. However, there are two speaker groups in Pini who do labialize: firstly, speakers with a close affinity to neighboring Pana villages, i.e. women who married into Northern Samo households in Pini and the people who are in a close relationship with them, that is, their children and husbands. Statistical evaluation of dichotomized data shows a strong positive correlation between labialization and a “Pana‐affinity” score above average for speakers of Northern Samo in Pini. Second, members of the Serme clan also show a strong tendency for labialization compared to members of other clans in the Pini speech community. The Serme clan is generally held to be conservative and traditional as they are the founders and ruling family of Pini. The general picture that emerges from these observations in Pini is that labialization is nowadays preserved by conservative traditionalists that stick to a variable that reflects an older version of contemporary Northern Samo. Additionally, the general tendency to abandon this pronunciation is countered by speakers with closer bonds to the neighboring contact language Pana. How then is labialization realized in Pana? The correlation of social network data with recorded pronunciation variables tells a slightly different story: the variable “labialization” correlates with speakers’ social integration into the social fabric of Donon: the better individuals are integrated, the higher is their incidence of labialized obstruents. This suggests that what counts as marginalized pronunciation in Pini is (still) part of the linguistic norm in Donon (Beyer 2015: 246, 247). If we assume that this is generally the case in Pana‐dominated speech communities, it becomes clear why Northern Samo speakers of Pini with Pana‐ affinity line up with local conservative groups in respect of this variation. The famous research by James and Lesley Milroy (1985, 1987) derived from Social Network Analysis (SNA)12 that central actors stick to established vernacular speech norms,

Contact among African Languages  659 while non‐central actors easily pick up new, often prestigious ways of speaking and therefore are likely to lead ongoing language change. If we follow this argument, and it seems that the impact of social networks on individual linguistic behavior in an African rural multilingual environment is in principle not very different from what the Milroys described for Belfast (Beyer and Schreiber 2017), then further predictions can be made. Thus, non‐labializing speakers in Donon are often marginal in social network terms and they also share other social attributes. That is, speakers who occupy a non‐central position in their local community networks are often educated people (who have been at far‐away boarding schools) and/or people who speak also Jula and French (i.e. non‐labializing languages) quite fluently. Others have been abroad for a long time or are of foreign origin entirely. Although many of these people are not central to their current social networks they enjoy a certain prestige (part of an intellectual elite) in the village. It is therefore highly conceivable that the Pana vernacular norm is changing towards their mainly non‐labializing pronunciation in the future, just as in neighboring Pini. As for Pini, the social network status of the “labializers” is not clearly marginal or central. Serme traditionalists and speakers with Pana affinity merely display a normal distribution in terms of connectedness in the social life of the village. Therefore, it seems like labialization became a phonological marker of Serme clan members and people with closer Pana contacts while the majority of speakers in the village adhere to a pronunciation norm that is only just developing in the Pana of Donon. In Pini, this may be also due to the impact of French that is gaining ground as a prestige language. This sketch of data allows for a differentiated picture of language change in Souroudougou: various contact languages, possibly at different times, are responsible for established contact‐ induced changes on all linguistic levels. Currently ongoing change is led by non‐central actors who introduce items from the current most prestigious languages Jula and French. Extrapolating the synchronic sociolinguistic account into the past would suggest that socially marginalized speakers use elements from prestigious or otherwise dominant contact languages (e.g. from the Pana perspective: Samo at one point in time, possibly later Marka and currently Jula) and accordingly reshape the established vernacular speech norm. Given the long‐standing multilingual set‐up with very flexible identity boundaries in Souroudougou and the fact that normative control is only maintained by local community networks, the development of a local areal language type with numerous local contact features is indeed inevitable.

3.2  Contact‐induced Variation in Ngaoundéré The northern parts of Cameroon are among the most linguistically diverse areas worldwide. Against the backdrop of this multitude of indigenous languages (mainly from the Chadic and Adamawa families), one can recognize Fulfulde as the dominant lingua franca over the entire area. It is also the first language of the traditional noble authority in the urban centers, particularly in the regional capitals Maroua, Garoua and, from the beginning of the nineteenth century, in Ngaoundéré (Connell 2017: 115, 116; Markgraf 2017: 94, 95). Used in such a multilingual setting, Fulfulde is highly exposed to contact‐induced variation introduced by L2 speakers from various linguistic backgrounds. The range of the Fulfulde varieties reaches from the high prestige dialect of Diamaré (fulfulde laamnde “pure Fulfulde”) around the town of Maroua in the Far North via an L1‐variety named Standard Fulfulde in the towns of Garoua and Ngaoundéré to an L2‐variety used outside the Fula community sometimes glossed as bilkiire “childish gibberish” (Kramer 2017: 32). In fact, laamnde and bilkiire can be regarded as the two poles of a continuum of socially relevant speech forms that in reality do not or no longer exist as speakers who exclusively use linguistic features associated with just

660  Klaus Beyer one particular Fulfulde variety can hardly be found. Rather, language use by Fulfulde speakers in northern Cameroon moves along a scale between the two poles comprising a broad range of linguistic variation showing speakers’ proficiency in several dialectal and/or sociolectal registers. This variation in Fulfulde is in the focus of a recently started research project on contact‐ induced language change in multilingual Ngaoundéré (Beyer and Kramer 2017). The questions we ask concern norm development of spoken Fulfulde in different professional groups who at the same time belong to different types of social networks. What we are interested in is the sort of retention, acquisition, and replacement of linguistic variables that cannot be completely accounted for in terms of corporate group membership. For instance, if a speaker belongs to the group of ethnic Fula in Ngaoundéré his or her Fulfulde version would include a higher frequency of noun‐class related pronominal forms than is found with other non‐ethnic Fulfulde speakers (Kramer 2017: 44). While such a correlation between ethnic membership and variable choice is hardly surprising, one also finds relational factors that play a role for the variable choices. Kramer (2017: 46–8) describes a group of eight motorcycle mechanics working together in an open roadside garage in Ngaoundéré where she collected social network data, attributive data, and data on linguistic variation. From her analysis it became clear that it is the language use of one eminently central actor which particularly reflects the garage’s norm of language use. That is, the variable use of the garage actors has a tendency to converge with the central actor’s variable choices which also display the most extreme values regarding the tendencies of the garage group’s linguistic norms. In her conclusion she states that speakers’ positions within social networks and the number of relationships they are involved in are crucial factors for their linguistic behavior (Kramer 2017: 48). Interestingly, when looking at her control group, i.e. mostly motor taxi drivers who are garage customers, no specific group norm seems to exist. This is possibly due to the fact that motor taxi drivers in general maintain less tightly knit social networks then the ones observed in the garage. This assumption relies on the observation that motor taxi drivers constantly move around all over town, meet many different people and customers and are generally not in such a structured working environment as, e.g. the garage mechanics. What is more, working as motor taxi driver is not a very prestigious job. Kramer’s observations thus suggest that in speech communities without norm‐enforcing institutions speakers like the motor taxi drivers tend to promote increasing variation in variable use and heterogeneous linguistic practices. This is somehow in contrast with findings from Northern and Western countries where loose network structures are said to trigger a shift of speakers’ linguistic behavior towards an overarching standardized norm (e.g. Milroy 1987: 190–1). But such a standardized norm of Fulfulde, to which actors in less tight‐knit networks would shift to in Ngaoundéré is, for the time being, non‐existent nor does it seem to be developing. Whether the assumptions of loose‐knit social networks and heterogeneous variable use in the group of motor taxi drivers really hold forms the focus of another part of the Ngaoundéré project. Beyer (2018) describes and compares personal or ego‐centered networks of motor taxi drivers in the city of Ngaoundéré and correlates these data with a subset of the linguistic variables Kramer used in her garage study (see Table 33.1). As the description of the social ties between all members of a fixed network (e.g. the garage mechanics) is not possible with motor taxi drivers for obvious reasons, the Ego‐ Centered Network (ECN) approach is the tool of choice. For the set‐up of an ECN, the researcher compiles information from one actor only. This actor provides linguistic data and information about people s/he is related within relevant domains. In the course of the interview the actor would thus provide names of other people, the alteri, s/he, for example, lives in a household with, spends leisure time with, has professional connections to, etc. The individual is furthermore questioned about his/her own and the alteri’s attributes (e.g. age,

Contact among African Languages  661 Table 33.1  Linguistic variables and their values as used in the Ngaoundéré study (adapted from Kramer 2018).

Variable [DOG] [AGR]

[PL.CL]

Lexeme used for the concept dog agreement classes or number/animacy distinction

Number distinction marked by paired noun form classes or plural marker

“Bilkiire” Ngaoundéré

“Standard” Garoua

“Laamnde” Maroua

goyru

ɓoosaaru

rawaandu

Number distinction: o ɗo naŋga mo ‘he is catching it (the dog)’ animacy distinction: o ɗo naŋga ŋga ‘he is catching it (the dog)’ Use of a fixed plural marker or numerals: ɓoggo ɓoggolji ‘rope/ ropes’ paɗe‐paɗeeji ‘shoe/ shoes’ hooseere – hooseereˀen ‘mountain/ mountains’ debbo‐debbo ɗiɗi ‘woman, two women’

Agreement classes: o ɗo naŋga ndu ‘he is catching it (the dog)’

Use of paired noun form classes marked by suffixes: ɓoggol – ɓoggi ‘rope/ropes’ faɗo – paɗe ‘shoe/shoes’ hooseere‐kooseeje ‘mountain/ mountains’ debbo‐rewɓe ɗiɗo ‘woman/two women’

gender, ethno‐linguistic background, etc.), the frequencies of meetings between him/her and them and the relations among the alteri, i.e. whether they know each other and how often they meet. Of course, such an ECN has not the same explanatory power as the social network of a fixed group where all actors are questioned, but it is a highly appreciated statistically‐based shortcut for assessing large and highly moveable groups, like the motor taxi drivers of Ngaoundéré (Jansen 1999: 73–8; Wasserman and Faust 1994: 53). It is assumed that with a large enough number of ECNs (n ≥ 50) the overall network characteristics of the group they represent emerge clearly (Wolf 2010). In the first year of this study in Ngaoundéré, 52 ECNs plus concomitant linguistic data of the ego‐actors have been collected, transcribed, and analyzed. From the analyzed networks the average density of the personal network of an motor taxi driver in Ngaoundéré was calculated at 62.6% which, surprisingly, is much higher than the average density of 38.04% for the 10 garage networks throughout Ngaoundéré which Kramer extended her study to recently (Kramer 2018). However, the residuals of the density measure in the ECN’s (17.9% < 62.6% < 100%) show that the variation there is much wider than in the garage networks (28.6% < 38% < 50%). This indicates the more heterogeneous character of the motor taxi driver networks that seem to be, at least partly, due to the less stable work environment. What can we see in terms of correlations between motor taxi drivers’ ECN structures and variable use? Generally speaking, the density of the motor taxi driver networks has an impact on the use of the variables inasmuch as dense networks show more variation then loosely knit ones. That is, the higher the density of an ECN the more variation in the use of the [DOG] variable is recorded (p = 0.016). The picture is, however, much more complicated:

662  Klaus Beyer the [DOG] variable also co‐varies with the “ethnic origin” of the motor taxi drivers. When a taxi driver has an ethnic background with either the Laamnde or Standard Fulfulde variety as L1 he is likely to use rawaandu or ɓoosaaru respectively. Moreover, ECN density also correlates very robustly (p = 0.005) with the homogeneity of the networks in terms of contacts with other ethno‐linguistic groups,13 showing that a higher density of an ECN usually results in more homogeneity. The first analysis of the data further suggests that the variable [AGR] does not reflect ECN measures but correlates with attributes of the motor taxi drivers: there is a strong correlation tendency (p = 0.053) between the time an actor has passed in the city and his use of agreement markers. The longer he lives in Ngaoundéré, the more likely he is to use the reduced set of agreement marking known from Standard Fulfulde. Nevertheless, a connection with network measure is also seen in the robust correlation (p= 0.023) between duration in the city and network size which tells us, the longer an actor lives in the city the more extended his personal network becomes. From the linguistic variables investigated it seems that the coding of plurals [PL.CL] is the only one that does not correlate in any clear manner with either attributes of the motor taxi drivers MTDs or any of the ECN‐measures. Although these first quite impressionistic correlations do not allow for any clear‐cut generalizations, a first hypothesis about linguistic behavior may be formulated: motor taxi drivers, who are part of comparatively dense networks, experience less centralized structures and therefore do not develop stable group internal linguistic norms (see Figure 33.1, ECN 1+2). The flip side being that motor taxi drivers with loose networks, most of which are also highly centralized (see Figure  33.1, ECN 3+4), adhere to the internal norm of their personal network. Furthermore, the longer a motor taxi driver lives in the city the more extended his network gets but the less normative influence from this immediate network is exerted and he adapts more readily to his wider surroundings, at least in terms of agreement marking. Whether and how this linguistic behavior of motor taxi drivers has a lasting effect on Fulfulde norm development in the city in general is yet unclear. A first prediction would be that the motor taxi drivers serve as a constant reservoir for a wide and heterogeneous variety of linguistic elements but that it is not through them, that lasting normative effects will be exerted. What we do know already from the comparative analysis is that professional groups in the city of Ngaoundéré display quite different social network structures and that actors within these structures behave differently according to linguistic adaptation and norm acceptance. What also results from these first data is that linguistic variables behave quite differently and may be more or less prone to signal group membership. This however, cannot be predicted from the nature of the linguistic item but seems to be related to the structure of the involved languages and the degree of social indexicality they evince.

4 Conclusions Language contact is by now a well‐researched theme in African linguistics. The majority of contact research still concentrates on classification and reconstruction tasks but convergence processes nowadays also play an important role in model building. There have been important contributions to the empirical database of language‐contact issues and general theory development has benefited substantially from this. What is still lacking are detailed studies at the micro level that focus on the underlying social and cognitive processes that trigger contact‐induced language change. Given the widespread linguistic areas and the prevailing, indeed in some parts the rising multilingual competences of the majority of speakers of African languages, for instance, in urban centers, Africa provides a vast and promising field for just this kind of micro‐level contact research. This advantage may be an obstacle at the

Contact among African Languages  663 Ethnie: Pullo (Ad)

Ethnie: Kolé Density: 0.944444444444444

1

Density: 0.714285714285714

5

Elb_ei: 0

Elb_ei: 0.333333333333333 9

V_dog: b’oosaaru/rawaandu

V_dog: goyru/rawaandu

8 Dog%: 50

Dog%: 86

7

2

3

Agr%: 40

Agr%: 63

9

PI CI%: 50

7

PI CI%: 50

6 2

4

10 1

5

8

ethnic origin

ethnic origin

4

Bamileke Gidar Kole Mada Mbaya Pullo

Bamileke Longto Mada Mbum Pullo

3

ECN 1: MTD = 10; high density, very heterogeneous elb; no clear [DOG] preference, [AGR] and [PL.CL] only partly realized.

6

ECN 2: MTD = 9; still high density, 50% internal elb-ties; [DOG] ‘goyru’ at 86%, [AGR] percentage comparatively low, [PL.CL] like ECN 1 Ethnie: Gbaya

Ethnie: Meri (Diamare; NW)

9 Density: 0.3333333333333333

Density: 0.5111111111111111

2

Elb_ei: 0

Elb_ei: –0.4

V_dog: goyru

V_dog: rawaandu

Dog%: 100

Dog%: 100 Agr%: 20

8

10

PI CI%: 60

1

Agr%: 25 PI CI%: 0

5

11

5

3 4

1

6 ethnic origin Mbum Mufo Pullo

7

ethnic origin

2 3

ECN 3: MTD = 11, under average density, mostly same ethnic contacts, [DOG] ‘rawaandu’ at 100%, [AGR] very low, [PL.CL] in the range of the others.

Bamileke Gbaya Mbum

4

ECN 4: MTD = 5, low density, two ethnical external contacts, [DOG] ‘goyru’at 100%, [AGR] low, no [PL.CL] realization.

Figure 33.1  Ego‐Centered Networks in comparison: network measures and linguistic variables’ values in upper left corner; ethnic origin of alteri in lower left corner.

same time, as this kind of research requires far‐reaching language competence in two or even more languages as well as the readiness and the open‐mindedness to delve deep into multilingual communities and study contact‐induced language change as it happens. The fine‐grained speaker‐centered approaches to contact‐induced language change under multilingual conditions presented in Section 3 of this chapter also make clear that there are no real shortcuts to knowledge on this level of research. The advocated ethno‐methodological approach of recording linguistic data in natural speech contexts and correlating them with speakers’ social network data and social attributes is time consuming in terms of data treatment and demanding in terms of personal involvement with the speakers. But the rewards for such a commitment are very interesting insights into the microcosm of language uses and practices under conditions that only seem unusual from a Eurocentric point of view. From the viewpoint of theory building such empirical data are indispensable as a base from which induction may lead us, eventually, to an integrated theory of language contact.

664  Klaus Beyer It is also the richness of languages and the often incredible multilingual skills of speakers of African languages that are frequently referred to as an obstacle to the “development” of Africans and African societies (begging the question of what is meant by “development” in this context). However, it is my conviction that people from the African continent may use, to their advantage, the fact that they are mostly familiar with fluid language and identity concepts, that they are often used to picking up another language quite quickly (but not necessarily just easily), and that they have multiple strategies for integration and adaptation, not only on the linguistic level. In the opinion of the present author it is not the case that “development” presupposes some sort of linguistic unity. On the contrary, in a changing world, proficiency in multiple skills of adaptation and hybridization represent more an asset than a burden and the study of some of the overarching principles of this proficiency is clearly promising.

NOTES 1 This number is, of course, approximate. The number is always in flux because languages are living and dynamic entities and the definition of what a language is may also change. The numbers given here are from the latest web edition of the Ethnologue (Simons and Fenning 2018; Ethnologue: Languages of the World; online version: www.ethnologue.com.as) as accessed on June 11, 2018. 2 Meinhof (1910–11: 165) regarded Bantu languages as an offspring “von hamitischem Vater und nigritischer Mutter [from Hamitic father and Nigritic mother]”. 3 “Linguistic theory is concerned primarily with an ideal speaker‐listener, in a completely homogeneous speech‐community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors … in applying his knowledge of the language in actual performance” (Chomsky 1965: 3) 4 Already Schuchardt (1882) was convinced “Es gibt keine völlig ungemischte Sprache. [There is no completely unmixed language.]” (Translation mine) 5 Caprile (1978a, 1978b, 1979, 1982) and Jungraithmayr (1987). 6 In the remainder of this text, I follow the tentative classification of African languages as presented in Dimmendaal (2011: 407, 408) in giving short classificatory hints in brackets after a first mentioning of language names. However, as research goes on, this classification may be subject to revisions. 7 Such as Dixon: “…the idea of genetic relationship and the term ‘language family’ are used in quite different ways by Africanists and scholars working on languages from other parts of the world” (Dixon 1997:32, 33) or “The Niger‐Congo situation is a classic example of taking the IE‐type family tree as the only model of linguistic relationship and employing it willy‐nilly, without proper care and criteria” (Dixon 1997: 34). 8 Güldemann (2008: 170–3) gives a comprehensive overview of these and other authors’ views on areal features and convergent tendencies in their comparative data. 9 Comprehensive overviews of this kind of contributions can be found in Dimmendaal (2011) and Güldemann (2018). 10 Many thanks to the German Research Foundation (Deutsche Forschungsgemeinschaft) for funding this project from 2005 to 2009. Research was conducted by Henning Schreiber for the Mande part and Klaus Beyer for the Gur(unsi) part. 11 The main forces in the wider region throughout the last 300 years or so have been the Moose kingdoms, the Fulani Empire of Massina, and the Bambara Kingdom of Segou. While none of these ever seized power in Souroudougou they used the region as a “slave‐reservoir” (Hubbell 2001). 12 A comprehensive overview of the social network approach and its application to African contact settings can be found in Beyer and Schreiber (2017). 13 Homogeneity is measured by the EI‐index that calculates the proportion of external versus internal ties according to some predefined attribute, in this case ethno‐linguistic background. The number oscillates between ‐1 (all links internal = completely homogeneous) and +1 (all links external = highly heterogeneous).

Contact among African Languages  665

REFERENCES Bender, Lionel M. 1996. The Nilo‐Saharan Languages: A Comparative Essay. Munich: Lincom. Beyer, Klaus 2006. Das Pana im Netzwerk arealer Beziehungen: das Lexikon. In Dymitr Ibriszimow and Kerstin Winkelmann (eds.), Zwischen Bantu und Burkina. Festschrift für Gudrun Miehe zum 65. Geburtstag. Cologne: Rüdiger Köppe, pp. 9–22. Beyer, Klaus 2010. Language contact and change: social factors in an African rural environment, Journal of Language Contact 3: 131–52. Thema III. Beyer, Klaus 2015. Multilingual speakers in a West‐African contact zone: an integrated approach to contact‐induced language change. In Gerald Stell and Kofi Yakpo (eds.), Code‐ switching at the Crossroads between Structural and Sociolinguistic Perspectives. Berlin, New York: de Gruyter Mouton, pp. 237–58. Beyer, Klaus 2018. Motorcycle taxi drivers as the engine room of language change? A report from multilingual Ngaoundéré, Northern Cameroon. Paper held at WOCAL9, 25– 28 August 2018, Rabat, Morocco. Beyer, Klaus 2020. Language contact. In Gerrit J. Dimmendaal and Rainer Voßen (eds.) The Oxford Handbook of African languages. Oxford: Oxford University Press, pp. 858–70. Beyer, Klaus and Raija Kramer (eds.) 2017. Language Contact and Change under Multilingual Conditions. Case studies from Africa (FAB 24 – 2012 (Frankfurter Afrikanistische Blätter)). Cologne: Rüdiger Köppe. Beyer, Klaus and Henning Schreiber 2013. Intermingling speech groups. Morpho‐ syntactic outcomes of language contact in a linguistic area in Burkina Faso, West Africa. In Isabelle Léglise and Claudine Chamoreau (eds.), The Interplay of Variation and Change in Contact Settings – Morphosyntactic Studies. Amsterdam: Benjamins, pp. 107–34. Beyer, Klaus and Henning Schreiber 2017. Social network approach in African sociolinguistics. In Oxford Research Encyclopedia of Linguistics. Available at http://linguistics.oxfordre.com/ view/10.1093/ acrefore/9780199384655.001.0001/acrefore‐978 0199384655‐e‐236. Beyer, Klaus and Henning Schreiber forthcoming. Driving Forces of Language Contact. A Case Study in the Upper Sourou

(Language Contact in Afrika 7). Cologne: Rüdiger Köppe. Caprile, Jean‐Pierre 1978a. Contact de langues et contacts de cultures, vol. 1: Démographie linguistique: approche quantitative. Paris: SELAF. Caprile, Jean‐Pierre 1978b. Contact de langues et contacts de cultures, vol. 2: La situation du Tchad: approche globale au niveau national. Paris: SELAF. Caprile, Jean‐Pierre 1979. Contact de langues et contacts de cultures, vol. 3: La création lexicale spontanée en Afrique centrale par emprunt au français. Paris: SELAF. Caprile, Jean‐Pierre 1982. Contact de langues et contacts de cultures, vol. 4: L’expansion des langues africaines: peul, sango, kikongo, ciluba, swahili. Paris: SELAF. Childs, Tucker G. 2010. Language contact in Africa: a selected review. In Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley‐ Blackwell, pp. 695–713. Chomsky, Noam 1965. Aspects of the Theory of Syntax, Cambridge, MA: MIT Press. Connell, Bruce 2017. Language history and language contact in the Mambiloid region in Cameroon. In Klaus Beyer and Raija Kramer (eds.), Language Contact and Change under Multilingual Conditions. Case Studies from Africa. Cologne: Rüdiger Köppe, pp. 11–129. Diallo, Mohamadou 1988. Eléments de systématique et de dialectologie du marka‐kan (Burkina Faso) (Université Stendhal III Grenoble Ph.D. dissertation). Dik, Simon C. 1978. Functional Grammar (North‐ Holland Linguistic Series 37). Amsterdam: North‐Holland. Dik, Simon C. 1997a. The Theory of Functional Grammar. Part 1: The Structure of the Clause (Kees Hengeveld, ed., 2nd edition) (Functional Grammar Series 20). Berlin: de Gruyter Mouton. Dik, Simon C. 1997b. The Theory of Functional Grammar. Part 2: Complex and derived constructions (Kees Hengeveld ed., 2nd edition) (Functional Grammar Series 20). Berlin: de Gruyter Mouton. Dimmendaal, Gerrit J. 2001. Areal diffusion versus genetic inheritance: an African perspective. In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative

666  Klaus Beyer Linguistics. Oxford: Oxford University Press, pp. 359–92. Dimmendaal, Gerrit J. 2011. Historical Linguistics and the Comparative Study of African Languages. Amsterdam: Benjamins. Dimmendaal, Gerrit J. 2018. Historical linguistics in an African context. A brief state of the art. In Augustine Agwuele and Adams Bodomo (eds.), The Routledge Handbook of African Linguistics. Abingdon: Routledge, pp. 33–54. Dixon, Robert M. W. 1997. The Rise and Fall of Languages. Cambridge: Cambridge University Press. Ehret, Christopher 2001. A Historical‐Comparative Reconstruction of Nilo‐Saharan. Cologne: Rüdiger Köppe. Greenberg, Joseph H. 1959. Africa as a linguistic area. In William R. Bascom and Melville J. Herskovits (eds.) 1959. Continuity and Change in African Cultures. Chicago: University of Chicago Press. pp. 15–27. Greenberg, Joseph H. 1963. The Languages of Africa. The Hague: de Gruyter Mouton. Greenberg, Joseph H. 1983. Some areal characteristics of African languages. In Ivan R. Dihoff (ed.), Current Approaches to African Linguistics, vol. 1. Dordrecht: Foris, pp. 3–21. Güldemann, Tom 2008. The Macro‐Sudan belt. In Bernd Heine and Derek Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, pp.151–85. Güldemann, Tom 2018. The Languages and Linguistics of Africa. Berlin: de Gruyter. Güldemann, Tom and Rainer Voßen 2000. Khoisan. In Bernd Heine and Derek Nurse (eds.), African Languages: An Introduction. Cambridge: Cambridge University Press. Haspelmath, Martin 2004. How hopeless is genealogical linguistics, and how advanced is areal linguistics? Studies in Language 28: 209–23. Haspelmath, Martin 2008. Loanword typology: steps toward a systematic cross‐linguistic study of lexical sility. In Thomas Stolz, Dik Bakker, and Rosa Salas Palomo (eds.), Aspects of Language Contact: New Theoretical, Methodological and Empirical Findings with Special Focus on Romancisation Processes. Berlin: de Gruyter Mouton, pp. 43–62. Heine, Bernd 1968. Afrikanische Verkehrssprachen. Cologne: Infratest. Heine, Bernd 1973. Pidgin‐Sprachen im Bantu‐Bereich. Berlin: Dietrich Reimer. Heine, Bernd and Tania Kuteva 2001. Convergence and divergence in the development of African languages.

In Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press, pp. 393–411. Heine, Bernd and Tania Kuteva 2003. On contact‐induced grammaticalization. Studies in Language 27.3: 529–72. Heine, Bernd and Derek Nurse 2008. A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Hubbell, Andrew 2001. A view of the slave trade from the margin: Souroudougou in the late nineteenth‐century slave trade of the Niger bend, Journal of African History 42: 25–47. Jansen, Dorothea 1999. Einführung in die Netzwerkanalyse: Grundlagen, Methoden, Forschungsbeispiele. Opladen: Leske and Budrich. Jungraithmayr, Hermann 1987. Contact de langues et contacts de cultures, vol. 5: Langues tchadiques et langues non‐tchadiques en contact en Afrique centrale. Paris: SELAF. Kembo‐Sure. 2000. Suba ethnolinguistic revival. In Rainer Voßen, Angelika Mietzner and Antje Meißner “Mehr als nur Worte.” Afrikanistische Beiträge zum 65. Geburtstag von Franz Rottland. Cologne: Rüdiger Köppe Verlag, 639–73. Kossmann, Marten 2005. Berber Loanwords in Hausa. Cologne: Rüdiger Köppe. Köhler, Oswin 1975. Geschichte und Probleme der Gliederung der Sprachen Afrikas: von den Anfängen bis zur Gegenwart. In Hermann Baumann (ed.), Die Völker Afrikas und ihre traditionellen Kulturen, Teil 1: Allgemeiner Teil und südliches Afrika (Studien zur Kulturkunde 34). Wiesbaden: Franz Steiner, pp. 135–373. Kramer, Raija L. 2017. On canines, classes and concordances: linguistic variation and norm development in Fulfulde varieties of Ngaoundéré (Northern Cameroon). In Klaus Beyer and Raija L. Kramer (eds.), Language Contact and Change under Multilingual Conditions (Case Studies from Africa FAB 24 – 2012). Cologne: Rüdiger Köppe, pp. 29–50. Kramer, Raija L. 2018. The role of central actors in distribution processes of linguistic variants: a multiple group analysis of motorcycle garages in Ngaoundéré (Cameroon). Paper presented at WOCAL9, 25–28 August 2018, Rabat, Morocco. Lacroix, Pierre Francis 1971. L’ensemble songhay‐ jerma: problèmes et thèmes de travail. In Maurice Houis (ed.), Actes du 8e Congrès international de linguistique africaine, Abidjan 24–28 mars 1969. Abidjan: Université d’Abidjan, pp. 87–99.

Contact among African Languages  667 Lüpke, Friederike and Anne Storch 2013. Repertoires and Choices in African Languages. Berlin: de Gruyter Mouton. Manessy, Gabriel 1969. Les langues gurunsi. Essai d’application de la méthode comparative à une groupe de langues voltaïques. Paris: SELAF. Manessy, Gabriel 1979. Contribution à la classification généalogique des langues voltaïques. Paris: SELAF. Markgraf, Holger W. 2017. Variation and linguistic change in Mbum. In Klaus Beyer and Raija L. Kramer (eds.), Language Contact and Change under Multilingual Conditions. Case studies from Africa. Cologne: Rüdiger Köppe, pp. 93–109. Meinhof, Carl 1910–11. Sudansprachen und Hamitensprachen. Zeitschrift für Kolonialsprachen 1: 161–6. Meinhof, Carl 1912. Die Sprachen der Hamiten. Hamburg: L. Friederichsen. Meinhof, Carl 1932. Introduction to the Phonology of the Bantu Languages. Berlin: D. Reimer/E. Vohsen. Milroy, Lesley 1987. Language and Social Networks. Oxford: Blackwell Milroy, James and Lesley Milroy 1985. Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339–84. Mous, Marten 2001. Ma’a as an ethnoregister of Mbugu. In Derek Nurse (ed.), Historical Language Contact in Africa. SUGIA 16/7. Cologne: Rüdiger Köppe, pp. 293–320. Mous, Maarten 2003. The Making of a Mixed Language: The Case of Ma’a/Mbugu. Amsterdam: Benjamins. Muysken, Pieter 2008. Introduction. In Pieter Muysken (ed.), From Linguistic Areas to Areal Linguistics. Amsterdam: Benjamins, pp. 1–23. Myers‐Scotton, Carol 1983. The negotiation of identities in conversation: a theory of Markedness and code choice. International Journal of the Sociology of Language 44: 115–36. Myers‐Scotton, Carol 1993a. Social Motivations for Codeswitching: Evidence from Africa. Oxford: Oxford University Pess. Myers‐Scotton, Carol 1993b. Duelling Languages: Grammatical Structure in Codeswitching. Oxford: Oxford University Press. Myers‐Scotton, Carol 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Myers‐Scotton, Carol and Janice L. Jake 2000. Four types of morpheme: evidence from aphasia, code switching, and second‐language acquisition. Linguistics 38: 1053–1100.

Myers‐Scotton, Carol and Janice L. Jake 2016. The 4‐M model revisited: codeswitching and morpheme election at the abstract level. International Journal of Bilingualism 21: 340–66. Nicolaï, Robert 1990. Parentés linguistiques (à propos du songhay). Paris: CNRS. Nicolaï, Robert 2003. La force de choses ou l’épreuve “nilo‐saharienne”: questions sur les reconstructions archéologiques et l’évolution des langues. Cologne: Rüdiger Köppe. Nicolaï, Robert 2009. Language contact, areality, and history: the Songhay question revisited, SUGIA 20: 187–208. Pugach, Sara 2018. A short history of African language studies in the nineteenth and early twentieth centuries, with an emphasis on German contributions. In Augustine Agwuele and Adams Bodomo (eds.), The Routledge Handbook of African Linguistics. New York: Routledge, pp. 15–32. Reh, Mechthild 2000. Migration, polygamy and interethnic marriage as causes of language birth: the Southern Lwoo languages. In Rainer Voßen, Angelika Mietzner, and Antje Meißner (eds.), “Mehr als nur Worte …” Afrikanistische Beiträge zum 65. Geburtstag von Franz Rottland. Cologne: Rüdiger Köppe, pp. 583–95. Rottland, Franz and Okoth Okombo 1992. Language shift among the Suba of Kenya. In Matthias Brenzinger (ed.), Language Death: Factual and Theoretical Explorations with Special Reference to East Africa. Berlin: de Gruyter, pp. 273–84. Schreiber, Henning 2008. Eine historische Phonologie der Niger‐Volta‐Sprachen. Cologne: Rüdiger Köppe. Schuchardt, Hugo 1882. “Zur afrikanischen Sprachmischung,’ Das Ausland, 867–9. Scotton, Carol 1971. Towards a linguistic theory of choosing a lingua franca. Studies in African Linguistics 1, Supplement 2: 109–29. Simons, Gary F. and Charles D. Fenning (eds.). 2018. Ethnologue: Languages of the World, 21st edition. Dallas, Texas: SIL International. Available at www.ethnologue.com. Souag, Lameen 2012. The subclassification of Songhay and its historical implications. Journal of African Languages and Linguistics 33.2: 181–213. Speke, John H. 1863. Journal of the Discovery of the Source of the Nile. London: Blackwoods. Swadesh, Morris 1955. Towards greater accuracy in lexicostatistic dating. International Journal of American Linguistics 21: 121–37.

668  Klaus Beyer Tadmor, Uri, Martin Haspelmath, and Bradley Taylor 2010. Borrowability and the notion of basic vocabulary. Diachronica 27.2: 226–46. Thomason, Sarah G. 2008. Social and linguistic factors as predictors of contact‐induced change. Journal of Language Contact 2.1: 42–65. Thomason, Sarah. G. and Terrence Kaufman 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Wasserman, Stanley and Katherine Faust, 1994. Social Network Analysis: Methods and Applications. Cambridge: Cambridge University Press.

Westermann, Diedrich 1927. Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. Berlin: de Gruyter. Westermann, Diedrich 1935. Charakter und Einteilung der Sudansprachen. Africa 8.2: 129–48. Williamson, Kay 1989. Niger‐Congo Overview. In John Bendor‐Samuel (ed.), The Niger‐Congo Languages. A Classification and Description of Africa’s Largest Language Family. Lanham, MD: University Press of America, pp. 3–45. Wolf, Christof 2010. Egozentrierte Netzwerke: Datenerhebung und Datenanalyse. In Christian Stegbauer, and Roger Haeussling (eds.), Handbuch Netzwerkforschung. Berlin: Springer, pp. 471–83.

34 Contact and Siberian Languages BRIGITTE PAKENDORF

1  Introduction: The Languages and Peoples of Siberia Siberia is the vast geographic area that dominates the Eurasian landmass, bordering on the Ural Mountains in the west, the Arctic Sea in the north, the Sea of Okhotsk and the Pacific Ocean in the east, and northern China, Mongolia, and Kazakhstan in the south. Owing to its severe climatic and ecological conditions, Siberia is extremely underpopulated. Such low population density may have precluded frequent contact among the indigenous ethnolinguistic groups, especially in the past (cf. Stern 2005a: 290). Siberia is therefore not the first region of the world that comes to mind when studying language contact; nevertheless, the indigenous languages show several structural similarities, leading Anderson (2004, 2006) to speak of a “Siberian linguistic macro‐area.” Over 30 languages belonging to eight language families plus one isolate, Nivkh, are spoken in Siberia (Table 34.1, Map 34.1). The language families found in Siberia are (following a rough west to east orientation): Uralic, Yeniseic (nowadays represented by only one highly endangered language, Ket), Turkic, Tungusic, Mongolic, Yukaghir, Chukotko‐ Kamchatkan, and Eskimo‐Aleut. The Yukaghir family (of which nowadays only two moribund languages survive, Tundra Yukaghir and Kolyma Yukaghir) might possibly be distantly related to the Uralic languages (cf. references in Maslova 2003: 1), while Turkic, Tungusic, and Mongolic are sometimes classified as belonging to the Altaic language family (cf. Georg et al. 1999). In addition to the isolate Nivkh, a further isolate, Ainu, used to be spoken in southern Sakhalin, on the Kurile Islands, and on the southernmost tip of Kamchatka. However, following World War II all Ainu‐speakers moved to Japan (de Graaf 1992: 186). The indigenous groups of mainland Siberia were for the most part nomadic hunters and gatherers or semi‐sedentary fishermen. Along the Pacific coast and the Sea of Okhotsk, a number of groups were sedentary hunters of large sea mammals. In the southern steppe zone, on the other hand, cattle and horse pastoralism prevailed; this mode of subsistence was imported to northeastern Siberia in relatively recent times by the Turkic‐speaking Sakha (Yakuts). Russians first entered Siberia in the late sixteenth century, with garrisoned forts established on the Irtysh river in 1586 and 1587, on the Yenisey river in 1604, on the middle Lena in 1632, and on the Anadyr river in 1649 (Forsyth 1992: 34, 36, 79). Further small outposts were scattered in between to aid in the collection of fur tax. During the first centuries of colonization, Russian interference in the life of the indigenous peoples consisted predominantly in the collection of fur tax, the conscription of indigenous peoples into providing transportation for

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

670  Brigitte Pakendorf Table 34.1  The extant languages1 of Siberia and their linguistic affiliation. Family

Languages

Uralic Yeniseic Turkic

Khanty, Mansi, Nenets, Enets, Nganasan, Selkup Ket Siberian Tatar, Chulym Turkic, Tuvan, Tofa, Khakas, Shor, Altai, Sakha (Yakut), Dolgan Buryat Evenki, Even, Negidal, Udihe, Oroč, Nanai, Orok (Ul’ta), Ulča Kolyma Yukaghir, Tundra Yukaghir Chukchi, Kerek, Alutor, Koryak, Itelmen2

Mongolic Tungusic Yukaghir Chukotko‐ Kamchatkan Eskimo‐Aleut Isolate

Central Siberian (Chaplino) Yupik, Naukan Yupik, Aleut Nivkh

Russian officials, as well as superficial Christianization (Gernet 2007: 69–72; Slezkine 1994: 23–4, 32, 43–4, 48–53). Although by the end of the seventeenth century there may have been as many Russian settlers as indigenous peoples in Siberia, these immigrants were concentrated in the more fertile southern districts of western Siberia (Forsyth 1992: 100). In the northern and eastern regions Russians were scarce and often outnumbered by the local people (Forsyth 1992: 101; Stern 2005a: 292). Therefore, a knowledge of Russian among the indigenous groups was not very widespread during the tsarist period of colonization (cf. Matić 2008: 100; Burykin 1996: 994). That situation changed, however, after the establishment of Soviet rule in the 1920s. In the initial years the Soviet state encouraged the maintenance of the indigenous languages, and a number of orthographies were created for the unwritten languages of Siberia. However, at a later period, especially in the 1960s and 1970s, language policies changed drastically, and children of indigenous minority peoples were forcibly taken to boarding schools where they were forbidden to speak their native languages. Furthermore, in the 1950s and 1960s, small settlements all over Siberia were closed down and their inhabitants relocated to larger, and often ethnically and linguistically mixed settlements, leading to increased use of Russian in daily life (see for example Krupnik and Chlenov 2007 for a description of such relocations among the Yupik). In addition, after World War II large numbers of settlers from the European parts of the Soviet Union (especially Russians, Ukrainians, and Belorussians) came to Siberia to exploit the natural resources, so that the indigenous peoples were greatly outnumbered by the settlers (Forsyth 1992: 360, 361, 405). All of these factors led to a large‐scale Russification of all spheres of life (Helimski 1997: 77; de Graaf 1992: 190, 191; Anderson 2005: 125–7; Pakendorf and Aralova 2018). Nowadays, the majority of Siberian indigenous languages are moribund, with only a few elderly speakers remaining, and no more acquisition by children (Vaxtin 2001: 163–80; see Kazakevič and Parfënova 2000: 283–5; Nikolaeva and Tolskaya 2001: 25–6; Maslova 2003: 2; Morgounova 2007: 193; Harrison and Anderson 2008: 245–6; Oskolskaya and Stoynova 2013; Siegl 2013: 18–21; Gruzdeva 2015: 156–8; and Pakendorf and Aralova 2018 for individual linguistic groups). Only relatively isolated dialect communities, such as that of the Lamunkhin dialect of Even spoken in central Yakutia (Pakendorf 2009), and a few of the larger ethnic groups, for instance the Turkic‐speaking Sakha (Pakendorf 2007: 2) or the Samoyedic‐speaking Nenets (Ljublinskaja 2000: 312; Vaxtin 2001: 163), have been able to maintain their heritage language in a viable state.

Map 34.1  The approximate distribution of the languages of Siberia (map © MPI‐EVA).

672  Brigitte Pakendorf

2 Russian Influence on the Indigenous Languages of Siberia As mentioned in Section  1, several factors have led to the widespread use of Russian among speakers of indigenous Siberian languages: first, since Russian was the predominant language in the Soviet Union, and is the language used in practically all spheres of public life in the Russian Federation, a good knowledge of Russian was and is expected to lead to upward social mobility and better job chances (Comrie 1989: 146; Kazakevič and Parfënova 2000: 288). Second, Russian functions as a lingua franca between individuals from diverse ethnolinguistic groups, and is used as the medium of communication in mixed marriages, even when it is not the first language of either spouse (Comrie 1989: 146; Pakendorf and Aralova 2018). Furthermore, since the late 1930s schooling has mainly been in Russian, which has in many cases led to a complete break in transmission of the native language. Last but not least, speakers of minority languages were frequently encouraged, more or less officially, to give up their language for a bigger one, often Russian (Comrie 1989: 148; Kibrik 1991: 10). It therefore comes as no surprise that the indigenous languages of Siberia show marked Russian influence. All of them exhibit a large number of lexical copies3 from Russian, with phonological differences depending on the time of copying. In the early, pre‐revolutionary period of contact, relatively few items were copied into the indigenous languages; these were predominantly designations of novel cultural items such as “bread” or “tea” and were adapted to the phonological system of the recipient language. During the Soviet era, on the other hand, a large number of Russian copies entered the indigenous languages, mostly without any phonological adaptation (Comrie 1996: 36; Kaksin 1999: 221–2; Grenoble 2000: 106; Nevskaja 2000: 285; Malchukov 2003: 237; Matić 2008: 103–4; Pakendorf and Novgorodov 2009: 510; Vajda 2009: 481, 484). In addition to importing a large number of lexical items from Russian, the indigenous languages of Siberia have also undergone structural changes that can be traced to Russian influence. Thus, a shift can be observed in the use of some cases, for example the use of the instrumental instead of the dative case to mark the overt agent of passive constructions in Evenki and Khakas (Gladkova 1991: 68; Grenoble 2000: 109; Anderson 2005: 172), the use of the dative instead of the allative to mark the addressee of verbs of speech in Evenki 4 (Gladkova 1991: 68; Grenoble 2000: 109), as well as the development of dative case‐marked experiencer subjects and the extension of the dative case to mark direct objects in Ket (which lacks the accusative case used for this purpose in Russian; Minayeva 2003: 48, 50–1). In Nivkh, Russian influence has led to the development of number agreement on nouns after numerals and on verbs after plural subjects, as well as to new imperative forms partly calqued on Russian (Gruzdeva 2015: 164–5, 168–73). The most salient structural changes undergone by Siberian languages in contact with Russian are in the domain of syntax. Thus, a shift toward a less strict verb‐final word order has been noted in some Tungusic languages (Malchukov 2003: 241; Grenoble 2000: 107–8; Gladkova 1991: 68), in Nivkh (Gruzdeva 2015: 175–7), in Khakas (Anderson 2005: 222), and in Dolgan (Stapert 2013: 262–5). Instead of the previously widespread use of parataxis, coordinate sentences joined with conjunctions copied from Russian have been documented in Samoyedic languages (Bátori 1980: 144), in Evenki (Grenoble 2000: 115, 2009: 150–3), and in Dolgan (Stapert 2013: 292–3). Finite subordinate clause constructions copied from Russian are increasingly replacing the indigenous use of case‐ marked participles or converbs (cf. Anderson 2004: 69–72), for example in the Turkic

Contact and Siberian Languages  673 languages Shor (Nevskaja 2000: 286), Khakas (Anderson 2005: 196–221), Tofa (Harrison and Anderson 2008: 260–1), and Dolgan (Stapert 2013: 301–2, 305), in the Tungusic languages (Malchukov 2003: 241; Grenoble 2000: 116–18; Gladkova 1991: 68), in Yukaghir (Matić 2008: 117–19), and in Enets (Sorokina 1991: 66–7; Khanina and Shluinsky 2008: 71–3). These copied constructions make use of indigenous adverbials as complementizers or conjunctions, but use of conjunctions and complementizers copied from Russian has been documented as well (1a). The formation of relative clauses with the use of interrogative pronouns as relativizers (1b) has been described for Evenki (Malchukov 2003: 241), for Khakas (Anderson 2005: 205–9), and for Tofa (Harrison and Anderson 261). Interestingly, Forest Enets appears to be developing finite relative clauses not with an interrogative pronoun, but with a demonstrative functioning as relativizer (Khanina and Shluinsky 2008: 70–1). (1) a. Yukaghir (Matić 2008: ex. 33; taken from Nikolaeva 2004: 29.49) jesli  Germanija  kejdej‐te‐j […]  taŋnugi er‐če if   Germany  advance‐FUT‐INTR.3SG    then   bad‐ATTR modol oː‐te‐j life COP‐FUT‐INTR.3SG ‘If Germany wins […] then life will be bad …’ cf. Russian Jesli  Germanija    pobedit […]  žizn’  budet   ploxoj if   Germany  win.FUT.3SG  life   be.FUT.3SG  bad.INS.F ‘If Germany wins, life will be bad.’ cf: uninfluenced Yukaghir (Matić 2008: ex. 31); taken from Nikolaeva 2004: 37.4–5) touke čugø  l’e‐de‐jne […]   odul‐ŋin   qon‐te‐jek dog trace  COP‐3‐DS.COND.CVB Yukaghir‐DAT      go‐FUT‐2SG.INTR ‘If there are dog traces there […] you will marry a Yukaghir.’ b. Evenki (Malchukov 2003: ex. 6b) i‐le   hurkeken  suru‐re‐n   gorot‐tu … where‐ALL  boy  go‐NFUT‐3SG   town‐DAT ‘In the town where the boy is going …’ cf: Russian v   gorode,   kuda     idët    mal’čik … in town.PREP where.ALL go.PRS.3SG boy ‘In the town where the boy is going …’ cf: uninfluenced Evenki (Malchukov 2003: ex. 6a) hurkeken suru‐mečin‐du‐n    gorot‐tu … boy    go‐FUT.PTCP‐DAT‐POSS.3SG  town‐DAT ‘In the town where the boy is going …’ Note that not only the use of conjunctions and relative pronouns has been copied from Russian, but so has the use of finite verbs in subordinate clauses. Thus, the impact of Russian on the languages of Siberia is leading to a gradual typological shift (cf. Anderson 2017: 645). However, while Russian influence has changed the indigenous languages of Siberia, indigenous languages have also had an impact on Russian spoken in the region. For instance, several cases of shift of Russian settler communities to indigenous languages, especially Sakha, are known (Sunderland 1996: 815; Stern 2009: 392), and phonological influence of

674  Brigitte Pakendorf indigenous languages on individual Russian dialects has also been recorded (Sunderland 1996: 815–16; Krasovicky and Sappok 2000; Schweitzer et al. 2013: 424).

3  Pidgins and Mixed Languages in Siberia Only two Russian‐based pidgins have been recorded in Siberia: Chinese Pidgin Russian (also known as Siberian pidgin, Kyakhta pidgin, or the Maimachin speech) spoken previously in the Chinese–Russian border town of Kyakhta as well as along the Lower Amur, and Taimyr Pidgin Russian (also known as Govorka) spoken on the Taimyr Peninsula (Stern 2005a; Perekhvalskaya 2013). In addition, in the nineteenth century a number of trade jargons may have existed in Chukotka involving Chukchi, Yupik, and English, which were used for communication between Chukchi and Yupik, as well as with sailors of whaling or expedition ships (de Reuse 1994: 319–29). The scarcity of pidgins in Siberia as compared to other colonies can be explained by the fact that the Russians did not relocate individuals from different ethnolinguistic groups for purposes of forced labor, so that there was no occasion for a system of interethnic communication to arise spontaneously (Stern 2005a: 289). And by the time people were resettled in linguistically mixed villages in the mid twentieth century, access to Standard Russian as a lingua franca was ensured through obligatory schooling in Russian. Chinese Pidgin Russian was initially the language used by Chinese and Russian traders in the trading towns of Kyakhta and Maimachin from the late eighteenth to the middle of the twentieth century (Stern 2005b: 178; Perekhvalskaya 2013: 69). A derivative of this pidgin was also spoken in Harbin, and it later spread to the Lower Amur region (Perekhavalskaya 2013: 70), where it played a role in the development of the pidginized Russian spoken by local Tungusic peoples (Khasanova 2000: 182, 193; Perekhvalskaya, forthcoming). Chinese Pidgin Russian is characterized by large‐scale insertion of epenthetic vowels to maintain the CV syllable structure characteristic of Chinese, loss of case and gender‐marking, loss of number agreement, and a complete lack of inflection on verbs, which instead are used in the Russian imperative form. There are three tense‐aspect markers that follow the verb: budu/budi (future), la (perfective), and esa/esi/ju (perfective or evidential); habitual present and past imperfective are zero‐marked. Budu/budi and esa/esi are inflected forms of the Russian copula ‘to be’, whereas ju is of Chinese origin (Perekhvalskaya 2013: 72). The lexicon is predominantly of Russian origin, with a significant number of Chinese words as well as some items from Mongolian and Tungusic languages. There is a significant amount of variation among the sources, however, depending on the speaker’s first language, and only about 100 words appear to make up a common core of lexicon (Perekhvalskaya 2013: 74–5). In contrast to Chinese Pidgin Russian, which developed as a trade language, Taimyr Pidgin Russian was predominantly developed and used for interethnic communication by Nganasans and the ethnically mixed communities that had developed along the Khatanga Trading Way. These latter comprised Russians, Sakha, and Evenks, and were the core of what later became the Dolgan ethnic group (Stapert 2013: 77–8). During winter extensive visits of nomadic groups such as the Nganasans in the settlements along the trading way would have favored the development as an intergroup language (Stern 2009: 393).5 Nowadays the lexicon of Taimyr Pidgin Russian consists mainly of Russian words; however, previously there may have been a large number of lexical items from Dolgan (Ubrjatova 1985: 68). This pidgin is characterized by a lack of case marking, with one predominant postposition mesto ‘place’ marking non‐core

Contact and Siberian Languages  675 arguments. A sociative marker meste (derived from Russian vmeste ‘with’) also exists; often mesto and meste are used interchangeably. A further postposition toroba (derived from Russian storona ‘side’) marks location (Stern 2005a: 301). In contrast to Chinese Pidgin Russian, in which verbs are uninflected, Taimyr Pidgin Russian shows some verbal inflection. Even in the “basilectal” system, which has been less influenced by Standard Russian, verbs take person marking; however, there is no strict agreement with the subject. Rather, the third person singular and first person plural forms predominate, while second person singular forms are rarely used (Stern 2005a: 309); third singular forms replace second person singular and third plural forms (Stern 2009: 384). Furthermore, in the past tense the gender distinction is lost, with 95% of verbs with a singular subject taking the masculine past tense form ‐l and the remaining 5% taking the plural form ‐li (Stern 2009: 382). Another difference between the two pidgins is that in Taimyr Pidgin Russian personal pronouns are based on the Russian genitive‐accusative forms, while in Chinese Pidgin Russian they derive from Russian possessive pronouns. Some of the salient differences between the two pidgins are illustrated in the following examples, with glosses adapted according to the Standard Russian form. (2) a. Chinese Pidgin Russian (Perekhvalskaya 2013: 74, ex. 19) jevó dúmaj majá jevó céna 3SG.POSS think[IMP] 1SG.POSS 3SG.POSS price daváj give[IMP] ‘He thinks that I give him the (real) price.’ b. (Perekhvalskaya 2013: 75, 1st sentence of text) kurica jajcy kupi‐la butyka apuskaj‐la chicken egg buy[IMP]‐PFV bottle put.down[IMP]‐PFV ‘… he bought chicken eggs and put them into a bottle.’ (3) a. Taimyr Pidgin Russian (Stern 2005a: ex. 56) taperja  menja  budem šamanit’ now 1SG[ACC]  will[1PL]  act.as.shaman[INF] ‘Now I will act as shaman.’ b. (Stern 2005a: ex. 10) nganasan tut baba utrom   mesto govorit in.the.morning Nganasan here woman place  say[PRS.3SG] ‘On the following morning that Nganasan says to his wife.’ Only one contact language in Siberia emerged as the result of relocation of peoples for labor purposes: Copper Island Aleut (CIA). This mixed language with a predominantly Aleut lexicon is characterized by Aleut noun inflection, derivational morphology, and nonfinite verb inflection, but by Russian finite verb morphology and pronouns (Thomason 1997: 457, 460). It arose on Copper Island, one of the Commander Islands off the coast of Kamchatka, which was uninhabited when discovered in 1741. In 1826 the Russian‐American Company settled Aleuts on the Commander Islands to work in the seal‐slaughtering trade along with Russian employees. A population called “creoles”6 arose at an early stage of the island’s settlement out of the union of Aleut women and Russian men (Thomason 1997: 451). These creoles were a socially and economically distinct group – they had a different legal status from and were better off economically than the Aleuts, but were looked down upon socially by both the Russians and the Aleuts since they were of illegitimate birth, at least in the early period (Thomason 1997: 453–4).

676  Brigitte Pakendorf Like other Aleut dialects, CIA has only two cases (absolutive and relative), possessive suffixes, singular, dual and plural number on nouns, and no gender distinctions. It has two sets of pronouns, derived from Aleut and Russian, which are used in distinct constructions: the Aleut pronouns are restricted to reflexive verbs, while Russian pronouns occur as subject markers, and in their accusative form have replaced the original Aleut objective conjugation of the verb (Golovko 1996: 70–1). The most notable difference between CIA and other Aleut dialects is the system of finite verbal inflection, which in CIA derives entirely from Russian. In the present tense, verbs take Russian portmanteau suffixes for each person‐number combination; in contrast to the nominal system, a dual number is lacking for verbs. In the past tense, the Russian past tense marker –l is used (Thomason 1997: 458–9). The following examples demonstrate the use of Russian pronouns and finite verb markers in CIA (4a, 5a) in comparison with Bering Island Aleut (4b, 5b). (4) a.

Copper Island Aleut (Golovko 1996: ex. 18) ˆ ona hixta‐it čto ona 3SG.NOM.F say–PRS.3SG that 3SG.NOM.F Russian: ona govor‐it čto ona 3SG.NOM.F say‐PRS.3SG that 3SG.NOM.F ‘She says that she loves him.’ b. Bering Island Aleut (Golovko 1996: ex. 19) ˆ ilaxta‐ku‐u love‐REAL‐3SG.OBJ.3SG.SBJ ‘S/he loves him/her/it.’

ego 3SG.ACC.M ego 3SG.ACC.M

ˆ ilaxta‐it love‐PRS.3SG ljub‐it love‐PRS.3SG

(5) a.

Copper Island Aleut (Golovko 1996: ex. 20) ˆ ty menja hamayaaxta‐iš 2SG.NOM 1SG.ACC ask‐PRS.2SG Russian: ty menja sprašiva‐eš 2SG.NOM 1SG.ACC ask‐PRS.2SG ‘You ask me.’ b. Bering Island Aleut (Golovko 1996: ex. 20) ˆ ˆ xt ting ahmayaaxta‐ku‐ 1SG.OBJ ask‐REAL‐PRS.2SG ‘You are asking me.’ CIA must have arisen between the period of initial settlement of Copper Island in 1826 and approximately 1900. It most probably arose before the demise of the Russian‐American company in 1867, which led to the departure of most of the Russians from the Commander Islands and to the end of the special social and legal status of the creoles (Thomason 1997: 461, 465). This mixed language must therefore have arisen in a very short time, in at most two generations. It probably did not arise as a pidgin, because neither the Aleut nor the Russian component is simplified. Not much is known about the use of Aleut and Russian on Copper Island in the early years of its settlement; however, the creole population was probably fluent in both languages, and it may well be that the long‐term Russian settlers knew Aleut (Thomason 1997: 462–3). The most likely explanation for the development of CIA is that it arose in a setting of bilingual code‐switching, with some “creative decisions” by the speakers themselves as to what form the final product would take (Thomason 1997: 464–5; Golovko 2003: 190–8). In this, CIA differs from Taimyr Pidgin Russian and Chinese Pidgin Russian, which arose as a means of communication in the absence of a common language between the groups in contact.

Contact and Siberian Languages  677

4  Language Contact Among the Indigenous Languages Little is known about contact between different ethnolinguistic groups before Russian colonization, which started at the turn of the sixteenth and seventeenth centuries in western Siberia. Sporadic warfare and territorial conflicts, exacerbated by the upheavals following Russian colonization, are known to have taken place between different peoples of Siberia (Forsyth 1992: 11, 58, 80; de Reuse 1994: 296; Slezkine 1994: 27–8); these often resulted in the capture of women from the defeated enemy (Forsyth 1992: 67; Slezkine 1994: 6, 44). Some trade relations existed in the eighteenth and nineteenth centuries between the nomadic reindeer‐herding Chukchi and their neighbors, from the Yukaghirs, Evens, and Sakha (Yakuts) in the west to the Yupik in the east (de Reuse 1994: 296, 307; Maslova and Vaxtin 1996: 999), as well as between the coastal Chukchi and Koryaks and their reindeer‐breeding compatriots from the interior (Forsyth 1992: 72). In the nineteenth century, the Turkic language Sakha played an important role as a vehicular language in large areas of northeastern Siberia (Wurm 1996: 976), while in Chukotka and Kamchatka Chukchi was in use for interethnic communication by Yupik, Evens, and Kereks (de Reuse 1994: 296; Burykin 1996: 990). In the Lower Amur region a system of exogamic clans existed that encompassed Nivkh as well as Tungusic Nanai, Negidals, and Evenks; each group comprised clans of linguistically foreign origin. The factors that governed language choice in this area of continuous intermarriage are not known, although it appears to have been determined partly by the language spoken by the numerically preponderant group (Starcev 2014). For the Taimyr Peninsula, Khanina and Meyerhoff (2018) reconstruct Enets‐Nganasan and Enets‐Nenets bilingualism as well as a more restricted knowledge of Russian and Evenki by Enets for the second half of the nineteenth and early twentieth centuries. Multilingualism is also recorded for speakers of the Eskimo languages Naukan Yupik and Sireniki, who in addition to knowing Chukchi also spoke the Imaklik dialect of Inupiaq and Central Siberian (Chaplino) Yupik, respectively (de Reuse 1994: 306). Yukaghir–Even–Sakha–Chukchi quadrilingualism existed in northeastern Yakutia from the nineteenth century, and perhaps earlier, up to the 1940s (Maslova and Vaxtin 1996: 999). However, it is not known to what extent such multilingualism would have been characteristic of interethnic relations in precolonial times as well. Some cases of language shift have been documented, such as the shift of Samoyedic and Yeniseic speakers to Turkic languages in South Siberia, and the shift of Evenks to Buryat (Forsyth 1992: 23; Anderson 2004: 6; Slezkine 1994: 28; Cydendambaev 1981; Čimitdoržieva 2004). Nowadays, speakers of Evenki and Even dialects in the Republic of Sakha (Yakutia) are under strong influence from the locally dominant language Sakha, leading to numerous contact‐induced changes in the Tungusic languages and language shift to Sakha (Malchukov 2006; Pakendorf 2009). Lexical copying among different languages has been documented over the whole geographical area (cf. Anderson 2004: 21–4 for an overview and further references), and especially among the Mongolic, Tungusic, and Turkic languages. Some of these copies can be traced back to the proto‐languages of the respective families, while others are clearly the product of more recent contact between individual languages; occasionally, lexemes can be shown to have been copied into a language at different times in its history (cf. Anderson, 2020, for an extensive summary and references). Furthermore, Anderson (2004, 2006) speaks of a linguistic area with respect to the languages of Siberia (restricted to northeastern Siberia in Anderson 2017). Typological features well known to be shared by a number of the languages are a system of vowel harmony, agglutinative morphology, relatively large case systems, predominantly SOV word order, and the widespread use of converbs or case‐marked participles to mark subordination (Anderson 2004: 36–40, 65–69, 2006; Comrie 1981: 59, 71, 117, 244, 246, 258). Among other features described by Anderson as characterizing the

678  Brigitte Pakendorf Siberian linguistic area are a four‐way distinction between labial, alveodental, palatal, and velar nasals, a morphologically marked reciprocal voice, a distinction between a comitative and an instrumental case, and a distinction between a dative and an allative case (Anderson 2006: 268–73, 279–92). However, the distinction between an allative and a dative case proposed by Anderson appears to be characteristic of the Tungusic language family alone, not a widespread areal feature. Apart from the Tungusic languages, only a few languages at the margins of the geographic area show this distinction; in contrast, the majority of languages spoken in Siberia use only one case to mark both indirect objects, addressees of verbs of speech, and goals of motion. It might therefore be preferable to speak of the lack of a distinction between a dative and allative case as being typical of this area. A well‐described case of changes induced by contact with a neighboring indigenous language is the influence of Chukchi on neighboring Yupik languages (de Reuse 1994: 295–455). Central Siberian Yupik (CSY) copied over 200 lexemes from Chukchi, more than half of which are particles. The influx of these particles has had an impact on the use of the inherited Yupik postbases, i.e. derivational morphemes with very specific meanings. On the one hand, CSY has been enriched by the Chukchi particles, which frequently add extra “expressivity” or emphasize particular nuances of the postbases. On the other hand, the influx of Chukchi particles has also led to a loss of postbases overall as well as to a loss of productivity of retained postbases (de Reuse 1994: 421–52). For example, the language has several particles of Chukchi origin that express that the time is ripe to act, such as enta ‘let’s go, time to go!’, kergam ‘ready to act’, legan, weni ‘go right ahead’, and yeqay ‘let’s do it!’, and it is notable that the CSY postbase ‐yaghqaaghte‐ ‘to be time to V’ is rarely used (de Reuse 1994: 448). Although Ket, the last surviving language of the Yeniseic family, is still typologically radically different from the majority of Siberian languages, Vajda (2009: 484–91) suggests that it has undergone a typological shift in the direction of the surrounding Samoyedic7 languages and Evenki through centuries‐long contact. For instance, the initial tonal prosody has changed partially into a system of non‐phonemic word accent, with only the monosyllabic root retaining the full range of phonemic pitch distinctions. Similarly, Ket has developed postposed case suffixes and enclitics out of relational nouns, and the shift of the semantic head of the verb to one of the leftmost slots in the verb template led to the development of a more suffixing verb model: “The realignment of the phonological verb’s semantic head to the extreme left edge served to accommodate the original Yeniseian prefixing structure to the pattern of suffixal agglutination prevalent in all of the neighboring languages” (Vajda 2009: 491). This restructuring took place in the absence of copied morphology, and Ket has also shown itself to be quite resistant to the copying of lexemes: less than 10% of over 1,000 studied words could be shown to be copies, and among these, 90% were copied from Russian during the twentieth century (Vajda 2009: 481–3). One of the most striking contact‐induced changes to be found among Siberian languages is the copying of four verbal TAM plus subject‐agreement paradigms from the Turkic language Sakha (Yakut) into the Lamunkhin dialect of the Tungusic language Even, which is spoken in central Yakutia (Pakendorf 2009, 2015). The paradigms concerned are the necessitive (6a) and the assertive (6b), which are established copies, as well as the present indicative and the hypothetical, which are still in the process of being copied. It is highly probable that these paradigms were transferred in a process called “direct affix borrowing” by Seifart (2015), indicating that full bilinguals are able to identify suffixes in one language and productively use them with stems in the other (Pakendorf, 2019). (6) a. Lamunkhin Even (RDA_TPK_spirits_005) tar em‐niʤur DIST come‐ANT.CVB.PL

tar DIST

tor‐du earth‐DAT

Contact and Siberian Languages  679 e‐jekteːk‐kin kuːnin‐na NEG‐NEC.Y‐PRED.2SG.Y scream‐NEG.CVB ‘Having arrived, in that place you mustn’t shout.’ b. Lamunkhin Even (IVK_memories_154) amm‐u ọttọn karabi‐ń ʤa‐j father‐POSS.1SG DP.Y carbine.R‐AUG‐PRFL.SG ia‐j‐dag‐a nugaha‐j‐dag‐a=di HESIT‐CONN‐ASS.Y‐3SG.Y take.out‐CONN‐ASS.Y‐3SG.Y=EMPH.Y ‘Of course my father took out his gun.’ Interestingly, the assertive paradigm was also copied from Sakha into the (now probably extinct) Učur dialect of Evenki (Myreeva 1964: 51; Malchukov 2006: 126 (7)). Here, the mood has both the assertive as well as the presumptive meaning which it carries in Sakha (Pakendorf 2014: 289–90), in contrast to the purely assertive meaning found in Lamunkhin Even. Furthermore, in Učur Evenki the Sakha suffixes appear to be attached  to  truncated Evenki non‐future forms, as shown by the ‐r that precedes them (Myreeva 1964: 51). (7) Učur Evenki (Pakendorf 2014: ex. 1b, taken from Myreeva 1964: 51) suː goro‐li‐r.dakkit 2PL far‐VR‐ASS.2PL.Y ‘You probably went far.’ Other examples of changes in Evenki and Even dialects that were probably induced by contact with Sakha are the optional replacement of the bilabial fricative [β] with [b], patterns of consonant assimilation, the use of personal pronouns as possessive pronouns, the lack of agreement between modifiers and their head nouns, and patterns of argument marking of some verbs, such as the use of the dative instead of allative case for addressees of verbs of speech (Malchukov 2006)  –  a change that has also been attributed to Russian contact (Section 2). Likewise, the use of the stem beje instead of the reflexive pronoun meːn in logophoric function can be attributed to Sakha contact influence, since Sakha uses the possessive‐ marked pronoun beje ‘self’ in this function. It is unclear whether this usage in Even is a calque from the Sakha construction using the Even noun bej ‘man, human’, as suggested by Malchukov (2006), or whether beje in these constructions is perhaps an actual copy from Sakha. Further substance copies from Sakha are the ordinal suffix ‐(i)s, e.g. il‐is ‘the third’ instead of eastern Even il‐i, as well as numerous modal adverbs and particles, such as araj ‘suddenly’, badaga ‘probably’, and the very frequent discourse particle buolla (Malchukov 2006; Pakendorf 2009: 89; 2015: 70–1). Although Sakha is nowadays the dominant indigenous language in Yakutia and is exerting pressure on the neighboring Tungusic languages, it has itself undergone some contact‐induced changes under the influence of Evenki during its history. These are mainly of a structural kind: only 1% of over 1,400 lexemes in Sakha can be shown to have a probable Evenki origin (Pakendorf and Novgorodov 2009: 504–9), and these refer mainly to natural phenomena, such as χočo ‘valley’ or turaːχ ‘crow’. This lack of lexical copying from Evenki contrasts with the large numbers of copies of Mongolic origin in Sakha: 11–13% of lexemes (Pakendorf and Novgorodov 2009: 509) as well as 18 derivational and one inflectional suffix have a Mongolic origin (Pakendorf 2015). Among the structural changes in Sakha probably induced by Evenki contact are the loss of the Turkic genitive case, the development of an indefinite accusative meaning of the partitive case, the retention of a distinction between the comitative and instrumental, extensive pragmatic uses of possessive suffixes, the development of a suffix that marks terms for

680  Brigitte Pakendorf kin and friends who are not in a relationship to the speaker, and the development of a distinction between an immediate future and remote future imperative. In all of these cases, the Evenki influence was purely structural, with no actual forms being copied (Pakendorf 2007: 95–270). Interestingly, language shift of entire groups of Evenks to Sakha is highly improbable in light of genetic evidence (Pugach et al. 2016). On the other hand, Y‐chromosomal analyses indicate that only a small group of Sakha paternal ancestors settled on the Lena river 500–1,300 years ago (Pakendorf et al. 2006). It is thus possible that the small group of immigrating Sakha pastoralists were dependent on the indigenous Evenks, at least until they had adapted to the new environment. This might have led to a degree of bilingualism in Evenki among Sakha speakers, which might explain the contact‐ induced changes in Sakha in the absence of shift to Sakha by Evenki speakers (Pakendorf 2007: 317–23). Dolgan, a daughter of Sakha that developed along the Khatanga Trading Way out of the intermarriage and assimilation of Evenks, Russians, and Sakha (Stapert 2013: 77–8), has undergone further changes under Evenki contact influence. For instance, Dolgan appears to have calqued its frequent use of the ablative‐marked distal demonstrative onton to express coordination from Evenki, which uses the ablative‐marked distal demonstrative taduk in precisely the same manner (Stapert 2013: 282–6). Evenki influence is also likely to have played a role in the increased use of the habitual aspect in Dolgan as compared to Sakha. The habitual participle ‐AːččI occurs nearly exclusively as a verbal suffix in Dolgan (ex. 8a), whereas in Sakha it has a nominalizing function in nearly 25% of its uses (ex. 8b). Since the western dialects of Evenki, which would have played a role in the formation of Dolgan, make frequent use of the habitual aspect, Evenki influence may have increased the verbal use of this aspect in Dolgan (Stapert 2013: 209–38). (8) a. Dolgan (Stapert 2013: 216, ex. 6.10) on‐tu‐gun bieχ ile   kül‐eːčči‐bin that‐DER‐ACC.2SG always laugh‐HAB‐PRED.1SG really ‘I always really laugh at that.’ b. Sakha (Stapert 2013: 216, ex. 6.8) dʒe mama‐bar haːmaj tireχ well mother‐DAT.1SG the.most support buol‐an χaːl‐l‐ïm kömölöh‐öːččü AUX‐SQ.CVB remain‐PST‐POSS.1SG help‐HAB buol‐an χaːl‐l‐ïm AUX‐SQ.CVB remain‐PST‐POSS.1SG ‘Well I remained my mother’s biggest support, I remained her helper.’ A striking instance of Evenki influence in Dolgan is the restructuring of the system of kinship terminology. Traditional Sakha has a very extensive system of terms for brothers and sisters (which is now falling out of use), with seven different terms distinguishing the sex of ego and the sex as well as relative age of the sibling; only for the younger sister of a male and female ego is there no distinction in terms. Dolgan, in contrast, has only three terms: one for older brother, one for older sister, and one for younger siblings ­irrespective of their sex; the sex of ego is irrelevant. Whereas Dolgan uses inherited Sakha terms, the partitioning of the semantic space perfectly matches that of Evenki (9). Similar restructuring of Dolgan kinship terminology under Evenki influence can be shown to have taken place in the terms for uncles and aunts, parents‐in‐law, and spouses (Stapert 2013: 136–44).

Contact and Siberian Languages  681 (9) Sibling terms in Dolgan compared to Sakha and Evenki (from Stapert 2013: 136)

older brother of M older brother of F older sister of M older sister of F younger brother of M younger brother of F younger sister of M younger sister of F

Sakha

Dolgan

Evenki

biː ubaj eʤij aɣas ini surus balïs balïs

ubaj ubaj eʤij eʤij balïs balïs balïs balïs

akiːn akiːn ekiːn ekiːn nekuːn nekuːn nekuːn nekuːn

Sakha is not the only language of Siberia to have developed a distinction between an immediate future and a remote future imperative under probable Evenki and Even influence (Pakendorf 2013: 216–18). A similar distinction is made in the Samoyedic language Nganasan, in Dolgan, in Yukaghir, and in the Mongolic languages Buryat and Dagur. All of these languages are currently or were historically in contact with Evenki or Even, and they are all the sole members of their respective language families to make such a distinction. In contrast, the Southern Tungusic languages Nanai, Orok, and Ulča also have a remote future imperative, making an origin in the Tungusic family quite probable. The most direct evidence for Northern Tungusic contact comes from the Mongolic language Dagur, which has long been spoken in contact with Solon Evenki in Inner Mongolia. Dagur developed a so‐called “indirect imperative” with a meaning of delayed action and politeness, e.g. yau–gaːm–miny [go–PURP–POSS.1SG] ‘I will go later; let me go later!’ The suffix used for this future imperative is the purposive converb, and, as in purposive constructions, it can take reflexive possessive suffixes as agreement markers for the second person (Tsumagari 2003: 143–4, 146). The use of the purposive converb with the reflexive possessive suffix as a future imperative marker is clearly a copy of the future imperative construction found in Evenki (cf. Tsumagari 2003: 144), since the second person remote future imperative marker in Evenki is identical to the purposive converb suffix, and agreement is achieved by the reflexive possessive suffixes. However, in contrast to Evenki, in Dagur the future imperative uses the purposive converb plus possessive suffixes for all person‐number combinations. Evenki influence can also be shown to have played a role in other features of the verbal domain in Buryat and Dagur, namely in the development of verbal subject agreement and of person marking on converbs. While most modern‐day Mongolic languages lack verbal subject agreement, some show varying degrees of person marking. Notably, person‐number agreement is obligatory and marked by suffixes derived from personal pronouns in Buryat and Dagur. These languages have been in close contact with Evenki and its close sister Solon, respectively. In Evenki, verbs obligatorily agree in person and number with their subject, and it is probable that the optional subject‐agreement marking attested in historical Mongolic documents from the thirteenth to the eighteenth century became fixed in Buryat and Dagur under Evenki/Solon contact (Pakendorf 2013: 212–16). Similarly, Evenki influence is likely to have led to the development of person‐marked converbs in Buryat. Like other Tungusic languages, Evenki has an elaborate system of converbs that function in coordination and subordination and also participate in reference‐ tracking. Same‐subject (SS) converbs occur only in subordinate clauses with a subject coreferential with that of the main clause, different‐subject (DS) converbs occur only in subordinate clauses whose subject is non‐coreferential with that of the main clause, and variable‐subject (VS) converbs can occur both in subordinate clauses with a coreferential and in

682  Brigitte Pakendorf clauses with a non‐coreferential subject (Nedjalkov 1995: 445; 1997). SS converbs do not take any person agreement markers, with the exception of the plural suffix ‐l (10a). The DS and VS converbs, on the other hand, obligatorily agree in person and number with the subject of the subordinate clause. This is accomplished by the use of possessive suffixes when the subordinate subject is non‐coreferential with the main clause subject (10b), and by the use of reflexive possessive suffixes when they are coreferential (i.e. with VS converbs; 10c). (10) a. Evenki (Nedjalkov 1995: ex. 7, 8a, 8b) ʤu‐la‐ver eme‐mi‐l house‐LOC‐PRFL.PL come‐TEMP.CVB‐PL ‘Having come home they ate.’ b. Turu‐du bi‐ŋesi‐n Tura‐DAT be‐SIM.CVB‐POSS.3SG ‘I knew that when s/he was/lived in Tura.’ c. Turu‐du bi‐ŋesi‐vi Tura‐DAT be‐SIM.CVB–PRFL

ʤep‐čo‐tin eat‐PST‐3PL tara‐ve that‐DEF.ACC

sa‐ča‐v know‐PST‐ POSS.1SG

tara‐ve that‐DEF.ACC

sa‐ča‐v know‐PST‐ POSS.1SG

‘I knew that when I was/lived in Tura.’ In Buryat the converbal system functions in a manner very similar to that in Evenki. Thus, the converbs occurring only or predominantly in SS constructions do not take person marking (Skribnik 1988: 143; 2003: 117; 11a). The remaining converbs take possessive subject‐ agreement markers when they occur in subordinate clauses with a non‐coreferential subject (11b), or reflexive possessive person markers when the subjects are coreferential (Poppe 1960: 70; Skribnik 1988: 149, 11c). (11) a. Buryat (Skribnik 2003: 116–17; my glosses) tedener‐te öxibüːd‐iːn’ tuhal‐xajaː those‐DAT children‐ help‐FIN.CVB POSS.3PL ‘Their children have come to them in order to help.’ b. tende xüre‐že ošo‐tor‐nai there reach‐IPF.CVB go‐TERM.CVB‐ POSS.1PL duːha‐xa johotoi end‐FUT.PTCP probably ‘By the time we get there, the war will surely be over.’ c. Butedmaː teren‐iːji tani‐xalaːr‐aː B. that.OBL‐ACC recognize‐SUCC. CVB‐PRFL ‘Recognizing him, Butedmaa was glad.’

 jere‐ŋxei come‐RES.PTCP dain enemy.OBL

baldaːn ?

bajarla‐ša‐ba be.glad‐INTS‐ TERM

The similarity of the Buryat converbal system to that of Evenki is striking. The same type of subject‐agreement suffixes fulfill the same syntactic role in both languages. In contrast to Evenki, Buryat did not inherit this system from its Mongolic ancestor, making Evenki contact influence in its development highly likely. This is most probably due to language shift from Evenks to Buryat, as documented by the presence of a number of Buryat clan names that are of Evenk origin as well as by phonological changes in Buryat that can be traced to Evenki influence (Cydendambaev 1981; Čimitdoržieva 2004).

Contact and Siberian Languages  683

5 Conclusions This brief sketch of language contact influences in the vast area of Siberia has illustrated that contact situations can be multi‐layered. Currently ongoing changes in the languages of Siberia are due to the influence of Russian and, in certain areas, of Sakha, both of which are politically dominant; unfortunately, this dominance is leading to a large‐scale shift to Russian, and occasionally to Sakha. In addition to the influence exerted by politically dominant languages, over the centuries the indigenous languages have been undergoing changes brought about by contact with their neighbors. Unfortunately, not much is known about the prehistoric contact between the indigenous peoples of Siberia, making it difficult to draw conclusions from these changes. In some cases, they are probably due to substrate influence resulting from language shift, as in the case of Evenki influence in Buryat. Whether in other cases contact influence may be due to long‐term multilingualism is hard to establish for certain. However, in the example of Sakha–Evenki contact, previous bilingualism of Sakha speakers in Evenki is a possibility. More studies involving both fine‐scaled molecular anthropological and linguistic analyses of contact in Siberia are therefore necessary to elucidate how these languages changed under different kinds of contact. It is notable that several of the known cases of contact involve Evenki. Evenks were traditionally highly mobile nomadic hunters who used domesticated reindeer for transport, and they were and are in contact with speakers of very many different languages. This explains why they may have played the role of “vectors of diffusion” of at least some of the features that characterize the Siberian macro‐area (Anderson 2006: 294), although the spread of the Northern Tungusic languages over the vast area they occupy today may have taken place quite recently, not more than 600 or 700 years ago (Janhunen 1996: 171).

Abbreviations ACC = accusative, ALL = allative, ANT = anterior, ASS = assertive, ATTR = attributive, AUG = augmentative, AUX = auxiliary, COND = conditional, CONN = connective, COP = copula, CVB = converb, DAT = dative, DEF = definite, DER = derivation = DIST = distal (demonstrative), DP = discourse particle, DS = different subject, EMPH = emphatic, F = feminine = FIN = final, FUT = future, HAB = habitual, HESIT = hesitative, IMP = imperative, INF = infinitive, INS = instrumental, INTR = intransivity, INTS = intensive, IPF = imperfective, LOC = locative, M = masculine, NEC = necessitive, NEG = negative, NFUT = non‐future, NOM = nominative, OBJ = object, OBL = oblique, PFV = perfective, PL = plural, POSS = possessive, PRED = predicative, PREP = prepositional case, PRFL = reflex possessive, PRS = present, PST = past, PTCP = participle, PURP = purposive, R = Russian copy, REAL = realis, RES = resultative, SBJ = subject, SG = singular = SIM = simultaneous, SQ = sequential, SUCC = successive, TEMP = temporal, TERM terminative, VR = verbalizer, Y = Sakha (Yakut) copy.

NOTES 1 Note that borders between dialects and languages are notoriously fuzzy; thus, glottolog.org counts four different Khanty languages (accessed September 25, 2018). Furthermore, as pointed out on p. 670, several of these languages are on the verge of extinction. 2 Although Itelmen is generally classified as belonging to the Chukotko‐Kamchatkan languages (e.g. Comrie 1981: 240), this is not undisputed; an alternative hypothesis suggests that the similarities with Chukchi and Koryak are due to areal influences (cf. Georg and Volodin 1999: 224–41).

684  Brigitte Pakendorf 3 Given the diverse meanings of the word “borrowing” in the literature on language contact, I prefer to speak of “copying” (cf. Johanson 1992: 175). 4 This development has also been attributed to Sakha influence (Malchukov 2006: 127). 5 Note that this contradicts an earlier account of the development of this pidgin, according to which Russians were probably not directly involved in its development, since “[u]p to the beginning of the 20th century only few Nganasans had direct contacts with Russians” (Stern 2005a: 291). In contrast, Stern (2009: 393) describes lengthy visits of Nganasans in the trading way settlements, and argues that “[t]he conditions for a convergence of TPR [Taimyr Pidgin Russian] and Peasants’ Russian may have been especially favorable, where the Zatundra peasants formed the majority of the settlement’s population and the Nganasans set up their winter camps right beside the settlers’ huts.” 6 Note that the term “creole” referred only to the peoples’ mixed ancestry: Copper Island Aleut is not a creole, but a mixed language. 7 Samoyedic is a major branch of the Uralic family.

REFERENCES Anderson, Gregory D. S. 2004. The languages of Central Siberia: introduction and overview. In Edward J. Vajda (ed.), Languages and Prehistory of Central Siberia. Amsterdam: Benjamins, pp. 1–119. Anderson, Gregory D. S. 2005. Language Contact in South Central Siberia. Wiesbaden: Harrassowitz. Anderson, Gregory D. S. 2006. Towards a typology of the Siberian linguistic area. In Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas: Convergence in Historical and Typological Perspective. New York: Palgrave Macmillan, pp. 266–300. Anderson, Gregory D. S. 2017. The changing profile of case marking in the Northeastern Siberia Area. In Raymond Hickey (ed.), Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, pp. 627–50. Anderson, Gregory D.S. 2020. Form and pattern borrowing across Siberian Turkic, Mongolic, and Tungusic languages. In Martine Robbeets and Alexander Savelyev (eds), The Oxford Guide to the Transeurasian Languages. Oxford: Oxford University Press, pp. 715–725. Bátori, István 1980. Russen und Finnougrier: Kontakt der Völker und Kontakt der Sprachen. Wiesbaden: Harrassowitz. Burykin, Aleksej A. 1996. Ethnic composition of the population, ethno‐cultural contacts and languages of interethnic communication in the northeast of the Asian coastal areas of the Pacific Ocean (Stephen A. Wurm, trans.). In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of

Intercultural Communication in the Pacific, Asia, and the Americas. Berlin: de Gruyter Mouton, pp. 989–98. ̌ C  imitdoržieva, G. N. 2004. Vlijanie tunguso‐ man’čžurskix jazykov na burjatskie govory [The influence of Tungusic languages on dialects of Buryat]. In Istorija i Vnešnie Svjazi Burjatskogo Jazyka [The history and external contacts of the Buryat language]. Ulan‐Ude: Izdatel’stvo Burjatskogo naučnogo centra SO RAN, pp. 96–107. Comrie, Bernard 1981. The Languages of the Soviet Union. Cambridge: Cambridge University Press. Comrie, Bernard 1989. The spread of Russian among the other speech communities of the USSR. Canadian Review of Studies in Nationalism 16: 145–52. Comrie, Bernard 1996. Language contact in northeastern Siberia (Chukotka and Kamchatka). In Ernst H. Jahr and Ingvild Broch (eds.), Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: de Gruyter Mouton, pp. 33– 45. Cydendambaev, C. B. 1981. Zametki ob ėtničeskix i jazykovyx kontaktax burjat i ėvenkov. [Notes on the ethnic and linguistic contacts of Buryats and Evenks] In Jazyki i Fol’klor Narodov Severa [Languages and folklore of the peoples of the North]. Novosibirsk: Izdatel’stvo “Nauka” Sibirskoe otdelenie, pp. 70–91. de Graaf, Tjeerd 1992. The small languages of Sakhalin. International Journal of the Sociology of Language 94: 185–200. de Reuse, Willem J. 1994. Siberian Yupik Eskimo: The Language and Its Contacts with Chukchi. Salt Lake City: University of Utah Press.

Contact and Siberian Languages  685 Forsyth, James 1992. A History of the Peoples of Siberia: Russia’s North Asian Colony 1581–1990. Cambridge: Cambridge University Press. Georg, Stephan and Alexander P. Volodin 1999. Die Itelmenische Sprache. Wiesbaden: Harrassowitz. Georg, Stephan, Peter A. Michalove, Alexis Manaster Ramer, and Paul J. Sidwell 1999. Telling general linguists about Altaic. Journal of Linguistics 35: 65–98. Gernet, Katharina 2007. Evenen – Jäger, Rentierhirten, Fischer. Zur Geschichte eines Nordostsibirischen Volkes im Russischen Zarenreich. Wiesbaden: Harrassowitz. Gladkova, N. I. 1991. Vlijanie russkogo jazyka na ėvenkijskij i ėvenskij [The influence of Russian on Evenki and Even] In Russkij Jazyk i Jazyki Narodov Krajnego Severa: Problemy Opisanija Kontaktnyx Javlenij. Tezisy Dokladov Vsesojuznoj Naučno‐Praktičeskoj Konferencii (Leningrad, 12–14 marta 1991 g.) [Russian and the languages of the far north: problems describing contact phenomena. Abstracts from the All‐Union Scientific‐Practical Conference (Leningrad, March 12–14, 1991)]. Leningrad: Leningradskoe otdelenie Instituta jazykoznanija AN SSSR, pp. 67–8. Golovko, Evgenij V. 1996. A case of nongenetic development in the Arctic area: the contribution of Aleut and Russian to the formation of Copper Island Aleut. In Ernst H. Jahr and Ingvild Broch (eds.), Language Contact in the Arctic: Northern Pidgins and Contact Languages. Berlin: Mouto, pp. 63–77. Golovko, Evgenij V. 2003. Language contact and group identity: the role of “folk” linguistic engineering. In Yaron Matras and Peter Bakker (eds.), The Mixed Language Debate: Theoretical and Empirical Advances. Berlin: de Gruyter Mouton, pp. 177–207. Grenoble, Lenore A. 2000. Morphosyntactic change: the impact of Russian on Evenki. In Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in Contact. Amsterdam: Rodopi, pp. 105–20. Grenoble, Lenore 2009. Losing it in Siberia: assessing the impact of contact. Proceedings from the Annual Meeting of the Chicago Linguistic Society 45.1: 143–59. Gruzdeva, Ekaterina 2015. Sociolinguistic and linguistic outcomes of Nivkh‐Russian language contact. In Christel Stolz (ed.), Language Empires in Comparative Perspective (Colonial and Postcolonial Linguistics 6). Berlin: de Gruyter, pp. 153–81. Harrison, K. David and Gregory D. S. Anderson 2008. Tofa language change and terminal

generation speakers. In K. David Harrison, David S. Rood, and Arienne Dwyer (eds.), Lessons from Documented Endangered Languages (Typological Studies in Language (TSL) 78). Amsterdam: Benjamins, pp. 243–70. Helimski, Eugene 1997. Factors of Russianization in Siberia and linguo‐ecological strategies. In Hiroshi Shoji and Juha Janhunen (eds.), Northern Minority Languages: Problems of Survival. Papers Presented at the Eighteenth Taniguchi International Symposium: Division of Ethnology. Osaka: Osaka: National Museum of Ethnology, pp. 77–91. Janhunen, Juha 1996. Manchuria: An Ethnic History. Helsinki: The Finno‐Ugrian Society. Johanson, Lars 1992. Strukturelle Faktoren in türkischen Sprachkontakten. In Sitzungsberichte der Wissenschaftlichen Gesellschaft an der Johann‐Wolfgang‐Goethe‐Universität Frankfurt am Main. Stuttgart: Steiner, pp. 169–299. Kaksin, A. D. 1999. Xantyjskie zaimstvovanija iz russkogo jazyka v oblasti leksiki, svjazannoj s bytom i xozjajstvennoj dejatel’nost’ju (na materiale kazymskogo dialekta) [Khanty borrowings from Russian in the domain of everyday life and household (based on the Kazym dialect)]. In Natal’ja N. Širobokova (ed.), Jazyki Korennyx Narodov Sibiri: Sbornik Naučnyx Trudov [The languages of the indigenous peoples of Siberia: collection of scientific papers]. Novosibirsk: Sibirskij xronograf, pp. 22–4. Kazakevič, Ol’ga A. and O. S. Parfënova 2000. Jazykovaja i ėtnokul’turnaja situacija v Krasnosel’kupskom rajone Jamalo‐Neneckogo avtonomnogo okruga [The linguistic and ethno‐cultural situation in the Krasnoselkupsk district of the Yamalo‐Nenets Autonomous Region]. In V. J. Mixal’čenko (ed.), Jazyki Rossijskoj Federacii i Novogo Zarubež’ja: Status i Funkcii [The languages of the Russian Federation and the former Soviet Union: status and functions]. Moscow: Ėditorial URSS, pp. 270–303. Khanina, Olesya and Miriam Meyerhoff 2018. A case‐study in historical sociolinguistics beyond Europe: reconstructing patterns of multilingualism of a language community in Siberia. Journal of Historical Sociolinguistics 4.2: 1–31. Khanina, Olesya and Andrey Shluinsky 2008. Finite structures in Forest Enets subordination: a case study of language change under strong Russian influence. In Edward Vajda (ed.), Subordination and Coordination Strategies in North Asian Languages. Amsterdam: John Benjamins, pp. 63–75.

686  Brigitte Pakendorf Khasanova, Marina 2000. The Lower Amur languages in contact with Russian. In Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in Contact. Amsterdam: Rodopi, pp. 179– 85. Kibrik, Aleksandr E. 1991. O faktorax, otricatel’no vlijajuščix na žiznesposobnost’ jazykov maločislennyx narodov [Factors with a negative effect on the viability of minority languages]. In Russkij Jazyk i Jazyki Narodov Krajnego Severa. Problemy Opisanija Kontaktnyx Javlenij: Tezisy Dokladov Vsesojuznoj Naučno‐ Praktičeskoj Konferencii (Leningrad, 12–14 marta 1991 g.) [Russian and the languages of the far north: problems describing contact phenomena. Abstracts from the All‐Union Scientific‐Practical Conference (Leningrad, March 12–14, 1991)]. Leningrad: Leningradskoe otdelenie Instituta jazykoznanija AN SSSR, pp. 8–10. Krasovicky, Aleksander and Christian Sappok 2000. The isolated Russian dialectal system in contact with Tungus languages in Siberia and the Far East. In Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in contact,. Amsterdam: Rodopi, pp. 199–207. Krupnik, Igor and Mikhail Chlenov 2007. The end of “Eskimo land”: Yupik relocation in Chukotka, 1958–1959. Études/Inuit/Studies 31.1–2: 59–81. doi:10.7202/019715ar Ljublinskaja, M. D. 2000. Jazykovaja situacija na Tajmyre [The language situation on the Taimyr Peninsula]. In V. J. Mixal’čenko (ed.), Jazyki Rossijskoj Federacii i Novogo Zarubež’ja. Status i Funkcii [The languages of the Russian Federation and the former Soviet Union: status and functions]. Moscow: ėditorial URSS, pp. 309–14. Malchukov, Andrej L. 2003. Russian interference in Tungusic languages in an areal‐typological perspective. In P. Sture Ureland (ed.), Convergence and Divergence of European Languages. Berlin: Logos Verlag, pp. 235–49. Malchukov, Andrej L. 2006. Yakut interference in North‐Tungusic languages. In Henrik Boeschoeten and Lars Johanson (eds.), Turkic Languages in Contact. Wiesbaden: Harrassowitz, pp. 122–38. Maslova, Elena 2003. Tundra Yukaghir. Munich: LINCOM. Maslova, Elena and Nikolaj B. Vaxtin 1996. The far north‐east of Russia. In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia, and the Americas. Berlin: de Gruyter Mouton, pp. 999–1001.

Matić, Dejan 2008. Russian influence on the Kolyma Yukagir morphosyntax. In Antje Casaretto and Silvia Kutscher (eds.), Sprachkontakt, Synchron und Diachron. Aachen: Shaker, pp. 97–128. Minayeva, Vera 2003. Russian grammatical interference in Ket. Sprachtypologie und Universalienforschung (STUF) 56: 40–54. Morgounova, Daria 2007. Language, identities and ideologies of the past and present Chukotka. Études/Inuit/Studies 31.1–2: 183–200. doi:10.7202/019721ar Myreeva, A. N. 1964. Očerk govora učurskix èvenkov [A sketch of the dialect of the Učur Evenks]. In A. V. Romanova and A. N. Myreeva (eds.), Očerki Učurskogo, majskogo i tottinskogo govorov [Sketches of the Učur, Maja, and Totta dialect]. Moscow, Leningrad: Izdatel’stvo “Nauka,” pp. 6–74. Nedjalkov, Igor V. 1995. Converbs in Evenki. In Martin Haspelmath and Ekkehard König (eds.), Converbs in Cross‐Linguistic Perspective. Berlin: de Gruyter Mouton, pp. 441–63. Nevskaja, Irina 2000. Shor–Russian contact features. In Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in Contact. Amsterdam: Rodopi, pp. 283–98. Nikolaeva, Irina 2004. Online Kolyma Yukaghir Documentation. Available at https://www.sgr. fi/yukaghir/. Nikolaeva, Irina and Tolskaya Maria 2001. A Grammar of Udihe. Berlin: de Gruyter Mouton. Oskolskaya, Sofia and Natasha Stoynova 2013. The la‐form: Russian verbs in Nanai speech. Eesti ja soome‐ugri keeleteaduse ajakiri. Journal of Estonian and Finno‐Ugric Linguistics 4.2. 99–116. doi:10.12697/jeful.2013.4.2.05 Pakendorf, Brigitte 2007. Contact in the Prehistory of the Sakha (Yakuts): Linguistic and Genetic Perspectives. Utrecht: LOT. Pakendorf, Brigitte 2009. Intensive contact and the copying of paradigms: an Ėven dialect in contact with Sakha (Yakut). Journal of Language Contact Varia 2: 85–110. Pakendorf, Brigitte 2013. The interplay of language‐internal variation and contact influence in language change. In Isabelle Léglise and Claudine Chamoreau (eds.), The Interplay of Variation and Change in Contact Settings (SILV 12). Amsterdam: Benjamins, pp. 199–227. Pakendorf, Brigitte 2014. Paradigm copying in Tungusic: the Lamunkhin dialect of Ėven and beyond. In Martine Robbeets and Walter Bisang (eds.), Paradigm Change in Historical

Contact and Siberian Languages  687 Reconstruction: The Transeurasian Languages and Beyond. Amsterdam: Benjamins, pp. 287–310. Pakendorf, Brigitte 2015. A comparison of copied morphemes in Sakha (Yakut) and Ėven. In Francesco Gardani, Peter Arkadiev, and Nino Amiridze (eds.), Borrowed Morphology. Berlin: de Gruyter Mouton, pp. 157–87. Pakendorf, Brigitte. 2019. Direct copying of inflectional paradigms: Evidence from Lamunkhin Even. Language, 95/3: e364–e380. doi:10.1353/lan.2019.0063. Pakendorf, Brigitte and Natalia Aralova 2018. The endangered state of Negidal: a field report. Language Documentation and Conservation. Available at https:// scholarspace.manoa.hawaii.edu/bitstream/ 10125/24760/1/pakendorf_aralova.pdf. Pakendorf, Brigitte and Innokentij N. Novgorodov 2009. Loanwords in Sakha (Yakut), a Turkic language of Siberia. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter Mouton, pp. 496–524. Pakendorf, Brigitte, Innokentij N. Novgorodov, Vladimir L. Osakovskij, Albina P. Danilova, Artur P. Protod’jakonov, and Mark Stoneking 2006. Investigating the effects of prehistoric migrations in Siberia: genetic variation and the origins of Yakuts. Human Genetics 120: 334–53. Perekhvalskaya, Elena 2013. Chinese Pidgin Russian. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), Atlas of Pidgin and Creole Language Structures, vol. 3: Contact Languages Based on Languages from Africa, Asia, Australia, and the Americas. Oxford: Oxford University Press, pp. 69–76. Perekhvalskaya, Elena forthcoming. Analyzing modern Chinese Pidgin Russian: Variability and the feature pool theory. In Diana Forker and Lenore A. Grenoble (eds.), Language Contact in the Territory of the Former Soviet Union. Amsterdam: Benjamins. Poppe, Nicholas N. 1960. Buriat Grammar. Bloomington: Indiana University. Pugach, Irina, Rostislav Matveev, Viktor Spitsyn, Sergey Makarov, Innokentiy Novgorodov, Vladimir Osakovsky, Mark Stoneking and Brigitte Pakendorf 2016. The complex admixture history and recent southern origins of Siberian populations. Molecular Biology and Evolution 33.7: 1777–95. doi:10.1093/molbev/ msw055 Schweitzer, Peter P., Evgeniy V. Golovko, and Nikolai B. Vakhtin 2013. Mixed communities

in the Russian North; or, why are there no “creoles” in Siberia? Ethnohistory 60.3: 419–38. doi:10.1215/00141801‐2140749 Seifart, Frank 2015. Direct and indirect affix borrowing. Language 91.3: 511–32. doi:10.1353/ lan.2015.0044. Siegl, Florian 2013. The sociolinguistic status quo on the Taimyr Peninsula. Études finno‐ ougriennes 45. doi:10.4000/efo.2472. http:// journals.openedition.org/efo/2472 Skribnik, Elena K. 1988. Polipredikativnye Sintetičeskie Predloženija v Burjatskom Jazyke [Polypredicative synthetic sentences in Buryat]. Novosibirsk: Izdatel’stvo “Nauka,” Sibirskoe otdelenie. Skribnik, Elena K. 2003. Buryat. In Juha Janhunen (ed.), The Mongolic Languages. London: Routledge, pp. 102–28. Slezkine, Yuri 1994. Arctic Mirrors: Russia and the Small Peoples of the North. Ithaca, NY: Cornell University Press. Sorokina, I. P. 1991. Vlijanie russkogo jazyka na ėneckij [The influence of Russian on Enets]. In Russkij Jazyk i Jazyki Narodov Krajnego Severa. Problemy Opisanija Kontaktnyx Javlenij: Tezisy Dokladov Vsesojuznoj Naučno‐Praktičeskoj Konferencii (Leningrad, 12–14 marta 1991 g.) [Russian and the languages of the far north: problems describing contact phenomena. Abstracts from the All‐Union Scientific‐Practical Conference (Leningrad, March 12–14, 1991)]. Leningrad: Leningradskoe otdelenie Instituta jazykoznanija AN SSSR, pp. 65–7. Stapert, Eugénie 2013. Contact‐induced Change in Dolgan: An Investigation into the Role of Linguistic Data for the Reconstruction of a People’s (pre)history. Utrecht: LOT. Starcev, Anatolij Fedorovič 2014. Problema etnogeneza i etničeskoj istorii Negidal’cev [Problems of the ethnogenesis and ethnic history of the Negidals]. In Anatolij Fedorovič Starcev, Dmitrij Viktorovič Jančev and V. L. Larin (eds.), Istorija i kul’tura Negidal’cev: istoriko‐etnografičeskie očerki [The history and culture of the Negidals: historical and ethnographic sketches]. Vladivostok: Dal’nauka, pp. 22–41. Stern, Dieter 2005a. Taimyr Pidgin Russian (Govorka). Russian Linguistics 29: 289–318. Stern, Dieter 2005b. Myths and facts about the Kyakhta Trade Pidgin. Journal of Pidgin and Creole Languages 20: 175– 87. Stern, Dieter 2009. The Taimyr Pidgin Russian morphology enigma. International Journal of

688  Brigitte Pakendorf Bilingualism 13.3: 378–95. doi:10.1177/1367006909346624 Sunderland, Willard 1996. Russians into Iakuts? “Going Native” and Problems of Russian National Identity in the Siberian North, 1870s–1914. Slavic Review 55.4: 806–25. doi:10.2307/2501239 Thomason, Sarah G. 1997. Mednyj Aleut. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 449–68. Tsumagari, Toshiro 2003. Dagur. In Juha Janhunen (ed.), The Mongolic Languages. London: Routledge, pp. 129–53. Ubrjatova, Elizaveta I. 1985. Jazyk Noril’skix Dolgan [The language of the Norilsk Dolgans]. Novosibirsk: Izdatel’stvo “Nauka.” Sibirskoe Otdelenie.

Vajda, Edward J. 2009. Loanwords in Ket. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook. Berlin: de Gruyter de Gruyter Mouton, pp. 471–95. Vaxtin, Nikolaj B. 2001. Jazyki narodov Severa v XX veke: Očerki jazykovogo sdviga [The languages of the peoples of the North in the twentieth century: sketches of language shift]. St Petersburg: Evropejskij Universitet v Sankt‐Peterburge. Wurm, Stephen A. 1996. Indigenous lingue franche and bilingualism in Siberia (beginning of the twentieth century). In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia, and the Americas. Berlin: de Gruyter Mouton, pp. 975–8.

35 Language Contact Sino‐Russian ZYGMUNT FRAJZYNGIER, NATALIA GURIAN, AND SERGEI KARPENKO

1 Introduction The main aim of this study is to describe the forms and functions of the language of first‐ generation speakers of a contact language. The term “contact language” refers only to the language of speakers who move into a territory where another language is dominant. The study focuses on two questions: (i) What are the formal features used by contact language speakers? and (ii) What functions are coded by these formal features? At the end of this study we look at the implications of our findings. The language that we are studying is spoken by speakers of northern varieties of Chinese who have moved into Russia, referred further in this study as Sino‐Russian. These speakers have not had any formal instruction in Russian. The importance of this study is manifold: 1. The initial situation that we study must have been prevalent in first‐generation contact language speakers throughout the history of human migrations. The situation that we observe now is thus a living laboratory of what must have happened time and again in the course of human history. 2. Because we study a language that is not inherited from parent languages (McWhorter 2011), our findings contribute to ongoing theoretical discussions regarding the nature of languages, in particular to the question of whether there are certain inherent principles pertaining to components of genetically transmitted language structure or whether a language is a product of creative solving of communicative needs. Our study is thus another take on the issues addressed in Klein and Perdue (1997). 3. Our study contributes to the understanding of the differences between the lexicon and other coding features, such as inflectional features and prosodic features. 4. Our study also contributes to the understanding of how grammatical systems emerge in languages. 5. Our study also points to the need for changing perspectives with respect to the controversial issue of simplicity/complexity. McWhorter (2011) claims that creoles have simpler grammar when compared with their lexifier languages. This may be true for creoles, but it is irrelevant to the study of the first‐generation language. The only grammatical system that is available to the first‐generation of contact

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

690  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko language speakers is the grammar of their native languages. In the process of communication, these speakers strive to communicate functions that are encoded in the grammatical systems of their native language. Whether they are successful or not depends fully on the formal coding features they have access to. The formal features that are available to speakers without formal education in the dominant language are lexical items of the lexifier language, free grammatical morphemes, and possibly prosodic features. As our study demonstrates, inflectional features, derivational features, and coding through linear order in the lexifier language are not as readily available to the first generation of contact language speakers, and the functions coded in the lexifier language are not available at all. Contact language speakers are thus faced with a lack of formal coding features for expressing the functions coded in their native language. The discovery of functions coded in the lexifier language comes much later, if it comes at all. Our study indicates that the relevant comparison with respect to the simplicity/complexity issue for first‐generation speakers (not for creole speakers) is between the grammatical system they use in contact communication and the grammatical system of their native language, not the grammatical system of the lexifier language. An additional importance of this study is that it involves contact between two languages that are typologically different both in the formal features available and in the functions coded in each language. Chinese has a tonal system and Russian has a stress system. Russian has morphologically marked distinctions between lexical categories, and Mandarin does not have such distinctions. Russian has a rich system of inflectional morphology, including inflection by case, gender, and number for all nonverbal categories, i.e. nouns, numerals, adjectives, demonstratives, and determiners. Mandarin has no inflectional system and no distinction of gender (in the spoken language), number (other than in pronouns), or case in the nominal system. Russian has a rich system of inflectional marking of person, gender, and number of the subject on the verb. Mandarin does not code person, gender, and number of the subject on the verb. With respect to grammatical relations Russian codes the distinction between subject, direct object, and indirect object. Grammatical relations in Mandarin are a controversial issue, with some claiming existence of the categories subject and object (see chapters in Guangshun, Djamouri, and Peyraube 2015) and others denying the existence of such categories in the language (LaPolla 1990, 2009). Most scholars agree that Mandarin codes the category of topic in every clause, while Russian does not seem to have such an obligatory category. If we find forms and functions that occur in the Sino‐Russian contact language and do not occur in either Russian or Mandarin, these will require an explanation. One explanation, for those who assume the existence of certain genetically transmitted elements of language structure, would be that the existence of these forms is evidence for their assumptions. They would need, of course, to explain why neither Russian nor Mandarin has retained these forms. For proponents of some cognitive linguistic approaches, the existence of such forms would be viewed as evidence that cognitive cause‐effect relationships are at play. For those who embrace such approaches the same question would need to be answered, namely why the cognitive cause‐effect relationships that are at play in Sino‐Russian idiolects are not at play in either Russian or Chinese. For participants in the present study, the existence of such forms would be evidence for the emergence of functions that is not driven by external cognitive factors, construals of situations or relationships, but rather by the evolution of grammatical systems.

2  State of the Art The discussion of the state of the art is divided into two parts: the state of the art with respect to the study of the initial stages and the state of the art with respect to the study of the language as spoken by Chinese immigrants in Russia.

Language Contact  691

2.1  Initial Stages Although language contact has become a major field in linguistics over the last 30 years or so, there are very few studies dealing with the initial stages of language contact. Muysken (2018) does not include the initial stages of language contact as an issue to be studied. Klein and Perdue (1997) remains the most thorough study of the initial stages. Klein and Purdue label the language that emerges from contact between immigrants and host communities as “basic variety” and see in it some evidence for the existence of certain elements of an I‐language, i.e. the genetically transmitted elements of language structure.

2.2 “Russian” as Spoken by Chinese Speakers in the Russian Far East We put the term “Russian” in inverted commas because there is really no evidence that the language is really Russian. The only Russian elements in this language are the vocabulary, the phonetic inventory, and some elements of Russian phonotactics. Studies of Russian as spoken by Chinese speakers in Russia are rather numerous and have a long scholarly tradition. Most often these studies refer to the language as “Russian‐ Chinese Pidgin,” as explained later, a problematic label for the language that we describe in this study. As reported in Oglezn’eva (2007b), the first linguistic descriptions of the Russian‐ Chinese Pidgin date from the end of the nineteenth century, in the works of Alexandrov (1884) and Schuchardt’s (1884) comments on this work. For the history of subsequent scholarship, see Oglezn’eva (2007a). Contemporary works include Shprintsyn (1968), based on his fieldwork from the 1930s; Perekhval’skaya (2003a, 2003b, 2004, 2006, 2008, 2010, 2014); Oglezn’eva (2007a and 2007b, 2009, 2010); Fedorova (2011); Shapiro (2012); and, more recently, Zhdanova (2016). According to Oglezn’eva (2007b), Chinese merchants at the end of the nineteenth century were required to learn Russian and pass an exam before they were allowed to trade with their Russian counterparts. Oglezn’eva states that at one time this language was spoken by about 1 million people. We have no evidence that this language was ever used as the language of communication between Chinese speakers. Moreover, travelers to the area at that time reported that the Russian‐Chinese Pidgin was used only by Chinese merchants when interacting with Russians, while their Russian counterparts spoke Russian, not Russian‐Chinese Pidgin. Hence, the use of the “Russian‐Chinese Pidgin” was different from the use of pidgins in multilingual societies, where speakers use the pidgin between themselves and not just when interacting with speakers of the lexifier language. The Russian‐ Chinese Pidgin of the beginning of the twentieth century has almost completely disappeared as a result of changed political situations. Some elements of this language may be preserved in the language of erstwhile speakers of Nanai (Gold) and other Tungusic languages whose original speakers shifted to Chinese (Perekhval’skaya 2010). The language that we are studying is quite different from what has been persistently referred to as Russian‐Chinese Pidgin. Our study points to a language usage that is quite different from pidgins as known across the world.1 The speech of any given speaker not only differs from that of other speakers of the language but also exhibits a large degree of variation within the speech of the individual speaker. Our data indicate that every speaker develops his or her own idiolect to communicate with Russian speakers. The Sino‐Russian idiolects are not subsequently unified in communication with the speakers of other idiolects because the communication with the speakers of other idiolects is conducted in Mandarin Chinese rather than in some variety of Russian. So, the interesting question is whether there are similarities between the various idiolects and, if so, what are the reasons for these similarities. One reason may be that some speakers have retained the same feature(s) used in Mandarin. Another reason may be that some, but not all, speakers have discovered and acquired the same features of Russian. But what if

692  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko there are idiolectic similarities that do not occur in either Mandarin or Russian? And where do those come from? The question of such differences is equally interesting to the question of similarities: Why do different speakers come up with different idiolects when, for all speakers, the source (Mandarin) and the presumed target (Russian) are the same? In what follows we first discuss the formal features occurring in the Sino‐Russian idiolects and the functions coded by these features. The formal features observed include: phonological features, i.e. the inventory of phonetic and underlying segments; phonotactics; phonological rules; and prosodic features, mainly changes in intonation and pauses. The formal features also include lexical categories and individual lexical items.

3 Phonology Previous scholarship on the phonology of Russian‐Chinese Pidgin has been rather limited, focusing mainly on how Chinese speakers realized Russian segments. The previous scholarship did notice constraints on word‐final consonantal endings – as early as Schuchardt (1884) (Oglezn’eva 2007b) – and vacillation in assigning voicing and some r/l alternations (Zhdanova 2016: 63–5; Shapiro 2012). The discussion of the phonological system of the Sino‐Russian idiolects that follows is based on Frajzyngier, Gurian, and Karpenko (forthcoming). The Russian segmental inventory is largely replicated in the Sino‐Russian idiolects. This is most clearly evidenced by the presence of segments that occur in Russian but not in Mandarin, cf. voiced and voiceless obstruents (Mandarin does not have the voiced‐voiceless contrast), palatalized, and non‐palatalized obstruents (Mandarin does not have such a distinction), and the presence of both trills and laterals. The effects of Mandarin on the segmental inventory of Sino‐Russian idiolects are minor when compared to the effects of Russian. We have noticed just a few cases of the presence of retroflex obstruents. We did not observe any effects of the aspirated‐non‐ aspirated distinction in the Sino‐Russian inventory other than its secondary effects on the choice of voicing features. The Sino‐Russian speakers perceive equally well the place of articulation of Russian segments. The manner of articulation of Sino‐Russian speakers differs from Standard Russian in the following way: speakers of Sino‐Russian perceive that the feature voice, which has no distinctive role in Mandarin, plays an important role in Russian. Speakers of Sino‐Russian do not know, however, which target words in Russian have underlying voiced segments and which have underlying voiceless segments. Consequently, the choice of voicing or voiceless features does not follow the Russian pattern. Nevertheless, the assignment of the feature voice is not completely random: labial voiced stops in Russian are most often devoiced in Sino‐Russian. Dental stops are sometimes devoiced and sometimes voiced. Voiced continuants are often reproduced as such, as are voiceless continuants. There are no traces of the aspirated/non‐aspirated distinction, which plays a contrastive role in Mandarin. The phonotactics of Sino‐Russian idiolects are almost completely dominated by the phonotactics of Russian. Word‐initial and word‐medial consonant clusters are abundant. Mandarin does not allow consonant clusters in any position. Mandarin does not allow consonants other than dental and velar nasals in word‐final position. Russian allows all consonants in word‐final position. Speakers of Sino‐Russian idiolects sometimes preserve Russian word‐final consonants without any changes and sometimes add an epenthetic vowel, often within the same idiolect. Deletion of the word‐final consonant in Sino‐Russian idiolects is very rare. We assume that the motivation for preserving word‐final consonants is the need to preserve the phonetic integrity of the word, which in turn is motivated by the communicative need, cf. the need to be understood.

Language Contact  693 It is not clear that speakers of Sino‐Russian idiolects have a phonological system with a finite number of underlying segments, a finite set of rules determining the realization of the underlying segments, a finite set of phonotactic constraints, and a set of rules to override the phonotactic constraints. There is plenty of evidence that, for the words they use, Sino‐Russian speakers know the places of articulation for the segments necessary to convey a given word. Outside the place of articulation, there exists a considerable margin in the realization of segments. With respect to the phonotactics of clusters, Sino‐Russian speakers do not follow the constraint that clusters of obstruents must have the same value for the feature [±voice]. The violation of this constraint is in harmony with the fact that the voiced‐voiceless distinction is not a feature of Mandarin, and for Mandarin speakers it is a secondary characteristic of segments. In the data that we have gathered, there appears to be no evidence of the existence of universal principles of the phonological systems. The speakers of Sino‐Russian idiolects produce the segments that occur in Russian and have largely adopted the phonotactics of Russian. They have reanalyzed the feature system of Russian by creating a hierarchy whereby the place of articulation is the determining distinctive feature of segments, with all other features being subordinate. Notice that the preservation of the feature (palatalized) provides additional support for the place of articulation occupying the top place in the hierarchy of features. In their language production, the speakers of Sino‐Russian idiolects are motivated by communicative needs and strive to produce lexical items in a form that will be understood by their Russian listeners. Quite unexpectedly, it turns out that the place of articulation of segments constituting a word is a sufficient tool to achieve the desired result. According to Zhdanova (2016: 63–5), Perekhval’skaya (2013: 334) asserts that speakers of Chinese have a tendency to treat every syllable in Russian words as a tone‐bearing unit and that consequently every syllable is treated as stressed. We did not observe this tendency in our data. In our data, the Sino‐Russian speakers usually reproduce correctly the Russian stress assignments for individual words. The results of our study of phonology are at variance with those obtained in Klein and Perdue (1997) in that we have found relatively little influence of the source language (Chinese) on the Sino‐Russian idiolects.

4  The Category “Word” The category “word” is a highly controversial category in linguistic theory because it is not possible to provide a single definition that would have cross‐linguistic validity. Some even question the usefulness of the term. In many languages there is no term corresponding to “word.” It is entirely possible, however, to provide a satisfactory definition in individual languages based on some formal characteristics. In Russian, such a characteristic is the main stress, only one per word. In Mandarin, the category “word” is much more difficult to define (Chao 1968: ch.3, 135–93). One of the most intriguing findings of our study of the Sino‐Russian idiolects is that speakers of Mandarin reproduce, and hence perceive, the entities which must be characterized as words. These are entities whose boundaries are phonologically marked. The phonological markers of these entities in Russian are stress markers. The evidence that stress is the main criterion by which a word is identified as such is that the Sino‐Russian idiolects, in the overwhelming majority of cases, assign stress in a manner faithfully corresponding to that of Russian. This is a non‐trivial fact given that stress in Russian is lexically determined and Chinese is a tonal, not stress‐based, language. The categoriality of a lexical item in Mandarin is, to a large extent, defined by the grammatical morphemes surrounding the open‐class word. The categoriality of a lexical

694  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko item in Russian is determined by the morphological markers attached to it. The ­categoriality of lexical items in Sino‐Russian idiolects is determined, if at all, by traces of morphological markers used in Russian. We explicitly use the term “traces” as the markers have the phonological shape of morphemes in Russian but do not necessarily carry the function that the corresponding morphemes carry in Russian. For example, many verbs in Sino‐Russian idiolects end in the form ‐ti, which corresponds to Russian infinitives ending in the palatalized consonant ‐t’. These endings indicate merely that the form is a verb. They are used in various modal, temporal, and aspectual functions, with and without subjects, in main and subordinate clauses. Here are a few examples, each containing a verb with the marker ‐ti and each representing a different modality, time, aspectual value, and person of the subject: (single slash / indicates a shorter pause and double slash // indicates a longer pause). Past event, completed Sveta (1) pɑ′tᴐm then ′pʊtɪm be

′svɑtɪ// call:INF ′mɑʧkә/ boy

′pɑpʃkɑ// grandmother k PREP:to

klә k to PREP:to nɪ′mʊ// 3SG:DAT

nɪ′mʊ// 3SG:DAT

‘Then [they] call[ed] grandmother to be with the boy’2 [klǝ] is an erroneous attempt to pronounce [k nimu]. This particular speaker often replaces n with l (Frajzyngier, Gurian, and Karpenko, forthcoming). Future event, not‐completed: Sveta (2) yɑ ʊ′hɑʒɪvɑt’ 1SG take care:INF ‘I’ll take care of you.’

tɪ′b’ɑ// 2SG:ACC

Present time ′dʊmɑt’ ‘think’, and the infinitive ‘delɑt’ ‘do’: Sveta (3) pʌ′tᴐm ә// ′ɛtɑ ′mɑʧkǝ ′tʊmɑt’/ ʃtɔ delɑt’ then DEM boy think:INF what do:INF ‘And then this boy thought what he should do with…’

s/ PREP:with

It is important to note the following: The ending ‐ti or ‐t’ always codes the lexical item as belonging to the category verb. Not every instantiation of the verb, however, ends in ‐ti or ‐t’. Some instantiations of the verb may have the Russian imperative or subjunctive ending, or even inflectional coding of the tense and person. No other lexical categories are morphologically distinguishable, i.e. there is no morphological distinction between nouns, adjectives, numerals, and adverbs. In the following example, the form zna ‘know’ carries no morphological indicators telling the listener that it is a verb. A Russian‐speaking listener can deduce the meaning of the form from the three segments included. Note that the form stʊ′ʧɑt’e ‘knock’ ends in ‐t’e, which resembles the Russian infinitive ending t’. Sveta (4) ɑ pɑ′tᴐm ′vᴐlkə/ znɑ dɑ/ ′mɑmɑ ʊʃ CONJ:a then wolf:NOM know yes mother:F:NOM already ʊʃ′lɑ dɑ// ɑ ᴐn tʊ′dɑ  stʊ′ʧɑt’e// leave:SG:F:PAST:PRF yes CONJ:a 3SG there knock:INF ‘And then, the wolf knew that the mother had left and he knocked there.’

Language Contact  695 In other examples verbs end in the related form s’e: Sveta (5) ´tɔʒə ´nɑdɑ uprɑv´l’ɑt’e// ku´dɑ tu´dɑ also need direct where there ‘[I] also have to direct [people] where they are to work.’

rɑ´bɔdɑs’e// work

Although not all verbs end in the form that resembles the Russian infinitive, many do. Moreover, no noun ever ends in this form. This indicates several things: (1) Sino‐Russian speakers perceive some Russian forms as verbs. This is remarkable in itself, as in Chinese there are no inflectional forms that would always allow the identification of the lexical form as a verb. (ii) Having perceived some forms as verbs, the speakers of Sino‐Russian idiolects most often mark these forms by a marker corresponding to the infinitive in Russian. This is remarkable in that Chinese speakers apparently consider it important to mark the form as a verb. The use of the infinitive form of the verb is what Klein and Perdue (1997) observed in their study of “basic variety.” Frajzyngier (2008), when working with speakers of Gidar without any formal education in their native language, noticed that speakers readily produced the infinitive form of the verb when asked for the citation form. The conclusion from these observations and from Klein and Perdue (1997) is that if a language has a distinct infinitive form, this form is perceived as a marker of the category verb by speakers without any education in the language in which the form occurs.

5  Free Grammatical Morphemes The Sino‐Russian idiolects contain free grammatical markers belonging to what in Russian are different formal and functional sets. Here is a list of such forms. Prepositions. One set of independent morphemes consists of prepositions: k ‘to’, u ‘at’, s ‘with’, pɑ ‘in, on’, ´cerəsə ‘through, by’, v ‘in’. These prepositions have functions in the Sino‐ Russian idiolects that differ somewhat from their functions in Russian. We discuss these functions later in this study. Conjunctions. This set consists of the markers i (coordinated conjunction), a (counter‐ expectation), and ili ‘or’. Again, the functions of these markers only partially overlap with their functions in Standard Russian. Pronouns. The full set of Russian independent pronouns in the singular and plural is represented in the Sino‐Russian idiolects. The use of the pronouns, however, differs significantly from their use in Russian. Demonstratives and determiners. This set, which is very rich in Russian, is reduced in Sino‐ Russian idiolects to ′ɛtɑ, and an occasional use of the forms takoj ‘such’. The reduction of the set of demonstratives in Russian to just one form, ′ɛtɑ, is remarkable given that the Sino‐Russian idiolects represent a product of the independent emergence of various languages from the same lexical source with the same first‐language knowledge. Note that Mandarin has at least two demonstratives, the proximate zhe and the remote na, which are often used as determiners of nouns. The form ′ɛtɑ in Sino‐Russian is seldom used as a determiner of nouns. Most often it is used as a comment marker. The properties of lexical items used by the speakers of Sino‐Russian idiolects differ significantly from their properties in Russian in that their meaning consists only of reference to entities or activities in the real or imagined world and is completely deprived of any of the grammatical properties that they have in Russian. As an illustration consider the verb

696  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko regulirovat’, which in Russian means ‘direct, manage’ and takes the entities that are managed in instrumental case. In a Sino‐Russian idiolect, the entity that is managed is preceded by the preposition s ‘with’: Sveta (6) ɑ s ´rusk’ɪm// ´nɑdɑ/ CONJ:a PREP:with Russian need ‘And one needs to manage the Russians.’

r’egu´r’ɪrɑvɑt’ɪ// manage:INF

Consider now the modal verb nada ‘need to do, have to’, which requires dative case marking on the noun or pronoun that represents the participant that has to perform an action. In Sino‐ Russian idiolects there is no case marking, and the noun representing the the agent is marked in the same way as participant representing the patient: (7) ɪ´ʃt͜ʃɔ ´nɑdɑ/ mɑ´ʃənʌ ɑtkru´z’ɪt’ɪ// yɑ ´tɔʒə ´nɑdɑ nayt’ɪ// also need truck load:INF 1SG also need find:INF ‘It’s also necessary [to find] the truck to be unloaded. I must also find [it].’ Consider also the verb pɑzdrɑ´vl’ɑt’ ‘greet, congratulate’. In Russian, this verb requires subject coding on the verb, regardless of the presence of the nominal or pronominal subject, and often has a complement as direct object, marked by the accusative case. The addition of any other noun phrase requires some coding for case on that noun phrase or a preposition before the noun. In Sino‐Russian idiolects, this verb can occur without a noun phrase that would correspond to the subject in English. This phenomenon is quite frequent with any verb in Chinese (Chao 1968, Frajzyngier and Liu, 2020). The verb pɑzdrɑ´vl’ɑt’ can occur with one noun functioning as an object and still another noun without any marking: (8) pɑzrɑ′vl’ɑ:t’e ′truk ′truku pɑ′dɑrɑkə/ congratulate:INF each other gift ‘[We/one] greet each other [with] gift[s].’ And as a final example consider the verb ʊhɑʒɪvɑt’ ‘take care’, which in Russian requires the preposition za before the non‐subject participant. In Sino‐Russian this verb can occur with two noun phrases, the second of which is marked by the accusative case marker used in Standard Russian: (example 2, repeated). Sveta (9) yɑ ʊ′hɑʒ ɪvɑt’ tɪ ʹb’ɑ 1SG take care:INF 2SG:ACC ‘I’ll take care of you.’ The speakers of Sino‐Russian idiolects use nouns in a way that is different from the use of nouns in both Chinese and Russian grammatical systems. In Chinese, most nouns have a “numeral classifier,” in many environments. In the Sino‐Russian idiolects nouns occur without any classifiers. In comparison to Chinese, the use of nouns in Sino‐Russian idiolects is reduced to their referential function. The speakers of Sino‐Russian idiolects have not developed anything resembling the case systems that are intricately linked with nominal categories in Russian. The properties of lexical items in the Sino‐Russian idiolects point to an important fact: not only do the speakers of Chinese perceive and reproduce words, but they also use words only in their referential function, i.e. the function of referring to an event or an entity in the real or an imagined world. These words are completely bereft of any grammatical properties.

Language Contact  697

6  Inflectional and Derivational Morphology Not a single variety of Sino‐Russian that we have recorded contained an instantiation of the productive use of derivational morphology, i.e. of deriving one lexical category from another or of deriving one lexical item from another within the same lexical category. With respect to inflectional morphology the situation is a bit complicated. There are numerous instantiations of words with inflectional markers, nouns, pronouns, adjectives, and verbs, which are used in the way in which they are used in Russian. Thus, the first person singular subject in the present tense is often marked by inflectional markers as it is in Russian: Sveta (10) nu kɑk rɑs dɑm fs’ɔ es’ fs’ɔ well right there everything be everything yɑ rɑ′bɔdɑyu// 1sg work:1SG:PRES ‘There is everything right there, [where] I work.’ Possessive pronoun agreeing in gender with the noun: Sveta (11) ′ɛtʌ ′tɔʒə mɑ′yɑ rɑ′bɔdʌ// dem also 1SG:F:POSS work:F ‘That also is my work.’ Another speaker does not show the agreement in gender between the possessive pronoun and the possessum: Boris (12) ′drasv’ɪd’e// moy/ ′im’e/ hello my:1SG:POSS name ‘Hello, my name is xxx.’ (anonymized) The Standard Russian form of the possessive pronoun in the above clause would be majo ‘1sg.n’, as imya is a neuter singular noun in Russian and as such requires neuter agreement. The occasional use of Standard Russian inflectional markers appears to be the result of learning and reproducing inflected forms rather than the result of productive use of inflectional processes. Consider the following utterances from the same speaker: Ivan (13) tam there

′rɨbǝ fish:PL?

rɑ′v’ɪrᴧ// catch:PAST

ʻ[I] caught fish there.ʼ In Standard Russian, the form used here would be the accusative singular form rybu. In the above example, the form ′rɨbǝ ‘fish’ looks like the genitive plural in Standard Russian. In the following example, the form ′drʊg ‘friend’ looks like the Russian nominative singular, while the Russian equivalent of example (14) would require the genitive plural form druz’ej. Moreover, the possessor in example (14) would have to be a prepositional phrase, u m’en’a ‘at me’, in Standard Russian:

698  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko Ivan ′mnᴐgә// many

(14) yɑ tam drʊg 1SG THERE friend ʻI have many friends there.ʼ

Some speakers of Sino‐Russian have acquired some inflectional forms and use them consistently. Thus Sveta, a speaker who spent 18 years in Russia uses the first person singular present correctly most of the time: Sveta (15) ya l’ub′l’u ′rusk’ɪye// 1SG love:1SG:PRES Russian:PL ‘I love Russians.’ (note that the speaker used the form rusk’ɪye ‘Russian’, equivalent of the Russian nominative, rather than the correct Russian form ruskix ‘Russian:GEN:PL’) Correct usage of one inflected form does not indicate correct usage of the same inflectional marker in another context. In the sentence below, in one verb the first person singular present is correctly marked by inflectional form ʃt͜ʃ’ɪ´tɑyu ‘count:1SG:PRES’ and in another verb it is not, namely in mne´rəcɪ ‘measure’, whose form may be closest to the Russian infinitive form: Sveta (16) ɪ´ʃt͜ʃɔ mne′rəcɪ in addition measure:INF ‘I also measure and count ...’

′tɔʒə// also

ɪ CONJ:i

ʃt͜ʃ’ɪ′tɑyu// count:1SG:PRES

Other inflectional forms of the verb have not been acquired, and their use does not agree with the Standard Russian use. In the following example, instead of the Standard Russian form rastut ‘grow:3PL:PRES’ the speaker uses the form rastutsə, which appears to have a reflexive marker, a form not attested in Standard Russian: Sveta (17) ′esv’ɪ n’e tɑpɑ′fl’ɑt’e/ ′plɔxɑ rɑ′stutsə// if NEG add:INF badly grow:PL:PRES: REFL ‘If [the fertilizer] isn’t added, [the vegetation] grows badly.’ The use of inflectional forms in Sino‐Russian thus does not provide evidence that the speakers have acquired knowledge of the functions of the inflectional markers. This use represents a reproduction of “chunks” i.e. inflected forms viewed as single unmarked forms, as they were heard and memorized. This is similar to what Klein and Perdue (1997) reported for their speakers of “basic variety.” It is entirely possible that, over time, the speakers will discover the patterns in such chunks and will eventually discover their functions, and that this may lead to the acquisition of the inflectional systems. But this has not happened yet in the first generation of Sino‐Russian speakers.

7  Linear Orders Linear orders in Russian are of several types: coding by position, where position with respect to some reference point has a function; coding by relative order, i.e. where the position 1 of the sequence X Y codes the same function as the sequence YX, except that the constituent in

Language Contact  699 the scope of the function changes. In Mandarin, the linear order is only of the relative type, i.e. the position of X before or after Y. Functions of linear orders differ significantly in Russian and Mandarin. In Russian, linear order codes pragmatic functions including the distinction between new and old information and focus (with proper intonation pattern). In Chinese, linear order is a coding means for the topic‐comment distinction (Chao 1968: 70; LaPolla 2009). In Sino‐Russian idiolects, there are no markers of lexical categories, hence there is no point of reference for the coding by position, hence there is no coding by position. Only the relative position is available as a coding means (Frajzyngier 2011; Frajzyngier with Shay 2016). But it will be shown later that even that possibility is not exploited in the Sino‐Russian idiolects. The linear order of lexical categories is not a coding means for any function in the Sino‐Russian speech of the first‐generation.

8  The Form of the Utterance in Sino‐Russian One of the common features of the Sino‐Russian texts that we have analyzed is the structure of the utterance. An utterance is an entity that is preceded and followed by a relatively long pause. In our transcriptions, new utterances are represented as new lines. In a few cases an utterance is unusually long and therefore consists of two lines. An utterance may have no internal breaks, save for very short breaks between the words. The break between utterances is marked by the single slash /. An unusually long break is marked by // in our transcription. An utterance may thus be simple, with no breaks, or may be complex consisting of two or more segments marked by the longer breaks //. Here are a few examples. Simple utterance: Nina (18) xɑrɑ′ʃɔ s’e svɑ′jɑ// good all one’s own ‘It’s fine, everybody is familiar.’ Complex utterance: Nina (19) dɑ ′nrɑɪtsɑ/ ′ɔʃɛ kɑ′n’ɛ pr’ɪ′vɨkәl’ɪ yes please:PRES already of course get used ‘Yes, [I] like [it]. [We] already got used here of course.’

s’jo all

′z’z’es’ɪ/ here

Nina (20) s’e l’et’ɪ dɔs nʧɔ// all people also not bad ‘All people are also not bad.’ Note that we are explicitly avoiding the term “sentence” as we are unable to discern the formal characteristics of such a putative entity (see also Mithun 2005). We use the term “clause” as the smallest unit that has a modal value, even if not overtly marked, rather than the frequently used definition of clause as an entity consisting of a verb and its arguments. The shorter and longer pauses within the utterance tell the listener nothing more than to interpret the first part of the utterance in connection with the second part. How to interpret the relationship between these parts may be further specified by particles, usually at the beginning or at the end or at both the beginning and the end, of the first part and the second

700  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko part of the utterance. For many utterances, the relationship between the utterances is underspecified by the speaker and it is left to the listener to decide on this matter.

9  Functions Coded in Sino‐Russian Idiolects The discussion of the functions coded in the Sino‐Russian idiolects is confined to relationships actually coded by some formal means that we were able to detect. We also mention some functions that the reader may expect to be coded, but for which we did not find any evidence.

9.1 Modalities The assertive modality is the unmarked modality in Sino‐Russian idiolects. There is nothing remarkable about it, as assertion is the unmarked modality in hundreds of languages, including Russian and Chinese. This modality is illustrated by all affirmative and negative clauses without other overt markers of modality. Evidence for the existence of the assertive modality is provided by the existence of markers that change the assertive modality value. Some of these markers involve hedging on the truth of the proposition, i.e. doubt in truth. Hedging is coded by a variety of means. One of them is the phrase kɑk skɑ´zɑt’ lit. ‘how to say’: Sveta (21) ɪ´ʃt͜ʃɔ// kɑk skɑ´zɑt’/ ɪ´ʃt͜ʃɔ also how say:INF also ‘How to say, and what…’

ʃ’tɔ// what

Sveta (22) ´ɔy // kɑk oh how say:INF ‘Oh, how to say.’

skɑ´zɑt’ɪ//

The use of the verb ‘to say’ to code hedging on the truth of the proposition is a common means found across languages (Frajzyngier 1996). It appears that Sino‐Russian speakers have also adopted the Russian particle da ‘yes’ as a  marker within the domain of epistemic modality, through which the speaker reinforces the assertive value of the proposition to follow or of a proposition that was produced in the preceding discourse: Nina (23) dɑ/ xɑrɑ′ʃɔ ′z’z’es’ɪ dɑ/ yes good here yes ‘Yes, it is good here. People are all OK.’

′l’jut’ɪ people

Sveta (24) kɑk u vɑs ′tᴐʒə/ like PREP:at 2PL:ACC also ‘It is just like in you[r country], it is.’

dɑ// yes

s’e all

nɑr′mɑlnɑ// OK

Language Contact  701

9.2  Modality of Obligation We have very few instances of the imperative modality, i.e. a wish with the expectation of immediate realization. Some speakers use the forms of Russian imperatives, as is the case with the verb ʌtk′rᴐj ‘open’ in the following example: Sveta (25) ʌtk′rᴐj/ ʌtk′rᴐj jɑ// ′pɑpʃkʌ// open open 1SG grandma ‘Open up, open up, I am [your] grandma.’ The modality of obligation without the expectation of immediate realization is marked by the auxiliary nada ‘need, have to’, which can occur in the clause several times, e.g. before the noun phrase that represents the entity that has the obligation to do something and also before the predicate or after the verbal predicate. The verbal predicate after the auxiliary nada has the Russian infinitive form: Sveta (26) ′nɑdɑ/ kra′sət’ɪ// need decorate ‘Everything should be beautiful/ decorated.’ Sveta (27) ′nɑdɑ ʒə′n’ɪxə/ ′nɑdɑ ɑtprɑv′l’yɑt’ɪ/ need groom need send:INF nɑ mɑ′ʃɨn’e/ vstrə′cɑt’ n’e′v’estɑ// PREP car:LOC meet:INF bride ‘The groom has to send a person by car to meet the bride.’

celɑ′v’ek person

Sveta (28) ′tᴐʒə pɑdɑ′r’ɪ tsv’etə// pɑdɑ′r’ɪ sv’e′tə// ′nɑdɑ/ ′tᴐʒə also give flowers:PL give flowers:PL need also ′t’en’g’ɪ/ ′tᴐʒə es’ ɑb’yɑzɑt’el’nɑ/ money also be obligatorily ‘[One has to] give flowers (two times), it also obligatory [to give] money.’ Sveta (29) ɑ pɑ′tᴐm/ ′nɑdɑ ′dv’erɑ/ ′nɑdɑ stu′cɑt’ CONJ:a then need door need knock ‘And then it’s necessary to knock at the door.’

′dv’er// door

′nɑdɑ/ need

Multiple coding of the modality of obligation has its parallel in multiple coding of the prohibitive modality, to be discussed in next section. (30) ɑ pɑ′tᴐm/ ′nɑdɑ// ′dv’erɑ/ ′nɑdɑ CONJ:a then need door need ‘And then it’s necessary to knock at the door.’

stu′cɑt’ ′dv’er/ knock door

′nɑdɑ/ need

702  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko ′esɪ zak / ɑt′krət’/ ′nɑdɑ t’en’gɪ/ if close(d)[error] open need money ‘If you need to open the door you are asked for the money.’ ′nɑdɑ xɑ′rᴐʃɪy ′slᴐvɑ gɑvɑ′rɪt’e// ′ʃdᴐbə need good word say:INF in order ‘One must say nice words to have the door opened.’

sprɑ′sɪ// ask

ɑt′krət’ open:INF

′dv’erɑ// door

9.3  Prohibitive Modality Prohibitive modality is coded by the forms nɪl’ɪ′z’yɑ ‘PROHIB’. Unlike its Russian source, the form nɪl’ɪ′zyɑ can be repeated several times within the utterance: its first occurrence may be before the proposition, and then before and after the predicate: Sveta (31) kɑg′dɑ ′svɑt’bɑ// nɪl’ɪ′z’yɑ// ʒе′n’ɪxə ɪ// kɑ when wedding PROHIB groom CONJ:i nɪl’ɪ′z’yɑ/ vstrə′cɑt’e // nɪl’ɪ′z’yɑ// PROHIB meet:INF PROHIB ‘Before the wedding the groom and bride may not meet.’

sɑ′za// n’e′v’estɑ how say:INF bride

The repetition of deontic markers across a clause has its parallels in Chinese. In Mandarin, repetition of prohibitive forms such as 不能 buneng, ‘not be able to’ or 不行 buxing ‘cannot’, is possible and occurs often, especially in spoken language. The repetition of modal markers is found across unrelated languages, e.g. repetition of interrogative and negative markers in spoken varieties of Swedish (Rosenkvist 2015) and repetition of negative markers in varieties of Central African languages, including Chadic (Frajzyngier with Shay 2002; Dryer 2009). These facts would indicate the existence of similar strategies across languages in the coding of modality.

9.4  Tense or Aspectual Distinctions Russian has a rich aspectual system that includes the distinction between perfective and imperfective and a variety of functions, sometimes referred to as Aktionsarten, that change the aspectual value of the verb. The Chinese aspectual system consists of several functions (Chao 1968: 245–54) which, however, do not overlap with the perfective‐imperfective distinction of Russian or with different Aktionsarten of Russian. The speakers of Sino‐Russian idiolects choose the lexical forms that code the perfective or imperfective distinction in Russian, but we have not observed a productive use of morphological markers to change aspectual values: Perfective: Lida (32) ´ɛtə‐ǝ‐ǝ// la´s’ɪǝ s’asə‐ə// ´t’ɪngǝ t’ɪ´ʒɔlə se´s’as// DEM Russia now   money heavy now ´ɛta‐ǝ‐ǝ// ´lubl’ɪ/ ´s’ɪlʼnɔ u´padǝ// DEM rubles strongly fall ‘The currency [situation] in Russia is difficult now; rubles fell strongly.’

Language Contact  703 Imperfective: Sveta (33) ɑ pɑ′tᴐm/ ′nɑdɑ ′dv’erɑ/ ′nɑdɑ CONJ:a then need door need ‘And then it’s necessary to knock on the door.’

stu′cɑt’ knock

′dv’er// door

′nɑdɑ need

Imperfective and perfective: Sveta (34) ′nɑdɑ xɑ′rᴐʃɪy ′slᴐvɑ gɑvɑ′r’ɪt’e// ′ʃdᴐbə ɑt′krət’ need good words say:IMPF in order to open:PRF ‘One must say nice words in order to have the door opened.’

′dv’erɑ// door

Russian has inflectional and periphrastic coding of tenses, past, present, and future. Chinese has no formal coding of tenses. It does code aspectual distinctions, but these distinctions are different from those of Russian. The Sino‐Russian idiolects do not appear to have a tense system. There is occasional use of Russian verbs in the future tense form, but they do not represent productive use. Clauses whose temporal reference implies the future have the same verbal forms as clauses that imply time of speech or reference to the past. Future time reference: Sveta (35) jɑ ʊ′hɑʒɪvɑt’ tɪ′b’ɑ// 1SG take care:INF 2SG:ACC ‘I will take care of you.’ Past‐time reference: Note that in the following example the same forms of the verb are deployed, once in the past‐time reference and another time in the infinitive, and both cases the verb has the infinitive ending: Sveta (past tense and infinitive) (36) pʌ′tᴐm ә// ′ɛtɑ ′mɑʧkǝ ′tʊmɑt’/ ʃtɔ delɑt’ then  DEM boy think:INF what do:INF ‘Then, this boy thought about what he should do with …’

s/ with

Sveta (past tense) (37) ′zɑɪʧɪk / ′nɛt rabbit no, ‘Rabbit: no, mother said …’

′mɑmɑ mother

ɡʌ′rɪ say:INF

The most frequent tense form in Sino‐Russian idiolects is marked in the way it is marked in Russian. Here is the first person singular present tense: Nina (38) n’ɪ ′znɑj// l’ɪ′tse ′dɑsә n’ɪ ′znɑju NEG know:1SG:PRES face even NEG know:1SG:PRES u′ʒɛ/ tɑm nɔ ′dɔlɡɑ/ already there well long ‘I don’t know. By now, I don’t even recognize faces [of those people].’

ʃtʃɑs now

704  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko Compare the past tense, which does not correspond to the past tense in Russian. The form  used by the speaker is pr’ɪ′vәkәl’ɪ instead of the third person singular feminine pr’ɪ′vәkla: Nina (39) da l’ju′z’e xɑrɑ′ʃɔ / jɑ z’z’es’jɑ yes people good 1SG here ‘Yes, the people are good, I’m used to living here.’

u’ʒɛ already

pr’ɪ′vәkәl’ɪ// get used

The use of the present tense form, however, does not necessarily carry the function of the present tense, as illustrated by the following sentence, where the verb bʌn’ɪ′mayu ‘understand:1SG:PRES’ actually refers to a past‐time event: Boris (40) ɪ// ya// tkag′da pr’ɪ′exar v ra′s’iu/ ɛ′dʌ// CONJ:I 1sg when come: SG:PAST:PRF PREP:in Russia DEM pʌ ′rusk’ɪ ɪ′zɪgə// ′tɔʒə  n’ɪ′cyo n’i bʌn’ɪ′mayu// in Russian language also  nothing neg understand:1:SG:PRES ‘And when I came to Russia, I couldn’t understand anything in Russian.’ The language of the first‐generation speakers does not carry any of the aspectual functions of Chinese, nor does it carry any of the aspectual or tense functions of Russian. It has also not encoded any functions of its own in the domains of tense and aspect in the grammatical system.

9.5  Relationships Between the Predicate and Noun Phrases Russian makes a distinction between the categories subject, marked by the nominative case on the noun or pronoun and the inflectional marking on the verb and all other relations marked by case marking and/or prepositions. While use of the nominal or pronominal subject in Russian is not obligatory, the subject must be coded on the verb. Russian has a case‐ marking system on all nominal categories. The category object is marked through the accusative case. The system of case marking enables the coding of a rich set of semantic relations between the predicate and noun phrases, which interacts with the inherent properties of verbs. Russian also has a set of prepositions that interact with the properties of verbs and with case marking on nouns, thereby increasing the coding possibilities. Mandarin Chinese, according to some, has a category subject and object, but according to others it does not (LaPolla 2009). There are no formal means in Mandarin to distinguish subject from other grammatical relations. Similarly, the existence of the category object in contemporary Mandarin is problematic (LaPolla 2009). Even if one accepts the existence of the categories subject and object, verbs that correspond to transitive verbs in other languages can occur without any noun phrase. In the Sino‐Russian idiolects we find that the clausal structure very much resembles the clausal structure of Mandarin in that the verb can occur without any noun phrase, with one noun phrase, or with two noun phrases, without any formal marking of grammatical relations and without any marking of semantic relations. Here is an example where the first clause has no nominal argument. The second clause has a nominal argument ′rusk’i cɪlu′wek ‘Russian man’ but its role is not overtly marked:

Language Contact  705 Boris (41) ′prɔsə ′zz’esə // ə‐ə tzara′bɔdə ɪ dam ′rusk’i ʧɪlə′wekə just here work CONJ:i there Russian man gruz‐ əə ′kruʧɪkə// load … loader ‘I just work here. And there [are] Russian people, [including] loaders.’ The only coding of semantic relations is through prepositions: Boris (42) ɪ// ya// tkag′da pr’ɪ′exar v ra′s’iu/ ɛ′dʌ// CONJ:i 1SG when come: SG:PAST:PRF PREP:in Russia DEM pʌ ′rusk’ɪ ɪ′zɪgə// ′tɔʒə n’ɪ′cyo n’ɪ bʌn’ɪ′mayu// in Russian language also nothing NEG understand:1SG:PRES ‘And when I came to Russia, I couldn’t understand anything in Russian.’ The absence of the coding of grammatical and semantic relations of the noun phrase does not necessarily mean that the speakers of Sino‐Russian idiolect repeat the pattern of Chinese. It may be that they did simply not encode these relations in the grammatical system of Sino‐ Russian and that the similarity with the absence of coding in Chinese is a coincidence.

9.6  Topic‐comment Distinction and the Comment Clause Most scholars of Chinese postulate that topic‐comment functions are important in Chinese grammar. The topic‐comment distinction in Mandarin is coded by relative order, whereby the utterance‐initial position codes the topic and what follows is the comment. This coding means is to some degree replicated in the Sino‐Russian idiolects, although the function of this formal means cannot be interpreted as the traditional topic‐comment distinction. Consider the following utterance, where the first part, u m’in’e k’ɪ′day, presumably ‘at me, with respect to China’, cannot be considered the topic of the second part, ‘I live in the town of Donning’: Boris (43) u m’n’i k’ɪ′day// du′n’in PREP:at 1SG:ACC China Donning ‘In China, I live in the city of Dongning.’

′gɔrt town

ʒiwu// live:1SG:PRES

Our interpretation, as discussed earlier, is that the bipartite division of the utterance in Sino‐ Russian instructs the listener to interpret the second part in connection with the first part. However, as discussed further in the present chapter, several Sino‐Russian idiolects do have the category comment clause, which is distinct from the bipartite division of utterances.

9.7  Discourse Connectors We use the term “discourse connectors” to refer to markers or constructions whose function is to indicate semantic relations between the elements of discourse. These constructions can consist of a single free morpheme or a phrase. These markers include: ′tᴐʒə ‘also’ and ɪ′ʃtʃɔ ‘still, also’, which are frequent components of utterances in discourse.

706  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko Sveta (44) ′tᴐʒə pɑdɑ′r’ɪ tsv’etə// pɑdɑ′r’ɪ sv’e′tə//   ′nɑdɑ/ ′tᴐʒə also give flowers:PL give flowers:PL need also ′t’en’g’ɪ/ ′tᴐʒə es’ ɑb’yɑzɑt’el’nɑ/ money also be necessarily ‘Give flowers, give flowers, and money, it is necessary [to give money].’ Boris (45) ɪ′ʃtʃɔ ′zz’esɨ/ ka′dɔrə xr’eba n’ɪ xr’eba// karba′sa also here  REL bread sausage bread NEG ′ɔʧɪn’ ′gusnɨ// very tasty ‘And one more thing, bread and sausage are very tasty here.’

wɔd/ PRS

9.8  Locative Predication Various Sino‐Russian idiolects appear to code a locative predication that has two subdomains: directional and non‐directional. With directional verbs, the complement is not marked by a preposition: Ivan (46) yɑ vᴐt 1SG PRS ʻI traveled to Russia.ʼ

′ɛtә// DEM

rᴧ′siya Russia

′esd’ɪrᴧ// travel:SG:PAST

With non‐directional verbs, the complement is marked by a preposition. In the following example, the first clause has a directional verb and no preposition before the locative complement. The second clause has a stative verb and a preposition before the locative complement: Boris (47) ja v ra′s’iu u′ʒɛ// ′vos’mə l’edə 1SG PREP:in Russia already eight years ‘I have already been working in Russia for eight years.’ Ivan (48) ɫʌ′siya ′ɫʌ′tɪ̇ fɔs′tɔk ′estɪɫʌ Russia Vladivostok travel:SG:PAST:IMPF nʌ ′mɔlә ′bïɫʌ PREP:on sea be:SG:PAST ʻI traveled to Russia, Vladivostok. I was at the sea.ʼ Sveta (stative locative): (49) yɑ z´d’esə/ 1SG here ‘I am here, at the sawmill.’

nɑ PREP:on

p’ɪɑ´rɑmɑ// sawmill

zəra′bɔdɨ// work

Language Contact  707 While these constructions have been found in various idiolects, they are not the norm for Sino‐Russian, as some speakers did use a preposition before the locative complement of a directional verb of movement: Boris (example already cited) (50) ɪ// ya// tkag′da pr’ɪ′exar v ra′s’iu/ ɛ′dʌ// CONJ:i 1SG when COME:SG:PAST:PRF PREP:in Russia DEM pʌ ′rusk’ɪ ɪ′zɪgə// ′tɔʒə    n’ɪ′cyo n’ii bʌn’ɪ′mayu// PREP:in Russian language also   nothing NEG understand:1SG:PRES ‘And when I came to Russia, I couldn’t understand anything in Russian.’

9.9  Presentative Function The presentative function (Hetzron 1971), i.e. the introduction of a proposition or another element of discourse appears to be coded by a variety of formal means in Sino‐Russian. Two of these involve the use of the verbs of perception smatri ‘look!’ and viʃ ‘see’: Sveta (51) ´mnɔgɑ// fs’e ´rɑsnəy many all different ‘There are a lot of different dishes.’ nu smɑt´r’ɪ// ´m’ɑ:sʌ well see:IMP meat ‘You see, [there is] meat, fish, chicken.’

es’n’ɪ be:INF

p´l’utɑ// dish:PL?

´ripʌ fish

´ku;rə// chicken:PL

Nina (52) v’ɪʃ / pɑstɑ′j’ɑna kl’ɪ′jenta xɑ′d’ɪ/ ʧu′ʧut’// see regular customer walk a little bit ‘You see, regular customers come sometimes.’ Nina (53) pɑ′ɛtɑ/ vɨʃ jɑ ′ɡ’enɑ skɑ′zɑ// so see 1SG Gena say u nɑʃə ɔ′ʒɛ dɑ ′stɑrәm ′z’z’es’ PREP 1PL:LOC already PREP:till old here ‘That’s why I said to Gena, “Let’s live here till the old age.”’

ɪdɑ′vɑj let’s

ʒit’// live

The evidence that the form codes the presentative function is provided by the fact that it occurs after a request to tell something to the listener.

10  Relationships Between Clauses The term “relationship between clauses” has in its scope what traditional grammars refer to as “complex sentences.” As per Frajzyngier with Shay (2016) a clause is the smallest constituent having a modal value.

708  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko

10.1 Parataxis The term “parataxis” refers to two clauses that are connected, where each of the clauses can otherwise occur as an independent clause. Parataxis in Sino‐Russian can be unmarked, or it can be marked by the marker ′tᴐʒə ‘also’, by the coordinated conjunction i, or by the counter‐ expectation conjunction a. With the exception of the marker ′tᴐʒə, the other three means of signaling parataxis are the same as those found in Russian. Unmarked parataxis: Nina (54) da/ l’ju′z’e xɑrɑ′ʃɔ / jɑ z’z’es’ jɑ yes people good 1SG here 1SG ‘Yes, the people are good, I’m used to living here.’

u’ʒɛ already

pr’ɪ′vәkәl’ɪ// get used:PL

Parataxis marked by the marker ′tᴐʒə ‘also’ Sveta (55) mɑ′ʃən ′nɑdɑ krɑ′s’əɪvɑ/ kɑk u vɑs car need beautiful like PREP:at 2PL:ACC ′tᴐʒə krɑ′səsɪ tsv’e′tə ′ɪl’ɪ ʃtᴐ tɑm also beautiful flowers:PL or what here ‘The car should be beautiful like in your [country], there should be nice flowers or whatever.’ The form ′tᴐʒə can connect an utterance in discourse as in the following example, where ′tᴐʒə occurs at the beginning of the utterance and then two more times inside the utterance: Sveta (56) ′tᴐʒə pɑdɑ′r’ɪ tsv’etə // pɑdɑ′r’ɪ sv’e′tə // ′nɑdɑ/ ′tᴐʒə also give flowers:pl give flowers:PL need also ′t’en’g’ɪ/ ′tᴐʒə es’  ɑb’yɑzɑt’el’nɑ/ money also be necessarily ‘Give flowers, give flowers, and money; it is necessary [to give money].’ Nina (57) ′nrɑɪtsɑ// da ɪ ′sɑm ′ɡlɑvnɑ ′vɔzdә xɑrɑ′ʃɔ/ tɑk xɑrɑ′ʃɔ// like  yes CONJ:i INTENS important air good so good ’[I] like [it here], and the most important, that the air is good.’ ′vɔzdu plɔx// air bad ‘The air is bad.’ i vɑ′dɑ CONJ:i water ‘And the water is so‐so.’

ʧ ′ʧut// a little bit

Lida (58) a ´lanʼʈʂʰɨ/ s’ɔ:// s’ɔ: mala´daja// ´ʐǝn’ǝʼts// CONJ:a earlier all all young:SG:F get married ǝ‐ǝ// ´wɔs’ɪmˊnats ˊl’eta// tvats  l’et// eh eighteen years twenty years ‘And earlier, they would get married young, eighteen or twenty years [old].’

Language Contact  709 The robust presence of Russian clausal conjunctions in the Sino‐Russian idiolect, in harmony with the phenomena observed across languages, indicates that clausal conjunctions are readily borrowed between languages.

10.2  Conditional Sentences Conditional sentences copy the Russian pattern, with the protasis clause marked by the ­particle ′eslɪ ‘if’ and the apodosis clause marked only by the pause that precedes it. The protasis and apodosis clauses have the same form of the verb, cf. the infinitive ending: Sveta (59) ′eslɪ ʧʊ′ʒᴐj// stʊ′ʧɑtɪ/ nɪl’zɑ ɑtk′rәʃʧ’ if stranger knock:INF PROHIB open:INF ‘If a stranger knocks, it is forbidden to open the door.’

′dvɛrɑ// door

The agent does not have to be mentioned in either the protasis or the apodosis clause: Sveta (60) ′esɪ ɑt′krət’/ ′nɑdɑ t’en’gɪ/ sprɑ′sɪ// if open need money ask ‘If/when [they] open the door [it is customary to be] asked for money’, or ‘If they open the door, they have to ask for money.’

10.3  Comment Clause In several Sino‐Russian idiolects there exists the function of comment clause, i.e. a comment on whatever preceded in the same sentence (Frajzyngier 2010). This clause is marked by the demonstrative ɛ′dʌ or ɛ′tʌ. The comment clause may be an apodosis of the temporal protasis ‘when’: Boris (example cited before) (61) ɪ// ya// tkag′da pr’ɪ′exar v ra′s’iu/ ɛ′dʌ// CONJ:i 1SG when arrive:SG:PAST:PRF PREP:in Russia this pʌ ′rusk’ɪ ɪ′zɪgə// ′tɔʒə  n’ɪ bʌn’ɪ′mayu// n’ɪ′cyo PREP:in Russian language also    nothing NEG understand:1:SG:PRES ‘And when I came to Russia, I couldn’t understand anything in Russian.’ The comment clause may be a comment on a topicalized noun phrase. As in Chinese, the topicalized noun phrase does not have to be a part of the proposition: Boris (62) s′trɔyka// ɪ′ʃtʃyo vət ɛ′da// ma′ʃɨnə ku′b’ir/ na construction again PRS DEM truck buy on ta′var ′vɔd’ɪtsʌ// cargo carry ‘Construction, [I] bought a truck, travel to the field, carry cargo.’

′pɔr’e field

′ed’ɪʒə// go

The comment may have in its scope the theme of the narrative, not necessarily the component directly preceding the comment clause:

710  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko Nina (63) pɑ′tɔ:m/ ′ɛtɑ l’yu′z’ej dɔs ɔ′ʒe ′skɔkɑ ′mnɔɡɑ then DEM people:PL:GEN also already how many l’et ′z’z’es’// years here ‘Then, as for people, we’ve been living here for so many years...’ Nina (64) ′ʧᴐta n’ɪ ′znaju// ′ɛta/ somehow NEG know:1SG:PRES DEM ‘Somehow I don’t know. Well, also people...’

l’ju′z’ej people:GEN

dɔs also

/m... mm

The comment clause may be a comment on a manner or temporal adjunct: Boris (65) i: // nu ′ʃtʃasə:/ s′gɔrə// ′srei ′gɔdə// CONJ:i WELL now soon next year ‘And now, soon, next year, there will be a baby.’

′ɛda// DEM

′b’eb’ɪ baby

′bud’ɪt// be:FUT

The comment clause may be a comment on a phrase marked by the presentative marker vət: Boris (66)

ɪ′ʃtʃyɔ vət ɛ′da// na ʒd’es’ɪ ra′bɔdəɔ/ ′nada also PRS DEM well here work need ra′bɔdə// work ‘And then, here the work is need, the work in Russia.’

v PREP:in

ra′s’iu Russia:ACC

Ivan: (67) yɑ vᴐt 1 SG PRS ʻI traveled to Russia.ʼ

′ɛtә// DEM

rᴧ′siya Russia

′esd’ɪrǝ// travel:SG:PAST

Comment clauses can also be marked by the confirmation particle da ‘yes’: Sveta (68) pɑ′tᴐm ftɑ′rᴐɪy d’en’/ dɑ// nu ʒе′n’ɪx then second day YES well groom n’e′v’estɑ zd’esə/ bride here ‘Then, on the second day the groom and the bride are here.’

zd’esə/ here

The importance of the comment clauses in Sino‐Russian idiolects is that they do not occur in Russian in the same environments as in Sino‐Russian idiolects. In Chinese, the remote demonstrative na introduces the material in the dicto domain (Frajzyngier, Liu, and Ye (2020)), but it also does not occur in the same environments as the Sino‐Russian demonstrative. On the other hand, the comment clause is a frequent component cross‐linguistically (Frajzyngier 2010). Our interpretation is that the presence of the

Language Contact  711 comment clause is evidence of the deployment of similar means to solve similar communicative needs. Ultimately, it is motivated by the principle of functional transparency, which states that the role of every component of the utterance must be transparent to the listener (Frajzyngier 2004).

10.4  A Change of Function? There is a construction in several of the Sino‐Russian idiolects whose form is attested in Russian but not in Mandarin Chinese. The construction has the form Noun Neg Noun. In Russian, as in some other Slavic languages, this construction expresses the speaker’s uncertainty with respect to the categorial identity of the referent. The expression ryba n’e ryba, lit. ‘fish not fish’, for example, means that the object looks like a fish but that the speaker doubts that it is indeed a fish. In Sino‐Russian the form Noun Neg Noun was recorded several times: Boris (69) ɪ′ʃtʃɔ ′zz’esɨ/ ka′dɔrə xr’eba n’ɪ xr’eba //karba′sa also here which bread NEG bread sausage wɔd/ ′ɔʧɪn’ ′gusnɨ// PRES very tasty ‘And one more thing,3 the bread and sausage are very tasty here.’ Boris (70) ɪ// dam dru′z’ya n’ɪ CON:I ADV:LOC friend:pl NEG ′vm’est’ɪ/ pagava′l’ɪ// together talk ‘And there [are] friends with whom I talk.’

dru′zya// friend:PL

The form occurs in some Sinitic languages in the area (see Aisin‐Giorro 2004; we are grateful to Alain Peyraube, p.c., for pointing us towards this work). It is possible that the speakers of Sino‐Russian have learned the construction N Neg N and attributed to it a meaning that does not exist in Russian. This type of creative approach to forms in a foreign language is a frequent phenomenon in situations of language contact.

11 Conclusions 11.1 Lexicon The language of Sino‐Russian speakers contains just a few lexical items from Chinese. This is to be expected from the conditions under which Sino‐Russian is spoken, as the use of lexical items in Chinese would not serve the purpose of communication between Chinese and Russians in Russia. Russian participants in the communicative exchange are not expected to know Chinese. This quite natural phenomenon has at least one non‐trivial implication: recall that in the domain of functions, the Sino‐Russian idiolects display functions that do not occur in the Russian grammar and that potentially occur in the Chinese grammar. The presence of functions coded in the Chinese grammatical system, along with the avoidance of Chinese lexical items, indicates that the control of vocabulary is a conscious process on the part of speakers of Sino‐Russian, while the functions coded in Sino‐Russian may not be subject to similar controls. Recall that the speakers of Sino‐Russian perceive and reproduce words only in their referential function, i.e. the function of referring to an activity or an entity in the real or imagined

712  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko world. These words are completely bereft of any grammatical properties. Would this be a property of lexical items used by first‐generation contact speakers? Examples given by Klein and Perdue would indicate this, but whether this is always the case can be answered only when we have more studies of first‐generation speakers of other contact languages.

11.2 Functions Encoded in the First‐generation Contact Grammar The functions expressed in Sino‐Russian idiolects require more extensive study, a part of the current project conducted by the authors. Functions encoded in all Sino‐Russian idiolects include: 1. Comment clause. This structure, marked by the demonstrative eta, appears to be an innovation in Sino‐Russian, as compared to both Chinese and Russian, but potentially triggered by the topic‐comment structure of Chinese syntax. 2. Modality a. assertive: unmarked b. hedging on truth, marked by intonation and the form kɑk skɑ´zɑt’, lit. ‘how to say’ c. negative, marked by the particle ne d. modality of obligation, marked by nada ‘need’ e. modality of prohibition, marked by nel’zja. Markers of the modality of obligation and the modality of prohibition can be repeated several times in the utterance. 3. Verbal predicate. The language of Sino‐Russian idiolects often codes the verbal predicate as a distinct lexical category. It does not code the relationships between the predicate and participants. 4. Topic. Similarly to Mandarin, Sino‐Russian marks the topic of the proposition through clause‐initial position. 5. Locative complements of inherently directional verbs are often unmarked by prepositions. Locative complements of stative predications are often marked by prepositions. 6. Parataxis is coded by three formal means: juxtaposition of clauses with a pause in between; the Russian coordinated conjunction i; and the Russian counter‐expectation conjunction a. Both of these conjunctions have the same functions in Sino‐Russian that they have in Russian.

11.3  Simplicity/Complexity of Grammatical Systems Our study indicates that the use of Sino‐Russian idiolects is different from that of pidgins, including Russian based pidgins (see Stern 2005 and references there) because (i) they are not used in multiethnic communication, (ii) they are not used in communication among the speakers of Mandarin, and (iii) they are not used by the native speakers of the area. Sino‐ Russian idiolects are not transmitted from one speaker to another and they are not transmitted from one generation of speakers to another. Since they are not transmitted to the next generation, they cannot become creoles. The grammatical system that we observe indicates that this system is not the result of any kind of merger of two languages. In the case of Sino‐ Russian, the proper comparison is not between Sino‐Russian and Russian, the lexifier language, but rather between Sino‐Russian and Chinese. The speakers of Sino‐Russian have no knowledge of the Russian grammatical system, nor do they strive to reproduce it. When using Russian vocabulary, the speakers of Sino‐Russian do not have the proper formal means to express all functions encoded in the grammatical system of Chinese, the functions that they learned in normal first language acquisition. The question of the “complexity” or

Language Contact  713 “simplicity,” that is of concern to McWhorter (2011), should be restated in the following way: Does the language of the first‐generation speakers have enough formal means to express all the functions encoded in their native language? We cannot give a full and definite answer to this question for a very simple reason: A grammar of Chinese that would describe all functions encoded in the grammatical system has yet to be written. It appears, however, that only a fraction of the functions encoded in Chinese can be expressed through the limited formal means available to the speakers of Sino‐Russian.

Abbreviations ACC = accusative, DAT = dative, F = feminine, GEN = genitive, IMP =imperative, IMPF = imperfective, INF = infinitive, INSTR = instrumental, NEG = negative, PAST = past, PFV = perfective, PL = plural, PRS = presentative, PRES = present, SG = singular.

Acknowledgments Part of this work was accomplished when Frajzyngier held a Visiting Professorship at the School of Pedagogy of the Far Eastern Federal University. The work was completed thanks to the College Scholar Award to Frajzyngier from the College of Arts and Sciences, University of Colorado. We would like to thank our collaborators Aleksander Kravchuk, Svetlana Mikhailova, Natalia Visarchuk, and Yana Yanovskaya, who gathered some of the texts. We would also like to thank Wolfgang Klein and Randy LaPolla for important bibliographic information and Alain Peyraube for comments on Noun Neg Noun construction. Finally, we want to thank Erin Shay for her comments on the substance and the style of this chapter.

NOTES 1 The study is based on narratives produced by native speakers of northern varieties of Chinese. The speakers are referred to solely by first names (all names are pseudonyms): Anonymous – male, time spent in Russia unknown, age unknown, education unknown, area in China where he grew‐up also unknown. Judging from his message, most probably a businessman. Sveta – female, spent about 18 years in Russia, age 47, factory worker, speaker of Putonghua, the official language in Beijing. Nina  –  businesswoman, spent about 20 years in Russia, age 49, originally from Northern China. Lida  –  businesswoman, spent about 20 years in Russia, 44, originally from Northern China. Ivan – has been interacting with Russians for about 10 years in Suifenhe, a Chinese town not far from the Russian border; age unknown, male, possibly from Northern China. Boris – age unknown, spent about ten years in Russia. 2 The Sino‐Russian text is transcribed in broad phonetic transcription. As a result, identical lexemes may be represented in different forms, since there is no uniformity of pronunciation among the speakers and even the same speaker can pronounce a given word in different ways in the same recording. Stress is marked only on polysyllabic words. Monosyllabic words are all stressed, as they are in Russian. We do gloss Russian inflectional forms even though in Sino‐Russian the inflectional markers most often do not fulfill any function. By glossing them we provide the readers with an opportunity to decide for themselves whether the markers fulfill a function or not. Translations of individual words are approximate at best and more often than not represent the meaning of the word in Russian. Material that does not occur in the Sino‐Russian original but is required to make the translations grammatical in English is enclosed in square brackets [ …]. 3 The form ka′dɔrə ‘which’ represents the form katoryj, which in spoken Russian has, among others, the function of the unspecified reference marker ‘some’.

714  Zygmunt Frajzyngier, Natalia Gurian, and Sergei Karpenko

REFERENCES Aisin‐Giorro, Yingsheng (愛新覺羅·瀛生) 2004. Manyu Zashi 滿語雜識 [Essays on the Manchu language]. Beijing 北京: Xueyuan Press 學苑出 版社. Aleksandrov A. (Александров, А.) 1884. Маймачинское наречие. [Maimachin speech] Русский филологический вестник [Russian Philological Herald] 12. Bickerton, Derek 2016 [1981]. Roots of Language. Berlin: Language Science Press. Chao, Yuen Ren. 1968. Grammar of Spoken Chinese. Los Angeles: University of California Press. Dryer, Matthew S. 2009. Verb‐object‐negative order in central Africa. In Norbert Cyffer, Erwin Ebermann, and Georg Ziegelmeyer. Negation Patterns in West African Languages and Beyond. Amsterdam: Benjamins, pp. 307–62. Fedorova, K. 2011. Language contacts on the Russian‐Chinese border: the “second birth” of Russian‐Chinese trade pidgin. In Tul’si Bhambry, Clare Griffin, Titus Hjelm, Christopher Nicholson, and Olga Voronina (eds.), Perpetual Motion? Transformation and Transition in Central, Eastern Europe and Russia. London: School of Slavonic and East European Studies, University College London, pp. 72–84. Frajzyngier, Zygmunt 1986. Propositional characterization of categories. In Scott DeLancey and Russell Tomlin (eds.), Papers from the First Pacific Linguistic Conference, pp. 10–19. Frajzyngier, Zygmunt 2004. Principle of functional transparency in language structure and language change. In Zygmunt Frajzyngier, David Rood, and Adam Hodges (eds.), Linguistic Diversity and Language Theories. Amsterdam: Benjamins, pp. 259–83. Frajzyngier, Zygmunt 2008. A Grammar of Gidar. Frankfurt‐am‐Main: Peter Lang. Frajzyngier, Zygmunt 2010. A comment predication: crossing the boundaries between the simple and the complex sentence. In Isabelle Brill (ed.), Clause‐hierarchy and Clause‐linking: The Syntax and Pragmatics Interface. Amsterdam: John Benjamins, pp. 313–32. Frajzyngier, Zygmunt 2011. Les fonctions de l’ordre linéaire des constituants. Bulletin de la Société de Linguistique de Paris 107.1: 7–37. Frajzyngier, Zygmunt, Natalia Gurian, and Sergei Karpenko forthcoming. Emerging phonology under language contact: the case of

Sino‐Russian idiolects. Journal of Language Contact. Frajzyngier, Zygmunt, Meichun Liu, and Yingying Ye 2020. The reference system of Modern Mandarin. Australian Journal of Linguistics 40.1. Frajzyngier, Zygmunt with Erin Shay 2016. The Role of Functions in Syntax: A Unified Approach to Language Theory, Description, and Typology. Amsterdam: Benjamins. Guangshun, Cao, Redouane Djamouri, and Alain Peyraube (eds). 2015. Languages in Contact in North China. Historical and Synchronic Studies. Paris: École des Hautes Études en Science Sociales. Centre de Recherches Linguistiques sur l’Asie Orientale Hetzron, Robert 1971. Presentative function and presentative movement. Studies in African Linguistics, Supplement 2: 79–106. Klein, Wolfgang and Clive Perdue 1997. The basic variety (or: Couldn’t natural languages be much simpler?). Second Language Research 13.4: 301–47. LaPolla, Randy J. 1990. Grammatical relations in Chinese: synchronic and diachronic considerations (University of California, Berkeley Ph.D. dissertation). LaPolla, Randy J. 2009. Chinese as a topic‐ comment (not topic‐prominent and not SVO) language. In Janet Xing (ed.), Studies of Chinese Linguistics: Functional Approaches. Hong Kong: Hong Kong University Press, pp. 9–22. McWhorter, John H. 2011. Linguistic Simplicity and Complexity: Why do Languages Undress? Berlin: de Gruyter Mouton. Mithun, Marianne 2004. On the assumption of the sentence as the basic unit of syntactic structure. In Zygmunt Frajzyngier, David Rood, and Adam Hodges (eds.), Linguistic Diversity and Language Theories. Amsterdam: John Benjamins, pp. 169–83. Muysken, Pieter 2018. Language contact research: three challenges and opportunities. In Rajend Mesthrie and David Bradley (eds.), The Dynamics of Language. Plenary and Focus Lectures from the 20th International Congress of Linguists, Cape Town: University of Cape Town, pp. 154–60. Oglezn’eva, E. A. (Оглезнева Е. А.)) 2007a. Русско‐китайский пиджин. Опыт социолингвистического описания [Russian‐

Language Contact  715 Chinese Pidgin. A sociolinguistic description]. Blagoveshchensk: Amur National University. Oglezn’eva, E. A. (Оглезнева Е. А.) 2007b. Русско‐китайский пиджин на Дальнем Востоке: особенности функционирования и языковая специфика [Russian‐Chinese Pidgin in the Far East: characteristics of its functioning and linguistic features]. Известия РАН. Серия литературы и языка [RAN News. Literature and Language Series] 66.004: 35–52 Oglezn’eva, E. A. (Оглезнева Е. А.) 2009. Русский язык в восточном зарубежье: на материале русской речи в Харбине [Russian in the eastern borderland. On the basis of Russian spoken in Harbin]. Blagoveshchensk: Amur National University. Oglezn’eva, E. A. (Оглезнева Е. А.) 2010. Русско‐китайское взаимодействие на дальневосточных территориях России: историко‐лингвистический очерк [Russian‐ Chinese interaction in the far‐eastern territories of Russia. A historical and linguistic sketch]. Слово: Фольклорно‐ диалектологический альманах [Words. Folklore‐Dialectal Almanac] 8: 6–25. Perekhval’skaya, E. V. 2003а. Quantification in the Russian‐Chinese Pidgin. In Pirkko Suihkonen and Bernard Comrie (eds.), Collection of Papers of the International Symposium on Deictic Systems and Quantification in Languages Spoken in Europe and North and Central Asia. Udmurt State University, Max Planck Institute for Evolutionary Anthropology, pp. 153–63. Perekhval’skaya, E. V. (Перехвальская Е. В.)) 2003b. Выражение числа в русско‐китайском пиджине [Expression of number in Russian‐ Chinese Pidgin] Лингвистика. История. Perekhval’skaya, E. V. 2004. Clause structure in the Russian‐Chinese Pidgin. In Bernard. Comrie, Pirkko Suihkonen, and Vakery Solovyev (eds.), Proceedings of the International Symposium “Typology of Argument Structure and Grammatical Relations.” Kazan: Kazan State University and Max Planck Institute for Evolutionary Anthropology, pp. 94–7. Perekhval’skaya, E. V. (Перехвальская Е. В.) 2006. Сибирский пиджин (дальневосточный вариант). Формирование. История. Структура [Siberian Pidgin (Far‐eastern variety). Formation, history, structure] (Saint Petersburg doctoral dissertation).

Perekhval’skaya, E. V. (Перехвальская Е. В.) 2008. Русские пиджины [Russian Pidgins] Saint Petersburg: Aletea. Perekhval’skaya, E. V. 2010. (Перехвальская Е. В.) 2010. Современные формы русского пиджина: уссурийский вариант [Contempoary forms of Russian pidgins: Ussuriysk variant]. In Arto Mustajoki, Ekatarina Protassova and Nikolai Vakhtin (eds.), Slavica Helsingiensia 40 [Instrumentarium of Linguistics. Sociolinguistic Approaches to Non‐Standard Russian] Helsinki, pp. 166–78. Perekhval’skaya, E. V. (Перехвальская Е. В.)2013. Pусско‐китайский пиджин и pусский “инtepязык” [Russian‐Chinese pidgin and the Russian “interlanguage”] Труды института лингвистических исследований. [Studies of the Institute for Linguistic Research] Vol. 9, Part 3. St. Petersburg: Nauka. Perekhval’skaya, E. V. (Перехвальская Е. В.) 2014. Исследования по русским пиджинам. Сборник статей. [Studies on Russian pidgins. Collection of articles] Berlin: Direct‐Media, p. 224. Rosenkvist, Henrik 2015. Negative concord in four varieties of Swedish. Arkiv för nordisk filologi 130: 139–66. Schuchardt, S. G. (Шухарт Г.) 1884. Маймачинское наречие [Schuchardt, Hugo. Maimatchin speech.]. Русский филологический вестник [Russian Philological Herald] Vol. 13. Stern, Dieter 2005. Taimyr Pidgin Russian (Govorka). Russian Linguistics 29: 289–318. Shapiro, Roman 2012. Chinese Pidgin Russian. In Umberto Ansaldo ed. Pidgins and Creoles in Asia. Benjamins: 1–58. Shprintsyn, A. G. (Шпринцин, А. Г.) 1968. О русско‐китайском диалекте на Дальнем Востоке [On Russian‐Chinese dialect in the Far East [of Russia]] Страны и народы Востока [Regions and peoples of the East] VI. Мoscow. Zhdanova Nadezhda Aleksandovna, A. (Жданова Надежда Александровна) 2016. Современный русско‐китайский пиджин Забайкалья в структурно‐системном и коммуникативном аспектах (на материале речи китайского этнолекта) [Contemporary Russian‐Chinese Pidgin of Baykal Region. Structural‐systemic and communicative aspects. Based on the speech of Chinese ethnolect] (University of Ulan‐Ude University Ph.D. dissertation).

36 Language Contact and Australian Languages JILL VAUGHAN AND DEBBIE LOAKES

1 Introduction The language ecology of twenty‐first‐century Australia is one of significant diversity and rapid change. The continent is home to some 25 million people – almost 4% of whom are Aboriginal – using a great variety of language types: traditional Aboriginal languages and more recent arrivals from across the globe, contact varieties like creoles and mixed languages, and many embryonically localized Englishes. At the time of British invasion in 1788, 800 Aboriginal varieties were spoken across Australia, grouped into over 250 distinct languages. The linguistic effects of colonization on Aboriginal communities have been catastrophic: at present fewer than 20 of these languages are still learned by children (Koch and Nordlinger 2014; McConvell et al. 2005), and these are largely limited to communities across northern and central Australia where settlement and farming endeavors were less intensive (see Map 36.1). Recent years have seen revitalization efforts on behalf of traditional languages across the country (Walsh 2012) and especially in the southern states where the effects of colonization have been most drastic, but the picture remains one of widespread language shift and loss. Since colonization, various contact languages have emerged from encounters between traditional languages and English, and between traditional languages and other contact varieties. Many of these varieties are now spoken as the main language of large numbers of Aboriginal people in Australia, especially Australian Kriol and varieties of Aboriginal English (both L1 and L2). Aboriginal peoples’ linguistic repertoires are typically complex, often including traditional languages and English alongside such contact varieties, combined in diverse patterns of multilingualism which are responsive to shifts in domain, interlocutor and other sociocultural factors. A good deal of recent work in Australian language contact focuses on the colonial/postcolonial period, but of course language contact predates European arrival, and is not only limited to contact with English or English‐influenced varieties. Examples of other kinds of pre‐European encounters that had significant linguistic impacts include contact with Macassans who fished along coastal Arnhem Land in northern Australia for trepang (sea cucumber) from the late seventeenth century (Evans 1992; Urry and Walsh 1981; Walker and Zorc 1981),

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

1 Alyawarr 2 Anindilyakwa 3 Anmatyerr 4 Arrernte varieties 5 Burarra 6 Guugu Yimidhirr 7 Kalaw Kawaw Ya/Kalaw Lagaw Ya 8 Kunwinjku 9 Luritja 10 Martu Wanga 11 Murrinhpatha 12 Ngaanyatjarra 13 Pitjantjatjara 14 Tiwi 15 Warlpiri 16 Yolŋu Matha varieties 17 Yumplatok 18 Dyirbal 19 Gurindji Kriol 20 Light Warlpiri 21 Pintupi

Map 36.1  Approximate locations of Australian languages with more than 500 speakers (Australian Bureau of Statistics 2016)1 (1–17) and some other languages referred to in the chapter (18–21).2 Map: Jill Vaughan.

Language Contact and Australian Languages  719 and contact with Malay, Japanese, and Chinese pearlers in the early twentieth century (Hosokawa 1987) in Western Australia, as well as ­ between traditional Aboriginal ­languages, as we will see. A broad view of contact in the region shows that the nature of contact between pre‐existing Aboriginal languages has been quite distinct from that between Aboriginal languages and non‐Australian/contact varieties. Predictably, the linguistic outcomes are similarly distinct. This chapter addresses both pre‐ and postcolonial language contact, and we draw out some contrasts between these in Section 4, considering case studies of language contact situated at either end of the continuum. While the boundaries of language communities are in reality constantly in flux, traditional languages in Aboriginal Australia are ideologized by their speakers as primordially connected to particular tracts of land  –  the mythological and eternal homeland of that linguistic variety and its people. Language communities in the region have long been small by global standards, with some having a few thousand speakers and others as few as a hundred or so. This fact, along with high levels of individual and societal multilingualism and other cultural priorities, scaffolds the region’s linguistic diversity (Singer and Harris 2016; Singer 2018; Vaughan and Singer 2018). The connections between language and land have genuine effects in lived experience, although they can obscure the actual patterns of mobility of Aboriginal people in the past and the present. The outcomes of language contact in Australia are crucially shaped by such ideological constructions, as well as locally salient social categories and real‐life encounters and mobilities. However, work on contact in the region (and elsewhere) has not typically fully appreciated these interacting forces. Contact linguistics in Australia is a rapidly developing field, and this is an important and exciting point in its evolution. On the ground, in communities, the destructive processes of language shift and loss continue apace, yet the resilience and creativity of Aboriginal speakers and signers is evidenced in the observable linguistic consequences (language change and creativity amongst Aboriginal people in northern Australia is explored in detail in Simpson 2013). New varieties are born all the time, serving novel and evolving social goals, and meeting the needs of the diverse life projects of Aboriginal Australians. Language contact and innovation is not limited to small, remote communities, but is occurring every day in larger regional hubs and in urban settings too, in cities across the country. Concurrently, researchers are bringing new questions and tools to the study of language contact in the region, enabling the field to explore broader time‐scales and more diverse contexts and datasets than has been possible before. In this chapter, we use the term “Australian languages” to encapsulate traditional Aboriginal languages (both spoken and signed), as well as the full range of Australian contact languages, which include pidgins, creoles like Kriol, mixed languages like Light Warlpiri, and koines like Dhuwaya. To the broader category of “Australian varieties,” we add the many Aboriginal English dialects spoken across the country, which are also born of contact (e.g. Malcolm 2013). We endeavor to provide an updated account of contact in Australian languages, both in terms of new scenarios of contact on the ground and new research on the topic, as well as revisiting key long‐standing debates in the field. Evidently, language contact in such a large and diverse region covers extensive ground and we cannot hope to cover everything of importance. This chapter incorporates a fundamentally sociolinguistic perspective on contact by drawing on the growing body of recent work which centralizes these issues – including work on contact language use and social identity, child language in the classroom, local perspectives on contact, and two case studies from our own research into different aspects of language contact. For further complementary overviews of the field, we point readers particularly to Meakins (2014) and McConvell (2010).

720  Jill Vaughan and Debbie Loakes

2  Language Contact Outcomes In this section, we briefly summarize the categories of contact languages attested in the Australian context, noting their linguistic and social characteristics.

2.1  Pidgins and Creoles Australian pidgins are likely to have been historically underreported. Owing to their transient nature, pidgins may arise, be of use for a short time to meet some particular need, and then disappear without documentation. Those we do know of in the Australian context have been especially important for communication at colonial frontiers and along the seafront in trade and fishing. Most, but not all, have been English‐based. Known Australian pidgins include: New South Wales Pidgin (Troy 1990; Malcolm 2018): used by local Aboriginal people and ­settlers in the Sydney area from 1788, spread into the Pacific, and to Queensland (as Queensland Pidgin English). Evolved as an important lingua franca and provided the basis for northern Australian creoles and other varieties. Northern Territory Pidgin English (Mühlhäusler 1996; Shnukal 1988): mid nineteenth century, possibly a relexified form of Macassan (see below), on Cobourg Peninsula, Northern Territory. Western Australian Pidgin English (Mühlhäusler 1998; Mühlhäusler and McGregor 1996): used among Aboriginal people and settlers around Perth and Fremantle from 1829, fed into modern‐day Kimberley Kriol. Pidgin Ngarluma (Dench 1998): a nineteenth‐century pidgin from the Pilbara, Western Australia, possibly spread south from Ngarluma country by pearling luggers. Jargon Kaurna (Simpson 1996): a nineteenth‐century pidgin between local people and colonizers around Adelaide in South Australia. Macassan (Evans 1992; Urry and Walsh 1981; Walker and Zorc 1981): a late seventeenth‐ to twentieth‐century pidgin based on elements of languages of South Sulawesi, used along the northern coast of Australia and leaving linguistic traces in Arnhem Land. Broome Pearling Lugger Pidgin (Hosokawa 1987): an early twentieth‐century pidgin used between local Australian, Malay, Japanese, and Chinese pearlers in Broome, Western Australia.

While most of these pidgins fell out of use after their contexts of utility ceased to exist, varieties originating in the original New South Wales (NSW) and Queensland pidgins provided the foundations for various contemporary contact varieties in Australia. In parts of southern Australia these pidgins acrolectalized, becoming more like English, while in the very different language ecologies of the north they creolized into several related creole varieties. Where Aboriginal people from diverse language groups were (usually brutally) brought together on cattle stations, missions and reserves, and where children were grouped together in schools and dormitories, the pidgin became a lingua franca alongside traditional languages (Munro 2000), and soon a native language – Australian Kriol (see e.g. Harris 1991). As with creoles that arose elsewhere within European colonial expansions, the development of Kriol was motivated by communicative pressures created by new trajectories of population movement, often forced, resulting in disparate groups coming together in new communities. Kriol is English‐lexified, but a great deal of the structural, semantic, and phonological properties are attributable to influence from local Aboriginal substrate languages. Some are shared across regions and some more localized, such as the presence of inter‐dental stops (not present in other Kriols) and semantic categories in the pronominal

Language Contact and Australian Languages  721 system of Roper River Kriol, which are transferred from local languages of that region (Meakins 2014; Munro 2011; Sandefur 1979). Lexical borrowing from regional languages is also widespread for semantic categories such as kinship terms, plants, and animals. The term “Kriol” therefore takes in a number of mutually intelligible lects stretching across northern Australia, and debate surrounds the question of whether Kriol can be thought to exist on a “continuum,” with speech ranging from more basilectal forms (i.e. closer to traditional‐language substrates) to more acrolectal forms (more like Aboriginal or Standard Australian Englishes) (see e.g. Bundgaard‐Nielsen and Baker 2016; Schultze‐Berndt, Meakins and Angelo 2013). Kriol is now the Australian Aboriginal language with the largest speaker community, likely around 20,000 in total. It is in daily use as a first language, yet in many communities is not accorded status as a ‘real’ Aboriginal language as traditional languages are. Its use is strongly socially motivated, functioning as an in‐group language with complex sociopolitical indexicalities (Meakins 2014; Rhydwen 1995, 1996). Another important creole in the Australian context is Torres Strait Creole, spoken in the Torres Strait and the northwestern Cape York peninsula, and known to its speakers by various names, such as Yumplatok, Broken, and Pizin. The language originated in Melanesian Pidgin English and Kanaka English (Meakins 2014), and owes its roots in part to the pearling and trepang trades which created the need for an industry lingua franca. It is thought the earlier pidgin creolized independently at several locations in the Torres Strait islands from the late nineteenth century (Schnukal 1988). Torres Strait Creole’s partially shared ancestry with Australian Kriol is reflected in many of its linguistic forms, but local substrate influences distinguish parts of its lexicon and structure. Contemporary Torres Strait Creole is under‐researched, but census data put its speaker population at some 6,000 first‐language speakers (likely an underestimate), with tens of thousands more speaking it as an additional language (see also Angelo and McIntosh 2014).

2.2  Mixed Languages Unlike pidgins and creoles, mixed languages arise in contexts of community bilingualism where a shared language exists already, and usually where some identity‐related goal – beyond simple communicative requirements – is served by the emergence of a new variety (Meakins 2013, 2014). The linguistic admixture in the case of a mixed language involves the fusion of both lexical and structural elements from two languages, such that the resulting stable variety cannot be said to have just one single linguistic ancestor (Matras and Bakker 2003). Recent work on various mixed languages in the Australian context amply illustrates how language genesis may be socially motivated, and how code‐switching practices in one generation may contribute to a more stabilized variety in the next. Light Warlpiri, a mixed language from Lajamanu in Central Australia, has come to index membership in a “young Lajamanu Warlpiri” community of practice, and systematically combines elements from Warlpiri (nominals and case‐marking), English (some nominal elements), and Kriol/Aboriginal English (verbal structure) (O’Shannessy 2015). In the Victoria River District of northern Australia, the mixed language Gurindji Kriol mirrors the pattern of Light Warpiri, drawing nominal constructions from Gurindji and verbal material from Kriol. This variety found its origins in the context of the Gurindji workers’ strike and land‐rights struggle of the 1960s–1970s, and continues to express a connection to Gurindji identity (Meakins 2008b). The Australian context provides exciting opportunities to witness and understand the emergence of mixed languages, effectively in real time. This is of vital importance to the field, as mixed languages documented elsewhere have typically been well‐established for 100 or more years before descriptions were first undertaken (Meakins 2008a: 148).

722  Jill Vaughan and Debbie Loakes

2.3  Koines and New Dialects Koines are distinct from both creoles and mixed languages, as they are born of contact ­between closely related varieties, such as dialects of the same language. Koines are the stabilized result of mixing between lects, leveling of linguistic differences, and typically simplification (Siegel 1985). As with the genesis of many other Australian contact varieties, they emerge from new encounters between groups, changing mobilities and social upheaval, and so despite not (necessarily) involving influence from English or Kriol, all koines thus far described in Australia are nevertheless products of colonization. Koine varieties have predominantly been studied “post hoc,” generations after the initial mixing, and so it is often difficult or impossible to ascertain the linguistic input and the exact processes that have occurred as the contributing varieties may no longer exist, or may have undergone significant change in the intervening time (Trudgill et al. 2000: 67). Few koines have been described in the Australian context; it is likely that more exist but have not been reported (Amery 1993: 1). Dhuwaya, which developed in Yirrkala, Arnhem Land (Amery 1985, 1993), is the best described (in terms of its status as a koine), but koineization processes are also suggested to have occurred in languages such as Papunya Luritja (Heffernan 1984), Modern Djambarrpuyngu (Devlin 1985; Wilkinson 1991), Burarra (Vaughan 2017) and the Bininj Kun‐wok varieties Kunwinjku and Manyallaluk Mayali (Evans 2003a). Dhuwaya is a variety of Yolŋu Matha – a pandialectal cover term encapsulating some 30 language varieties spoken in northeastern Arnhem Land – but unlike other Yolŋu Matha varieties it has no clan‐specific affiliation and is spoken by virtually all Yirrkala community members. It emerged from prolonged contact between speakers of different Yolŋu Matha clanlects and originated at least in part from community baby‐talk registers (Amery 1985, 1993). Dhuwaya’s morphology is somewhat simplified, or arguably regularized, as compared to the original dialects. Although the language has often been referred to as a lingua franca, Amery considers it rather to be a “communilect” – a linguistic variety specific to a particular community, as distinct from other varieties of the same language. Unlike a lingua franca, Dhuwaya was never needed to facilitate communication between speakers of different languages. Instead, it served the social goal of highlighting “solidarity within the peer group” (Amery 1985: 128), and is characteristic of Yirrkala and its surrounding homelands.

2.4  Restructured Traditional Languages A further category of language contact outcome pertains to languages which have undergone internal changes following contact with English, Kriol, or another (not closely related) traditional language. Often these restructured varieties come to serve new goals within changing social structures and patterns of mobility. A few key examples are provided in this section, but many more exist in the literature, and more still are likely to be as yet undescribed. Given that so many speakers of Australian languages are multilingual with good command of English and/or Kriol, it is inevitable that traditional languages across the region will show the influence of contact with these newer arrivals, and many have undergone significant restructuring. Examples can be found of these influences within most, if not all, linguistic subsystems of traditional languages, but most striking are changes to the lexicon and the grammatical system. Borrowings from English/Kriol are attested in many descriptions of traditional languages, although few studies center solely on this phenomenon. In fact, it is likely that language descriptions under‐report the extent of borrowings in actual use due to their (understandable) focus on more classical or “pure” registers of the language (in the sense of linguistic stock). Of course, borrowings in the other direction are also commonplace, for example from local languages to Kriol.

Language Contact and Australian Languages  723 A number of studies address structural changes in traditional languages under the influence of English, where the lexical material remains largely unchanged. Many focus on changes in the case system of Pama‐Nyungan languages like Dyirbal (Schmidt 1985) and Warlpiri (Bavin and Shopen 1985), and exemplify common paths for morphosyntactic features resulting from language contact. Languages may develop such features as optional ergativity and more fixed, English‐like word order (see also Meakins and O’Shannessy 2010). Meakins (2014: 404–5) also identifies changes in possessive constructions as a common change in traditional‐language restructuring in Australia, highlighting especially the collapse of the alienable/inalienable split in nominals. The role of code‐switching in borrowings and structural change is a matter of some debate, but it is widely acknowledged that code‐switching is at least one “processual mechanism” of contact‐induced language change (e.g. Backus 2005; Thomason 2001: 131–6). While a small body of work on code‐switching in the Australian context has emerged (see e.g. Disbray 2009; McConvell 1985, 1988; Mushin 2010; O’Keeffe 2016; Singer and Harris 2016; Vaughan 2018b, as well as, e.g. Meakins 2011, 2012 and O’Shannessy 2006, 2012 on code‐switching and mixed language genesis), it is hoped that further diachronic work on the topic will illuminate these connections. Traditional‐language restructuring may be framed as a stage in the path to language obsolescence, as in the case of Schmidt’s (1985) work on Dyirbal. But restructuring need not foreshadow language loss. It may even add new dimensions to the social life of a language, as with Areyonga Teenage Pitjantjatjara and its related secret “short‐way” register (Langlois 2004, 2006); these are youth varieties of the Western Desert language Pitjantjatjara which draw many features from English and index teenage culture. Similarly, Murriny Kardu Kigay, a “slang” variety of Murrinhpatha spoken at Wadeye in the Daly River region of northern Australia, borrows and reinterprets English vocabulary and serves as an in‐group language for young men in the community (Mansfield 2014). Of course, restructuring of traditional languages may also occur as a result of contact with languages other than English or Kriol. Examples of contact between traditional languages abound, although historical‐comparative linguistic approaches have been characterized by various premature judgments, heated debates, and methodological disputes (see Koch 2014 for a summary). It has been suggested that languages in the Australian region have undergone exceptionally high rates of borrowing (e.g. Dixon 1970; Heath 1981), but this claim is subject to much debate (e.g. Black 1997; Bowern and Atkinson 2012). It is likely that high levels of long‐standing multilingualism in the region (Brandl and Walsh 1982; Evans 2007) have produced contact effects, but the precise nature of these has yet to be resolved (cf. Harvey 2011, see further discussion in Section 4.2). Pintupi‐Luritja is an example of a variety born of contact between traditional languages within a sociopolitical context that was able to be documented (unlike most historical work) (Holcombe 2004a). Pintupi‐Luritja is primarily based in the Pintupi language, but exhibits features of neighboring languages such as Warlpiri and Anmatyerr. It developed as a lingua franca and a communilect (Heffernan 1984) among speakers of different languages who moved (or were driven unwillingly by colonists) to Papunya, Haasts Bluff, and Amunturrngu (Mt. Liebig) in Central Australia around the mid twentieth century. In more recent decades, it has come to serve an important role in the construction of a wider regional identity connected to the Indigenous polity of the Haasts Bluff Land Trust (Holcombe 2004b). We will see a further example of contact between traditional languages from Arnhem Land in Section 4.2.

2.5  Aboriginal Englishes The eradication of Indigenous languages in many parts of Australia, and the changing nature of the linguistic landscape more generally, means that large numbers of Aboriginal Australians have a greater reliance on English in their everyday lives (Eades 2013a; Malcolm

724  Jill Vaughan and Debbie Loakes 2013, 2018). The variety of English spoken by Aboriginal people, Aboriginal English, is different in many ways from mainstream varieties, for example in the lexicon, grammar, sound system, timing, politeness strategies, and body language (see, for example, Adams 2017; Butcher 2008; Eades 2013a; Malcolm 2018; also see Section  4.1 of this chapter). For some speakers Aboriginal English is an L2 (spoken alongside Kriol and/or other traditional languages), whereas for others it is their L1. These speakers may also be bidialectal, speaking Standard Australian English in addition (see e.g. Oliver and Nguyen 2017; Malcolm 2018), as well as bicultural (Eades 2013b). The fact that Aboriginal English is divergent from the standard, to varying degrees, makes sense in light of the fact that “Aboriginal people had a different path toward English … [t]hey learnt it by picking it up, pidginizing it [and this] eventually became Aboriginal English” (Malcolm 2011). Eades (2013a: 417) describes various pathways (note the plural) stating that in some parts of Australia, Aboriginal English likely developed from the decreolization of contact languages, whereas in other places development occurred in the opposite direction, essentially an “Aboriginalization of English.” While this is the case, Aboriginal English in Australia is dynamic and changing according to the needs of the communities in which it exists and, as Eades (2013b: 1) notes, has “[c]ontinuing influences from traditional Aboriginal languages and cultures,” as well as from English. Like Kriol, Aboriginal English is often described as being on a continuum, ranging from acrolectal (or light – closer to the standard) and basilectal (heavy – furthest from the standard (see e.g. Butcher 2008: 635)). Because of the variability for both individuals and communities, Eades (2013b: 4) uses the phrase “Aboriginal ways of using English,” rather than Aboriginal English, to reflect the dynamic nature of individual and community ­language use. Aboriginal English is considered a contact variety (Malcolm 2013: 44). Its existence and spread arose through various means: contact between European people in the first ­instance (Malcolm 1994, 2013) and, in many places, due to attitudes, interventions, and policies after colonial invasion which prevented Aboriginal Australians from using their traditional languages. Fesl (1993) and Walsh (1993), for example, describe how Aboriginal languages were considered inferior to English, and so “English only” policies were encouraged and intergenerational language loss subsequently occurred. Today attitudes towards Aboriginal English are variable. Malcolm describes how Aboriginal English is sometimes called “broken” (2018: 216), “bad English,” or “lazy” (2018: 26). Eades (2013b: 18) describes how many people do not know that Aboriginal English is a specific, r­ ule‐ governed variety, often referring to it as an “uneducated” form of the standard. This lack of understanding is known to be problematic for intercultural communication, particularly in situations where Standard Australian English is seen as the prestigious norm, such as in legal contexts (Eades 2013b; Walsh 1994) and in education settings (e.g. Malcolm 1994; Wigglesworth, Simpson, and Loakes 2011, see Section 3). There are also, however, highly positive attitudes to Aboriginal English within Australia, both amongst speakers of the variety and amongst the non‐Indigenous mainstream population. Aboriginal English is often used for the expression and transmission of cultural identity; Malcolm (2018: 24) gives recent examples and describes how this can reinforce positive attitudes. Similarly, Oliver and Nguyen (2017) give examples of how Aboriginal English is used in social media, also reinforcing positive attitudes amongst speakers. Aboriginal English is continuing to be more visible in mainstream Australian media via songs, television and movies which have broad appeal. Some recent examples are the cartoon series Little J and Big Cuz (see e.g. Owen 2019), and the successful rapper Baker Boy, who sings in the Aboriginal language Yolngu Matha, as well as in Aboriginal English (see e.g. Vadovac 2017). Further examples of the varied and changing attitudes to creoles and Aboriginal English in Australia are discussed by Eades and Siegel (1999).

Language Contact and Australian Languages  725

3  Contact and the Classroom The way that Aboriginal languages and varieties come into contact with the monolingual English‐dominated education system is one way that the ripple effects of colonization continue to impact communities of Aboriginal people around Australia. In this section, we address some of the cultural, political, and educationally problematic ways in which children and families are let down by a general lack of understanding about Aboriginal contact languages and multilingualism in Australia today. As has been reported by various researchers over the years, in a report about the use of languages in classrooms back in the 1950s, UNESCO (1953: 6) stated that “we take it as axiomatic … that the best medium for teaching is the mother tongue of the pupil.” However, this is certainly not the case in 2020 in Australia, nor in numerous other places where contact languages are spoken (see e.g. Macqueen et  al. 2018; Renganathan and Kral 2017; Wigglesworth and Simpson 2018; Steele and Wigglesworth 2017). Many Aboriginal children in Australia start out multilingual – their early years are spent with one or more traditional languages, or Kriol, or a mixture of these (see e.g. Walton 1993; Simpson and Wigglesworth 2008). So they have “home” languages, and experience with the “school” language, English, often begins in the classroom (Wigglesworth and Simpson 2018). For Indigenous children growing up in many parts of Australia this exposure to the standard (mainstream) language tends to occur when they are five or six years of age and in their first years of schooling (Macqueen et al. 2018: 6). With vast numbers of Aboriginal children going to school in an unfamiliar language, L2 support would seem an obvious response, however they are rarely provided with additional learning support (Angelo and Hudson 2018). Kriol, for example, is often seen in the classroom as a “deficient” form of the standard rather than as a second language (Wigglesworth and Billington 2013; Wigglesworth, Billington, and Loakes 2013), which echoes the situation faced in many other parts of the world when creoles come into contact with the education system. Siegel (2007) provides an overview of how creole and creole speakers fare in education systems around the world. Children who speak creole‐based languages have been called “invisible L2 learners” (Angelo and Hudson 2017). This is because they may be perceived as having fluency in English due to shared lexical items (i.e. words for nouns may be shared in Kriol and English, given Kriol is an English‐based lexifier) but this can mask some of the fundamental grammatical and structural differences between such languages. Angelo and Hudson (2017) point out that this might even be invisible to classroom teachers (also see Siegel 1999). A resulting issue for Aboriginal students in the education system has been national standardized testing, which is conducted in English only; there is no recognition of the fact that many of these students do not have English as an L1 (Simpson, Caffery, and McConvell 2009; Wigglesworth, Simpson, and Loakes 2011). This standardized testing, the National Assessment Program – Literacy and Numeracy began in Australia in 2008, and has since been shown to be culturally and linguistically inappropriate for many Aboriginal children whose first languages may be Kriol, traditional languages, Aboriginal English, or perhaps a combination of these (see e.g. Macqueen et  al. 2018). For example, Wigglesworth, Simpson, and Loakes (2011) analyzed NAPLAN tests and showed the tests expected unattainable levels of cultural knowledge simply not accessible in many Aboriginal communities (i.e. highly urban concepts such as letterboxes and cinemas). Another example is that multiple choice tests which have only one “correct” answer in Standard Australian English may be ambiguous for Kriol speaking children (i.e. more than one response could be correct because of grammatical differences across the languages). After the first year of NAPLAN testing in Australia, Aboriginal children in remote areas were unsurprisingly shown to have the lowest test scores across the country, and a policy decision was made to teach solely in English for the first four hours of a school day. This

726  Jill Vaughan and Debbie Loakes change effectively scrapped bilingual education, including Kriol to English programs which occurred in some regions (Simpson, Caffery, and McConvell 2009; Devlin 2011); this is despite the fact that programs such as this are consistently shown to be an effective way for children to access education and improve literacy in their home and school languages (Siegel 2007). As pointed out by Simpson et al. (2009: 12) “supporters of the decision caricatured bilingual programs as programs to revive Indigenous languages, ignoring their function of giving children access to information through their mother tongues.” This decision has also meant that in some places, home languages have been completely banned in schools (Edmonds‐Wathen 2015). Macqueen et al. (2018) also discuss some of the personal, social, and political issues these tests cause to children and teachers in Aboriginal communities. The diminishing support for Kriols (and traditional languages) in the education system, and their invisibility in many respects, extends the picture painted some decades ago by Walton (1993: 83) who stated that there are “structural and institutional constraints on Aboriginal [children’s] access to education” in the Australian context. This has been more recently echoed by Edmonds‐Wathen (2015: 48) who explains that there are “assessment expectations that entire cohorts of Indigenous language speaking students are unable to meet … [which] institutionalizes both student and teacher failure.” Malcolm (2013: 195) also paints a bleak picture of the state of the education system in Australia when it comes to Aboriginal Australians, stating that while language situations may be complex “this is no excuse for the anomalies in present practice … [which effectively misrepresent] … Indigenous students’ language state.” See also Wigglesworth and Simpson (2018) for a recent overview of this topic. Going beyond these institutional concerns, it is important to end this section with a focus on the rich language knowledge that children bring to school. Feller and Vaughan (2017: 177) describe the concept of translanguaging – which is similar to the concept of code‐switching, but “but differs fundamentally in its core assumption that a bilingual individual’s linguistic resources make meaning in ways not clearly assignable to one ‘code’ or another.” They note the value of translanguaging in the classroom, as it effectively respects the individual and the complex language skills each student comes to school with, with social and politically defined expectations about language set aside (Feller and Vaughan 2017: 176–7). Similarly, Oliver and Nguyen (2017) describe translanguaging amongst young Aboriginal students on social media, noting that mixing between Aboriginal English and Standard Australian English was beneficial for language development in general; they also recommend the introduction and encouragement of translanguaging practices in the classroom to ensure learning is maximized. Translanguaging is thus a way of using language with “social creativity” (Simpson 2013) but this tends not to be prized or understood in largely monolingual Australian classrooms (Devlin 2011; Macqueen et  al. 2018), nor by policy makers and government more generally (Simpson 2013).

4  Contact Case Studies Language contact is often depicted as a large‐scale process acting upon entire language systems, rather than something that occurs in the minds of speakers/signers, and within the interactions of day‐to‐day life. Languages do not exist outside of the linguistic repertoires of speakers/signers, and we overlook important forces in language contact and change when we neglect to engage with this fuller picture, the “total linguistic fact” (Silverstein 1985; Woolard 2008). To this end, we include here case studies of language contact from our own work in two contrasting contexts in Australia: Aboriginal English varieties in Victoria in Australia’s southeast, and contact in the traditional languages of Arnhem Land, in remote northern Australia. These different scenarios highlight contextual factors within which

Language Contact and Australian Languages  727 contact can play out, and how these factors may shape eventual linguistic outcomes. They are chosen because they give a fuller picture of the extremes of language contact in Australia, and highlight the diversity we find across the continent today.

4.1  Aboriginal Englishes in Victoria In this section we focus on the most “acrolectal” form of a contact variety in Australia – Aboriginal English spoken as an L1 in Victoria (Southeastern Australia). As mentioned (Section 1), traditional languages are mostly in revival mode in this part of the country, and so Aboriginal people tend to have to rely on English here. Work on Aboriginal English in Victoria is quite limited, although there is some research describing its use in various locations. Malcolm (2018: 4) notes that linguistic work has been carried out on Melbourne Aboriginal English by Fesl (1977), and on Aboriginal English in Goulburn Valley by McKenry (1995). The focus in this section is to explain some of the reasons Aboriginal English in Victoria sounds different from the mainstream, by paying attention to some lexical examples, as well as focusing on some new descriptive phonetic work. As well as the above academic work mentioned, Adams (2014: 4) led a paper on Aboriginal English (and culture) called Koorified, which gives an impressive overview of linguistic features for health workers, with data collated by Aboriginal people in Victoria, and further validated by other Aboriginal people. In this publication, there is a recognition of the fact that L1 English‐speaking Aboriginal people use it in a different way from non‐Aboriginal people, and the work is designed to minimize intercultural misunderstandings, as well as provide cultural safety for Aboriginal people in the health system. As noted by Butcher (2008), Aboriginal English uses loanwords from traditional Aboriginal languages, specialized lexical items relating to kinship, and some inherited and retained features from earlier/colonial forms of English. Koorified gives various examples of such words and phrases from Aboriginal English in Victoria, which may also overlap with Aboriginal English usage elsewhere in Australia. Some examples, presented by Adams (2014) are: auntie girl – a female the same age or younger boorai – baby cultural man – a man with expertise about his culture deadly – something good or excellent gammon – joking, fake Koorified  –  make something developed by non‐Aboriginal people suitable for Aboriginal people mob – family or people shame – stigma and embarrassment story teller – someone experienced in passing on oral culture The phonetics of Aboriginal English is another important area for study because, as described by Malcolm (2018: Section  3.1) “[t]he most observable identifier of Aboriginal English to most Australian English speakers is the way it sounds.” Despite this, detailed phonetic descriptions of any type of Aboriginal English have been uncommon. In the remainder of this section, we describe new findings about the sound system of Aboriginal English in Victoria based on acoustic phonetic work led by Loakes and colleagues. These studies have looked at both consonant variability and vowel variation, contributing to the growing body of research on what specific linguistic details make Aboriginal English unique. In the consonant study (Loakes, McDougall, Clothier, Hajek, and Fletcher 2018), the focus was on intervocalic and coda /t/ variation in a small regional town called Warrnambool in

728  Jill Vaughan and Debbie Loakes the southwest of Victoria. A group of Aboriginal English speakers and a group of mainstream Australian English listeners took part in a production study, and /t/ was analyzed acoustically. Results showed that the mainstream Australian English speakers used canonical aspirated /t/, as well as variants which can be best described as “breathy”  –  this includes pre‐aspirated stops, fricated variants, and hyperlengthened affricated variants (50% of tokens in this sample were lengthened /ts/). The Aboriginal English speakers, on the other hand, used many more variants than the mainstream Australian English group, and individuals had much more variability as well. While the Aboriginal English speakers also used canonical aspirated /t/ to some extent, various other consonants were also observed – pre‐ glottalized stops, full glottal stops, ejectives, partially voiced tokens, approximants, and taps. The Aboriginal English sample had a majority of tokens which can best be described as “glottal”‐like in nature, and also had variants previously unattested for Australian English; coda glottal stops before a pause, and ejectives. Ejectives have been shown to occur in other Englishes, and have also been linked to voice quality (laryngeal setting) and sociolinguistic behavior more generally (see e.g. Gordeeva and Scobbie 2013). Relatedly, Butcher (2018: 631) notes that there is a “distinctly and recognizably ‘Aboriginal’ voice quality” used by Aboriginal English speakers. Loakes et  al. (2018) accordingly interpret the frequency of “breathy” versus “glottal” tokens across the speaker groups as partially driven by underlying phonatory setting differences, and partially driven by indexicality. This consonant study offers a description of /t/ variability in Aboriginal English compared with the mainstream variety, linking it to higher order linguistic differences amongst the groups, and providing a reason for audible differences between them. In the vowel study, Loakes (2018), compares vowel variation across two Aboriginal English groups in regional Victoria; the southwestern city of Warrnambool and the northwestern city of Mildura. These communities are approximately six hours apart by road and not linked in any way, thus the study allows some analysis of regional dialect variability amongst Aboriginal Englishes, which has not previously been carried out on the sound system. The study focused on lax vowels, which have undergone dramatic change in the mainstream variety in a relatively short period of time (Cox and Palethorpe 2008) and also included a focus on the DRESS and TRAP vowels prelaterally, as there is a merger in progress (on TRAP) in the mainstream communities in the southern parts of Victoria (e.g. Loakes, Hajek, and Fletcher 2018). Results of the vowel study showed some compression of vowel spaces, particularly in height, when compared with other recent work on Australian English (e.g. Elvin, Williams, and Escudero 2016). Close front vowels were shown to be not as high as in the mainstream community, and the TRAP and STRUT vowels in particular were less open than has been reported (Elvin et  al. 2016; Cox and Palethorpe 2008) for mainstream speakers. This study also showed a very small amount of regional variation as well, with the Aboriginal English female speakers from Mildura having fronter vowels than the male speakers. Thinking about the vowel spaces as a whole, results can be interpreted as the Aboriginal speakers not taking part in the same “innovations” that are occurring in Australian English. Similar results and a similar interpretation have been shown by Butcher and Anderson (2008) for a small group of L2 Aboriginal English speakers in another part of Australia, who also had compressed vowel spaces compared to mainstream speakers, so it is interesting that an L1 variety patterns to some extent with an L2 group elsewhere. Another interesting finding shown by Loakes (2018) was that the Warrnambool Aboriginal speakers merged DRESS and TRAP completely before /l/, and so did many of the Mildura speakers. Merger is unreported for the mainstream Australian English speakers in Mildura, so this finding shows a greater entrenchment of a sound change in the Aboriginal English groups compared to the mainstream groups. Contextualizing these findings to the literature more generally, Clopper (2014) talks about phenomena like this amongst bidialectal people, which these Aboriginal English speakers

Language Contact and Australian Languages  729 ultimately are. She notes that people exposed to multiple varieties generally have more ­variable distributions, and so the ways in which production of forms plays out is potentially more variable too – sound changes can entrench further (as seen here), and there is also the potential for leveling, and for innovations (as seen in the consonant study where ejectives and glottal stops were observed). The findings described in this section make sense in light of Malcolm’s observations (2018: 1) that Aboriginal English has no “unifying force of standardization,” so variability occurs more prevalently. In this section we have seen some marked examples of differences between L1 Aboriginal English in Victoria (an especially “acrolectal” variety) and the mainstream Australian English community. We have seen some features which contribute to the two varieties sounding different – by choice of lexical items on the one hand, and by variability in vowels and consonants (and likely voice quality) on the other. Such descriptions add to the so far quite limited knowledge we have about features of Aboriginal English that make it stand apart from the mainstream.

4.2  Language Contact in North‐central Arnhem Land Turning now to a geographically and linguistically distinct region, the case study in this ­section focuses on Arnhem Land, an Aboriginal‐owned territory on Australia’s remote north‐central coast, which is home to many cultural groups who communicate within a highly multilingual and multilectal language ecology. The region’s traditional languages are  divided by a significant boundary: between the large Pama‐Nyungan family which covers some four fifths of Australia, and several language families grouped under the non‐Pama‐Nyungan umbrella (see Map  36.2). Arnhem Land is acknowledged to be a “hotspot” for linguistic diversity: its languages represent several diverse language families and individual linguistic repertoires can draw from as many as eight traditional languages, as well as from English, a range of contact varieties like Kriol, and local alternate sign language (Singer 2018; Singer and Harris 2016; Vaughan 2018a, 2018b). The language ecology reflects aspects of what has been termed “small‐scale” or “egalitarian” multilingualism, whereby many languages with small speaker communities co‐exist without being subject to a rigid sociopolitical hierarchy of language use (e.g. François 2012; Lüpke 2016). Although

PAMA-NYUNGAN 1 Djinang/Wurlaki 2 Ritharrngu 3 YoIŋu Matha varieties

NON-PAMA-NYUNGAN 4 Burarra 5 Gurr-goni 6 Na-kara 7 Ndjébbana 8 Ngandi 9 Nunggubuyu (Wubuy) 10 Rembarrnga 11 Warndarrang

Map 36.2  Approximate locations of Arnhem Land languages discussed in this section. Map: Jill Vaughan.

730  Jill Vaughan and Debbie Loakes Arnhem Land’s language ecology is undergoing significant changes, this persisting “regional system” of multilingualism (Sorensen 1967; Aikhenvald 2002; Epps 2008; Epps and Stenzel 2013) is likely to show significant reflexes of pre‐contact patterns of language use, as English and Kriol have more marginal roles here than in much of the rest of Aboriginal Australia. In this sense, there is great potential in the region to explore the role of long‐standing multilingualism in shaping language contact outcomes. Contemporary Arnhem Land provides examples of a diverse range of contact scenarios, from significant long‐term diffusion (including across the Pama‐Nyungan/non‐Pama‐Nyungan boundary) within a regional Sprachbund, to the more recent emergence of creoles, koines, and localized Englishes. We focus here especially on north‐central Arnhem and the region surrounding Maningrida, a largely Aboriginal community of some 2,500 inhabitants. Maningrida serves as a regional hub, and community members regularly travel between Maningrida, small outstation settlements, and Darwin, the state capital 500 km to the west. Some 15 languages  –  both Pama‐Nyungan and non‐Pama‐Nyungan  –  are represented in daily interactions in the community (alongside local auxiliary sign language), and multilingual strategies such as code‐switching and receptive multilingualism (Singer 2018; ten Thije and Zeevaert 2007) are commonplace (Vaughan 2018b). Considerable attention has been given to the processes of traditional‐language contact in the region, most significantly by Jeffrey Heath in his work on structural (1978) and lexical (1981) diffusion, but also through work on the shape of particular Arnhem Land languages (e.g. Harvey 2012; Waters 1989; Wilkinson 1991), and in more general treatments such as Dixon (2002), Bowern and Koch (2004), McConvell (2010), and Harvey (2011, 2013). (Work on stabilized varieties born of traditional‐language contact in the region has already been discussed in Section  2.1, especially Amery’s (1985, 1993) work on Dhuwaya, a koine variety resulting from prolonged contact between different Yolŋu Matha varieties.) Heath describes the extensive and long‐term diffusion of features like phonological systems, grammatical morphemes, morphophonemic patterns and lexicon amongst various Pama‐Nyungan (Yolngu) and non‐Pama‐Nyungan languages, especially Ritharrngu, Ngandi, Nunggubuyu (Wubuy), and Warndarang, which he credits largely to demographic factors such as the rate of extensive cross‐linguistic marriage (1978: 14–21, 1981). Diffusion is understood to have operated both directly (resulting, for example, in the creation of new phonemes or the borrowing of bound grammatical morphemes and lexicon) and indirectly (such as in the development of obligatory bound pronouns motivated by similar patterns in adjacent ­languages). Indirect borrowing often results in functional convergence, but not necessarily one‐to‐one morphemic intertranslatability. Heath also notes that certain linguistic subsystems and classes of morphemes appear to be particularly resistant to diffusion (e.g. inflectional verbal suffixes, pronominal prefix complexes, and demonstratives). In the Maningrida region specifically, various outcomes of contact between traditional languages are proposed (see, e.g. Carew 2017; Coleman n.d.; Cysouw 2005; Dixon 2002; Elkin 1961; Green 1987, 1995; Mushin and Simpson 2008; Waters 1989). These include: • the diffusion of contemporary/precontemporary tense marking (probably) across Maningridan languages, and further east to Djinang/Wurlaki • the development of preposed pronominal enclitics in Djinang/Wurlaki through contact with prefixing regions (Rembarrnga and Burarra) • serial verb constructions across Yolngu languages and Maningridan languages • the neutralization of laminal distinctions in Djinang/Wurlaki through the influence of languages to the west • the southwards diffusion of Yolngu patrimoiety terms (elsewhere matrilineally cycled) • calquing using phrasal templates for natural species (e.g. echidna: spike loc‐back) across Maningridan languages.

Language Contact and Australian Languages  731 Many of these proposals are subject to ongoing debate and require a good deal of further attention. What is clear is that special local conditions appear to scaffold diffusion in the region; these include marriage patterns which produce intrafamilial bi/multilingualism, the small size of language groups, multi‐directional social exchange, the operation of egalitarian multilingualism, and language ideologies (see e.g. Heath 1978, 1981; Meakins and Pensalfini, forthcoming; Singer and Harris 2016; cf. Harvey 2011). It is not clear, however, that levels of diffusion in the region, and in Australia more generally, are significantly higher than other places in the world with similar sociocultural conditions. This idea has been suggested in particular by Dixon in various works (e.g. 1997, 2001, 2002), and indeed high loan rates have also been attributed to hunter‐gatherer groups more generally (Dixon various; Nettle 1999). More recent broad‐scale empirical work, however, has shown that such generalizations are untenable (e.g. Black 2006; Bowern 2006; Evans 2005, Harvey 2011) and that rates are highly dependent on local factors (Bowern et al. 2011). As discussed in Section 2.1, significant contact has occurred between traditional varieties and English, following colonization, and later also between traditional varieties and ­­stabilized contact varieties like Kriol. One outcome of this intersection is the range of code‐ switching practices attested across Australia. Arnhem Land equally evidences many forms of code‐switching, and interesting local instantiations have been observed. While our focus in this section is on contact between traditional languages, code‐switching between traditional languages and English/Kriol provides an interesting site of contrast. Code‐switching in the Maningrida region may be broadly grouped into switching between a traditional language and English or Kriol on the one hand, and switching between two (or more) traditional languages on the other. The former kind of code‐switching is commonplace, especially in Maningrida community itself. Vaughan (2018b) profiles switching in public settings between Burarra (the language with the largest speaker community in Maningrida) and English and finds that the practice is unremarkable and typically socially unmarked, but a core characteristic of communication in community “hybrid spaces” (e.g. Gutiérrez et al. 1999). While certain tendencies were observable in the shape of the switching, no hard‐and‐fast rules applied. Instead, speakers tended to “soft assemble” their linguistic resources (García and Li Wei 2014: 25; Thelen and Smith 1994) – that is, features from across the repertoire were assembled in the moment to fit the particular demands of the situation at hand (context, interlocutors, speaker goals, etc.), and patterns were flexibly adapted as new communicative needs arose. The result was predominantly code‐switching of the intrasentential type, that is switching between languages within the clause by inserting elements of one language within the matrix of another, or switching between the grammatical structures of each language. Burarra‐English code‐switching seems to have much in common with mixing practices attested in other contact zones in Australia where English and/or Kriol form part of Indigenous speakers’ repertoires. Code‐switching between traditional languages, on the other hand, is quite different in its shape and function. Examples in the literature are vanishingly rare, and those accounts that do exist are often limited to brief statements in introductory sections of grammars of specific languages (notable exceptions include McConvell 1988, 2002; Evans 2010; Bradley 1988; Rumsey 2018: 95). It is difficult to know whether this is because code‐switching rarely occurred or because it was not attended to by researchers. In the Maningrida context, Elwell’s work on multilingualism (1977, 1982) mentions code‐switching only briefly, and gives no actual transcribed data to illustrate the nature of the switches. More contemporary accounts and ongoing research in Maningrida indicate that switching between traditional languages is attested, but less frequently than the other type. When this kind of code‐switching does occur in Maningrida, striking differences may be observed. Code‐switching between traditional languages is typically highly “marked” serving readily observable communicative goals, and switches tend to occur outside the clause, i.e. intersententially (Vaughan 2018b).

732  Jill Vaughan and Debbie Loakes Differences of this kind have been noted elsewhere, for example in the Amazonian Vaupés region, where Indigenous contact has resulted in language maintenance, grammatical diffusion, and limited lexical borrowing, while colonially mediated contact has tended towards language shift, code‐switching, and lexical borrowing (Epps 2018; Vaughan and Singer 2018). Further afield in Arnhem Land, at Warruwi community, Singer and Harris (2016) note that code‐switching may be avoided by the use of the receptive multilingual mode. Why do these differences exist? From a purely typological perspective, code‐switching might be expected to be “easier” between traditional languages that share similar features through shared inheritance and diffusion (i.e. via the operation of “congruence”/ “equivalence” constraints, e.g. Weinreich 1964; Myers‐Scotton 1993, 1995; Poplack 1980), and yet traditional‐language switching is rarer and occurs most commonly outside the clause level. Various socio‐psychological and ideological factors help explain this, contributing to creating a kind of “situational” congruence (or non‐congruence) between codes (Sebba 1998) that may be more significant than structural constraints. Language ideologies play a key role here: a guiding principle of language choice in the region is the priority to use one’s father’s language, the patrilect (Vaughan and Singer 2018), and code‐switching between traditional languages is seen to undermine this fundamental convention. Although the patrilect is understood to be inherent, its primacy must nevertheless be continually performed and constructed in interaction. The shape of code‐switching is one reflex of the various ways in which ideologies scaffold the establishment and maintenance of linguistic difference in the Maningrida region. These ideologies may be seen at work in other sites of language contact, such as in the push‐and‐pull between regional diffusion and differentiation across and within lects (e.g. Evans 2003b; Garde 2008; Vaughan 2018a). Language practice in north‐ central Arnhem Land provides a rare and fascinating window into an enduring post‐contact egalitarian language ecology. Changing mobilities have contributed in some communities like Maningrida to a “scaling up” of existing multilingual practices which, alongside the ongoing effects of newer linguistic arrivals, create space for new kinds of language contact to arise and stabilize.

5 Conclusion In this chapter, we focused on language contact in Australia. We saw that the story of Australian language contact begins well before colonization (for example, contact between different Aboriginal languages, contact with Macassan fishermen and other groups, as discussed in Section 1). However, the most rapid changes have occurred in communities since colonization, which has impacted these communities and their languages variably. In Section 2, we profiled language contact outcomes that are attested in the Australian context: pidgins and creoles, mixed languages, koines and new dialects, restructured traditional languages, and Aboriginal Englishes. In Section 3, we considered the ongoing difficulties caused by language contact and the education system, finding that misunderstandings about anything other than English‐ medium instruction are still perpetuated across the majority of the country. We looked at the effect this has on highly multilingual children and young speakers of contact varieties who are taught and tested in English, when English is an L2 which they may have only been exposed to once they started school. In Section 4, we saw two distinct case studies about language contact in Australia. In the first, we focused on two small locations within Victoria, showing that phonetic variation was extremely variable in L1 Aboriginal English, as has also been described for L2 varieties previously. We also saw that there was previously unreported regional variation in vowel systems in the southeast. The second case study focused on diffusion and code‐switching in

Language Contact and Australian Languages  733 Arnhem Land in Australia’s remote north, showing the breadth of language experiences in that region and marking it as a rich and evolving linguistic landscape. It may be argued that we are now at a crossroads of sorts when it comes to language contact in Australia with regard to the evolving linguistic landscape, the kinds of questions being asked by linguists, and the methodologies used for analysis and description. We can reflect back on the fact that recent language change has been very rapid in the region, and what we know about historical language contact in this country contrasts sharply to the continually emerging and evolving landscape we are now observing. In this chapter we have particularly sought to emphasize the importance of attending to the social, the ideological, and the emotional in language contact; it is only by engaging with this “total linguistic fact” that we can fully understand the drivers behind, and the outcomes following the complex encounters of speakers/signers and their language systems.

NOTES 1 Note, however, that census data is usually incomplete and does not reflect many contact varieties or community language naming conventions (see Morphy 2002, 2007; Simpson 2008). 2 Kriol and Aboriginal English are not included here as they are spoken across very large areas.

REFERENCES Adams, Karen (ed.) 2014. Koorified: Aboriginal Communication and Well‐being. Fitzroy, Vic.: The Public Health and Research Unit, Victorian Aboriginal Community Controlled Health Organisation (VACCHO) and the School of Nursing and Midwifery, La Trobe University. Aikhenvald, Alexandra Y. 2002. Language Contact in Amazonia. Oxford: Oxford University Press Amery, Rob 1985. A new diglossia: contemporary speech varieties at Yirrkala in North East Arnhem Land (Australian National University Master’s dissertation). Amery, Rob 1993. An Australian koine: Dhuwaya, a variety of Yolŋu Matha spoken at Yirrkala in Northeast Arnhem Land. International Journal for the Sociology of Language 99: 45–64. Angelo, Denise and Catherine Hudson 2018. Dangerous conversations: teacher‐student interactions with unidentified English language learners. In Gillian Wigglesworth, Jane Simpson, and Jill Vaughan (eds.), Language Practices of Indigenous Children and Youth. Basingstoke: Palgrave Macmillan pp. 207–36. Angelo, Denise and Sophie McIntosh 2014. Anomalous data about Aboriginal and Torres

Strait Islander language ecologies. In Elke Stracke (ed.), Intersections: Applied Linguistics as a Meeting Place. Newcastle‐upon‐Tyne: Cambridge Scholars Publishing, pp. 270–93. Backus, Ad 2005. Codeswitching and language change: one thing leads to another? International Journal of Bilingualism 9(3–4), 307–40. Bavin, Edith and Tim Shopen 1985. Warlpiri and English: Language in contact. In Michael Clyne (ed.), Australia, Meeting Place of Languages (Pacific Linguistics C‐92). Canberra: Australian National University, pp. 81–94. Black, Paul 1997. Lexicostatistics and Australian languages: problems and prospects. In Darrell Tryon and Michael Walsh (eds.), Boundary Rider: Essays in Honour of Geoffrey O’Grady (Pacific Linguistics C‐136). Canberra: Australian National University, pp. 51–69. Black, Paul 2006. Equilibrium Theory applied to Top End Australian languages. In Keith Allan (ed.), Selected Papers from the 2005 Conference of the Australian Linguistic Society. Available at www.als.asn.au Bowern, Claire 2006 Another look at Australia as a linguistic area. In Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic

734  Jill Vaughan and Debbie Loakes Areas: Convergence in Historical and Typological Perspective. Basingstoke: Palgrave Macmillan, pp. 244–65. Bowern, Claire and Quentin Atkinson 2012. Computational phylogenetics and the internal structure of Pama‐Nyungan. Language 88.4: 817–45. Bowern, Claire, Patience Epps, Russel Gray, Jane Hill, Keith Hunley, Patrick McConvell and Jason Zentz 2011. Does lateral transmission obscure inheritance in hunter‐gatherer languages? PLoS ONE 6.9: e25195. Bowern, Claire and Harold Koch (eds.) 2004. Australian Languages: Classification and the Comparative Method. Amsterdam: Benjamins. Bradley, John 1988. Yanyuwa: “Men speak one way, women speak another.” Aboriginal Linguistics 1: 126–34. Brandl, M. M. and Michael Walsh 1982. Speakers of many tongues: toward understanding multilingualism among Aboriginal Australians. International Journal of the Sociology of Language 36: 71–81. Bundgaard‐Nielsen, Rikke and Brett Baker 2016. Fact or furphy? The continuum in Kriol. In Felicity Meakins and Carmel O’Shannessy (eds.), Loss and Renewal. Berlin: de Gruyter Mouton, pp. 177–216. Butcher, Andrew 2008. Linguistic aspects of Australian Aboriginal English. Clinical Linguistics and Phonetics 22.8: 625–42. Butcher, Andrew and Victoria Anderson 2008. The vowels of Australian Aboriginal English. Proceedings of the 9th Annual Conference of the International Speech Communication Association, 22–26 September 2008, pp. 347–350. Brisbane: Curran Associates publisher. Carew, Margaret 2017. Gun‐nartpa grammar. Available at http://call.batchelor.edu.au/ gun‐nartpa‐grammar/ Clopper, Cynthia 2014. Sound change in the individual: effects of exposure on cross‐dialect speech processing. Laboratory Phonology 5.1: 69–90. Coleman, Carolyn. n.d. Maningrida comparative wordlists. Unpublished notes. Cox, Felicity and Sallyanne Palethorpe 2008. Reversal of short front vowel raising in Australian English Proceedings of the 9th Annual Conference of the International Speech Communication Association, 22–26 September 2008. Brisbane: Curran Associates, pp. 342–5. Cysouw, Michael 2005. Morphology in the wrong place: a survey of preposed enclitics. In

Wolfgang U. Dressler, Dieter Kastovsky, Oskar E. Pfeiffer, and Franz, Rainer (eds.), Morphology and its Demarcations: Selected Papers from the 11th Morphology Meeting, Vienna, February 2004. Amsterdam: Benjamins, pp. 17–38. Dench, Alan 1998. Pidgin Ngarluma: An Indigenous Contact Language in North Western Australia. Journal of Pidgin and Creole Languages 13: 1–61. Devlin, Brian 1985. Language maintenance in a North East Arnhem Land settlement (Columbia University Ph.D. dissertation). Devlin, Brian 2011. The status and future of bilingual education for remote indigenous students in the Northern Territory. Australian Review of Applied Linguistics 34.3: 260–79. Disbray, Samantha 2009. More than one way to catch a frog: Children’s discourse in a language contact setting (University of Melbourne Ph.D. dissertation). Dixon, Robert M. W. 1970. Languages of the Cairns rain forest region. In Stephen A. Wurm and Donald C. Laycock (eds.), Pacific Linguistic Studies in Honour of Arthur Capell (Pacific Linguistics C‐13). Canberra: Australian National University, pp. 651–87. Dixon, Robert M. W. 1997. The Rise and Fall of Languages. Cambridge: Cambridge University Press. Dixon, Robert M. W. 2001. The Australian linguistic area. In Alexandra Y. Aikhenvald and R. M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press, pp. 64–104. Dixon, Robert M. W. 2002. Australian Languages: Their Nature and Development. Cambridge: Cambridge University Press. Eades, Diana 2013a. Aboriginal English. In Harold Koch and Rachel Nordlinger (eds.) The Languages and Linguistics of Australia: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 417–47. Eades, Diana 2013b. Aboriginal Ways of Using English. Canberra: Aboriginal Studies Press. Eades, Diana and Jeff Siegel 1999. Changing attitudes towards Australian Creoles and Aboriginal English. In John R. Rickford and Suzanne Romaine (eds.) Creole Genesis, Attitudes and Discourse: Studies Celebrating Charlene J. Sato. Amsterdam: Benjamins, pp. 265–78. Edmonds‐Wathen, Cris 2015. Indigenous language speaking students learning

Language Contact and Australian Languages  735 mathematics in English: expectations of and for teachers. Australian Journal of Indigenous Education 44.1: 48–58. Elkin, A. P. 1961. The Yabuduruwa. Oceania 31.3: 166–209. Elvin, Jaydene, Daniel Williams, and Paola Escudero 2016. Dynamic acoustic properties of monophthongs and diphthongs in Western Sydney Australian English. Journal of the Acoustical Society of America 140.1: 576–81. Elwell, Vanessa 1977. Multilingualism and lingua francas among Australian Aborigines: a case study of Maningrida (Australian National University, Canberra honours thesis). Elwell, Vanessa 1982. Some social factors affecting multilingualism among Aboriginal Australians: a case study of Maningrida. International Journal of the Sociology of Language 36: 83–104. Epps, Patience 2008. A Grammar of Hup. Berlin: de Gruyter Mouton. Epps, Patience 2018. Contrasting linguistic ecologies: indigenous and colonially mediated language contact in northwest Amazonia. Language and Communication 62: 156–69. Epps, Patience and Kristine Stenzel (eds.) 2013. Upper Rio Negro: Cultural and Linguistic Interaction in Northwestern Amazonia. Rio de Janeiro: Museu do Índio‐FUNAI. Available at www. museunacional.ufrj.br/ppgas/livros_ele.html. Evans, Nicholas 1992. Macassan loanwords in top end languages. Australian Journal of Linguistics 12.1: 45–91. Evans, Nicholas 2003a. Bininj Gun‐Wok: A Pan‐dialectal Grammar of Mayali, Kunwinjku and Kune (No. 541). Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, Australian National University. Evans, Nicholas 2003b. Context, culture, and structuration in the languages of Australia. Annual Review of Anthropology 32.1: 13–40. Evans, Nicholas 2005. Australian languages reconsidered: a review of Dixon (2002). Oceanic Linguistics 44: 242–86. Evans, Nicholas 2007. Warramurrungunji undone: Australian languages in the 51st millennium. In Matthias Brenzinger (ed.), Language Diversity Endangered. Berlin: de Gruyter Mouton, pp. 342–73. Evans, Nicholas 2010. A tale of many tongues: polyglot narrative in North Australian oral traditions. In Brett J. Baker, Llana. Mushin, Mark Harvey and Rod Gardner (eds.), Indigenous Language and Social Identity. Papers in Honour of Michael Walsh. Canberra: Pacific

Linguistics, College of Asia and the Pacific, The Australian National University, pp. 275–95. Feller, Nayalin P. and Jill Vaughan. 2017. Language practices of Guarani children in a community‐based bilingual school. In Gillian Wigglesworth, Jane Simpson and Jill Vaughan (eds), From Home to School: Language Practices of Indigenous and Minority Children and Youth. Basingstoke: Palgrave Macmillan, pp. 173–204. Fesl, Eve M. D. 1977. Melbourne Aboriginal English (Melbourne: Monash University BA honours thesis). Fesl, Eve M. D. 1993. Conned! St. Lucia: University of Queensland Press. François, Alexandre 2012. The dynamics of linguistic diversity: egalitarian multilingualism and power imbalance among northern Vanuatu languages. International Journal of the Sociology of Language 2012(214): 85–110. García, Ofelia and Li Wei 2014. The translanguaging turn and its impact. In Ofelia García and Li Wei (ed.), Translanguaging: Language, Bilingualism and Education. Basingstoke: Palgrave Macmillan, pp. 19–45. Garde, Murray 2008. Kun‐dangwok: “clan lects” and Ausbau in western Arnhem Land. International Journal of the Sociology of Language 191: 141–69. Gordeeva, Olga B. and James M. Scobbie 2013. A phonetically versatile contrast: Pulmonic and glottalic voicelessness in Scottish English obstruents and voice quality. Journal of the International Phonetic Association 43.3: 249–71. Green, Rebecca 1987. A Sketch Grammar of Burarra. Canberra: Australian National University. Green, Rebecca 1995. A Grammar of Gurr‐goni (North‐central Arnhem Land). Canberra: Australian National University. Gutiérrez, Kris D., Patricia Baquedano‐López, and Carlos Tejeda 1999. Rethinking diversity: hybridity and hybrid language practices in the third space. Mind, Culture, and Activity 6.4: 286–303. Harris, John W. 1991. Kriol – the creation of a new language. In Suzanne Romaine (ed.), Language in Australia. Cambridge, Cambridge University Press, pp. 195–203. Harvey, Mark 2011. Lexical change in pre‐ colonial Australia. Diachronica, 28.3: 345–81. Harvey, Mark 2012. Warndarrang and Marra: a diffusional or genetic relationship? Australian Journal of Linguistics 32.3: 327–60.

736  Jill Vaughan and Debbie Loakes Harvey, Mark 2013. Reconstructing long‐term limits on diffusion in Australia. Anthropological Linguistics 55.2: 158–83. Heath, Jeffrey 1978. Linguistic Diffusion in Arnhem Land. Canberra: AIATSIS. Heath, Jeffrey 1981. A case of intensive lexical diffusion: Arnhem Land, Australia. Language 57: 335–67. Heffernan, John 1984. Papunya Luritja language notes. Northern Territory: Papunya Literature Production Centre. Holcombe, Sarah 2004a. The politico‐historical construction of the Pintupi Luritja and the concept of tribe. Oceania 74.4: 257–75. Holcombe, Sarah 2004b. Socio‐political Perspectives on Localism and Regionalism in the Pintupi Luritja Region of Central Australia: Implications for Service Delivery and Governance. Canberra: Centre for Aboriginal Economic Policy Research. Hosokawa, Komei 1987. Malay talk on boat: an account of Broome Pearling Lugger Pidgin. In Donald Laycock and Werner Winter (eds.), A World of Language: Papers Presented to Professor S. A. Wurm on his 65th Birthday (Pacific Linguistics C‐100). Canberra: Australian National University, pp. 287–96. Koch, Harold 2014. Historical relations among the Australian languages. In Harold Koch and Rachel Nordlinger (eds.), The Languages and Linguistics of Australia: A Comprehensive Guide. Berlin: Gruyter, pp. 23–89. Koch, Harold and Rachel Nordlinger 2014. The languages of Australia in linguistic research: context and issues. In Harold Koch and Rachel Nordlinger (eds.), The Languages and Linguistics of Australia: A Comprehensive Guide. Berlin: de Gruyter, pp. 3–21. Langlois, Annie 2004. Alive and Kicking: Areyonga Teenage Pitjantjatjara (Pacific Linguistics 561). Canberra: Australian National University. Langlois, Annie 2006. Wordplay in Teenage Pitjantjatjara. Australian Journal of Linguistics 26.2: 181–92. Loakes, Debbie 2018. The shape of Aboriginal English vowels in two Victorian communities. Talk presented at the Linguistic Discoveries: The CoEDL Shape of Language Showcase, November 8. Melbourne: University of Melbourne. Loakes, Debbie, John Hajek, and Janet Fletcher 2018. Can you t[æ]l I’m from M[æ]lbourne? An overview of the DRESS and TRAP vowels before /l/ as a regional accent marker in Australian English. English World Wide 38.1: 29–49.

Loakes, Debbie, Kirsty McDougall, Josh Clothier, John Hajek, and Janet Fletcher 2018. Sociophonetic variability of post‐vocalic /t/ in Aboriginal and mainstream Australian English, Proceedings of the 17th Speech Science and Technology Conference (SST2018) ASSTA: Cooegee, Sydney 1–4. Lüpke, Friederike 2016. Uncovering small‐scale multilingualism. Critical Multilingualism Studies 4.2: 35–74. Macqueen, Susy, Ute Knoch, Gillian Wigglesworth, Rachel Nordlinger, Ruth Singer, Tim McNamara, Rhianna Brickle 2019. The impact of national standardized literacy and numeracy testing on children and teaching staff in remote Australian Indigenous communities. Language Testing 36.2: 265–287. Malcolm, Ian G. 1994. Discourse and discourse strategies in Australian Aboriginal English. World Englishes 13.3: 289–306. Malcolm, Ian G. 2011. Issues in English language assessment of indigenous Australians. Language Assessment Quarterly 8.2: 190–9. Malcolm, Ian G. 2013. The ownership of Aboriginal English in Australia. World Englishes 32.1: 42–53. Malcolm, Ian G. 2018. Australian Aboriginal English: Change and Continuity in an Adopted Language. Berlin: de Gruyter Mouton. Mansfield, John 2014. Polysynthetic sociolinguistics (Australian National University Ph.D. dissertation). Matras, Yaron and Peter Bakker 2003. The study of mixed languages. In Yaron Matras and Peter Bakker (eds.), The Mixed Language Debate: Theoretical and Empirical Advances. Berlin: de Gruyter Mouton, pp. 1–20. McConvell, Patrick 1985. Domains and codeswitching among bilingual Aborigines. In Michael Clyne (ed.), Australia, Meeting Place of Languages (Pacific Linguistics C‐92). Canberra: Australian National University, pp. 95–125. McConvell, Patrick 1988. Mix‐im‐up: Aboriginal codeswitching old and new. In. Monica Heller (ed.), Codeswitching: Anthropological and Sociolinguistic Perspectives (Contributions to the Sociology of Language 48). Berlin: de Gruyter Mouton, pp. 97–150. McConvell, Patrick 2002. Mix‐Im‐Up speech and emergent mixed languages in Indigenous Australia. Proceedings of SALSA 2001 (Symposium on Language and Society). Texas Linguistic Forum 44.1: 328–49.

Language Contact and Australian Languages  737 McConvell, Patrick 2010. Contact and indigenous languages in Australia. In Raymond Hickey (ed.), The Handbook of Language Contact, 1st edition. Oxford: Wiley‐Blackwell, pp. 770–94. McConvell, Patrick, Doug Marmion and Sally McNicol 2005. National Indigenous Languages Survey Report. Canberra: Australian Institute of Aboriginal and Torres Strait Islander Studies. McKenry, Rosemary 1995. Academic success for speakers of Koorie English: the need for teacher intervention (University of Melbourne M.Ed. dissertation). Meakins, Felicity 2008a. Case‐marking in Contact: The development and function of case morphology in Gurindji Kriol, an Australian mixed language (University of Melbourne Ph.D. dissertation). Meakins, Felicity 2008b. Land, language and identity: the socio‐political origins of Gurindji Kriol. In Naomi Nagy and Miriam Meyerhoff (eds.), Social Lives in Language – Sociolinguistics and Multilingual Speech Communities: Celebrating the Work of Gillian Sankoff. Amsterdam: Benjamins, pp. 69–94. Meakins, Felicity 2011. Case‐marking in Contact: The Development and Function of Case Morphology in Gurindji Kriol. Amsterdam: Benjamins. Meakins, Felicity 2012. Which mix? – code‐ switching or a mixed language – Gurindji Kriol. Journal of Pidgin and Creole Languages 27.1: 105–40. Meakins, Felicity 2013. Mixed languages. In Yaron Matras and Peter Bakker (eds.), Contact Languages: A Comprehensive Guide (Language Contact and Bilingualism 6). Berlin: de Gruyter Mouton, pp. 159–228. Meakins, Felicity 2014. Language contact varieties. In Harold Koch and Rachel Nordlinger (eds.), The Languages and Linguistics of Australia: A Comprehensive Guide. Berlin: de Gruyter Mouton, pp. 361–411. Meakins, Felicity and Carmel O’Shannessy 2010. Ordering arguments about: word order and discourse motivations in the development and use of the ergative marker in two Australian mixed languages. Lingua 120.7: 1693–1713. Meakins, Felicity and Rob Pensalfini forthcoming. Holding the mirror up to converted languages: two grammars, one lexicon. International Journal of Bilingualism. Morphy, Frances 2002. When systems collide: the 2001 census at a Northern Territory outstation. In David F. Martin, Frances Morphy, Will G.

Sanders, and John Taylor (eds.), Making Sense of the Census: Observations of the 2001 Enumeration in Remote Aboriginal Australia (CAEPR Research Monograph No. 22). Canberra: ANU EPress, pp. 29–75. Morphy, Frances (ed.) 2007. Agency, Contingency and Census Process: Observations of the 2006 Indigenous Enumeration Strategy in Remote Aboriginal Australia (CAEPR Research Monograph No. 28). Canberra: ANU EPress. Mühlhäusler, Peter 1996. Post‐contact Aboriginal languages in the Northern Territory. In Stephen Wurm, Peter Mühlhäusler, and Darrel Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas, vol. 2.1. Berlin: de Gruyter Mouton, pp. 123–32. Mühlhäusler, Peter 1998. Pidgins, creoles and post‐contact Aboriginal languages in Western Australia. In Peter Mühlhäusler (ed.), Papers in Pidgin and Creole Linguistics No. 5 (Pacific Linguistics A‐91). Canberra: Australian National University, pp. 1–33. Mühlhäusler, Peter and William McGregor 1996. Post‐contact languages of Western Australia. In Stephen Wurm, Peter Mühlhäusler, and Darrel Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas, vol. 2.1. Berlin: de Gruyter Mouton, pp. 101–21. Munro, Jennifer 2000. Kriol on the move: a case of language spread and shift in Northern Australia. In Jeff Siegel (ed.), Processes of Studies from Australia and the South Pacific. Saint‐Laurent, Quebec: Fides, pp. 245–70. Munro, Jennifer 2011. Roper River Aboriginal language features in Australian Kriol: considering semantic features. In Claire Lefebvre (ed.), Creoles, their Substrates, and Language Typology. Amsterdam: Benjamins, pp. 461–87. Mushin, Ilana 2010. Code‐switching as an interactional resource in Garrwa/Kriol talk‐in‐ interaction. Australian Journal of Linguistics 30.4: 471–96. Mushin, Ilana and Jane Simpson 2008. Free to bound to free? Interactions between pragmatics and syntax in the development of Australian pronominal systems. Language 84.3: 566–96. Myers‐Scotton, Carol 1993. Duelling Languages: Grammatical Structure in Codeswitching. Oxford: Clarendon Press. Myers‐Scotton, Carol 1995. A lexically based model of codeswitching. In Lesley Milroy and

738  Jill Vaughan and Debbie Loakes Pieter Muysken (eds.), One Speaker, Two Languages: Cross‐disciplinary Perspectives on Codeswitching. Cambridge: Cambridge University Press, pp. 233–56. Nettle, Daniel 1999. Linguistic Diversity. Oxford: Oxford University Press. O’Keeffe, Isabel 2016. Multilingual manyardi/ kun‐borrk: manifestations of multilingualism in the classical song traditions of western Arnhem Land (University of Melbourne Ph.D. dissertation). Oliver, Rhonda and Bich Nguyen 2017. Translanguaging on Facebook: exploring Australian Aboriginal multilingual competence in technology‐enhanced environments and its pedagogical implications. The Canadian Modern Language Review 73.4: 463–87. O’Shannessy, Carmel 2006. Language Contact and Children’s Bilingual Language Acquisition: Learning a Mixed Language and Warlpiri in Northern Australia (University of Sydney Ph.D. dissertation). O’Shannessy, Carmel 2012. The role of code‐ switched input to children in the origin of a new mixed language. Linguistics 50.2: 305–40. O’Shannessy, Carmel 2015. Multilingual children increase language differentiation by indexing communities of practice. First Language, 35.4–5: 305–26. Owen, Kirsty 2019. Supporting indigenous children’s transition to school. Teacher Magazine. Melbourne: Australian Council for Educational Research (ACER). Available at www.teacher magazine.com.au/articles/supporting‐indigenous‐ childrens‐transition‐to‐school Poplack, Shana 1980. Sometimes I’ll start a sentence in Spanish Y TERMINO EN ESPANOL: Toward a typology of codeswitching. Linguistics 18: 581–618. Renganathan, Sumathi and Inge Kral 2017. Comparative analysis of language and education policies for indigenous minorities in Australia and Malaysia. International Journal of Multicultural Education 20.1: 138–56. Rhydwen, Mari 1995. Kriol is the colour of Thursday. International Journal of the Sociology of Language 113: 113–19. Rhydwen, Mari 1996. Writing on the Backs of Blacks: Voice, Literacy and Community in Kriol Fieldwork. Brisbane: University of Queensland Press. Rumsey, Alan 2018. The sociocultural dynamics of indigenous multilingualism in northwestern Australia. Language and Communication 62: 91–101.

Sandefur, John 1979. An Australian Creole in the Northern Territory: A Description of Ngukurr‐ Bamyili Dialects (Part 1) (Work Papers of SIL‐ AAB Series B Volume 3). Darwin: Summer Institute of Linguistics. Schmidt, Annette 1985. Young People’s Dyirbal: An Example of Language Death from Australia. Cambridge: Cambridge University Press. Shnukal, Anna 1988 Broken: An Introduction to the Creole Language of Torres Strait (Pacific Linguistics C‐107). Canberra: Australian National University. Schultze‐Berndt, Eva, Felicity Meakins, and Denise Angelo 2013. Kriol. In Susanne Michaelis, Philip Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 1. Oxford: Oxford University Press, pp. 241–51. Sebba, Mark 1998. A congruence approach to the syntax of codeswitching. International Journal of Bilingualism 2.1: 1–19. Siegel, Jeff 1985. Koines and koineization. Language in Society 14.3: 357–78. Siegel, Jeff 2007. Creoles and minority dialects in education: an update. Language and Education 21.1: 66–86. Silverstein, Michael 1985 Language and the culture of gender: at the intersection of structure, usage, and ideology. In Elizabeth Mertz and Richard Parmentier (eds.), Semiotic Mediation. Orlando: Academic Press, pp. 219–60. Simpson, Jane 1996. Early language contact varieties in South Australia. Australian Journal of Linguistics 16.2: 169–207. Simpson, Jane 2008. Language landscapes of children in remote Australia. In Rob Amery and David Nash (eds.), Inaugural Indigenous languages conference (ILC). Adelaide: University of Adelaide, pp. 51–60. Simpson, Jane. 2013. What’s done and what’s said: language attitudes, public language activities and every day talk in the Northern Territory of Australia. Journal of Multilingual and Multicultural Development 34.4: 383–98. Simpson, Jane, Jo Caffery, and Patrick McConvell 2009. Gaps in Australia’s indigenous language policy: dismantling bilingual education in the Northern Territory. AIATSIS Research Discussion Paper No. 24. Canberra: Australian Institute of Aboriginal and Torres Strait Islander Studies. Simpson, Jane and Gillian Wigglesworth 2008. Children’s Language and Multilingualism:

Language Contact and Australian Languages  739 Indigenous Language Use at Home and School. London: Continuum. Singer, Ruth 2018. A small speech community with many small languages: the role of receptive multilingualism in supporting linguistic diversity at Warruwi Community (Australia). Language and Communication 62: 102–18. Singer, Ruth and Salome Harris 2016. What practices and ideologies support small‐scale multilingualism? A case study of Warruwi Community, northern Australia. International Journal of the Sociology of Language 241: 163–208. Sorensen Jr., Arthur P. 1967. Multilingualism in the northwest Amazon. American Anthropologist 69: 670–84. Steele, Carly and Gillian Wigglesworth 2018. Teaching English as an additional language or dialect to young learners in indigenous contexts. In Rhonda Oliver and Bich Nguyen (eds.), Teaching Young Second Language Learners: Practices in Different Classroom Contexts. London: Routledge, pp. 69–88. ten Thije, Jan D. and Ludger Zeevaert (eds.) 2007. Receptive Multilingualism: Linguistic Analyses, Language Policies and Didactic Concepts. Amsterdam: Benjamins. Thelen, Esther and Linda B. Smith 1994. A Dynamic Systems Approach to the Development of Cognition and Action. Cambridge, MA: MIT Press. Thomason, Sarah G. 2001. Language Contact. An Introduction. Washington: Georgetown University Press. Troy, Jakelin 1990. Australian Aboriginal Contact with the English Language in New South Wales: 1788 to 1845 (Pacific Linguistics B‐103). Canberra: Australian National University. Trudgill, Peter, Elizabeth Gordon, Gillian Lewis, and Margaret Maclagan 2000. Determinism in new‐ dialect formation and the genesis of New Zealand English. Journal of Linguistics 36: 299–318. UNESCO 1953. The Use of Vernacular Languages in Education. Monographs on Fundamental Education, VIII. Paris: UNESCO. Urry, James and Michael Walsh 1981. The lost “Macassar language” of northern Australia. Aboriginal History 5.2: 91–108. Vadovac, Neda 2017. Arnhem Land performer inspiring kids with goal to be first to rap in indigenous language. ABC News September 12. Available at www.abc.net.au/ news/2017‐05‐21/ arnhem‐land‐musician‐wants‐to‐be‐first‐to‐ rap‐in‐language/8543752

Vaughan, Jill 2017. Disputed dialects: Dimensionalisation of variation in Burarra (northern Australia). Paper presented at Structural and Developmental Aspects of Bidialectalism workshop, Tromsø (25–6 October 2017). Available at www.academia. edu/37707791/ Disputed_dialects_Dimensionalisation_of_ variation_in_Burarra Vaughan, Jill 2018a. “We talk in saltwater words”: dimensionalisation of dialectal variation in multilingual Arnhem Land. Language and Communication, 62, 119–32. Vaughan, Jill 2018b. Translanguaging and hybrid spaces: boundaries and beyond in north‐ central Arnhem Land. In Gerardo Mazzaferro (ed.), Translanguaging as Everyday Practice. Cham: Springer, pp. 125–48. Vaughan, Jill and Ruth Singer. 2018. Indigenous multilingualisms past and present. Language and Communication 62: 83–90. Walker, Alan and R. David Zorc 1981. Austronesian loanwords in Yolngu‐Matha of northeast Arnhem Land. Aboriginal History: 109–34. Walsh, Michael 1993. Languages and their status in Aboriginal Australia. In Colin Yallop and Michael Walsh (eds.), Language and Culture in Aboriginal Australia. Canberra: Aboriginal Studies Press, pp. 1–14. Walsh, Michael 1994. Interactional styles in the courtroom: an example from northern Australia. In John Gibbons (ed.), Language and the Law. London: Longman, pp, 217–33. Walsh, Michael 2012. Re‐awakening Languages: Theory and Practice in the Revitalisation of Indigenous Languages. Sydney: Sydney University Press. Walton, Chris 1993. Aboriginal education in Northern Australia: a case study of literacy policies and practices. In Peter Freebody and Anthony R. Welch (eds.), Knowledge, Culture and Power: International Perspectives on Literacy as Policy and Practice. London: Falmer Press, pp. 57–83. Waters, Bruce E. 1989. Djinang and Djinba: A Grammatical and Historical Perspective. Canberra: Pacific Linguistics. Weinreich, Uriel 1964 [1953]. Languages in Contact. The Hague: de Gruyter Mouton. Wigglesworth, Gillian and Rosey Billington 2013. Teaching creole‐speaking children: issues, concerns and resolutions for the classroom. Australian Review of Applied Linguistics, 36.3: 234–49.

740  Jill Vaughan and Debbie Loakes Wigglesworth, Gillian, Rosey Billington, and Debbie Loakes 2013. Creole speakers and standard language education. Language and Linguistics Compass 7.7: 388–97. Wigglesworth, Gillian and Jane Simpson 2018. Going to school in a different world. In Gillian Wigglesworth, Jane Simpson, and Jill Vaughan (eds.), Language Practices of Indigenous Children and Youth Basingstoke: Palgrave Macmillan, pp. 1–20. Wigglesworth, Gillian, Jane Simpson, and Debbie Loakes 2011. Naplan language assessments for

indigenous children in remote communities. Australian Review of Applied Linguistics 34.3: 320–43. Wilkinson, M. (1991). Djambarrpuyngu: A Yolngu variety of Northern Australia (University of Sydney Ph.D. dissertation). Woolard, Kathryn A. 2008. Why dat now? Linguistic‐anthropological contributions to the explanation of sociolinguistic icons and change. Journal of Sociolinguistics 12.4: 432–52.

37 Contact Languages of the Pacific JEFF SIEGEL

With over 1,000 indigenous languages and a recent history of colonial exploitation, the Pacific region has provided a fertile context for the growth of contact languages. This chapter first describes new languages (pidgins and creoles) and then new dialects (koines and indigenized varieties) that have emerged in the Pacific as the result of language contact. For the purpose of this chapter, the Pacific region is defined as including only small island ­countries and territories (thus excluding Australia, New Zealand, and the countries of Asia and the Americas that border the Pacific Ocean). Also, due to space limitations, the chapter concentrates only on lexicon and morphosyntax.

1  New Languages: Pidgins Pidgins are new languages that develop out of a need for a medium of communication among people who do not share a common language  –  for example, between traditional trading partners or among plantation laborers from diverse geographic origins. Most of the forms in the lexicon of the new language come from one of the languages in the contact situation, called the “lexifier” (or sometimes the “superstrate”) – often the language of the group in control of the area where contact occurs. However, the meanings or functions of some of the lexical forms of the pidgin, as well as the phonology and grammatical rules, are different from those of the lexifier. First, they appear to be greatly reduced or simplified in comparison, and second, they may resemble those of one or more of the other languages in contact, usually referred to as the “substrate languages.” The first step in the development of a pidgin is when people use their own individual ways of communicating. Speakers of the lexifier may simplify their language or use a simplified register, referred to as “foreigner talk.” Speakers of other languages use words and phrases they have learned from the lexifier, but with little if any grammatical morphology, or with overgeneralized rules, as in the early stages of second language acquisition. The combination of these individual ways of communicating in which some conventions have emerged is called a “jargon” or “pre‐pidgin.”

1.1  Indigenous Pre‐pidgins Several pre‐pidgins lexified by indigenous languages have been recorded in the Pacific. Jargon Fijian (Siegel 1987) was derived in part from the foreigner talk that Fijians used to communicate with Tongans and other neighbors (Geraghty 1978). Simplified Motu (Dutton 1985; 1997) was used by the Motu people on the south coast of New Guinea around Port

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

742  Jeff Siegel Moresby to talk to visitors or trading partners from other language areas. Both of these ­pre‐pidgins were later used for communication with European explorers, traders, beachcombers, or missionaries – Jargon Fijian from the early 1800s (Siegel 1987) and Simplified Motu from 1874 (Dutton 1985, 1997). (Note that the term “European” as used in this chapter includes American and Australian.) Hawaiian Maritime Pidgin (Day 1987) or Jargon Hawaiian (Siegel 1987; Roberts 2005) emerged after 1778 when Hawai‘i became a stopover for large numbers of European ships involved in the fur trade between the northwest coast of America and China. This was preceded by a “foreigner‐talk variety of Tahitian” (Drechsel 2014: 287) that emerged in the late 1760s in the Society Islands as a result of contact with French and British explorers. According to Drechsel, (1999, 2007, 2014), it was this pre‐pidgin that was the forerunner of Jargon Hawaiian and also spread to the Marquesas and New Zealand, forming the basis of a Polynesian based “South Seas Jargon” (Clark 1979). The linguistic features of some of these pre‐pidgins have been ascertained from a variety of historical sources, including word and phrase lists collected by early explorers and the writings of the first missionaries (see Dutton 1985; 1997; Day 1987; Roberts 1995, 2005). Here they are exemplified by Jargon Fijian. The features of Jargon Fijian included the use of only independent forms of pronouns (INDEP), rather than subject referencing (SRP) or possessive (POSS) pronouns; the overgeneralized use of sa (based on a Fijian aspect marker sa) as a predicate marker (PM); a lack of verbal morphology (such as transitive marking); and the use of adverbs such as  malua ‘slowly, by‐and‐by’ instead of tense or aspect markers (Siegel 1987: 115–19). An example is:1 (1)

malua sa laGo later PM go ‘I’ll come later.’

mai DIR

koeau. 1SG.INDEP

In Standard Fijian (SF), this would be: (2)

au 1SG.SRP

na FUT

lako go

mai DIR

e LOC

muri. following

However, as is generally the case in pre‐pidgins, these features were not found consistently, and there was a great deal of individual variation – for example in basic constituent ordering between SV and VS, and in the ordering of the possessor and possessum in genitive constructions.

1.2  Stable Restricted Pidgins If groups in contact continue to use a pre‐pidgin with each other, and especially if speakers of other languages start to use the pre‐pidgin as a lingua franca among themselves, norms begin to stabilize. Individual variation may decrease, some variants may be eliminated, and further communicative conventions may develop. The end result is a “stable pidgin.” Although relatively stable, this type of pidgin is still restricted in both its grammatical development and its functional range. In Fiji, the process of stabilization began as a result of the introduction of small, European‐ owned plantations in the 1860s. Jargon Fijian was used in the earlier plantations, where the laborers were Fijians. However, from 1865 to 1911, 27,000 indentured laborers were imported from islands in the New Hebrides (now Vanuatu), the Gilberts (now Kiribati), the Solomons, and New Guinea. Jargon Fijian continued to be the language used to run these plantations,

Contact Languages of the Pacific  743 but it also became the lingua franca among the Pacific laborers, who spoke scores of different languages. Thus, in the late 1860s and early 1870s, Jargon Fijian stabilized to become Plantation Pidgin Fijian (PPF; Siegel 1987). Most of the features of PPF were similar to those of Jargon Fijian, except that they were used more consistently, and a great deal of the variation was eliminated  –  for example, constituent ordering stabilized as SVO and possessum possessor. Also, some additional ­features conventionalized, such as fusion of the third person inalienable possessive suffix onto nouns that are normally inalienably possessed. These features can be seen in the following examples of PPF, compared to Standard Fijian (SF): (3)

(4)

a.

PPF:

b.

SF:

na ligana koiau sa musi. ART hand 1SG.INDEP PM hurt ‘My hand hurts.’ sā mosi na liga‐qu. ASP hurt DEF hand‐1SG.POSS

a. PPF: kokoya sa 3SG.INDEP PM ‘He cut down the branch.’ b. SF: e musu‐ka 3SG.SRP cut‐TR.OM

musu na cut ART

tabana. branch

na DEF

(o PRP

taba‐na branch‐3SG.POSS

koya). 3SG

Compared to SF, however, PPF still had virtually no grammatical morphology and demonstrated a great deal of reduction. For example, SF has between 70 and 135 pronouns, depending on how they are counted. There are subject marking, objective and independent sets, and four different possessive sets, each set with pronouns indicating person, the inclusive/exclusive distinction, and number (singular, dual, paucal, and plural). In PPF, there were only six ­pronouns – first, second, and third person, singular and plural. Most of the Pacific Island laborers who survived their contracts in Fiji returned to their home islands. There are reports that in the late 1890s and early 1900s, Pidgin Fijian was being used as a lingua franca in the western Pacific, where most of the laborers returned to (Siegel 1992). However, this was soon displaced by an English‐lexified pidgin (see pp. 745–48). Those Pacific Islanders who stayed in Fiji were most often married to indigenous Fijian women and their children spoke the usual Fijian of their communities, not PPF. But in addition to workers imported from Pacific islands, from 1879 to 1916 over 60,000 indentured laborers were imported from India. About a third of these worked on small plantations alongside Pacific Islanders. There they learned PPF, and it became the language of interethnic communication both on and off the plantations. Unlike the Pacific Islanders, the indentured Indians came mostly as family groups, not single men, and when their contracts expired, the nearly 40% who stayed on did not integrate into the Fijian community and spoke Fiji Hindi (see p. 755). However, they continued to speak Pidgin Fijian for ­interethnic communication, and the language is still spoken today for this purpose (Siegel 1987). Current Pidgin Fijian (CPF) spoken by Indo‐Fijians has been influenced by their language and differs from PPF in possessor possessum ordering, optional SOV  ordering, and the use of the SF demonstrative (o)qō for the third person singular ­pronoun – for example: (5)

a. CPF: koiau na 1SG.INDEP ART ‘My father called him.’

tamana father

oqō 3SG

sa  kaDi. PM   call

744  Jeff Siegel b. SF:

e 3SG.SRP

kaci‐vi call‐TR

koya na tama‐qu. 3SG.OBJ DEF father‐1SG.POSS

Another stable restricted pidgin language also emerged from the plantation system in Fiji. From 1882, much larger plantations were established by the Colonial Sugar Refining Company, employing nearly all‐Indian labor. The majority of Fiji’s Indian immigrants worked on these plantations, where the policy was to use Hindustani, the lingua franca of northern India (the vernacular form of what is now called Hindi). But many European overseers spoke a pidginized form of this language, including some features of Bazaar Hindustani, a pidgin spoken in Calcutta (now Kolkata), the main port of embarkation for most of the laborers, who came from North India. After 1903, however, 42% of the laborers embarked from Madras (now Chennai) and were from South India. They spoke Dravidian languages and were mostly not familiar with the Hindustani lingua franca. However, when they arrived on the plantations, they quickly had to learn the Jargon Hindustani spoken there to communicate with the overseers and the North Indian laborers. Thus, when the pre‐pidgin was used as the medium of communication among the laborers themselves, it began to stabilize and became Pidgin Hindustani (Siegel 1987). This pidgin also came to be used with Fijians and other groups outside the large plantations, and today remains as a language of interethnic communication, alongside Pidgin Fijian, see Siegel (2013a, 2013b). Back to other areas of the Pacific, the first missionaries to arrive in Port Moresby in 1874 were soon followed by visitors and settlers from diverse areas such as China, Singapore, India, the Philippines, America, Europe, and various Pacific islands. The Motu people used Simplified Motu to communicate with these people as they had done with other outsiders. When this pre‐pidgin became the lingua franca among the visitors and settlers themselves, a stable pidgin Motu began to emerge. In 1884, the Protectorate of British New Guinea was established, becoming a colony in 1888. The new colonial government adopted this form of Motu as the unofficial language of administration, and used it to establish law and order and government control throughout the colony. The Armed Native Constabulary was set up in 1890, and the association of the language with this police force led to the name Police Motu. As the language further stabilized, it was influenced by Papuan substrate languages, especially Koita and Kiwai, and by the pidginized English spoken by some of the original police recruits. Eventually Police Motu became the lingua franca among speakers of the many indigenous languages in the colony (for more details, see Dutton 1985, 1997). When the colony became the Australian Territory of Papua after World War I, Police Motu was still unofficially used for administration. With the start of World War II, it was given official recognition and codified for use in published and broadcast media. In 1971, when it had approximately 150,000 speakers, the name was changed to Hiri Motu, because of the view (incorrect, as it turns out) that it was the direct descendant of the language the Motu people used for the hiri, traditional annual trading voyages to the Gulf of Papua. In 1975, the Territory of Papua along with the Protectorate of New Guinea on the northern part of the island (also administered by Australia) became the independent country of Papua New Guinea, and Hiri Motu along with Tok Pisin (see pp. 747–48) became the two unofficial national languages. However, in recent years, the number of speakers of Hiri Motu has been decreasing, and it is being displaced by Tok Pisin. In Hawai‘i, as in Fiji, a stable restricted pidgin appears to have emerged with the establishment of a European‐run plantation system. In the earliest plantations, from 1835 to the 1850s, the laborers were Hawaiians, and Jargon Hawaiian was used as the plantation language. From 1852 to 1876, approximately 2,000 Chinese contract laborers were imported,

Contact Languages of the Pacific  745 and reports show that they also used this form of Hawaiian for communication (Roberts 2005). Some stabilization may have started during this period, but a few years later larger numbers of Chinese laborers were imported (approximately 37,000 from 1877 to 1897), as well as indentured workers from other countries, including Portugal (10,000 from 1878 to 1887), Germany (1,050 from 1882 to 1884), and other Pacific islands (2,450 from 1877 to 1887). Thus, according to this view, a stable Pidgin Hawaiian emerged in the late 1870s or early 1880s when the pre‐pidgin was used as a lingua franca among all these language groups on the plantations. (For a detailed description the linguistic features of Pidgin Hawaiian, see Roberts 1995, 2013a, 2013b.) Pidgin Hawaiian remained as the plantation language and an important lingua franca elsewhere until the early 1900s, when it was displaced by the developing English‐lexifier pidgin (see pp. 750–51). An alternative scenario for the stabilization of Pidgin Hawaiian is presented by Drechsel (2014). His view is that Pidgin Hawaiian was a variety of a stable Maritime Polynesian Pidgin (MPP) that developed in the eastern Pacific in the late 1700s and early 1800s. This was the result of the Polynesian based pre‐pidgin being used more widely among Tahitians, Hawaiians, Marquesans, Māoris, and Europeans involved in increased maritime activities such as the sandalwood trade and whaling. At this time, other stable local varieties of MPP also developed, including Pidgin Māori and Pidgin Marquesan.

1.3  Other Indigenous Pidgins Several indigenous trading pidgins existed before European contact in what is now Papua New Guinea. Two of these were the actual Hiri Trading Languages used by the Motu people in the traditional hiri expeditions mentioned on p. 744. These were based on the languages of their trading partners: the Eleman languages, and Koriki (Dutton 1983, 1997). Several other trading pidgins, based on Papuan (non‐Austronesian) languages have existed in the Sepik region of northern New Guinea. The best‐known of these is Yimas Pidgin (Foley 1988). This is unlike other pidgins in that it has relatively complex morphology, though still not as complex as that of the main lexifier language, Yimas. Yimas Pidgin is also interesting in that it has multiple varieties, each used by a different clan with trading rights to villages speaking a particular language. These varieties include Yimas‐Arafundi (with at least three separate dialects, depending on the village dialect of Arafundi), Yimas‐Alamblak, and Yimas‐ Karawari (see Williams 1993, 2000; Foley 2013a, 2013b). A now extinct Yimas‐Iatamul variety existed as well. Other trading pidgins or pre‐pidgins reported in the Sepik area are Iatamul Jargon, Kwoma Pidgin, Hauna Trade Language, and Arafundi‐Enga Pidgin (see Williams 1993; Mühlhäusler et al. 1996). In the Papua region, speakers of the Austronesian language, Mekeo, devised “trade jargons” to use in commercial transactions with non‐Austronesian Kunimaipa speakers (Jones 1996). Similar to Yimas Pidgin, this Mekeo Trading Language had three varieties – Imunga, Ioi, and Maipa – each belonging to a particular family with inherited trade contacts.

1.4  Melanesian Pidgin The best‐known and most widely spoken contact language in the Pacific is Melanesian Pidgin (MP), with over 4 million speakers. The first stage in the development of the language was in the early 1800s when Europeans began to have frequent contact with Pacific Islanders as the result of whaling in the region, followed by trading in sandalwood and bêche‐de‐mer. In some places, the Europeans attempted to use the local language, but in others, they used foreigner talk registers of English, or existing English‐lexified contact varieties, such as New

746  Jeff Siegel South Wales Pidgin English and Chinese Pidgin English, that they had learned through previous trading encounters (see Baker and Mühlhäusler 1996). Pacific Islanders then picked up aspects of these varieties and also introduced their own second‐language versions of the English they were exposed to. As a result, a pre‐pidgin emerged that was used across the Pacific. This is often referred to as South Seas Jargon (Clark 1979). Some features of South Seas Jargon are illustrated in following examples (from Clark 1979: 29–30, 37): (6)

a. Only he got using all the same pigeon. (Gilbert Islands, 1860) b. Me saba plenty. (Gilbert Islands, 1860) c. Canoe too little, by and bye broke – All man go away, canoe gone, very good me stop. (Lifu (Loyalty Islands), 1850) d. He too much bad man. (Kosrae, 1860)

The features include got meaning ‘have’, all the same ‘like, similar to’, saba (= savvy) ‘know’, plenty ‘a lot (of)’, by‐and‐by ‘later, in the future’, stop ‘stay, be at’, and too much ‘very’. The second stage of development came with the beginning of the Pacific labor trade in 1863. Islanders, mainly from the Gilbert Islands, New Hebrides, and Solomon Islands were recruited to work on plantations in Queensland (Australia) and later in Fiji, as described on pp. 743–43. Unlike in Fiji, however, English (or rather the English‐based pre‐pidgin) was the language used to run the plantations in Queensland. The pre‐pidgin then became the lingua franca among the linguistically diverse laborers, and with continued use, a stable pidgin language began to develop – early MP.2 Like other restricted pidgins, this had a small vocabulary, mainly based on English, no grammatical inflections, and only a few grammatical rules. Some examples (from Keesing 1988: pp. 43–5) are: (7)

a. White man allsame woman, he no savee fight. (Kolombangara (western Solomon Islands), 1880) b. Suppose me come along school, by‐and‐by me no savee fight. ‘If I come to school, I won’t be able to fight.’ (Bundaberg, Queensland, 1886) c. Me no care, me no belong this fellow place, man here no good – rogue. (Tanna, New Hebrides, 1877)

These examples show the continued use of allsame from all the same to mean ‘like, similar to’ (as in example 6a); savee (saba in example 6b) meaning ‘know how to’ and extended to mean ‘be able to’, and by‐and‐by (6c) used to indicate the future. Also shown is the emergence of along as a general locative preposition. All these features have correspondences in modern MP. On the other hand, the use of the word bad in the pre‐pidgin (6d) dropped out while no good was retained. Labor recruiting for plantations in German‐controlled Samoa began in 1878. Since many of the recruits had already worked in Queensland, early MP was transported to Samoa. From 1879, large numbers were also recruited from the German‐controlled New Guinea Islands (especially eastern New Britain, New Ireland and nearby small islands). After 1885, however, laborers from the New Hebrides or Solomons were no longer recruited for Samoa, and early MP began to diverge into two slightly different varieties – one spoken in Queensland and one in Samoa. Over 62,000 Pacific Islanders went to Queensland between 1863 and 1904, and more than 10,000 to Samoa between 1878 and 1913. Twentieth‐century descriptions of Queensland Canefields English (Dutton 1980) and Samoa Plantation Pidgin (Mühlhäusler 1978a) give some indication of the varieties of pidgin spoken by these laborers.

Contact Languages of the Pacific  747 The third stage of development of MP began when the laborers’ contracts finished and they returned to their home islands, bringing the developing pidgin with them. The pidgin spread rapidly, functioning as a lingua franca. It was also used by the large‐scale internal labor force that worked on plantations in German New Guinea, the New Hebrides, and Solomon Islands at the turn of the century. In each of these countries, early MP further stabilized and expanded under the influence of the local indigenous languages. Today, there are three dialects, differing mainly in vocabulary and a few grammatical rules: Papua New Guinea Tok Pisin, Vanuatu Bislama, and Solomon Islands Pijin. (For lexical and grammatical differences between the dialects, see Siegel 1998, 2008; Tryon and Charpentier 2004. For Tok Pisin, see also Smith 2008a, 2008b, 2013; Smith and Siegel 2013a, 2013b, and for Bislama, Meyerhoff 2013a, 2013b, 2013c). After MP had stabilized into its separate varieties and spread through the islands, it began to be used for new functions, such as imported religion. Tok Pisin was developed into a written language by missionaries in the 1930s, and later used in newspapers and radio broadcasting. Similar developments occurred with Bislama and Pijin. As its use was extended into new areas, MP changed linguistically to become more complex – e.g. acquiring more vocabulary and more grammatical morphology. Thus, in both function and structure, MP became what is called an “expanded pidgin.” As MP was expanding, it acquired many of the grammatical features shared by its substrate languages, which almost all belong to several closely related groups of Austronesian languages, together referred to as Central Eastern Oceanic (CEO; Lynch, Ross, and Crowley 2002). These features came into MP via individuals either creating new variants by transferring properties from their CEO mother tongue, or selecting (subconsciously) the existing variants that were similar to properties in their mother tongue. The presence of corresponding features in the mother tongues of a large proportion of the individuals led to the reinforcement of the transferred and selected variants, and these became incorporated into the expanding pidgin (see Siegel 2008). Keesing (1988) identifies seven “core syntactic structures” of the CEO substrate languages that are found in MP, expressed with forms from the lexifier, English. These are: a. b. c. d. e. f. g.

srp in the verb phrase transitive suffix on verbs adjectives functioning as stative verbs preverbal causative marker postnominal possessive marker third plural pronoun used as a plural marker exclusiveness and dual number marked in the pronoun system

Features (a), (b), and (f) can be seen in the following examples from Bislama: (8)

a. Man ya i man DET SRP.3SG ‘This man stole the money.’ b. Ol woman oli PL woman SRP.3PL

stil‐im steal‐TR

mane. money

kat‐em CUT‐TR

taro. taro

‘The women cut the taro.’ The SRP i is derived from the English word he. The plural SRP oli, used mostly for human subjects, is derived from the third person plural pronoun ol (> all) plus the already existing

748  Jeff Siegel marker i. Ol is also used prenominally as the plural marker, but it is now obsolete as the third person plural pronoun in Bislama, having been replaced by olgeta. A preverbal causative marker (feature d) is shown in this example with mek‐ (from make) in Pijin (Jourdan 2002: 135): (9)

mifala 1PL.EXCL

nating NEG

mek‐rere CAUS‐be.ready

yet yet

fo for

disfala this

gogo trip

blong POSS

mifala. 1PL.EXCL

‘We have not yet prepared anything for our trip.’ The postnominal possessive marker (feature e) in Tok Pisin is bilong (< English belong): (10) mi 1SG

luk‐im see‐TR

haus house

bilong POSS

papa father

bilong POSS

yu. 2SG

‘I saw your father’s house.’ These features are all the consequence of one type functional transfer – i.e. “functionalization” (Siegel 2015: 169). Adjectives functioning as stative verbs (feature c), illustrated in Bislama in (11), is the result of transfer of a morphosyntactic strategy (Siegel 2015: 162). (11) Ol PL

gel girl

blong from

Malo Malo

oli SRP.3PL

bin PST

naes nice

tumas. very

‘The girls from Malo were very nice.’ The pronoun system (feature g) results from the transfer of grammatical categories (Siegel 2015: 165). Bislama, for example, adopted the CEO pronominal categories, indicating dual (and trial), as well as the inclusive/exclusive distinction. These are indicated with forms derived from English pronouns as well as numerals, tu ‘two’ or tri ‘three’, and ‐fala, a ­pronominal plural marker, derived from fellow (Table 37.1).

1.5  Other English‐lexified Pidgins At least three other Pacific pidgins lexified by English have been described in the literature, but these did not stabilize or expand to the extent that MP did. The first is Papuan Pidgin English (Mühläusler 1978b), spoken from the 1880s well into the 1900s, mainly in dealings of indigenous people with English speakers in Papua (British New Guinea). It was most closely linked to the English‐lexified pidgins used in Queensland and the Torres Strait, but was also influenced by early MP. The reasons that Papuan Pidgin English never stabilized or expanded were twofold: First there were no large‐scale plantations or other industries with laborers from diverse locations. Second, as mentioned on p. 744, Police Motu was promoted by the colonial administration and eventually displaced the English‐lexified pidgin. Table 37.1  Bislama pronoun system.

First person inclusive First person exclusive Second person Third person

Singular

Dual

Trial

Plural

– mi yu hem/em

yumitu(fala) mitufala yutufala tufala

yumitrifala mitrifala yutrifala trifala

yumi mifala yufala olgeta

Contact Languages of the Pacific  749 Another variety is Nauru Pidgin English (Siegel 1990). This was spoken at least up until the 1980s on the tiny island of Nauru, primarily in commercial interactions between Chinese shopkeepers and indigenous Nauruans or temporary residents, mainly from Kiribati and Tuvalu (formerly the Gilbert and Ellice Islands) but also from Fiji, the Solomon Islands, Australia, New Zealand, Britain, India, and the Philippines. The origins of the pidgin may go back to the phosphate industry which began in 1908. Laborers were imported from both China and other Pacific islands, at first mostly from the Micronesian Caroline and Marshall Islands, with small numbers also from New Guinea, and later (in the 1950s) primarily from the Gilbert and Ellice Islands. Nauru Pidgin English has a mixture of distinctive features from Chinese Pidgin English – such as the numeral classifier piecee – and MP – e.g. the ‐pela/‐ fala adjective marker. It also has at least 17 features shared by the two varieties, plus some unique features, such as lexical items from Cantonese (e.g. yáuh ‘there is [existential], have’), Kiribati (e.g. tekimoa ‘thief’), and Nauruan (e.g. kumo ‘pig, pork’). The third English‐lexified pidgin, Ngatikese Pidgin or Ngatikese Men’s Language, originated under very different circumstances (Tryon and Charpentier 2004: 145–9). Ngatik is a small island in the Sapwuahfik Atoll, approximately 140 kilometers southwest of Pohnpei (formerly Ponape, now part of the Federated States of Micronesia). In 1837, the crew of the trading ship Lambton, with the help of some Ponapeans, landed on Ngatik, and massacred nearly the entire adult male population. After the massacre, the island was populated by men from nearby islands, especially Pohnpei, and by European sailors and beachcombers. It seems that the Pacific‐wide English‐lexified pre‐pidgin spoken in Micronesia at the time was used for communication among the diverse population, and this developed into what is called Ngatikese Pidgin. Today the first language of the 500 residents of the island is the Sapwuahfik dialect of Ponapean, and this pidgin has very restricted use. It is spoken only by adult males especially when they are involved in communal activities, such as fishing, although it is said to be understood by women and children. With regard to its linguistic features, the pronouns, demonstratives, TMA markers, prepositions, and articles are derived from English while the nouns, verbs, and adjectives are from both English and Ponapean, although more commonly the latter. This gives the language the appearance of more a mixed language (e.g. Matras and Bakker 2003) than a typical pidgin or creole.

2  New Languages: Creoles In some contexts, people in a newly emerging mixed community use a pidgin on a daily basis, and some of them shift to it as their primary language, which they speak to their children. Because of this extended use, the pidgin would already be expanded or in the process of expanding. Thus, children growing up in this context acquire the expanded pidgin as their mother tongue, and it becomes their community language. At this stage it is then called a “creole.” Like any other vernacular language, a creole has a community of native speakers and complete range of informal functions, and as well, it has a full lexicon and a complex set of grammatical rules. Some controversy exists, however, over applying term “creole” to some contact languages. In Melanesia, for example, many people have recently been marrying outside their traditional language groups, especially in urban areas. So often the common language of the parents is a variety of MP, and this is what their children acquire as their first language. Because of this nativization (the process of a pidgin becoming a native language), some linguists say that MP is now a creole, emphasizing that it has thousands of first‐language speakers and has the functions and grammatical features found in typical creoles. However, others say that MP is still an expanded pidgin, pointing out that more than 90% of its speakers still use their ancestral language and learn MP as a second rather than a first language. In

750  Jeff Siegel contrast, creole‐speaking populations generally have shifted from their ancestral languages and are monolingual. Furthermore, MP is not the vernacular language of any distinct, newly emerged community.

2.1  English‐lexified Creole Languages Hawai‘i Creole A clear example of a creole in the Pacific region is Hawai‘i Creole (generally called “Pidgin” by its speakers). The early history of Hawai‘i and the concurrent use of Jargon Hawaiian and South Seas Jargon were described on p. 742. The sugar industry expanded rapidly in the last quarter of the nineteenth century, and imported large numbers of laborers from other countries. In 1884, there were approximately 18,200 Chinese, 10,000 Portuguese, 6,600 “other Caucasians,” 100 Japanese, and 1,400 others living in Hawai‘i, in addition to 40,000 Hawaiians and 4,200 “Part-Hawaiians” (Reinecke 1969: 42). Pidgin Hawaiian continued to dominate on the plantations at this time, but the first generation of immigrants (G1) continued to maintain their own languages (Roberts 2005). Because the different ethnic groups were segregated on the plantations, the locally born children of immigrant laborers (the G2) acquired their parents’ language and did not socialize with other children until they started school. There they learned the languages of their classmates from other ethnic groups, including Hawaiian or Pidgin Hawaiian, as well as some English. However, off the plantations was a different story. Varieties of an English pre‐pidgin, with features of both South Seas Jargon and Chinese Pidgin English, were being used for interethnic communication in Honolulu and other urban areas, and a distinct Hawai‘i Pidgin English (HPE) began to stabilize. Early HPE was characterized by the features typical of a restricted pidgin: no inflections on nouns or verbs; no copula, existential marker, or complementizers; the use of adverbs (such as by‐and‐by and all time) instead of tense or aspect markers; and a single preverbal negator no. Examples from before 1899 (Roberts 2005: 249, 150, 163) illustrate these features and some similarities to South Seas Jargon: (12) a. Melican man he too much smart. ‘Americans are too smart.’ b. Ae (yes), he only boy now, no got sense. By’n by he man, he good. ‘Yes, he is just a boy now, lacking wisdom. When he is a man, he will be good.’ c. Today go court house buy license, go church make marry, all same haole [Caucasian] style. ‘Today I’m going to the court house to buy a license and going to a church to get married, just like Whites do.’ The situation began to change in the late 1880s with increased immigration from Japan. By 1890 there were over 12,600 Japanese in Hawai‘i and after 1900, a large number of G2 children of Japanese ethnicity entered the schools. In addition, in the first decade of the twentieth century there was an influx of laborers from Korea, Puerto Rico, Spain, and the Philippines. When the immigrant population was speaking a dozen or more mutually unintelligible languages, the English‐lexified pidgin HPE came to be used more widely as the language of interethnic communication, especially among the G2, many of whom had left the plantations. The next change occurred from around 1895 to the 1910s, when older G2 children and adults began to shift to HPE as their primary language (Roberts 1998, 2005). With this extension of use, HPE began to expand grammatically. Examples from the G2 show that by 1920, many grammatical morphemes had developed and were frequently used where lexical items or Ø were found in early HPE. These examples come from historical attestations

Contact Languages of the Pacific  751 provided by Roberts (1998, 2005) and data from interviews of a group of male speakers – seven locally born before 1905 and one foreign born in 1904 but arriving in Hawai‘i a year later (Bickerton 1977).3 First, there was the development of a TMA system, with bin V for past, go(n) V for future/ irrealis, and stay V or V‐ing for progressive: (13) This fella bin see. ‘This person saw (it/him/her).’ (1909; Roberts 2005: 180) (14) lawya gon teik, e. ‘Lawyers are going to take (money), aren’t they?’ (Bickerton 1977: 113) (15) a. b.

Wan taim wen we go hom in da nait dis ting stei flai ap. ‘Once when we went home at night these things were flying about.’ (Bickerton 1977: 18) maeshin shap hi teiking, si? ‘He’s apprenticed in a machine shop, see?’ (Bickerton 1977: 101)

Second, there was development of a copula stay in locatives, an existential (and possessive) marker get, and a complementizer for: (16) That time Sing Ping no stay, about 12 o’clock Sing Ping come home. ‘Sing Ping wasn’t here at that time; he came home about 12 o’clock.’ (1904; Roberts 1998: 23) (17) I believe get all black paint. ‘I believe there was just black paint.’ (1923; Roberts 2005: 176) (18) You speak you want one good Japanese man for make cook. ‘You said you wanted a good Japanese man to cook.’ (1905; Roberts 1998: 29) Evidence exists that the development of at least some of these expanded features of HPE was influenced by reinforcement or transfer from the substrate languages that were dominant at the time that the language was expanding. Of the immigrant languages, these were Portuguese and Cantonese. Although the number of speakers of Japanese was greater in the total population, in the all‐important G2 that first shifted to HPE, the numbers of Portuguese and Cantonese speakers were more significant. It is also clear that the Portuguese were the first group to abandon their ancestral language, followed by the Chinese (see Roberts 1998; Siegel 2000). One of the substrate‐influenced features appears to be the use of a single form get to express both existential and possessive. This occurs both in Portuguese (the verb ter or haver) and in Cantonese (yáuh). And the use of stay for both the locative copula and verbal auxiliary in progressives is parallel to that of Portuguese estar. (For more details and other features, see Siegel 2000, 2008; Sakoda and Siegel 2008a, 2008b, 2013; Velupillai 2013a, 2013b.)

752  Jeff Siegel From approximately 1905 to the 1920s, the number of children of the G3 increased rapidly. These were the children of the first locally born generation (G2) who had shifted to the pidgin as their primary language. Thus the children of the G3 heard the expanded HPE from their parents, and later from their peers, and acquired it as their first language – and in most cases, only language. Thus, Hawai‘i Creole emerged with the G3, the children of the locally born children of the original immigrants.4 Of course the G3’s creole was not exactly the same as their parents’ HPE, as the children regularized the still variable input, and adopted some features but not others.

2.1.1  Other English‐lexified Creoles Two other English‐lexified creoles developed in the Pacific under very different circumstances. The first was on Pitcairn Island in the central South Pacific. In 1792, after the famous mutiny on the Bounty, 9 mutineers (5 from England, 2 from Scotland, 1 from the USA, and 1 from St. Kitts in the Caribbean) settled on the uninhabited island with 6 men and 12 women from Tahiti (some were originally from Tubuai). The Tahitians were treated almost as slaves, and in 1794 the men revolted. The violence resulted in their deaths and left alive only four of the original mutineers. After a few years of peace, more violence erupted and that along with illness led to further deaths. In 1800, there was only one surviving man, John Adams, left on the island, with 10 women and 23 children. When Adams died in 1829, the population was 80. It is likely that the first children born on the island were bilingual in the mutineers’ language, English, and their mother’s language, Tahitian. But a pre‐pidgin also developed on Pitcairn that had a mixture of features of the English dialects of the mutineers, the Tahitian of the women, South Seas Jargon used by visiting whalers, and perhaps an Atlantic contact variety, spoken by the mutineer from the Caribbean. It is possible that the first island‐born generation stabilized this variety and adopted it as their primary language, at which time it expanded. When it became the first language of the second generation of children born on the island, it could then be classified as a creole. But throughout its history, this creole has differed from others in that a large proportion of its speakers have been literate in English, and exposed to the Bible and religious texts (Ingram and Mühlhäusler 2008: 272). In 1856, all of the 194 people living on Pitcairn moved to Norfolk Island, a larger island near Australia that had been a penal colony. A few families returned to Pitcairn in 1859 and others in 1864. The splitting of the community resulted in two different varieties of the language: Pitcairn (now sometimes spelled Pitkern), and Norfolk (sometimes spelled Norf’k). The precise differences between the modern varieties are not well understood, despite fairly recent studies (Kållgård 1993 on Pitcairn; Ingram and Mühlhäusler 2008, and Mühlhäusler 2008, 2013a, 2013b on Norfolk; Mühlhäusler 2013c on both). Nevertheless, since Pitcairn Island remains a fairly isolated colony of Great Britain, while Norfolk Island has been a territory of Australia since 1901, it is thought that Norfolk is closer to English because its speakers have more frequent contact with tourists and other outsiders. Today, speakers of Norfolk all know and use standard English, and reserve Norfolk for informal in‐group communication. However, both Pitcairn and Norfolk are not as commonly spoken as before and the language as a whole is considered endangered (Mühlhäusler 2008: 579). The origins of the other English‐lexified creole appear to have been similar to those of Pitcairn/Norfolk. It developed in the Bonin (Ogasawara) Islands, which were uninhabited until 1830, after 20 settlers arrived: 5 Europeans (3 English speakers, 1 Genoese, and 1 Danish) and 15 Pacific Islanders (10 men and 5 women), mostly from Hawai‘i but also from Tahiti and the Marquesas (Long 1999, 2007). Later settlers spoke dozens of languages, from Europe, the Pacific, and Asia. Reports indicate, however, that the lingua franca was English, although a list of words collected there by shipwrecked Japanese

Contact Languages of the Pacific  753 sailors in 1840 contained items of both Hawaiian and English origins. Because of many visiting ships after the initial settlement, it is assumed that the settlers were also exposed to the South Seas Jargon that was used all over the Pacific at that time. Thus, it is thought that an English‐lexified contact variety developed among the first generation of settlers, influenced by Hawaiian (perhaps Jargon Hawaiian, see p. 742) and South Seas Jargon. This became the first language of either the first or second generation born on the island, and thus a creole. After Japan colonized the islands in the 1870s, the original settlers continued to speak “English” along with Japanese, and they maintained their language into the twentieth century. However, with increased travel away from the islands, more education in English, and the presence of the American military after World War II, the language became closer and closer to mainstream English.

2.2  Pidgins and Creoles Lexified by Other European Languages Although German‐lexified contact languages are rare in the world, one did emerge in German New Guinea: Rabaul Creole German, also known as Unserdeutsch (Volker 1991). This developed at a boarding school for mixed‐race children established by Catholic missionaries in 1897 at Vunapope, near the capital, Rabaul. These were the children of local women who had relations with men from other parts of the world, including Germany, Australia, Micronesia, China, Ambon, and the Philippines. The children came to the school at a young age, knowing a bit of their mother’s language and an early form of Tok Pisin. But at school they were taught German and that was the only language they were allowed to speak. The German‐lexified contact language seems to have resulted from the students using German words in Tok Pisin sentences. This relexified Tok Pisin stabilized quickly in the relatively isolated dormitories, and remained as an in‐group language, even though the students eventually progressed in standard German. When the students left school and became adults, many of them married other former students and settled around the Vunapope mission. They continued to use their own form of German, which they called Unserdeutsch (‘our German’), among themselves, and spoke a local colloquial form, which they called Normaldeutsch (‘normal German’), with other Germans. This continued even after Australia took over in 1914. However, with the coming of World War II, the situation began to change; teaching of German was prohibited and the community became more mobile. At the time of independence, most of the community moved to Australia, and more than a thousand of their descendants live in southeastern Queensland. In several ways, Unserdeutsch is not a typical creole. First, it did not arise from a pidgin or pre‐pidgin needed as a medium of wider communication among speakers of different languages, since the students already had such a medium in Tok Pisin. Second, its speakers were bilingual in a more standard form of German. But it is typical of a creole in that it filled the need for a distinctive in‐group language for a newly emerged mixed community. (See also Maitz and Volker 2017.) France has played a much greater role in the Pacific than Germany, yet French‐lexified contact languages are also rare. A French‐lexified pidgin is said to have been used in plantations in the New Caledonia region in the latter half of the nineteenth century after the French took over the islands (Corne and Hollyman 1996), but little is known about it. More is known about the French‐lexified creole of the French territory of New Caledonia, Tayo, also known in French as Patois de St‐Louis or just Patois. It is currently spoken primarily in the village of St‐Louis, about 15 kilometers from the capital, Nouméa. It has about 2,000 speakers, half of these the permanent inhabitants of St‐Louis and the other half former residents who live elsewhere around the territory. Information on the origins of Tayo comes from Ehrhart and Corne (1996). (See also Ehrhart and Revis 2013a, 2013b.)

754  Jeff Siegel St‐Louis was established at an uninhabited site by French Marist missionaries in 1860. It was to be a village for new converts and a training center for catechists. From 1860 to 1880, speakers of as many as 20 different Melanesian languages were attracted to St‐Louis, but three mutually unintelligible languages were the most common: Cèmuhî, Drubea, and Xârâcùù. The Melanesian settlers were exposed to French in varying degrees at the mission, where they went to school and church and worked in the sawmill, rice paddies, sugarcane fields, and vegetable gardens. During the first 20 years of the settlement, a French‐lexified pidgin began to develop as the lingua franca. This pidgin became more and more important as the medium of communication at St‐Louis, especially among the first locally born generation. According to oral tradition, this generation was bilingual in the pidgin and the Melanesian language of their group, but the next generation, especially those born after around 1920, acquired the pidgin as their first language and had only passive knowledge of their parents’ and grandparents’ first languages. This is when the creole now known as Tayo became established in the community. (Note that this matches the three generational shift scenario that occurred in Hawai‘i.) Today there are very few speakers of Melanesian languages left in St‐Louis. As in Melanesian Pidgin and Hawai‘i Creole, the expanded features of Tayo demonstrate influence from the substrate languages. With regard to the verb phrase, Tayo has preverbal TMA markers derived from French forms but their functions correspond to markers that occur in all three of the major substrate languages or at least in the two most dominant ones when expansion was taking place: Cèmuhî and Drubea. This is the result of another type of functional transfer: “refunctionalization” (Siegel 2015: 171). These include a future maker va (from va, the most common form of the French verb aller ‘to go’), a progressive aspect marker aṅtraṅde (from en train de ‘in the process of’), and a past ­accomplished or completive marker fini (the past participle of finir ‘finish’).5 The following examples come from Ehrhart (1993): (19)

a. b. c.

boṅ la va raṅtre    se   swar … (p. 220) good 3SG FUT come.home  DEM  evening ‘OK, she will come home this evening …’ mwaṅche chokola (p. 118) nu aṅtraṅde 1PL PROG eat chocolate ‘We are eating chocolate.’ pi kaṅ sola fini labure later sola plaṅte mais (p. 246) and when 3PL ACP plow earth 3PL plant maize ‘And when they had finished plowing the earth, they planted the maize.’

All three of the main substrate languages also have a preverbal marker of evidential modality, asserting the reality of the event or state being reported. Tayo appears to have a similar marker: ryaṅke (from French rien que ‘nothing but’) and its variant aṅke  –  for example: (20) sola aṅke fe aṅ graṅ barach 3PL EVID make a big barricade ‘They’ve (certainly) made a big barricade on the road.’6

si on

larut (p. 181) road

This contradicts the claim (McWhorter 2001) that evidential marking is not found in creoles.

Contact Languages of the Pacific  755

3  New Dialects Two types of new dialects may emerge in contact situations: mixed dialects (or koines), which result from dialect contact; and indigenized varieties, which result from language contact.7 Examples of each are found in the Pacific.

3.1 Koines A koine is a new dialect that results from a changed pattern of contact between dialects of the same language. For example, an “immigrant koine” may develop when speakers of different regional dialects relocate to a new location and together form a new community (Siegel 1985). The process of koine formation is called koineization, and it often leads to intermediate or “interdialect” forms, not found in any of the contributing dialects (Trudgill 1986). In the first stage of koine formation, the pre‐koine, the various dialects in contact are used concurrently for communication. A stable koine emerges much as a stable pidgin does – i.e. when leveling occurs, and some variants are eliminated while others are retained. Often the resulting koine is formally simpler than any of the contributing dialects – for example, in having fewer marked grammatical categories. Fiji Hindi (sometimes called Fiji Hindustani or Fiji Bāt), the informal language of most Indo‐Fijians, is an immigrant koine comprised mainly of features of several of the regional dialects of Hindi. These were spoken by the more than 45,000 indentured laborers who were brought to Fiji from North India between 1879 and 1916. Most of the lexicon and grammatical morphology of Fiji Hindi come from Eastern Hindi dialects such as Awadhi, and from Bihari dialects, mainly Bhojpuri. Other grammatical features come from the Pidgin Hindustani spoken on the sugar plantations, as mentioned on p. 744. Mixing is evident in the pronoun system, with forms from all three sources just mentioned. The pronoun system shows simplification as well, with the loss of the intimate second person pronouns, leaving only familiar and formal categories. Also, the system is more semantically transparent with log ‘people’ joined to the singular pronoun to form the plural for animates – e.g. ham 1SG, hamlog 1PL. An interdialect form is the second person possessive pronoun tumar (tohar, tuhar, or tum(h)ara in the contributing dialects). With regard to the lexicon, many items also come from Fijian, especially for names of local flora and fauna (such as dalo ‘taro’ and walu ‘kingfish’), and from English (such as room, towel, book, and reef). However, for some, semantic shift has occurred – for example, gate means ‘field or paddock’. Another immigrant koine began to emerge in Fiji in the early twentieth century among former plantation laborers from the Solomon Islands who stayed on in Fiji at the end of their contracts. This was known as Wai (Siegel 1987: 211–33). It was a mixture of dialects of North Malaitan. Although five separate languages are often distinguished today – Lau, To’abaita, Baelelea, Baegu, and Fataleka – they are mutually intelligible, and generally considered to be “major dialects or sublanguages” of a single language (Tryon and Hackman 1983: 27). Again, the pronoun system is illustrative, with a mixture of forms from different dialects and the loss of pronouns distinguishing dual and trial first person exclusive from the plural. An example of an intermediate form is raŋguva ‘grease’ (ragufa, raraŋa, or ndū la in the contributing dialects). It appears, however, that Wai did not stabilize before it disappeared when the Solomon Islanders shifted to the Fijian language as they became integrated into the Fijian community. A koine based on dialects of Japanese began to emerge in the Palauan islands in the western Pacific (described by Matsumoto and Britain 2003). Japan occupied the islands (now the Republic of Palau, or Belau) from 1914 to 1945, and there was a mass migration of Japanese workers. In 1941, there was a total of 23,980 Japanese, compared to only 6,514 Palauans

756  Jeff Siegel (Matsumoto and Britain 2003: 43). The workers came from 11 districts of Japan, representing all four major dialect regions: Eastern Honshū, Western Honshū , Kyū shū , and Ryū kyū (Okinawan). It appears that Palauan Japanese was a mixture of features of dialects from these regions. The main contributors, however, were the Eastern dialects (from the Kantô (Tokyo), Tōhoku, Hokkaido, and Tōkaidō districts), whose speakers were originally dominant in number. Because of a great deal of interaction with the Japanese, including intermarriage, Palauans, also learned the language, and their second‐language versions of Japanese also contributed to the mix. It appears that variants from Okinawan and other nonprestigious dialects were leveled out. But it is not clear if Palauan Japanese got past the pre‐koine stage because surviving speakers use variants from both Eastern and Western dialects for the same function. Another example of mixing of Japanese dialects occurred in Hawai‘i among the more than 200,000 Japanese imported as plantation laborers from 1884 to 1924 (described by Hiramoto 2006, 2010). Again, immigrants came from all four major dialect regions, but in contrast to Palau, the most dominant dialects were from the Western region, especially the Chū goku dialect. Also in contrast to Palau, other ethnic groups did not learn Japanese. Recordings of the first generation of immigrants reveal them using some lexical and ­morphosyntactic forms from dialects other than their own. Some speakers also used an interdialect form dakē for a conjunction (Eastern dakara and Western jakē), and loanwords from English, including the pronoun me. However, it appears that a stable koine did not emerge in the second generation before they shifted to Hawai‘i Creole as their primary language.

3.2  Indigenized Varieties An indigenized variety is a new dialect that develops when a language is transplanted to a new location and some aspects of its lexicon, phonology, and morphosyntax become influenced by local indigenous languages (thus, indigenized). Prototypical indigenized varieties have arisen in colonies where the colonial language has had widespread use in the education system, and has been learned as a second language by a large proportion of the population. Like an expanded pidgin, this type of indigenized variety is used in a multilingual environment and functions as a lingua franca for daily interactions among speakers of different languages. Such an indigenized variety may also have native speakers, like a creole. Unlike an expanded pidgin or creole, however, its grammatical rules are much closer to those of the lexifier (the colonial language). Nearly all research on such indigenized varieties has been done in countries around the world where English was the colonial language until independence in the latter half of the twentieth century – for example, Singapore, India, and the Philippines. Thus, the terms New Englishes, World Englishes, and Postcolonial Englishes are often used as well. Two such indigenized varieties have been documented in the Pacific: Fiji English (Siegel 1989; Mugler and Tent 2008; Tent and Mugler 2008; Tent 2013), and Papua New Guinea (PNG) English (Smith 1978, 1988).8 Fiji English has many lexical items from both Fijian (e.g. sulu ‘sarong’ and kasou ‘drunk’) and Fiji Hindi (e.g. roti ‘Indian flat bread’ and choro ‘steal’). It also has a pronoun system influenced by Fijian, with gang as a plural marker (e.g. you gang), and us gang as first person plural exclusive versus us two as first person dual inclusive. Another influence of Fijian is a preverbal intensifier full as in: (21)

The fella full sleeping over there. ‘The guy’s sound asleep over there.’ (Siegel 2008: 125)

Contact Languages of the Pacific  757 PNG English also has many items from Tok Pisin (e.g. singsing ‘traditional singing and dancing’, wantok ‘speaker of the same language’, and kaukau ‘sweet potato’), and from indigenous languages, most probably via Tok Pisin as well (e.g. bilum ‘string bag’ and buai ‘betelnut’). Both varieties have items from English with shifted meaning – e.g. Fiji English grog ‘kava’ and PNG English rascal ‘thug, criminal’. And they have many instances of grammatical shift – e.g. Fiji English broom the floor ‘sweep the floor’ and PNG English be aftering someone ‘be following someone’. Even though there has been hardly any contact between Fiji English and PNG English, they also share many morphosyntactic features, and these are found in other indigenized varieties as well. These features include regularization of plurals to include non‐count nouns (e.g. furnitures), phrasal verbs where simple verbs exist in standard varieties (e.g. cope up), no change of word order for questions, invariant question tags, variable copula/auxiliary use, and one used as an indefinite article. Some of these are illustrated in examples from Fiji English (Mugler and Tent 2008: 551–2): (22) a. Jone and them coming to the party tonight, èh? b. They should have one security guard up here at night sitting in one shed. A research project is currently underway to document the English spoken in several locations in Micronesia: Palau, Saipan, Kosrae, Guam, Kitibati, Nauru and the Marshall Islands. This research (Britain et al. forthcoming) will determine whether new indigenized varieties have emerged in this part of the Pacific as well. Another type of indigenized variety, which has similarity in origins to the creoles of Pitcairn and Bonin Islands (see pp. 752–53), is Palmerston Island English (Ehrhart and Hendery 2013; Hendery 2015). Palmerston Island, a remote atoll of the Cook Islands, was uninhabited when it was settled in 1863 by an Englishman, William Masters, his three Cook Islands wives, and a Portuguese‐speaking man, Jean‐Baptiste Fernández.9 Masters ruled the island autocratically and decreed that only English could be spoken. All speakers of the current variety, just over 50 of whom remain on the island, are the descendants of Masters and the Cooks Islands women. Palmerston Island English has linguistic features similar to those of other indigenized varieties, such as influence of the first language of some of its original speakers (in this case, Cook Islands Māori) – for example in lexicon (e.g. motu ‘islet’) and grammar (e.g. a singular/dual/plural distinction in pronouns, you/you two/you lot). There are also examples of grammatical shift, such as then you dough it ‘then you knead it’ (Ehrhart and Hendery 2013: 632).

4 Conclusion This chapter has shown the diverse origins and linguistic features of Pacific contact varieties. However, some commonality in their development can also be seen in various shared characteristics. These include formal simplicity compared to contributing languages  –  for example in having more regularity and fewer marked grammatical categories. Such features are the consequence of overgeneralization and lack of complexity that are found in early second language acquisition. Also, expanded pidgins and creoles commonly exhibit grammatical features resulting from transfer from substrate languages  –  a compensatory strategy of second‐language use.

758  Jeff Siegel

Abbreviations ACP = accomplished, ART =article, ASP = aspect, CAUS = causative, DEM = demonstrative, DIR = directional, EVID = evidential, EXCL = exclusive, FUT = future, LOC = locative, NEG = negative.

NOTES 1 Standard Fijian orthography is used in examples of both standard and pidginized Fijian: = /mb/, = /ð/, = /nd/, = /ŋ/, = /ŋg/. Capital letters are used for non‐prenasalized stops not found in standard Fijian: = [d], = [g]. See also the abbreviations listed at the end of this chapter. 2 According to Keesing (1988), a single, distinctive Pacific Nautical Pidgin developed in the central Pacific before the 1860s and this was the forerunner of Melanesian Pidgin. However, historical research does not back up this claim (see Baker and Mühlhäusler 1996). 3 Bickerton calls these “early creole speakers” (1977: 333) but clearly states that they were “nonmonolingual” as opposed to the “monolingual” creole speakers who were born later. Of course, it must be kept in mind that these early speakers were not recorded until the 1970s and could have been influenced by later developments in the language. 4 Note that this description of the origins of Hawai‘i Creole is based on the recent findings of Roberts (2005) and differs from that presented by Bickerton (e.g. 1981), see Siegel (2008). 5 The orthography used here is an adapted version of the lortograf‐linite system created by Baker and Hookoomsing (1987). The major differences between this system and IPA are that tch = [c] and ch = [ʃ]; also, either aṅ or aḿ = [ã] and oṅ or oḿ = [õ]. 6 This example was collected by Chris Corne before his death in 1999. (See Siegel 2008: 223.) 7 Of course, it is often difficult to distinguish whether two varieties are separate languages, or dialects of the same language. While there are some clear cases of language versus dialect, there is no precise linguistic dividing line that can distinguish them. Similarly, the other types of contact varieties described here are graded phenomena rather than essential categories. For example, there are ­pidgins, such as Hiri (Police) Motu, that fall somewhere between the restricted and expanded ­categories. And as we have seen with Melanesian Pidgin, it is sometimes hard to draw the line between an expanded pidgin and a creole. 8 With more than 4 million speakers, PNG English is certainly worthy of further research. 9 A small group of other Cook Islands men and women also settled initially on the island, and stayed for at least for the first few years.

REFERENCES Baker, Philip and Vinesh Y. Hookoomsing 1987. Diksyoner kreol morisyen: Dictionary of Mauritian Creole. Paris: L’Harmattan. Baker, Philip and Peter Mühlhäusler 1996. The origins and diffusion of Pidgin English in the Pacific. In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas. Berlin: de Gruyter Mouton, pp. 551–94.

Bickerton, Derek 1977.Change and Variation in Hawaiian English, vol. 2: Creole Syntax (Final report on National Science Foundation grant no. GS‐39748). Honolulu: Social Sciences and Linguistics Institute, University of Hawai‘i. Bickerton, Derek 1981. Roots of Language. Ann Arbor: Karoma. Britain, David, Kazuko Matsumoto, Dominique Hess, Tobias Leonhardt and Sara Lynch forthcoming. Micronesian Englishes. Berlin: de Gruyter Mouton.

Contact Languages of the Pacific  759 Clark, Ross 1979. In search of Beach‐la‐mar: towards a history of Pacific Pidgin English. Te Reo 22: 3–64. Corne, Chris and Jim Hollyman 1996. French in the South Pacific. In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas. Berlin: de Gruyter Mouton, pp. 251–63. Day, Richard R. 1987.Early pidginization in Hawaii. In Glenn G. Gilbert (ed.), Pidgin and Creole Languages: Essays in Memory of John E. Reinecke. Honolulu: University of Hawai‘i Press, pp. 163–76. Drechsel, Emanuel J. 1999. Language contact in the early colonial Pacific: evidence for a maritime Polynesian jargon or pidgin. In John R. Rickford and Suzanne Romaine (eds.), Creole Genesis, Attitudes and Discourse: Studies Celebrating Charlene J. Sato. Amsterdam: Benjamins, pp. 71–96. Drechsel, Emanuel J. 2007. Sociolinguistic‐ ethnohistorical observations on Maritime Polynesian Pidgin in Herman Melville’s two major semi‐autobiographical novels of the Pacific. Journal of Pidgin and Creole Languages 22.2: 231– 61. Drechsel, Emmanuel J. 2014. Language Contact in the Early Colonial Pacific: Maritime Polynesian Pidgin before Pidgin English. Cambridge: Cambridge University Press. Dutton, Tom [Thomas E.] 1980. Queensland Canefields English of the Late Nineteenth Century (Pacific Linguistics D‐29). Canberra: Australian National University. Dutton, Tom [Thomas E.] 1983. Birds of a feather: a pair of rare pidgins from the Gulf of Papua. In Ellen Woolford and William Washabaugh (eds.), The Social Context of Creolization. Ann Arbor,: Karoma, pp. 77–105. Dutton, Tom [Thomas E.] 1985. Police Motu: Iena sivarai [Its Story]. Port Moresby: University of Papua New Guinea Press. Dutton, Tom [Thomas E.] 1997. Hiri Motu. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective. Amsterdam: Benjamins, pp. 9–41. Ehrhart, Sabine 1993. Le créole français de St‐Louis (le tayo) en Nouvelle‐Calédonie. Hamburg: Buske. Ehrhart, Sabine and Chris Corne 1996. The creole language Tayo and language contact in the “Far South” region of New Caledonia. In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of

Intercultural Communication in the Pacific, Asia and the Americas. Berlin: de Gruyter Mouton, pp. 265–70. Ehrhart, Sabine and Rachel Hendery 2013. Palmerston Island English. In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://ewave‐ atlas.org/languages/66 Ehrhart, Sabine and Melanie Revis 2013a. Tayo structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at. http://apics‐ online.info/contributions/57 Ehrhart, Sabine and Melanie Revis 2013b. Tayo. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 2: Portuguese‐based, Spanish‐ based, and French‐based Languages. Oxford: Oxford University Press, pp. 271–82. Foley, William A. 1988. Language birth: the processes of pidginization and creolization. In Frederick J. Newmeyer (ed.), Linguistics: The Cambridge Survey, vol. 4: Language: The Socio‐ Cultural Context. Cambridge: Cambridge University Press, pp. 162–83. Foley, William A. 2013a. Yimas‐Arafundi Pidgin structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/69 Foley, William A. 2013b. Yimas‐Arafundi Pidgin. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber, (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 3: Contact Languages Based on Languages from Africa, Asia, Australia, and the Americas. Oxford: Oxford University Press, pp. 105–13. Geraghty, Paul 1978. Fijian dialect diversity and foreigner talk. In Albert J. Schütz (ed.), Fijian Language Studies: Borrowing and Pidginization (Bulletin of the Fiji Museum 4). Suva: Fiji Museum, pp. 51–67. Hendery, Rachel 2015. Palmerston Island English. In Jeffrey P. Williams, Edgar W. Schneider, Peter Trudgill, and Daniel Schreier (eds.),

760  Jeff Siegel Further Studies in the Lesser‐Known Varieties of English. Cambridge: Cambridge University Press, pp. 267–87. Hiramoto, Mie 2006. Dialect contact in Hawai‘i: the use of Japanese by plantation immigrants (University of Hawai‘i doctoral dissertation). Hiramoto, Mie 2010. Dialect contact and change in the northern Japanese plantation immigrants in Hawai‘i. Journal of Pidgin and Creole Languages 25.2: 229–62. Ingram, John and Peter Mühlhäusler 2008. Norfolk Island‐Pitcairn English: phonetics and phonology. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 267–91. Jones, Alan A. 1996. Privately owned Mekeo‐based trade languages. In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas. Berlin: de Gruyter Mouton, pp. 219–24. Jourdan, Christine 2002. Pijin: A Trilingual Cultural Dictionary (Pacific Linguistics 526). Canberra: Australian National University. Kållgård, Anders 1993. Present‐day Pitcairnese. English World‐Wide 14.1: 71–114. Keesing, Roger M. 1988. Melanesian Pidgin and the Oceanic Substrate. Stanford: Stanford University Press. Long, Daniel 1999. Evidence of an English contact language in the nineteenth century Bonin (Ogasawara) Islands. English World‐Wide 20.2: 251–86. Long, Daniel 2007. English in the Bonin (Ogasawara) Islands (Publications of the American Dialect Society 91). Durham, NC: Duke University Press. Lynch, John, Malcolm Ross, and Terry Crowley 2002. The Oceanic Languages. Richmond: Curzon. Maitz, Péter and Craig Alan Volker 2017. Documenting Unserdeutsch: reversing colonial amnesia. Journal of Pidgin and Creole Languages 32.2: 365–97. Matras, Yaron and Peter Bakker (eds.) 2003. The Mixed Language Debate: Theoretical and Empirical Advances. Berlin: de Gruyter Mouton. Matsumoto, Kazuko and David Britain 2003. Contact and obsolescence in a diaspora variety of Japanese: the case of Palau in Micronesia. Essex Research Reports in Linguistics 44: 38–75. McWhorter, John H. 2001. The world’s simplest grammars are creole grammars. Linguistic Typology 5.2–3: 125–66.

Meyerhoff, Miriam 2013a. Bislama structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/23 Meyerhoff, Miriam 2013b. Bislama. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 1: English‐based and Dutch‐based Languages. Oxford: Oxford University Press, pp. 223–31. Meyerhoff, Miriam. 2013c. Bislama. In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://ewave‐ atlas.org/languages/69 Michaelis, Susanne Maria, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) 2013a. The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at https://apics‐online.info Michaelis, Susanne Maria, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.) 2013b. The Atlas of Pidgin and Creole Language Structures, vol. 1: English‐based and Dutch‐based Languages. Oxford: Oxford University Press. Mugler, France and Jan Tent 2008. Fiji English: Morphology and syntax. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 546–67. Mühlhäusler, Peter 1978a. Samoan Plantation Pidgin. In Papers in Pidgin and Creole Linguistics no. 1 (Pacific Linguistics A‐54). Canberra: Australian National University, pp. 67–119. Mühlhäusler, Peter 1978b. Papua Pidgin English rediscovered. In Stephen A. Wurm and Lawrence Carrington (eds.), Second International Conference on Austronesian Linguistics: Proceedings (Pacific Linguistics C‐61). Canberra: Australian National University, pp. 1377–446. Mühlhäusler, Peter 2008. Norfolk Island‐Pitcairn English (Pitkern Norfolk): morphology and syntax. In Kate Burridge and Bernd Kortmann (eds.) 2008. Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 568–82. Mühlhäusler, Peter 2013a. Norf’k structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus

Contact Languages of the Pacific  761 Huber (eds.) 2013a. The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/24 Mühlhäusler, Peter 2013b. Norf’k. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 1: English‐based and Dutch‐based Languages. Oxford: Oxford University Press, pp. 232–40. Mühlhäusler, Peter 2013c. Norfolk Island/ Pitcairn English. In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://ewave‐ atlas.org/languages/65 Mühlhäusler, Peter, Thomas E. Dutton, Even Hovdhaugen, Jeffrey Williams, and Stephen A. Wurm 1996. Precolonial patterns of intercultural communication in the Pacific islands. In Stephen A. Wurm, Peter Mühlhäusler, and Darrell T. Tryon (eds.) 1996. Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas. Berlin: de Gruyter Mouton, pp. 401–31. Reinecke, John E. 1969 [1935]. Language and Dialect in Hawai‘i: A Sociolinguistic History to 1935. Honolulu: University of Hawai‘i Press. Roberts, Julian M. 1995. Pidgin Hawaiian: a sociohistorical study. Journal of Pidgin and Creole Languages 10: 1–56. Roberts, Sarah J. 1998. The role of diffusion in the genesis of Hawaiian Creole. Language 74: 1–39. Roberts, Sarah J. 2005. The emergence of Hawai‘i Creole English in the early twentieth century: the sociohistorical context of creole genesis (Stanford University doctoral dissertation). Roberts, Sarah J. 2013a. Pidgin Hawaiian structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/71 Roberts, Sarah J. 2013b. Pidgin Hawaiian In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds), The Atlas of Pidgin and Creole Language Structures, vol. 3: Contact Languages Based on Languages from Africa, Asia, Australia, and the Americas. Oxford: Oxford University Press, pp. 119–27.

Sakoda, Kent and Jeff Siegel (2008a). Hawai‘i Creole: phonology. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 210–33. Sakoda, Kent and Jeff Siegel (2008b). Hawai‘i Creole: morphology and syntax. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 514–45. Sakoda, Kent and Jeff Siegel. 2013. Hawaiʼi Creole. In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://ewave‐atlas.org/ languages/71 Siegel, Jeff 1985. Koines and koineization. Language in Society 14: 337–78. Siegel, Jeff 1987. Language Contact in a Plantation Environment: A Sociolinguistic History of Fiji. Cambridge: Cambridge University Press. Siegel, Jeff 1989. English in Fiji. World Englishes 8: 45–58. Siegel, Jeff 1990. Pidgin English in Nauru. Journal of Pidgin and Creole Languages 5.2: 157–86. Siegel, Jeff 1992. The transformation and spread of Pidgin Fijian. Language Sciences 14: 1–22. Siegel, Jeff 1998. Dialectal differences and substrate reinforcement in Melanesian Pidgin. Journal of Sociolinguistics 2: 347–73. Siegel, Jeff 2000. Substrate influence in Hawai‘i Creole English. Language in Society 29: 197–236. Siegel, Jeff 2008. The Emergence of Pidgin and Creole Languages. Oxford: Oxford University Press. Siegel, Jeff 2013a. Pidgin Hindustani structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://­apics‐online.info/ contributions/70 Siegel, Jeff 2013b. Pidgin Hindustani. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 1: English‐based and Dutch‐ based Languages. Oxford: Oxford University Press, pp. 114–18 Siegel, Jeff 2015. The role of substrate transfer in the development of grammatical morphology in language contact varieties. Word Structure 8.2: 160–83.

762  Jeff Siegel Smith, Anne‐Marie 1978. The Papua New Guinea Dialect of English (ERU Research Report 25). Port Moresby: University of Papua New Guinea. Smith, Anne‐Marie 1988. English in Papua New Guinea. World Englishes 7.3: 299–308. Smith, Geoff P. 2008a. Tok Pisin: phonology. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 188–209. Smith, Geoff P. 2008b. Tok Pisin: morphology and syntax. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 488–513. Smith, Geoff P. 2013. Tok Pisin. In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://ewave‐ atlas.org/languages/70 Smith, Geoff P. and Jeff Siegel 2013a. Tok Pisin structure dataset. Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/22, Smith, Geoff P. and Jeff Siegel 2013b. Tok Pisin. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 1: English‐based and Dutch‐based Languages. Oxford: Oxford University Press, pp, 214–22. Tent, Jan 2013. Pure Fiji English (basilectal FijiE). In Bernd Kortmann and Kerstin Lunkenheimer (eds.), The Electronic World Atlas of Varieties of English. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://ewave‐atlas.org/languages/68

Tent, Jan and France Mugler 2008. Fiji English: Phonology. In Kate Burridge and Bernd Kortmann (eds.), Varieties of English 3: The Pacific and Australasia. Berlin: de Gruyter Mouton, pp. 234–66. Trudgill, Peter 1986. Dialects in Contact. Oxford: Blackwell. Tryon, Darrell T. and Jean‐Michel Charpentier 2004. Pacific Pidgins and Creoles: Origins, Growth and Development. Berlin: de Gruyter Mouton. Tryon, Darrell T. and Brian D. Hackman 1983. Solomon Islands Languages: An Internal Classification (Pacific Linguistics C‐72). Canberra: Australian National University. Velupillai, Viveka 2013a. Hawai‘i Creole structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Available at http://apics‐ online.info/contributions/26 Velupillai, Viveka 2013b. Hawai‘i Creole. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber (eds.), The Atlas of Pidgin and Creole Language Structures, vol. 1: English‐based and Dutch‐based Languages. Oxford: Oxford University Press, pp. 252–61. Volker, Craig 1991. The birth and decline of Rabaul Creole German. Language and Linguistics in Melanesia 22.1–2: 143–56. Williams, Jeffrey P. 1993. Documenting the Papuan‐based pidgins of insular New Guinea. In Francis Byrne and John Holm (eds), Atlantic Meets Pacific: A Global View of Pidginization and Creolization. Amsterdam: Benjamins, pp. 355–63. Williams, Jeffrey P. 2000. Yimas‐Alamblak Tanim Tok: An indigenous trade pidgin of New Guinea. Journal of Pidgin and Creole Languages 15.1: 37–62.

Index

African Englishes Afro‐Saxon bilingualism, 386 background, 385 beginning of English in Africa, 386 Krio in Sierra Leone, 385 Southern and East Africa, 386 Black South African English, 388 DRESS vowel, 389 FACE vowel, 389 FOOT and GOOSE vowels, 389 GOAT vowel, 389 KIT and FLEECE vowels, 389 LOT and THOUGHT vowels, 389 NURSE vowel, 389 STRUT vowel, 389, 390 TRAP vowel, 388 Zulu and Xhosa influence, 390 contact versus non‐contact, 386–387 early pidginization, 385 English in Cameroon, 388 founding of Liberia, 385–386 language phyla in Africa, 386 Afro‐Asiatic, 386 Khoe‐San, 386 Niger‐Congo, 386 Nilo‐Saharan, 386 respective roles of Swahili and Fanakalo, 386 Sub‐Saharan English, 386 comparison with L2 varieties, 387 group learning of a superstrate, 386 influence of orthography, 391 lexical features, 388 literacy rates, 391 phonology, 387 features, 387 TRAP vowel, 388 vowel systems, 388

relationship to ‘World Englishes’, 386 spelling‐sound equivalence, 391 syntax, 387, 392 anti‐deletions, 396 examples, 397 input conditions, 398 morpheme insertion, 398 left dislocation, 393–394 pronoun gender conflation, 396 resumptive pronouns, 393 stative and habitual aspect, 395–396 West African Pidgin English, 387 African languages, 807 Afro‐Asiatic, 651 Bantu languages, 650 Berber, 651 borrowing, 654 code‐switching, 650 contact in Souroudougou, 656 ethno‐linguistic groups, 656–658 labialization, 658 lexical borrowings, 658 Pana, 658–659 Pini, 658–659 relative social status of speakers, 658 size of zone, 656 social network analysis, 659 contribution to a framework for language contact, 653–656 factors in language contact, 654 Greenberg’s African languages classification, 652 Heine and Kuteva’s model of grammaticalization on contact, 654 lingua francas, 650 linguistic situation, 649

The Handbook of Language Contact, Second Edition. Edited by Raymond Hickey. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

764 Index African languages (cont’d) Luo (Eastern Sudanic, Nilo‐Saharan), 655 calquing, 655 preverbal tense markers, 655 sociolinguistics of replication, 656 Macro‐Sudan belt, 652 key cross‐linguistic features, 652 Mbugu, 656 multilingual practices, 653 multilingualism, 651 Myers‐Scotton’s code‐switching model, 653 Ngaoundéré, 659 contact‐induced variation, 659 Fulfulde, 659 Ego‐Centred Network, 660 analysis of data, 662 correlations between factors, 661 occupation and speaker variation, 660 pronominal forms and noun‐classes, 660 Niger‐Congo phylum, 650 pidgins and/or creoles, 650 repertoires and choices, 653 research at present, 651 research history, 650 role of contact, 650 sociolinguistic micro‐studies, 656 Songhay, 651 speaker‐centred approaches, 663 relevance for theory building, 663 speaking languages, 653 theories about relationships, 650 Afroasiatic languages, 571 borrowing, 571 contact in Berber, 575 Arabic influence on Siwi, 576 Maghreb Arabic and Berber languages, 575 contact in Chadic, 576 additional argument marker, 582–583 benefactive, 582 causative, 582 alternative proposition, 579 assessment, 585 clausal coordination, 578 common origin with Berber, 577 counter‐presuppositional function, 578 free grammatical morphemes, 577–578 lexical borrowings, 577 logical conclusion, 578 logophoricity, 584 disjoint reference, 584 negative markers, 579–580 nominal plurality, 581 phonology tone, vowel harmony, 577 possessive subject suffixes, 581–582

pragmatics of tense and aspect, 580–581 preverbal position of object, 584–585 serial verb constructions, 583–584 split coding of person and number, 580 verbal extensions, 581 word order, 579 contact in Cushitic and Omotic, 576 marked nominative case, 576 verb‐final syntax, 576 contact in Semitic, 574 lexical borrowings, 574 Modern Hebrew Standard Average European?, 575 SVO syntax of Hebrew (possible contact origin), 575 SVO syntax of modern languages, 575 verb‐final syntax of Amharic and Tigrinya, 574 word order, 574 Egyptian, 573–574 contact with Cushitic languages, 573 internal changes, 573 Ge’ez, 575 labialized velars, 575 verb initial, 574 linguistic environment, 572 Old Egyptian, Late Egyptian, Coptic, 573–574 reconstruction of proto‐Afroasiatic grammatical system, 572 research general implications of individual studies, 586–587 prerequisites, 572–573 spread of grammatical elements, 586 time depth and migrations, 572 American English African American English, 371–372 accommodation, 364 local linguistic ecology, 371 substrate effects, 363 Wolfram’s Detroit study, 364 contact origins, 362 speaker age, 362 determining contact‐induced change, 362 development of regional varieties, 372–377 diversity of English, 377 historical paths, 377 outcomes of borrowing/imposition, 378 roles of immigrant languages, 372 Upper Midwest, 372, 377 dialect contact rhoticity, 364 local significance, 364 spread of R‐lessness, 364 discourse patterns, 376

Index  765 final laryngeal neutralization, 374–375 occurrence in heritage languages, 374–375 imposition, 363 borrowing, 363 internal vs. external factors, 362 Jewish English, 368–369 features, 368 phonology, 368 Yiddish influence, 368 koineization, 363 complexification, 363 founder principle, 363 new dialect formation, 363 path of development, 363 Latino Englishes, 369–371 Chicano English, 369 prepositional usage, 370 Spanish loans, 369–370 substrate influence, 369 loanwords, 365–366 foreign pronunciations, 365 impact on the language, 365 Japanese, 365 transmission, written or oral, 365 Native American English, 366–368 phonology, 367 Quinault English, 367 patterns of variation, 362 questions of timing, 364 recession of contact features, 364 role of bilingualism, 366 French in Louisiana, 366 German in Pennsylvania, 366 Spanish in the Southwest, 366 social vs. structural factors, 362 Upper Midwest TH, DH‐stopping, 373 German origin, 373 language‐internal accounts, 373 verbal particle with, 375–376 work on contact, 361 local level focus, 361 settler colonial variety, 361 Australian languages, 717 Aboriginal Englishes, 723–724 internal variation, 724 Victoria, 727–728 features, 727 phonetics, 727 case studies, 726 contact in education, 725–726 translanguaging, 726 contact in north‐central Arnhem Land, 729–732 code‐switching, 731 outcomes of contact, 730

intra‐Aboriginal contact, 725 koines / new dialects, 722 Dhuwaya, 722 morphology, 722 mixed languages, 721 Gurindji Kriol, 721 outcomes of contact, 720–724 pidgins and creoles, 720 Broome Pearling Lugger Pidgin, 720 Jargon Kaurna, 720 Kriol, 720–721 Macassan, 720 New South Wales Pidgin, 720 Northern Territory Pidgin English, 720 Pidgin Ngarluma, 720 Torres Strait Creole, 721 Western Australian Pidgin English, 720 research situation, 719 restructured traditional languages, 722–723 Dyirbal, 723 Pitjantjatjara, 723 Balkan area Balkanisms, 539–541 calques, 545 exhortations, 544 interjections, 544 linguistic levels, 539 key features, 540 historical depth, 541 onomatopoeia, 545 causes of convergence, 542–545 ’enhancement via contact’, 542 discourse‐related borrowings negation, 544 intensity of contact, 543 pidginization, 542 role of conversation, 544 speaker accommodation, 542 substratum/adstratum effects, 542 clusters of convergent languages, 546 history of features, 546 convergence, lexical and structural, 538 convergence, localized vs. broadly realized, 545–546 convergent languages, 539 cross‐linguistic similarities, 539 genetic affiliations, 539 history of contact, 537 language family, 539 language groupings, 539 languages included, 537 linguistic area definition, 539 terminology, 539

766 Index Caribbean Englishes and creoles African heritage, 403 assessment, 409 Berbice Dutch, 409 cline of creoleness, 414–415 contact parameters, 407–408 creole exceptionalism, 411 creolization and decreolization, 408–413 decreolization scenarios, 412 early Bajan (Barbados), 412 early contact situation, 409 European involvement, 404 European lexifier languages, 408 founder effect, 410 historical background, 405–406 morphosyntax, 413 copula, 413 lack of morphology, 413 negation, 413 object clauses, 413 preverbal markers, 413 post‐creole continuum, 412 pragmatics, 414 sociolinguistic context, 407–408 substrate thesis, 408 Africanisms, 408–409 uniformitarian framework, 411 universalist theories, 408 Celtic languages assessment of contact, 495 contact in Europe, 490 Celtic borrowings in Germanic, 490 contact in Ireland and Britain, 490–491 Old Norse borrowings, 491 contact with known languages, 490–491 contact with unknown languages, 489–490 Continental Celtic linguistic features, 493 conjugated prepositions, 494–495 object markers (infixing/suffixing), 495 polypersonal verbs, 495 relative verb forms, 495 VSO syntax, 494–495 historical background, 489–490 prehistory of Insular Celtic, 491 morphosyntax, 491 phonology, 491 substratum theories, 492 Afro‐Asiatic, 491 critics of theory, 492 suggested etymologies, 492 work by Theo Vennemann, 492 Basque, 492 Early Indo‐European areal diffusion, 304 criticism of unified proto‐language, 317

early contact in Europe, 304 areal considerations, 304 assessment, 318 Baltic and Slavic, 306 Celtic and Germanic lexical correspondences, 306 Greek and Latin, 313 bilingualism, 313–314 calques on Greek models, 315 contact outside of Rome, 314 mutual influence, 316–317 prestige of Greek, 314–315 role of religion, 315–316 Italian Peninsula, 309–313 Etruscan, 309 Latin and other Indo‐European languages, 309 Latin and Sabellian, 309 perfect structures, 312 Oscan, 311–312 Umbrian, 310–311 Italic and Celtic hypothesis, 305 Italic and Germanic, 305–306 North‐West Europe Celtic poetry, 307 stress shift, 307 possible Italic parallels, 308 Verner’s Law, 307 Wave Model (Schmidt), 304 Western Indo‐European languages, 304–306 role of contact, 303 Finno‐Ugric languages assessment of contact, 528 native versus borrowed elements, 528 contact and globalization, 529 Finnish and Estonian non‐native use, 530 immigrant languages in nation‐states, 530 minority languages, 529 multilingualism, 530 contact with majority language, 520 different branches, 519 earliest loanword strata, 525 phonological reconstruction, 526 Eastern Finnic peoples, 521 effects of nationalism, 522 Estonia contact with German, 521–522 Fennisms in Swedish, 524 Finland Finnish dialects, 521 position of Swedish, 522 Finnic and Saami languages, 521 Finnish influence on other languages, 654 Finnish Romani, 654 intra‐Finnic‐Ugric contact, 523

Index  767 Karelian, 521 minorities in Finno‐Ugric nation‐states, 521 multiple causation, 527–529 nation‐state languages, 520 outcomes of contact, 523 morphological influence, 524 phonological influence, 524 past contact German in the Eastern Baltic, 524 Russian influence on Finnish and Estonian, 522 situation in Russia, 521 size of group, 519 Swedish/Low German borrowings, 521 Uralic languages endangered status, 520 ‘Western’ languages, 521–523 Germanic languages breakup of Proto‐Germanic, 329 competition with Latin, 330 early uniformity, 329 first attestations of Germanic, 331 North, West and South Germanic affinities, 330 relationship to Roman culture, 330 contact varieties, 336 Afrikaans, 337 German in Namibia, 337 mixed languages, 337 multi‐ethnolects, 336 Pennsylvania German, 338 pidgins and creoles, 336 Dutch creoles in the Caribbean, 337 German in Eastern Europe, 338 Russenorsk, 336 recent contact due to immigration, 336 Yiddish, 338 Weinreich’s model, 338 contact with Finno‐Ugric, 326 early loans, 326 Germanic as a result of contact with Balto‐ Finnic, 327 Germanic influence on Finnic and Sámi, 326 North‐West Europe as a linguistic area, 327 contact with French in England, 333 early fragmentation, 323 formation of Proto‐Germanic, 324 founder population, 324 Germanic outside Europe, 334 colonial expansion, 334 Dutch in the American Midwest, 335 Dutch overseas colonies, 335 German in North and South America, 335 new varieties of English, 335 Norwegian and Swedish in the American Upper Midwest and Canada, 335 Germanic‐Brythonic contact, 333

historical background, 323 Indo‐European movement into Northern Europe, 324 Latin and Christianity, 334 migration period (200‐600 CE), 331 Alamanni, 332 appearance in the Baltic, 331 Baiovarii (Bavarians), 332 Gothic contact with Greek, 332 Langobards, 332 Mediterranean Germanic, 331 movement of the Vandals, 332 North Sea area, 332 Non‐Indo‐European influence, 325 loss of verbal categories, 326 nature of archaic vocabulary, 325 toponyms, 326 unassigned etymons, 325 Vennemann’s views, 326 Norse power (750‐1050), 333 contact in north/east of England, 333 Norse in Scotland/Ireland, 333 Northwest Indo‐European spread, 328 contact between Germanic, Celtic and Italic, 328 Germanic‐Celtic contact, 328 scholarship positions, 328 Wanderwörter, 329 Proto‐Indo‐European and DNA patterns, 324 Proto‐Indo‐European and shared vocabulary, 324 Proto‐Indo‐European and Yamnaya culture, 324 role of Hanseatic League, 334 scholarship of early stages, 323 the Kalmar Union, 334 history of English complexity of borrowing, 349 Indian loanwords, 349 culture and cuisine, 350 contact with immigrant languages, 346 Chinese, 346 Dutch, 346 French, 346 French_Huguenots, 346 Low German, 346 Old Norse, 346 contact‐induced change, 345 cultural influence and borrowing, 348 Ancient Greek, 349 Arabic loanwords, 348 dialect contact, 352 Cockney and Australian/New Zealand English, 353 Glasgow Scots, 352 Scots and English, 352

768 Index history of English (cont’d) Dictionary of the Scots Language, 345 emigration and spread of speakers, 347 lexical borrowings, 347 New Zealand, 347 North America, 347 South Africa, 348 French and Italian influence, 350 French influence nature of borrowings, 351 cuisine, 351 music, 351 Italian influence nature of borrowings, 351 lexical borrowing, 345 Oxford English Dictionary, 345 transition period (850‐1300), 353 change in typology, 353 possible causes, 354 role of Celtic, 356 role of Norse, 355, 357 simplification/complexification, 356 London English, 353 language change agents of change, 37 analogy, 94 causal factors, 36 complexification, 44 contact‐induced re‐interpretation, 102 stable vs. obsolescent languages, 231 Francoprovençal, 232 palatalization, 232 language death, 231 non‐threatened varieties, 232 drift, 43 external causes, 35 grammaticalization, 93 acceleration, 103 areal diffusion, 105 borrowing/replication, 93–94 change in typological profile, 105 Sri Lankan Portuguese, 105 Tamil, 106 conceptual strategy, 96–97 contact‐induced, 93 constraints, 101, 106 degrees of grammaticalization, 103 directionality, 106 L2>L1, 107 ordinary vs. replica, 96 sociolinguistic factors, 107 wh‐relativization, 97 discourse pragmatics, 95 grammaticalization area, 104–105

language internal, 94 Meillet, 94 paths, 102 polysemy copying, 99 calquing/loan translation, 99 predictions, 100 propelling force, 102 replication, 95, 107–108 speaker innovation, 95 spontaneous replication, 95 use of articles, 99 developmental stages, 100 stages, 101–102 use patterns, 96 Basque, 103 Bislama, 96 Molisean and Italian, 97, 100–101 Ottoman Turkish, 98 Pipil (El Salvador), 103 Turkish/Balkan Turkish, 98 Upper Sorbian, 100–101 Vetmbao, 96 inflectional simplification, 43 innovations, 35 learnability of structures, 35 metaphor and metonymy, 102 metatypy, 114 polysemy copying, 93 predictable vs. unpredictable effects, 230–231 Communication Accommodation Theory, 230 feature pools, 231 koineization, 230 Milton Keynes, 230 predictors, 38 Proto‐Indo‐European, 46 restructuring, 93 sources, 34 spread of innovation, 38 language contact accommodation, 242 dissociation, 242 social context, 242 adult second‐language learning, 44 areas Pacific Northwest Sprachbund, 36, 38 Vaupés region (Colombia), 37 area‐wide features, 36 attributing phenomena to contact, 17 authors Weinreich, 93 bilingualism and diglossia assessment, 63 code‐switching, 56–59 regularities, 58

Index  769 definitions, 52–53 degree of codification, 62 domains of use, 62 extended diglossia, 53–54 Ferguson’s view, 52, 53 functional compartmentalization, 55–56 genetic affiliation of varieties, 54–55 grammatical convergence, 61 Haitian Creole, 52 L and H varieties, 52, 53, 54 language genesis, 62–63 language maintenance, 62–63 language shift, 62–63 lexical insertions, 59–60 Latin and Spanish, 59 lexical realignment, 60 Maltese, 60, 61, 62 Modern Greek, 52 recipient and source language agentivity (Van Coetsem), 54 relative prestige of L variety, 53 research background, 52–53 societal bilingualism, 53–54 sound transfer, 52 Spanish in the United States, 55 Standard Arabic and French, 55, 59 Standard Arabic and vernacular varieties, 51 structural influence, 60–61 Swiss German, 52–53 Tunisian Arabic, 54, 56, 57 varieties of the same historical language, 54 borrowability, 41, 177–178 borrowing, 15, 93 basic meanings, 177 code‐switching, 174–175 definition of borrowing, 175 German lexicography, 175–176 degrees of bilingualism, 174 imposition, 20, 176–177 Norse elements in English, 215 lexical, 169 derivational blends, 171 etymons, 173 Japanese, 169 loan blends, 172 Haugen’s classification, 171 motivation, 172–173 role of prestige, 172–173 polygenesis, 174 French and English in the Middle Ages, 174 Zulu and South African English, 174 questions of agentivity, 169 non‐lexical, 170–171 affixes, 170 Middle English and French, 170

phonological assimilation, 176 semantic, 171–172 Old English and Latin, 171 Yiddish, 171 borrowing to resolve ambiguity, 15 category and exponence, 12 child bilingualism, 44–46 child first language acquisition, 67–85 Al‐Sayyid Bedouin Sign Language, 73 assessment, 85 Ban Khor Sign Language, 73 Bickerton’s views criticism, 70 Language Bioprogram Hypothesis, 70 changes in progress, 81–83 Caribbean scenarios, 82 Kam Muang and Standard Thai, 81 Papua New Guinea, 82 Rapa Nui and Chilean Spanish, 82 Turkish‐German bilinguals, 81–82 changes instigated by children, 83–85 influence on adults, 84–85 phonology, 83 children in contact situations, 69–70 creole genesis, 70 differing inputs, 68 emergence of koines, 67 Gurindji Kriol, 81 Hawai‘ian Creole English, 70 incremental changes, 71 Indian Sign Language, 73 Israeli Sign Language, 72 koines Dhuwaya, 77 Høyanger, 77 Milton Keynes, 76 Svalbard, 76 mixed languages, 68, 73 Gurindji Kriol, 75 Javindo, 75 Kriol, 73 Light Warlpiri, 73–75 new dialect formation, 76–81 Petjo, 75 Tiwi, 75 multiethnolects, 78 Copenhagen multietnolect, 78, 80 Gårdstenish, 78 Kiezdeutsch, 78, 79, 81 lexical borrowings, 79 Multicultural London English, 79, 80 Multicultural Paris French, 79, 80 Rinkeby Swedish, 78, 79, 80 Rosengård Swedish, 78, 79 Stadtteilsprache, 78 Straattaal, 79, 80

770 Index language contact (cont’d) new language creation, 69–75 Nicaraguan Sign Language, 72 Nigerian Pidgin English, 72 regularizing patterns, 71 sets of choices, 71 signed contact languages, 72–73 Tok Pisin, 71 code‐switching, 181–183 additional borrowing, 187–188 assessment, 194 diachronic perspective, 183–185 English–French–Latin, 182 French and Brussels Dutch, 192 French–Latin, 182 function as discourse strategy, 192 German–Latin, 182 Greek and Turkish (pre‐1922), 184 in Facebook exchanges, 182 Jersey French, 188 language change and language shift, 183–190 literary examples, 182 London Jamaican, 191 Matrix Language Frame critical assessment, 189 Matrix Language Frame (Myers‐Scotton), 189 mechanism of change, 188–189 methodological issues, 183 Middle English and Norman French, 184 pidginization and creolization, 191–192 pragmatic factors, 183 Quebec French, 183 reasons for occurrence, 181 relationship to borrowing, 184, 185–186 grammatical categories, 186–187 Greek Cypriot‐English, 187 Japanese and English, 186 Punjabi–English, 187 structural integration, 187 types of loans, 186–188 relevance to language change, 184 role of identity, 192–194 Syrian Arabic and English, 192 Russian–English in advertising, 182 shift‐change interface, 185 Dyirbal, 185 situation with Tariana, 190 sociolinguistic significance, 183 Spanish and English New Mexico, 183 Spanish–English in magazine for Hispanics, 182 stable varieties and issues of identity, 191–192 studies, 183–185 translanguaging, 182 triggering during contact, 181

Warlpiri and Aboriginal English, 191 word‐order considerations, 184 code‐switching and contact, 3–4 contact and complexity, 5 contact and globalisation, 8 contact in hindsight, 11–12 contact‐induced change, 45 contact‐induced linguistic change, 114 convergence scenarios, 15–16 defining the scope, 3 dialect contact, 6 dialect levelling, 243 regional, 243 supralocalization, 243 diasporic varieties, 6 excluded situations, 45 explanations in linguistics, 47 explanatory models, 34 explanatory value, 44 face‐to‐face, 34 filling gaps in paradigms, 15 geographical spread of languages, 16–17 grammatical borrowing, 35 grammaticalization, 5, 93–95 factors involved in change, 94 minor to major use pattern, 96 process versus outcome, 95–96 types of contact‐induced linguistic transfer, 94–95 high‐and low‐contact scenarios, 44 historical explanations, 36 history of contact phenomena, 11 innovations in borrowing language, 36 integration into borrowing language, 43 intensity and duration, 37 internal and external change, 33, 35 internal vs. external linguistic change, 9 koineization new towns, 249 Høyanger, 249–251 features, 249 input varieties, 249 local identity formation, 250 role of social factors, 250–251 tension between East and West Norwegian forms, 250 Milton Keynes, 251 absence of local dialect features, 253 data collection, 251 GOAT fronting, 252 model for acquisition, 252–253 MOUTH realizations, 253 population development, 251 principles of koineization, 251–252 relationship of adults’ to children’s speech, 252 situation in Norway, 249

Index  771 language change beyond the Western world, 234–235 language death, 5 languages Asia Minor Greek and Turkish, 45 Breton and French, 38 Bulgarian and Megleno‐Rumanian, 40 Comox and Wakashan, 46 Czech and German, 46 French loans in English, 38 Lithuanian and Finnic languages, 43 Ma’a and Shambala, 42 Media Lengua, 45 Mednyj Aleut, 45 Michif, 45 Mokkī, 39 Nisgha and English, 42 Ossetic and Georgian, 41 Quechua, 39 Yiddish‐English contact, 38 language shift, 37 case of shift, 155–156 category and exponence, 153–155 English influence on Irish, 161–164 linguistic parallels, 158–159 motivation, 150–152 nature of shift, 152 non‐transferred features, 159–160 prosody of transfer, 156–158 search for categorial equivalence, 153–156 tracing features to contact, 152–153 vernacular features, 160–161 lexical borrowing, 37 hierarchies, 177–178 nonbasic vocabulary, 41 linguistic areas, 4 linguistic convergence, 114–115 Amazigh Valley Berber‐Songhay mixed languages, 114, 124 Amazonian languages, 115 assessment, 125–126 bidirectional influence, 114 Creek/Muskogee, 115 Fagauvea, borrowing from Iaai, 117 Finno‐Livonian, 114 Hainan Cham/Tsat (Austronesian), 117 Lower Mississippi Sprachbund Quapaw, 117 Mayan languages, 124 metatypy, 114 Nguni languages (South Africa) Xhosa and Khoe languages, 116 Para‐Romani varieties, 114, 123–124 phonetic/phonological, 116–117 Qinghai Gansu Sprachbund, 116 Quechua and Aymaran, 114 research, 114–115

Stedsk (Netherlands), 114 Tetum Dili (East Timor), 117 Tojo‐ab’al (Mexico), 114, 124 linguistic landscapes, 281–298 assessment, 296–298 first wave studies, 282–283 quantitative approach, 283 globalisation, 281, 282 Hong Kong languages present, 287 Lennon walls, 287, 288–291 analysis results, 291–296 Chinese University of Hong Kong, 294, 295 language choice, 294 methodology, 291 use of romanised script, 295–296 multilingualism, 281–282 research background, 281–282 second‐wave studies, 283–285 qualitative approach, 284 superdiversity, 285 third‐wave studies, 285 identity construction, 285 Tokyo, 283 signage, 283 locus of contact, 10 markedness, 42–43 metatypy Dami (Papua New Guinea), 123 Takia (Papua New Guinea), 123 Waskia (Papua New Guinea), 123 mixed languages, 5 analysing ‘mixture’, 215–217 Berbice Creole, 205 code‐switching, 215 creoles Berbice Creole, 210 creolization as a process of change, 209 Palenquero, 209 possible mixed language status, 209, 211 expanded pidgins, 204–205 historical‐comparative method, 203–205 Ma’a, 206 metatypy, 203 mixed and unmixed creoles, 208–210 Karipuna Creole French (Brazil), 208 mixed and unmixed pidgins, 206–208 Chinese Pidgin English, 207–208 Hudson Bay Eskimo Pidgin, 207 Icelandic Pidgin Basque, 207 Pidgin French, 206 narrow interpretation, 205 pidgins analytic syntax, 207 lexifier language, 208 Trio‐Ndyuka Pidgin, 208

772 Index language contact (cont’d) pidgins and creoles, 206–210 relationship to normal transmission, 204 relationship to pidgins and creoles, 202–203 research, 205–206 research background, 201 semi‐creoles, 205, 210–211 Afrikaans, 210 Kikóngo, Kituba, 210–211 socio‐historical scenarios for genesis, 215 structural types, 211 F‐R mixed languages, 214 Sri Lankan Portuguese, 214 G‐L mixed languages, 213 INFL‐rest mixed languages, 214–215 Copper Island/Mednyj Aleut, 214–215 Hubner Mischsprache, 214–215 V‐N mixed languages, 213–214 Gurindji Kriol, 213–214 Michif, 213–214 model language, 93 naturalness, 94 neglect of distinctions, 13 new dialect formation, 243 determinism, 246–248 homogenization, 248–249 koineization, 243, 249 lingua franca, 243 New Zealand English, 243–246 appeal to markedness, 245 drift as a heuristic, 245 feature variability, 245 lack of regional variation, 246 Mobile Unit Recordings, 244–245 mobility and uniformity, 245 role of gender and class, 248–249 role of society and group identity, 246 stages according to Trudgill, 246–248 supraregionalization, 248 situation in present‐day Denmark, 248–249 tabula rasa view, 244 Trudgill’s views, 243 criticism, 250–251 reassessment, 247 role of immigration, 243 social factors, 248–249 South African Bhojpuri, 243–246 differences compared to Indian Bhojpuri, 244 settler community, 244 new varieties, 241 ethnolects and multiethnolects, 254–256 Accra (Ghana), 255–256 Multicultural London English, 254–255

diphthong realizations, 254–255 relationship of L1 and L2, 255–256 genesis scenarios, 249 nonbinary categories, 14 non‐transferred elements, 13–14 obsolescence, 5 older research positions, 5–6 overview literature, 2 paradigmatic regularization, 17–18 permeability of linguistic systems, 15 pluricentric languages, 6 relative rarity, 40 replica language, 93 replication, 93 grammatical, 93 lexical, 93 research situation, 2 role of imperfect learning, 45 role of substrates and superstrates, 9 direction of influence, 9–10 scope, 53 search for categorial equivalence, 13 social factors, 39, 47 sociolinguistic perspectives, 7 sociolinguistically informed documentation, 233–234 language description and documentation, 233–234 cultural context, 234 sociolinguistic variation, 221–227 assessment, 235 Garifuna, 227 indigenous minority languages, 222 linguistic factors, 224–225 principles of variation and change, 221–222 research background, 221–222 social factors, 226–227 sound change, 221 Uniformitarian Principle, 221 source language, 35 source of change, 33 structural borrowing, 33 structural transfer, 35–36 syntactic accommodation, 42 target language, 47 terminology in contact studies, 18–19 transfer in language shift, 12 transfer of fabric, 114–115 examples, 115 transfer of pattern, 114–115 examples, 115 Trudgill’s view of contact, 46 types of contact‐induced linguistic transfer, 116 typological change, 116 typological distance, 39–40

Index  773 typologically marked features, 36 typology, 4 ‘holistic’ approaches, 129 areality, 129 assessment, 150 Atlas of Pidgin and Creole Language Structures, 130 diffusion of features, 129–130 emergence of Gurindji Kriol and Light Warlpiri, 135 Gurindji complex predicates, 136 hybridity processes, 130 internal vs. external explanations, 142–143 linguistic convergence, 129–130 loss of gender Cappadocian Greek, 140, 141 diachronic pathways, 141–142 morphological profile, 135–139 morphology, 130–131 agglutinative vs. fusional, 130 augmentation and substitution, 132–133 borrowing, 133–134 Bora, 133–134 noun class and gender markers, 133 Resígaro, 133–134 borrowing stems, 132 borrowing, limits, 134–135 concomitant borrowing, 132 direct borrowing, 132–133 integration of borrowed stems, 131–132 Khoe languages, 131 isolating vs. synthetic, 130 loss of inflection, 131 overabundance, 132 ‘paralexification’, 131–132 relative transparency, 131 Trojan Horse borrowings, 132 morphosyntactic subsystem integrity, 135 multilingualism, 130 Murrinhpatha coverbs, 138–139 research background, 129–130 semantics metaphorical extensions, 131 syntax argument structure, 131 Turkish‐Greek convergence in Cappadocia, 140–141 universals of language, 129 Warlpiri complex predicates, 136 contact with Kriol, 137–138 World Atlas of Language Structures, 130 urban environments, 7–8, 11

urban scenarios Berlin dialect, 267 creative exploitation of resources, 274 features, 268 derivations from Turkish, 270 learner SVO, 271 parallels to heritage languages, 269–270 semantic innovation, 273 features of Turkish and Kurdish speakers, 273 immigration into Berlin, 267 Kiezdeutsch, 268–271 features, 271 linguistic landscape, 272 urban markets and contact, 271–274 contact‐induced and contact‐facilitated variation and change, 266 dialect leveling and change, 262–263 historical background, 261–262 language contact in Berlin, 267–268 majority language vernaculars, 264–266 metrolingualism, 266–267 multiethnic urban youth, 263–264 situation in Africa and Europe, 263–264 multilingual mixed languages, 264–266 national languages, 261 new contact dialects, 263 present‐day contact, 262 urban contact dialects, 263–266 Cameroonian French, 264 new vernaculars of a majority language, 264–266 Old Helsinki Slang, 264 status as jargons, 266–267 status of monolingualism, 265 verb third word order in Germanic languages, 265 variationist sociolinguistics, 222 actuation problem, 223 analytical dichotomies, 227–233 internal vs. external, 227–229 natural vs. unnatural, 229–230 Sognefjord Norwegian, 229 Bislama null subjects and objects, 225 constraints problem, 222 contact‐induced change, 224 context of conquest, 226 embedding problem, 223 evaluation problem, 223 intensity of contact, 224 inter‐group social interaction, 226 internally‐and externally‐motivated change, 224 local ecology, 224

774 Index language contact (cont’d) Mandarin contact with Southern Min, 226–227 forms in Taiwan, 226–227 phonetic distinctions, 227 orderly heterogeneity, 222 theories of language change, 223 transfer effects, 223 transition problem, 222–223 varieties of English, 7 vernacular universals, 7 word‐order changes, 47 linguistic area, 125 Balkans, 125 India, 125 Sprachbund, 125 linguistic convergence language creation, 123–124 morphosyntax, 117–121 Berbice Dutch Creole, 121 Cappadocian Greek, 117–118 Chamorro, 120 Dominican Creole French, 118–119 Dominican Island Carib, 118 Eastern Ijo, 121 Garifuna, 118 Hidalgo Nawa, 119 Mexicanero, 119 Middle English and Norman French, 117 Nahuatl, 119 Pearl Lagoon Basin Miskitu, 121 Resígaro, 118 Ulağaç, 118 Uto‐Aztecan, 119 Western Caribbean Creole English, 120 semantics and pragmatics, 121‐122 Cherokee, 121 expressions of politeness, 122 Khakas (Siberia), 121 kinship terms, 121 Natchez (Mississippi), 121 Polish, 121 South Siberian Turkic, 121 Southwestern Pomo (California), 122 Yiddish, 121 Mayan languages, 613 background, 613‐615 contact areas, 615‐616 contact with non‐Mayan languages, 618 core Lowland languages, 616 dialect mixing, 616‐617 directionality of influence, 616 genetic relatedness, 619

Highland and Lowland Maya, 614 Huehuetenango Sprachbund, 615 lexical borrowing, 618 linguistic diversity, 625 loanwords from Mixe‐Zoquean, 618, 619 Mesoamerica, 617‐620 mixed languages, 617‐618 pre‐existing linguistic similarity, 621 sharing by contact, 620 similarity between languages, 619 views on language diversification, 619 North American languages, 593 Ahousaht Nuuchahnulth clitic‐s, 605 Ahousaht Nuuchanulth, 602 Athabaskan‐Eyak‐Tlingit languages, 602 borrowed suffixes, 597 Central Pomo, 598 suffixes, 599 Chimakuan families, 597 Chimakuan languages, 607 Chimariko (isolate), 607 consonant inventories, 594 contact without documented history, 594 degrees of copying, 595 Eyak pronominal subject x‐‘I’, 603 frequential copying, 596 Haida agent/patient system, 604 Heiltsuk, 601 hierarchical patterns, 607 Iroquoian languages, 594 Karuk (isolate), 608 nominative/accusative base, 608 Kawaiisu, 599 Kwak’wala, 600 case structure, 600 linguistic areas, 593 loans from European languages, 593 Navajo, 602, 603 pronominal prefixes, 602 Northwest language families, 595 person hierarchies, 608 polysynthesis, 595 Pomoan languages, 593 rarity of affixes, 597 Salishan families, 597 Salishan languages, 607 Sm’algyax, 601 structure copying, 596 clitic structures, 600‐602 core arguments, 602‐609 hierarchical systems, 605‐609 lexical suffixes, 596‐598

Index  775 manner and direction, 598‐600 morphology, 596 semantics, 602‐605 sound patterns, 595 Tlingit pronominal prefix x‐‘I’, 603 Tuscarora, 595 Uto‐Aztecan languages, 595 Wakashan languages, 596, 605 Yana languages, 607 Yurok, 608 Pacific languages, 741 Bazaar Hindustani, 744 Bislama features, 747, 748 Chinese Pidgin English, 746 creoles, 749 Bonin (Ogasawara) Islands, 752 Hawai‘i Creole, 750 features, 750 Norfolk Island, 752 Pitcairn Island, 752 Rabaul Creole German, 753 Tayo features, 753 Tayo (Patois de St‐Louis), 753 Unserdeutsch, 753 Current Pidgin Fijian, 743 early Melanesian Pidgin lack of grammatical inflections, 746 Hawaiian Maritime Pidgin, 742 Hiri Trading Languages, 745 indigenized varieties, 755 Palmerston Island English, 757 Papua New Guinea English, 756 indigenous trading pidgins, 745 Jargon Fijian, 741 Jargon Hawaiian, 744 lexification by other European languages, 753‐754 Maritime Polynesian Pidgin, 745 Melanesian Pidgin, 745 later development, 746 substrate influence, 747 Nauru Pidgin English, 749 new dialects/koines, 755 Fiji Hindi, 755 Palauan Japanese, 756 Wai, 755 New South Wales Pidgin English, 746 Ngatikese Pidgin, 749 Pacific Nautical Pidgin, 745 Papuan Pidgin English, 748 pidgins, 741

Pijin features, 747 Plantation Pidgin Fijian, 743 Police Motu, 744 pre‐pidgins, jargons, 742 Simplified Motu, 742 South Seas Jargon, 746 features, 746 stable restricted pidgins, 744 Yimas Pidgin, 745 Portuguese‐lexified creoles antiquity, 479 Cape Verdean creole, 469 changing role of lexifier language, 470 continuing contact with Portuguese, 481 diffusion and speciation, 473‐477 ancestry of West African creoles, 474 Asian creoles, 474 Batavia (Jakarta), 476 Macau, 475 features, 475 Malaccan, 475 provenance of populations, 477 Upper Guinea, 474 disappearance in some cases, 480 documentation, 472 Sri Lanka, 472 work of Hugo Schuchardt, 472 foundational role, 477 Chinese Pidgin English, 479 Papiamentu, 478 Philippines Chabacano, 478 position of Jews from Brazil, 478 Suriname, 477 Saramaccan features, 477 Saramaccan, Sranan, Matawai, 477 geographical clusters, 472 geographical spread, 473 Guinea‐Bissauan creole, 472 historical background, 469, 470 origins, 472 single versus multiple origins, 473 position in creole research, 480 range of languages involved, 470 relationship to lexifier, 480 Sri Lanka Portuguese, 480 Romance languages adstratum influences Arabic, 437 influence on Sicilian, 437 lexical borrowings, 437 scientific loans, 437

776 Index Romance languages (cont’d) English, 438 French and Italian, 437 French on English, 439 Latin, 436‐437 domain‐specific loans, 436 influence on Western European languages, 436 New World languages, 438 Romancisation, 438‐439 assessment of contact, 444 Aymara/Quechua influence on Bolivian Spanish, 442 cocoliche (Spanish‐Italian contact in Montevideo), 443 code‐switching, 442‐443 contact background, 425 contact between Germanic and Latin, 431 contact vernaculars, 442‐443 contact with Eastern Germanic languages, 431 diaglossia and advergence, 443 French lingua franca during the Middle Ages, 443 western Andalusia, 444 differentiation of Latin, 425‐426 fronterizo (Spanish‐Portuguese contact), 443 Germanic loans in French, 431 Istro‐Romanian, 441 lexicon, 427 Basque loans in Spanish, 428 Celtic loans in Spanish, 428 Gaulish influence on French, 428 Media Lengua, 443 Michif, 443 morphology, 430 infinitival complementation in Romanian, 430‐431 infinitival complementation, absence in Southern Italy, 431 phonology, 428 /f/ > /h/ in Spanish, 429 gorgia toscana, 429 nasal assimilation, 428‐429 /u/ > /y/ in Gallo‐Romance, 430 Western Romance intervocalic voicing, 429‐430 role of diglossia, 425 simplification and complexification, 439 additive borrowing, 440, 441 child bilingualism, 440, 441 role of L2 learning, 440 Romance examples, 440‐441 structured accommodation, 441 Trudgill’s model, 439

substratum influences, 427 manifestation in later Romance, 427 position of early Latin, 427 superstratum influences, 431 French, 432 phonology, 433 later Germanic, 434 Romanian, 434 ecclesiastical loans, 435 structural change, 435 Turkish loans (from Arabic/Persian), 435 syntax case, 434 demonstratives, 433 indefinite subject pronoun, 433‐434 ordering of noun and adjective, 434 verb‐second rule, 433 typology of substratum/superstratum/ adstratum, 426 continuum, 426 Siberian languages, 669 conjunctions and complementizers from Russian, 673 contact among indigenous languages, 677 immediate and remote future imperative, 680 kinship terminology, 680 language and group Ket and Samoyedic languages, 678 language pairs Buryat and Evenki, 682 Chukchi and Yupik, 678 Dolgan and Evenki, 680 Sakha and Evenki, 680 Sakha and Lamunkhin Even, 679 lexical copying, 677 person marking on converbs, 681 verbal subject agreement, 681 copying (borrowing), 672 distribution of languages, 671 history of Russian involvement, 674 the Soviet era reasons for language loss, 674 pidgins and mixed languages, 674 Chinese Pidgin Russian, 674 Copper Island Aleut, 675 Taimyr Pidgin Russian, 674 Russian influence (on indigenous languages), 672 changes in case, 672 sentence coordination, 672 subordinate clauses, 672 word order changes, 672 Russification, 670 shift from Russian to Sakha, 673 structural change, 672, 679

Index  777 Sino‐Russian changes in function, 711 Chinese and Russian, 689 clause structure comments, 709‐711 conditionals, 709 parataxis, 708‐709 discourse connectors, 705‐706 emergent grammar, 690 first generation contact speakers, 689, 712 free grammatical morphemes, 690 idiolectal usage, 699 inflectional and derivational morphology, 697 levels of language, 689 lexicon, 711‐712 linear ordering, 698‐698 locative predication, 706‐707 modalities, 700 modality of obligation, 701‐702 phonology, 692‐693 phonotactics, 692 predicates and noun phrases, 704‐705 presentative function, 707 prohibitive modality, 702 Russian competence, 690 Russian‐Chinese Pidgin, 691 simplicity vs. complexity, 712 source languages grammar, 689 prosody, 689 stress patterns, 693 tense and aspect distinctions, 702 future reference, 703 perfective‐imperfective, 702 theoretical discussion of creoles, 689 topic‐comment phenomena, 699 utterance‐level phenomena, 699‐670 verbal syntax, 694 word‐level phenomena, 692 Slavic languages early contact, 504 Greek influence on Old Church Slavonic, 505 Old Church Slavic and written Russian, 505 Old Church Slavonic and the rise of Russian, 505 Polish and Russian, 505 Finno‐Ugric contact substrate in Russian, 503 comitative constructions, 504 influence on northern dialects, 504 loss of copula and verb ‘have’, 503 partitive genitive, 504

phonology, 504 postpositive definite articles, 504 Germanic contact, 502 East Germanic (Gothic) lexical borrowings, 502 historical expansion, 502 intra‐Slavic contact, 509 Czech and Slovak, 509 Russian, Belarusian and Trasjanka, 511 effects of the Polish‐Lithuanian Commonwealth, 511 features, 512 Russian, Ukrainian and Surzhyk, 510 linguistic features, 510 Iranian contact, 502 lexical borrowings, 502 prehistoric contact, 502‐503 role of trade, 512 Sino‐Russian contact, 513 Taimyr Pidgin Russian, 513 Russian in post‐Soviet successor states, 515 Estonia, 515 Russian‐lexified pidgin, 513 Slavic diasporas, 513 heritage languages, 514 contact with English, 514 United States (Alaska), 514 United States (mainland), 514 Israeli Russian, 514 West European contact, 506 Czech and German, 506 lexical borrowings, 506 phonological effects, 508 re‐Slavicization of Czech, 507 Polabian, 508 German influence on morphosyntax, 508 Polish, 508 Latin influence during the Renaissance, 508 situation of Silesian, 509 Sorbian, Upper and Lower, 507 calques from German, 507 South African Indian English creole classification, 150 language shift signal features, 165 South American indigenous languages, 625 borrowing avoidance, 626 chain borrowing, 626 extinction, 640 general languages (lenguas generales, línguas gerais), 634 language endangerment (endangered languages), 640 language shift, 640

778 Index South American indigenous languages (cont’d) lingua francas Chiquitano, 634 Kamayurá, 634 Kubeo, 634 Lingua Geral do Sul, 634 Makurap, 634 Nheengatú, 634 Tonocoté, 635 Tukano, 635 Waiwai, 635 linguistic acculturation, 626‐627 linguistic areas, 627 ‘Fuegian’ area, 633 Aikhenvald’s ‘language region’, 629 Amazonia, 628‐629 Andean area, 631 agglutination, 632 Caquetá‐Putumayo area, 629 Colombian‐Central American area, 630 Eastern South American Linguistic Area, 628 Ecuadoran‐Colombian Subarea area, 632 Gran Chaco (Chaco linguistic area), 632 Gran Chaco area directional verbal affixes, 633 grammatical gender, 633 possessive classifiers, 633 SVO word order, 633 Guaporé‐Mamoré area, 630‐631 macro areas, 627‐628 Northern Amazonia, 629‐630 post‐specification, 630 Southern Cone area, 632‐633 Southern Guiana area, 630 Tocantins‐Mearim Interfluvium area, 631 Upper Xingu area, 631 Vaupés area, 629 diffusion within area, 629 features, 629 Venezuelan‐Antillean area, 629‐630 verb agreement, 628 linguistic exogamy, 638‐639 loanwords, 626‐627 marriage patterns, 638‐639 mixed languages, 635 Callahuaya, 636 Kokama, 636‐637 Media Lengua, 635‐636 multilingualism, 638‐639 pidgins and creoles, 637 Akuntsú‐Kanoe ‘pidgin’, 637 Carib Pidgin‐Arawak Mixed Language, 637 Ndjuka‐Trio Pidgin, 637 sociocultural political factors, 638 Wanderwörter, 626

Spanish in the Pacific, 453 Chamorro, 457 comparison with Tagalog, 457 irrealis and future markers, 457 mixtures, 457 Spanish‐origin vocabulary, 456 change in Tagalog, 456 bilingual interaction, 457 conjunctions and discourse markers, 456‐457 discourse organization, 457 creoles Portuguese‐lexified creoles, 454 sociolinguistic situation for genesis, 459–460 Spanish‐lexified creoles, 460 structural definitions, 460 creolization, 459–462 cultural authenticity, 465 domains of usage, 464 early contact, 460 gender agreement, 463 Hiligaynon and Cebuano adstrates, 459 major locations, 455 Philippines, 455 the Marianas, 456 Mexican immigrants, 455 situation of Chamorro, 456 multilingual practices, 465 orthography for creoles, 464 role of Spanish in creole preservation, 464 sociohistorical background, 455‐456 sociolinguistic position, 465 verbal system, 460 degree of restructuring, 462 perfective/imperfective, 461 tense/mood/aspect markers, 461 Zamboanga Chabacano, 459–460 inclusive‐exclusive distinction, 460 indefinite‐definite articles, 460 Turkic languages, 551 contact areas, 553 Anatolia, 556 reduced presence of non‐Turkic languages, 556 Balkans, 556 extent of Turkish influence, 557 Gagauz, 557 Caucasus, 555‐556 research on Karachay‐Balkar, 556 Central Asia, 553‐554 Lithuania, 557 Tatars, 557 Northwestern Europe, 557‐558 recent waves of immigration, 558

Index  779 Siberia, 554 influence of Samoyedic and Yeniseic, 554 Transcaucasia and Iran, 555 Volga‐Kama, 554 influence of Tatar, 554 Western Ukraine, 557 copying of free function markers, 552 interfamily contacts, 551‐552 intrafamily contacts, 551 lexical copying, 552 structural and social factors, 552 Iranization of Uzbek, 553 Russian influence, 553 Turkicization of Central Asia, 553 types of copying, 558 adaptation of loans, 564‐565 Arabic‐Persian loans, 565 Russian loans, 565 affixes, 564 Arabic‐Persian loans and Islamic culture, 559, 560 Arabic‐Persian loans in Azeri, 559

Arabic‐Persian loans in Kazakh, 559, 560 Arabic‐Persian loans in Ottoman Turkish, 559 category loss on contact, 567 Chinese loans, 561‐562 combinatorial properties, 566 word order, 566 frequential properties, 567 function words, 563‐564 Greek and Armenian loans, 562 lexicon, 558 Arabic‐Persian loans, 559‐560 Slavic loans, 560‐561 Russian impact, 560 Mongolic loans, 561 situation in South Siberia, 561 phonological copies, 565 Uralic loans, 562 verbs, 563 borrowings into Iranian languages, 563 typological convergence of Persian with Turkic languages, 553 written forms, 558