Medieval English in a Multilingual Context : Current Methodologies and Approaches [1 ed.] 9783031309465, 9783031309472

Citation preview

NEW APPROACHES TO ENGLISH HISTORICAL LINGUISTICS

Medieval English in a Multilingual Context Current Methodologies and Approaches Edited by Sara M. Pons-Sanz Louise Sylvester

New Approaches to English Historical Linguistics

Series Editors

Sara M. Pons-Sanz English, Communication and Philosophy Cardiff University Cardiff, UK Louise Sylvester School of Humanities University of Westminster London, UK

The field of historical linguistics has traditionally been made up of the theoretical study of the various levels of linguistic analysis: phonology, morphology, syntax, vocabulary and semantics. However, scholars have increasingly become aware of the significance of other methods of applied/culturally aware research which were initially introduced to examine present day English, e.g. stylistics, sociolinguistics, pragmatics, code-switching and other language contact phenomena. This has produced exciting new avenues for exploration but has inevitably led to specialization and fragmentation within the field. This series brings together work in either one or several of these areas, thus enabling a dialogue within the new conceptualization of language study and English historical linguistics. The series includes descriptive and/or theoretical work on the history of English and the way in which it has been shaped by its contact with other languages in Britain and beyond. Much of the work published in the series is engaged in redefining the discipline and its boundaries.

Sara M. Pons-Sanz  •  Louise Sylvester Editors

Medieval English in a Multilingual Context Current Methodologies and Approaches

Editors Sara M. Pons-Sanz English, Communication and Philosophy Cardiff University Cardiff, UK

Louise Sylvester School of Humanities University of Westminster London, UK

ISSN 2946-4056     ISSN 2946-4064 (electronic) New Approaches to English Historical Linguistics ISBN 978-3-031-30946-5    ISBN 978-3-031-30947-2 (eBook) https://doi.org/10.1007/978-3-031-30947-2 © The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover contributor: HeritagePics / Alamy Stock Photo This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Paper in this product is recyclable.

Acknowledgements

We are grateful to many people and institutions who made this volume possible. We would first like to express our gratitude to the UK’s Arts and Humanities Research Council for funding the network Medieval English (ca 600–1500) in a multilingual context (MEMC) between 2018 and 2020 (AH/R00756X/1). It was a joy to be able to bring all the scholars represented in this book, and many others together for enlivening conversations about medieval multilingualism. We are also grateful to the Society for the Study of Medieval Languages and Literature, which supplied additional funding that enabled us to invite postgraduate students to participate in the various events associated with the network. We would like to thank the University of Westminster, which hosted a workshop on the impact of multilingualism on the lexis of medieval English in May 2018; the University of Seville, where we discussed the impact of multilingualism on the morphosyntax of medieval English in February 2019; the University of Bristol, which hosted a workshop on textual representations of multilingualism in medieval England in July 2019; and Cardiff University, where our final network conference was held over a span of two days in January 2020. Fellow medievalists, historical linguists and the administrative staff in all these universities were unfailingly helpful and kind. We are also grateful to everyone who gave papers and joined in the discussions at those events. We would like to thank all the network v

vi Acknowledgements

members who gave public lectures; and the National Archives, which hosted a Medieval Family Day in September 2019. We are particularly grateful to Paul Dryburgh of TNA for all his input and support, and to Charles Farris of Historic Royal Palaces, who arranged and hosted the family day and ran the MEMC website, including the ‘Research in focus’ blog. Special thanks to Cathy Scott at Palgrave Macmillan, who has shepherded this volume through every stage, and to the anonymous reviewers for helpful and pertinent comments offered in a spirit of collaborative scholarship. Finally, to our patient families, who do their best to understand the appeal of this fascinating area of scholarship and, even when they don’t quite get it, support us anyway.

Contents

1 I ntroduction  1 Sara M. Pons-Sanz and Louise Sylvester Part I Research Contexts  15 2 Contact  Theory and the History of English 17 Susan Fox, Anthony Grant, and Laura Wright 3 From  Original Sources to Linguistic Analysis: Tools and Datasets for the Investigation of Multilingualism in Medieval English 49 Carola Trips and Peter A. Stokes Part II Medieval Multilingualism and Lexical Change  93 4 Contact-Induced  Lexical Effects in Medieval English 95 Richard Dance, Philip Durkin, Carole Hough, and Heather Pagan

vii

viii Contents

5 The  West Germanic Heritage of Yorkshire English123 Arjen Versloot 6 Reframing  the Interaction between Native Terms and Loanwords: Some Data from Occupational Domains in Middle English159 Louise Sylvester and Megan Tiddeman 7 Cheapside  in Wales: Multilingualism and Textiles in Medieval Welsh Poetry187 Helen Fulton 8 Caxton’s  Linguistic and Literary Multilingualism: English, French and Dutch in the History of Jason213 Ad Putter Part III Medieval Multilingualism and Morphosyntactic Change 237 9 An  Overview of Contact-Induced Morphosyntactic Changes in Early English239 George Walkden, Juhani Klemola, and Thomas Rainsford 10 Traces  of Language Contact in Nominal Morphology of Late Northumbrian and Northern Middle English279 Elżbieta Adamczyk 11 Origin  and Spread of the Personal Pronoun They: La Estorie del Evangelie, a Case Study311 Marcelle Cole and Sara M. Pons-Sanz 12 Language  Contact Effects on Verb Semantic Classes: Lability in Early English and Old French343 Luisa García García and Richard Ingham

 Contents 

ix

13 Exploring  Norn: A Historical Heritage Language of the British Isles377 Kari Kinn and George Walkden Part IV Textual Manifestations of Medieval Multilingualism 405 14 Textual  and Codicological Manifestations of Multilingual Culture in Medieval England407 Ad Putter, Joanna Kopaczyk, and Venetia Bridges 15 Adapting  Winefride in Welsh, Latin and English441 David Callander 16 Let  Each One Tell its Own Story: Language Mixing in Four Copies of Amore Langueo 467 Mareike Keller and Annina Seiler 17 The  Materiality of the Manières de langage499 Emily Reed 18 A  fterword531 Sara M. Pons-Sanz and Louise Sylvester I ndex539

Notes on Contributors

Elzbieta Adamczyk  is Junior Professor of English Linguistics in the Department of English and American Studies at Bergische Universität Wuppertal. Her research interests focus on English historical linguistics, especially historical morphology, linguistic variation, comparative Germanic linguistics, historical multilingualism and language contact in medieval Europe. She is the author of a monograph on nominal morphology of early Germanic languages (Reshaping of the Nominal Inflection in Early Northern West Germanic, 2018) and a co-author of the eROThA repository (2014–2019). Venetia Bridges  is Associate Professor of Medieval Literature at Durham University. Her research focuses on issues of reception and transmission between the twelfth and fifteenth centuries, with particular interests in the representation of different languages in the various literary cultures of Britain and France. Among her publications is a major comparative study of Latin, Old French and English Alexander material (Medieval Narratives of Alexander the Great, 2018), and several articles discussing translatio studii in journals such as Mediaeval Studies and Studies in Philology. David Callander  is a senior lecturer in the School of Welsh at Cardiff University. He works comparatively on medieval Welsh, English and Latin literatures. His first book Dissonant Neighbours: Narrative Progress in Early Welsh and English Poetry (2019) is a contrastive study of narrative xi

xii 

Notes on Contributors

in Welsh and English verse before c. 1250. His second monograph Trawsffurfio’r Seintiau (forthcoming) examines and edits the trilingual (Latin, Welsh, English) seventeenth-­ century manuscript Beinecke Library, Osborn fb229. Marcelle Cole  is Assistant Professor in English Historical Linguistics at the Department of English and the Utrecht Institute of Linguistics at Utrecht University. Her research focuses on the historical development of English, particularly on the development of verbal morphology and pronoun phenomena in the Old and Middle English dialects. She has a special interest in Old Northumbrian and northern Middle English. Richard Dance is Professor of Early English in the University of Cambridge, and a Fellow of St Catharine’s College. He is the author of a number of books and articles about Old and Middle English language and literature, with a special interest in etymology and language contact, in particular the influence of the early Scandinavian languages on English vocabulary. His recent publications include Words Derived from Old Norse in Sir Gawain and the Green Knight: An Etymological Survey (2019), and he was principal investigator of the AHRC-­funded ‘The Gersum Project: the Scandinavian Influence on English Vocabulary’ (2016–20). Philip Durkin  is Deputy Chief Editor at the Oxford English Dictionary, where he has led the dictionary’s team of specialist etymology editors since the late 1990s. Among his publications are The Oxford Guide to Etymology (2009) and Borrowed Words: A History of Loanwords in English (2014). He is editor of The Oxford Handbook of Lexicography (2016) and is currently editing a handbook of etymology in the same series. Susan Fox  is a retired senior lecturer in Modern English Linguistics at the University of Bern. She is a sociolinguist whose research interests are language variation and change (especially in urban, multicultural contexts), language and dialect contact, the impact of immigration on language change and the language of adolescents from a variationist perspective. Her research has mainly focused on the social and historical circumstances that have led to the variety of English spoken in London today. Helen Fulton is Chair and Professor of Medieval Literature at the University of Bristol. She specialises in the literature and language of

  Notes on Contributors 

xiii

medieval Wales, and her recent publications include the co-edited Cambridge History of Welsh Literature (2019) and the co-edited Charlemagne in the Norse and Celtic Worlds (2022). Luisa García García  is Associate Professor in the English Department at the University of Seville. She studied Anglo-Germanic Philology at the University of Seville, and holds a PhD in Historical-Comparative Linguistics from the University of Cologne. Her research interests include morpho-syntactic change, Germanic linguistics and linguistic typology. She’s the author of Germanische Kausativbildung (2005). Anthony Grant is Professor of Historical Linguistics and Language Contact at Edge Hill University. He is a creolist and typologist who has worked on cases of intense relexification and on models of intimate language contact in Europe and elsewhere. He is particularly interested in languages of the Americas, Romani, Austronesian languages and on what language contact theory can tell us about the history of English and less well-described languages. His most recent book is the Oxford Handbook of Language Contact (2019). Carole Hough  is Professor of Onomastics at the University of Glasgow, where she has worked in the Department of English Language and Linguistics since 1995. A historical linguist with over 350 publications, she has particular interests in Old English, name studies and semantics. Among other executive roles, she is a former president of the International Council of Onomastic Sciences, the International Society of AngloSaxonists and the Society for Name Studies in Britain and Ireland. Her most recently funded project, Recovering the Earliest English Language in Scotland: Evidence from Place-Names, produced the online Berwickshire Place-Name Resource, available at www.gla.ac.uk/reels. Richard Ingham  is Visiting Professor at the University of Westminster (London) and was previously Mercator Professor in a DFG project at the University of Mannheim (English linguistics/diachrony). He was prior to that Professor of English Linguistics at Birmingham City University, and lectured at the University of Reading. His research areas have included the first language acquisition of English syntax and language contact between French and English in medieval England. His diachronic studies focus on bilingualism, sentence structure and negation.

xiv 

Notes on Contributors

Mareike Keller  is a junior researcher in linguistics at the University of Mannheim; she is the principal investigator of ‘The Heritage Speaker Lexicon: Dynamics and Interfaces’, a project funded by the German Research Foundation which explores lexical phenomena in heritage German. She has published books and articles on historical and contemporary language contact, focusing on structural aspects of code-switching (e.g. CodeSwitching: Unifying Contemporary and Historical Perspectives, 2020). Kari Kinn  is Professor of Scandinavian Linguistics at the University of Bergen. She specialises in comparative and diachronic syntax, and also has a strong interest in language contact. Her work has focused particularly on Old and Middle Norwegian, and on Norwegian as a heritage language in North and Latin America. She holds a PhD from the University of Oslo. Before taking up her current post, she held postdoctoral positions at the University of Cambridge and the University of Oslo. Juhani Klemola is Professor emeritus at Tampere University and Member of the Finnish Academy of Science and Letters. He holds a PhD in English Language and Linguistics from the University of Essex (1996). Klemola’s research interests include language contact, dialectology and corpus linguistics. He is co-author of English and Celtic in Contact (2008) and has published around thirty international peer-reviewed articles and book chapters, and eleven co-edited volumes, most recently The Oxford Handbook of World Englishes (2017), Changing English: Global and Local Perspectives (2017) and Corpora and the Changing Society: Studies in the Evolution of English (2020). Joanna Kopaczyk  is Professor of Scots and English Philology at the University of Glasgow. She is a historical linguist interested in corpus methods, formulaic language, the history of Scots and historical multilingualism. Her most recent co-edited collections include Historical Dialectology in the Digital Age (2019) and Applications of Pattern-Driven Methods in Corpus Linguistics (2018). She has also published on The Legal Language of Scottish Burghs (2013), and co-edited Binomials in the History of English (2017) and Communities of Practice in the History of English (2013). Heather Pagan  is a senior lecturer in English Language and Linguistics at the University of Westminster and an editor at the Anglo-Norman

  Notes on Contributors 

xv

Dictionary. Her research interests include medieval multilingual glossaries, language contact and textual editing. She is currently working on an edition of Garland’s Dictionarius with Annina Seiler and Christine Wallis. Sara M. Pons-Sanz is Reader in Language and Communication at Cardiff University. She led the AHRC-funded network ‘Medieval English (ca600–1500) in a Multilingual Context’ and co-led the AHRC-funded ‘The Gersum Project: The Scandinavian Influence on English Vocabulary’ (2016–20). She is the author of The Lexical Effects of Anglo-Scandinavian Linguistic Contact on Old English (2013), and other books and articles on medieval English. Ad Putter  is Fellow of the British Academy and Professor of Medieval English Literature at the University of Bristol, where he is the Director of the Centre for Medieval Studies. His publications include An Introduction to the Gawain Poet (1996), The Cambridge Companion to the Arthurian Legend, co-edited with Elizabeth Archibald (2009), and North Sea Crossings: The Literary Heritage of Anglo-Dutch Relations, 1066–1688, co-­ authored with Sjoerd Levelt (2021). Thomas Rainsford  is a lecturer (Akademischer Rat) at the Institut für Linguistik/Romanistik at the University of Stuttgart. His field of research is historical Romance linguistics with a focus on the history of French, Occitan and Italian. He is currently one of eight PIs in the DFG-funded Research Unit ‘Structuring the Input in Language Processing, Acquisition and Change’ (SILPAC) (FOR 5157), investigating links between lexical change and morphosyntactic change in the expression of motion events in French and Italian. Emily Reed  holds a PhD from the University of Sheffield (2020). Her thesis examined the pedagogical techniques of the Manières de langage in their multilingual contexts. She has since left academia to pursue a career in instructional design. Emily is also a circus artist, and is currently working on a performance piece that blends storytelling, physical theatre and medieval bestiaries. Annina Seiler is Wissenschaftliche Mitarbeiterin at the English Department of the University of Zurich. She holds a PhD in English Linguistics and is the author of The Scripting of the Germanic Languages

xvi 

Notes on Contributors

(2014). She teaches and researches on various topics in English historical linguistics with a focus on Old and Middle English. She is interested in questions relating to genres and text types, orality and literacy, historical multilingualism and the history of linguistic thought. She is currently working on an edition of Garland’s Dictionarius with Heather Pagan and Christine Wallis. Peter A. Stokes  is directeur d’études en humanités numériques et computationnelles appliquées à l’étude de l’écrit ancien at the École Pratique des Hautes Études—Université PSL (AOROC, UMR 8546), in Paris. He has (co-)led major research projects, including an ERC Starting Grant (‘DigiPal’) and two AHRC grants (‘Exon Domesday’ and ‘Models of Authority’). He is currently co-director of eScriptorium and a cluster coordinator for Biblissima+, and sits on the Bureau of the Comité International de Paléographie Latine and the Comité of Humanistica, the Francophone Association of Digital Humanities. Louise Sylvester  is Professor of English Language at the University of Westminster. She co-edited the Bilingual Thesaurus of Everyday Life in Medieval England and the multilingual database Lexis of Cloth and Clothing in Britain c700–1450. She has published widely on the effects of contact with French on the vocabulary of Middle English. Megan Tiddeman  studied French and Italian at the University of St Andrews as an undergraduate, and holds a PhD in Historical Linguistics from Aberystwyth University. She is currently a research fellow at the University of Westminster and has worked on two projects led by Professor Louise Sylvester: ‘Technical Language and Semantic Shift in Middle English’ and ‘The Semantics of Word Borrowing in Late Medieval English’. She has also previously worked for the Anglo-Norman Dictionary at Aberystwyth University. Carola Trips  is Professor of English Linguistics at the University of Mannheim. Her main research interests are language contact, diachronic syntax and morphology, lexical semantics and linguistic theory. Her current research investigates language change from a psycholinguistic perspective. She is the author of From OV to VO in Early Middle English (2002) and Lexical Semantics and Diachronic Morphology: The Development of -hood, -dom, and -ship in the History of English (2009).

  Notes on Contributors 

xvii

Arjen Versloot  is Professor of Germanic Linguistics at the University of Amsterdam and Professor for Frisian Studies at the University of Groningen. His area of expertise covers language variation and language change in (older) Germanic languages, particularly Frisian. George Walkden is Professor of English Linguistics and General Linguistics at the University of Konstanz. He works on language change and syntax, particularly in the context of the Germanic languages, and is editor-in-chief of the Journal of Historical Syntax. He is the author of Syntactic Reconstruction and Proto-Germanic (2014) and A History of English (with Michaela Hejná; 2022), as well as numerous papers in Germanic and general historical linguistics. Laura Wright  is a professor of English Language at the University of Cambridge. She is a historical sociolinguist who has researched the origins of Standard English and the history and development of London English, including the fate of London English taken to North America and elsewhere, such as the East India Company island of St Helena, South Atlantic, and on historical code-­switching. She has recently published research on Britain’s historic house names, and Gibraltar’s historic street names. Her most recent book, The Social Life of Words: A Historical Approach (2023), outlines lexical sociolinguistics as a field of study.

Abbreviations

Languages and Varieties A-Fr. Anglo-French Dan. Danish Du. Dutch EFri. East Frisian FA North Frisian from the islands of Föhr and Amrum Fr. French Fri. Frisian HG High German Ic Modern Icelandic Ir Irish L Latin MDu. Middle Dutch ME Middle English MLE Multicultural London English MLG Middle Low German NFri. North Frisian NGmc North Germanic Norw. Norwegian ODu. Old Dutch OE Old English OFr. Old French OFri. Old Frisian xix

xx Abbreviations

OIc Old Icelandic ON Old Norse OS Old Saxon OSwed. Old Swedish PDE Present-Day English PGmc Proto-Germanic PWGmc Proto-West Germanic Sat. East Frisian from the Saterland region St.E Standard English Swed. Swedish Sy. North Frisian from the island Sylt Wang. East Frisian from the island Wangerooge WFri. West Frisian WGmc West Germanic YE Yorkshire English

Linguistic Terms Acc. Accusative Adj. Adjective Adv. Adverb Dat. Dative Def. Definite (article) Det. Determiner Dial. Dialectal Fem. Feminine Gen. Genitive Impers. Impersonal IDG Indigenous strand Intr. Intransitive Masc. Masculine NCP Negative comparative particle Nom. Nominative NP Noun phrase Pl. Plural Pron. Pronoun Poss. Possessive/possessor PSP Prepositional secondary predicates Pret. Preterite

 Abbreviations 

Psych Psychological (verbs) Refl. Reflexive Sg. Singular V Verb V2 Verb-second order

Resources AND ASD BASICS BPNR BTh CSD DMF DMLBS DOE DOEC DSL eLALME GPC HTE HTR IIIF LAEME LexP MED MNW OED OCR PLAEME PCMEP PPCME2 TEI YD

Anglo-Norman Dictionary Anglo-Saxon Dictionary Borrowing of Argument Structure in Contact Situations Berwickshire Place-Name Resource Bilingual Thesaurus of Everyday Life in Medieval England Concise Scots Dictionary Dictionnaire du Moyen Français Dictionary of Medieval Latin from British Sources Dictionary of Old English Dictionary of Old English Corpus Dictionaries of the Scots Language Electronic Version of A Linguistic Atlas of Late Mediaeval English Geiriadur Prifysgol Cymru/Dictionary of the Welsh Language Historical Thesaurus of English Handwritten Text Recognition International Image Interoperability Framework Linguistic Atlas of Early Middle English Lexis of Cloth and Clothing database Middle English Dictionary Middelnederlandssch Woordenboek, Oxford English Dictionary Optical Character Recognition A Parsed Linguistic Atlas of Early Middle English Parsed Corpus of Middle English Poetry Penn-Helsinki Parsed Corpus of Middle English Text Encoding Initiative Yorkshire Dictionary of Dialect, Tradition and Folklore

xxi

xxii Abbreviations

General Add. Additional AHRC Arts and Humanities Research Council A.k.a. also known as Cent. century Cf. conferre (‘compare’) Ch. Chapter CL Category level E.g. exempli gratia (‘for example’) Fn. Footnote Fol./fols Folio(s) Forthc. Forthcoming L./ll. Line(s) Misc. Miscellaneous MS Manuscript r (superscript) recto (‘right [side]’) sth. something s.v./s.vv. sub verbo/verbis (‘under the word(s)’) US United States (of America) v (superscript) verso (‘left [side]’) viz. videlicet (‘namely; that is to say’) vs. versus

Particular Scholars or Works Estorie Li Manières Receuil Recuyell Ru2 T&K WSCp

La estorie del evangelie Aldred’s glosses to the Lindisfarne Gospels Manières de langage Le receuil des histoires de Troies (Lefèvre) The Recuyell of the Historyes of Troye (Caxton) Owun’s glosses to the Rushworth Gospels Thomason and Kaufman Text of the West-Saxon (or Old English) Gospels in Cambridge, Corpus Christi College, MS 140

List of Figures

Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 3.5 Fig. 3.6 Fig. 3.7

Fig. 3.8 Fig. 3.9 Fig. 5.1

London, British Library, MS Cotton Tiberius A VI, 99v56 Dendrograms of PSPs in M3 and M4 (Percillier 2022: 41) 69 Types of of-objects per ME sub-period (Percillier 2022: 44) 70 Word counts by ME sub-period for PPCME2, PLAEME and PCMEP (Percillier and Trips 2020) 72 Additions from PLAEME and PCMEP, text size shown in log10 (Percillier and Trips 2020) 73 Reflexive strategies in Middle English and Early Modern English per 1000 words (Peitsara 1997: 289) 76 Normalised frequencies of ‘simple’ and ‘self ’ reflexive marking strategies in the PPCME2 only (left), and the PCMEP/PLAEME/PPCME2 combined (right) (Percillier and Trips 2020) 77 Ratios of reflexive strategy use per 1000 verbs of the same etymological group in the combined data from the PCMEP/PLAEME/PPCME2 (Percillier and Trips 2020) 78 Ratios of reflexive strategy use per 1000 verbs of the same etymological group in the combined corpora accounting for grammatical person (Percillier and Trips 2020) 79 Etymological classification of the words identified as Norse loanwords in the YD142

xxiii

xxiv 

List of Figures

Fig. 13.1 Possessors in texts from the Isles and from Norway 392 Fig. 14.1 London, British Library, MS Harley 2253, fols 52v–53r415 Fig. 14.2 London, British Library, MS Harley 2253, fol. 52v (close-up on the bottom half of Column A) 416 Fig. 14.3 London, British Library, MS Harley 2253, fol. 53r (close-up of the top half of Column A) 418 Fig. 14.4 London, British Library, MS Harley 2253, fol. 76r (bottom part) 423 Fig. 15.1 Stemma of the Lives of Winefride 443 Fig. 16.1 Bal. fol. 34v (Oxford, Balliol College, MS 149) 480 v Fig. 16.2 Cam. fol. 146 . (Cambridge, University Library, MS Kk.4.24) 481 Fig. 16.3 Magd. fol. 154r (Oxford, Magdalen College, MS lat. 93) 481 Fig. 16.4 Trin. p. 191 (Dublin, Trinity College, MS 277) 481 Fig. 17.1 Distribution of manuscripts across Manières families 508 Fig. 17.2 Manuscript dimensions (cms)  515 Fig. 17.3 Layout types in the Manières corpus 518 Fig. 17.4 Oxford, All Souls College, MS 182 (OA), fol. 317r (close-up)522 Fig. 17.5 Cambridge, University Library, MS Add. 8870 (CA), fol. 34v (close-up) 523 Fig. 17.6 Cambridge, University Library, MS Ii.6.17 (CI), fol.103r (close-up)524

List of Tables

Table 3.1 Table 3.2 Table 5.1 Table 6.1 Table 6.2 Table 6.3 Table 6.4 Table 6.5 Table 6.6 Table 6.7 Table 6.8 Table 6.9

Sample from Reserve Lookup 66 Enrichment of the PPCME2, PLAEME, PCMEP with metadata 81 Evaluation of lexical items in English with respect to the question of a potential Old Norse origin 138 Distribution of terms across the category levels in Trade170 Numbers of categories in each of the category levels in Trade170 Numbers of terms per category at each of the category levels of Trade170 Terms in the category Trade and finance arranged according to language of origin 171 Terms in the subcategory Money arranged according to language of origin 174 Terms in the category Trader/dealer in herbs and spices arranged according to language of origin 174 Terms in the category Trader/dealer in grain arranged according to language of origin 175 Terms in the subcategory .Seller of meat arranged according to language of origin 175 Terms in the subcategory .Seller of pepper arranged according to language of origin 175

xxv

xxvi 

List of Tables

Table 6.10 Table 6.11 Table 6.12 Table 9.1 Table 10.1 Table 11.1 Table 11.2 Table 11.3 Table 11.4 Table 11.5 Table 11.6 Table 11.7 Table 11.8 Table 11.9 Table 11.10 Table 11.11 Table 11.12 Table 11.13 Table 11.14 Table 11.15 Table 11.16 Table 12.1 Table 12.2 Table 12.3 Table 12.4 Table 12.5 Table 13.1 Table 13.2 Table 16.1

Proportion of all subcategories which contain one or more mixed-­language compounds in Trade176 Number of word and senses for lexical pairs, trios and quads analysed, per domain 177 Outcome types 180 The double paradigm of present indicative of the verb ‘to be’ in West Saxon and Brittonic and the simple paradigms in the other Germanic languages 249 The distribution of analogical inflections in minor classes in Late Northumbrian and West Saxon Gospels 291 Þ-type and h-type pronouns in the versions of Estorie318 MS S, ll. 988–1974 (they): Clause type 326 MS S, ll. 988–1974 (they): Syntactic context 326 MS P (they): Clause type 327 MS P (they): Syntactic context 327 MS L (them): Syntactic role 327 MS S, ll. 988–1974 (they): Referent 329 MS P (they): Referent 329 MS L (them): Referent 330 MS S, ll. 988–1974 (they): Phonological environment 331 MS P (they): Phonological environment 332 MS L (them): Phonological environment 332 MS S, ll. 988–1974 (they): Number morphology 334 MS P (they): Number morphology 334 MS S, ll. 988–1974 (they): Lexical conditioning 335 MS P (they): Lexical conditioning 335 Labile and non-labile Middle English verbs in the destroy-class in MED352 Anglo-French destroy-class verbs 353 Psych-verb types as attested in ASD and DOE356 Psych-verb types as attested in MED entries, 1150–1450 358 Psych-verb categories as attested in AND entries (twelfth to fourteenth century) 360 Investigated texts from Orkney and Shetland 382 Investigated texts from Norway 390 Overview of code-switching in Amore langueo according to linguistic level of English insertion 475

1 Introduction Sara M. Pons-Sanz and Louise Sylvester

1 Context of the Volume This volume explores a growing interest in the multilingual culture of medieval Britain and its impact on the development of English. Until quite recently, scholarship on medieval literature was almost invariably focused on texts written in English; even where language-mixing was present in the texts, it was not a focus of interest for literary scholars. Equally, as is now well known, historical linguists generally discussed the results of the contact with Norse, Latin and French, but the continued coexistence of English with other languages as evidenced in the textual record had largely been ignored in the scholarship because of a long-held S. M. Pons-Sanz (*) Cardiff University, Cardiff, Wales, UK e-mail: [email protected] L. Sylvester University of Westminster, London, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_1

1

2 

S. M. Pons-Sanz and L. Sylvester

negative attitude towards language-mixing by speakers, as shown in modern sociolinguistic studies, and by researchers, who regarded it as signalling lack of linguistic competence or random idiosyncratic usage (Schendl 2000: 78–79). The multilingual situation of later medieval Britain became a focus for scholarship in the late 1990s when David Trotter hosted an international conference with the aim of exploring ‘the languages of post-Conquest medieval Britain from a deliberately multilingual perspective’, noting in his introduction to the resulting collection of papers from the conference that this was ‘the first gathering which attempted this for that period in British history’ (2000: 2). Trotter’s interest in the multilingual textual culture of later medieval Britain was ignited by his study of non-literary texts, and it is in the administrative culture that we continue to find evidence of language-mixing and even a mixed-­ language variety of business English. This variety has been extensively explored by Laura Wright, who observed that business texts produced in Britain during the Middle English and Early Modern period commonly mix Anglo-French and English, or Latin and English to produce a deliberate, formal register, suggesting that it had systematic rules which changed over time (Wright 1992, 1995, 2003 and 2011). Early work on the effects of language contact was mostly concerned with lexical borrowing, and here we may point to the strong influence on the field of the Anglo-Norman Dictionary, particularly the work of Rothwell, its first editor, who explored the effects of Anglo-French contact with Middle English in a series of papers (Rothwell 1975/76, 1980, 1991, 2000 and 2007). More controversial was Trotter’s opinion that many of those writing in the medieval period were familiar with two or more languages without any barrier to effective communication, and moreover the authors of the many multilingual glossaries ‘often seem uncertain as to which language is which’ (2000: 3; see also 2011). Hunt (2011) similarly argues that it is futile to attempt to draw clear distinctions of language identity with regard to lexis. Indeed, terms are ported across language barriers to create a technical register as English develops in the Middle English period (see, e.g., Sylvester 2018 and below). The interest in the multilingualism of medieval Britain has also led to a much greater understanding of the processes of standardisation through examinations of the role of the multilingual context in the creation of what

1 Introduction 

3

became Standard English (see Wright 2005 and 2013; the essays collected in Wright 2020; and Sylvester Forthcoming). Later work on the multilingualism of medieval Britain has branched out to focus on grammatical analyses of switch sites and the discourse functions of the different languages in a variety of text types (see below). This volume has grown out of the substantial body of recent work on the languages of medieval Britain, including Old and Middle English, Old Norse, Norn, Scots, French, Dutch and the Celtic languages, within a variety of disciplinary frameworks, such as historical linguistics, lexicology and lexicography, modern and historical sociolinguistics, contact linguistics, onomastics and stylistics. While earlier work focused on the various languages in isolation, more recently scholarship has turned to a range of topics involving influence and contact across languages, from the borrowing of argument structures from Anglo-French into Middle English (e.g. Stein and Trips 2014; Trips and Stein 2019; Trips 2020) to the place of Welsh in medieval English culture (Meecham-Jones 2008) and many more. Researchers have also received funding for large-scale projects investigating language contact in medieval Britain, such as the Bilingual Thesaurus project, funded by a grant from the Leverhulme Trust (2009–2012), which collected vocabulary relating to seven domains of everyday life in Middle English and Anglo-French with the aim of capturing the influence of Anglo-French on the written English that emerged following the Norman Conquest; and the Gersum project, funded by a grant from the Arts and Humanities Research Council (AHRC; 2016–2019; Dance et  al. 2019), examining the impact of AngloScandinavian linguistic contact on the lexicon of Middle English and its literary uses. The effect of this research has begun to be visible in the dictionaries of the various languages in play. While the dictionaries recording Middle English, Anglo-French and Medieval Latin are designed to collect the vocabularies of those languages only, the Anglo-Norman Dictionary has included terms labelled ‘M.E.’, while Anglo-French lexis in the Middle English Dictionary has also been the subject of study (see Trotter 2012). We also find in the Middle English Dictionary and the Oxford English Dictionary citations in which the matrix language may be French or Latin. Quotations from the mixed-language documents have begun to be included in the historical dictionaries. Wright (2011) looks

4 

S. M. Pons-Sanz and L. Sylvester

closely at the treatment of mixed-language texts which have begun to play a part in forming our picture of the linguistic culture of medieval Britain. The database of the Lexis of Cloth and Clothing in Britain project is the first multilingual lexicographical resource. Funded by the Arts and Humanities Research Council (2006–2011), it collects together the vocabulary for textiles, garments, accessories and armour in use in all the languages of the British Isles c. 700–1450. Its design was greatly influenced by Trotter’s approach to the linguistic culture of medieval Britain. There are, however, few if any discussions of methodologies across the various studies. The main approaches to medieval multilingualism have been taken from studies of contemporary mixed-language or diglossic cultures and so, in the earliest studies, medieval multilingualism was framed as code-switching (e.g. Davidson 2005; Hunt 2000; Schendl 2000, 2002, 2012, 2013; see also Schendl and Wright 2011, where the term is present in the titles of all but one of the chapters). This has begun to change. A conference held in Tampere in 2014 was entitled ‘Historical Code-Switching: The Next Step’ but, by the time of the publication of the papers (Pahta et al. 2018), the name had changed to Multilingual Practices in Language History: English and Beyond. This model is still a useful and productive one for the study of medieval texts, but the field has now broadened to draw on other approaches. The AHRC-funded network out of which this volume arises was intended to offer space for scholars working within the different linguistic subdisciplines to engage in discussions about the approaches and theoretical frameworks they are using. As part of our activities, we thus had workshops focusing on lexis, morphosyntax, and textual manifestations of multilingualism in the medieval period, respectively. This volume is planned to mirror this structure, and it also includes overviews of the current state of activity across the disciplines and approaches taken to the study of medieval multilingualism. A further important feature is that all the overview chapters and the majority of the case studies are co-written so that collaboration and sharing, often across disciplinary boundaries, underpin the entire endeavour. During our workshops, we were very keen to bring together scholars working in very different fields but with similar materials who had never worked together and were in some cases almost certainly unaware of each other’s work; for instance,

1 Introduction 

5

palaeographers met syntacticians, while sociolinguists met scholars interested in medieval glosses and textual pragmatics. This has resulted in unexpected and fruitful synergies, as evidenced for example in the chapters by Carola Trips and Peter Stokes, and Mareike Keller and Annina Seiler.

2 Contents and Structure of the Volume The essays included in this volume address key topics associated with the study of the impact that multilingualism had on the development of medieval English, in relation to the linguistic features themselves, as well as the broader sociohistorical and methodological context of this research. Thus, this volume is intended to be a helpful reference tool for both those generally interested in the topic as well as experts looking for more in-­ depth information about specific examples of the multifaceted effects of language contact on English and other languages it coexisted with in Britain. We hope that the two chapters that directly follow this introduction and the overviews that lead up to chapters on the transfer of lexis and morphosyntactic features, and the textual and material manifestations of multilingualism will be particularly helpful for non-specialists because of their scope and level of detail. The research-led chapters that follow the overviews offer detailed case studies of the opportunities that the careful and nuanced application of a wide range of approaches opens up for the understanding of the multilingual context of medieval Britain. Most of the chapters focus on the effects that the interaction between medieval English and other languages had on the former, while studies on medieval Welsh and Norn help us to place this interaction in a broader sociolinguistic and methodological context. It is precisely the volume’s broader methodological and theoretical background that the opening essay by Susan Fox, Anthony Grant and Laura Wright engages with. They first discuss key theoretical approaches to language contact, including the works by Weinreich, Hancock, Thomason and Kaufman, van Coetsem, Mufwene and Muysken. Trudgill’s views on the study of language and interdialect contact, particularly his New Dialect Formation schema, opens up the methodological dialogue to the benefits of bringing in advances in modern

6 

S. M. Pons-Sanz and L. Sylvester

sociolinguistics to the analysis of historical multilingualism. As a case study, the authors discuss how work on Multicultural London English can provide important insights into the interpretation of the role of medieval mixed-language writing in the development of Standard English. In their assessment of the tools that Digital Humanities can offer historical linguists in general and those interested in medieval multilingualism in particular, Trips and Stokes guide the reader through the different stages in a project of linguistic enquiry, from getting access to the manuscripts where the texts are preserved to manipulating the tagged textual data gathered from a number of corpora. They exemplify the advantages of these tools through the discussion of two articles focusing on various ways in which linguistic material from Old French shaped the argument structure of various constructions in Middle English (associated with, amongst others, the expression of reflexivity, secondary predicates and genitive / of-objects). Given the close connection of the enquiry into argument structure with the analysis of lexical and morphosyntactic transfer, this chapter leads the way to the overarching discussion of the various areas of research associated with the study of the lexical effects of medieval language contact in Britain. In their overview chapter, Richard Dance, Philip Durkin, Carole Hough and Heather Pagan discuss a number of approaches in the lexicological and lexicographic handling of lexical material, including the challenges involved in the etymological identification of loans, particularly when the donor and recipient languages are closely related; the help that onomastics can provide in this respect; and the difficulties that dictionaries face in establishing the dates of first attestation and, more generally, in arriving at inclusion criteria, as it is not always obvious which language a term should be associated with. Their discussion of the various factors that one needs to take into account when making etymological decisions is the broad background for Arjen Versloot’s reanalysis of the significance that one could (and, in his view, should) attribute to the existence of cognates in other West Germanic languages, particularly (Old) Frisian, one of English’s closest relatives, when trying to determine whether a term attested in English is

1 Introduction 

7

Norse-derived in the absence of clear phonological or morphological evidence. With Louise Sylvester and Megan Tiddeman’s chapter, the focus moves from Anglo-Scandinavian to Anglo-French linguistic contact, and from etymology to the process of integration and accommodation of loans into English. They analyse the interaction between native and borrowed terms in relation to the semantic hierarchy of a number of semantic domains, including terms referring to trade as a specific case study, with the help of the Historical Thesaurus of English (Kay et  al. 2020), the Bilingual Thesaurus of Everyday Life in Medieval England (Ingham and Sylvester 2017–) and additional datasets. Their work shows that the traditional understanding of the semantic relationship between existing terms and newcomers as one of competition needs to be reconsidered, as the development of specific technolects, among other variables, created the right conditions for the harmonious coexistence of terms of different origins at different levels of the semantic hierarchy, at times with concomitant semantic shift (Ingham et al. 2019). Helen Fulton offers a companion piece in which she explores the stylistic and sociolinguistic implications of the use of French-derived terms referring to textiles and luxury fabrics by medieval Welsh poets, following the rapid growth of urban trading activities. Her discussion provides a point of linguistic (Welsh instead of English) and methodological (broadly philological rather than semantic) comparison for Sylvester and Tiddeman’s piece, and thus further emphasises Fox, Grant and Wright’s argument that, in order to gain an in-depth understanding of the impact of language contact on the lexis of medieval English, one needs to place this investigation into a wider context of similar situations of language contact, besides adopting different yet complementary methodological approaches. In his exploration of the language of William Caxton’s translations (particularly his History of Jason), Ad Putter continues Fulton’s close scrutiny of textual evidence and individual choices, but shifts the focus from French to Dutch influence on medieval English. He makes clear that this is an area of research in need of the same systematic scholarly attention that the impact of Old Norse and Old French has received in recent times.

8 

S. M. Pons-Sanz and L. Sylvester

The overview chapter by George Walkden, Juhani Klemola and Thomas Rainsford channels the discussion towards the significance of language contact for our understanding of a number of morphosyntactic features of medieval English. After clarifying some terminological and theoretical points (e.g. the distinction between borrowing and imposition as channels of linguistic transfer, and between complexification and simplification as outcomes of language contact), the authors structure their overview around the contact that English established with Celtic languages, Old Norse and Old French. The sociolinguistic and sociohistorical context of each interaction is established before they review the current scholarly views on the extent to which language external factors can account for the double paradigm of the verb to be in Old English (beon vs. wesan) and the use of comparative nor for Anglo-Celtic contact; the use of th- as opposed to h-forms for the third person plural pronoun, the presence of -s as opposed to -th as the marker for third-person present indicative verbal forms, the increased syncretism of the nominal inflectional endings in the transition from Old to Middle English and verb-­ second word order for Anglo-Scandinavian contact; and various topics associated with word order and argument structure for Anglo-French contact. The four chapters that follow offer case studies on some of the specific areas of possible morphosyntactic influence outlined by the overview. ElŻbieta Adamczyk explores the extent to which Anglo-Scandinavian linguistic contact can be said to have induced or accelerated the changes in nominal inflectional morphology associated with case and grammatical gender markers that we see in tenth-century Old Northumbrian texts but are more frequently associated with later documents. Taking the distinction between complexification and simplification as her starting point, she concludes that, while the linguistic patterns in Old Northumbrian suggest contact-induced simplification, those in northern Middle English texts point instead towards complexification triggered by long-contact bilingualism. The following two chapters deal with different aspects of the interface between lexical and morphosyntactic transfer. Marcelle Cole and Sara Pons-Sanz examine the th- and h-forms for the third-person plural pronouns in the seven manuscripts of the early Middle English text La estorie

1 Introduction 

9

del evangelie in relation to textual attestation (in connection with each witness, as well as their manuscript and dialectal context), and syntactic, phonological and stylistic patterns. They conclude that, contrary to well-­ established views, the th-forms cannot be solely identified as Norse-­ derived but should rather be interpreted as the result of polygenesis, with the Old English demonstrative pronouns also playing a role in their use. Luisa García García and Richard Ingham focus instead on the influence that Anglo-French language contact could have had on argument structure, particularly in relation to lability. While previous studies have suggested that the spread of lability amongst native verbs during the Middle English period (e.g. those referring to change of state and change of location) could be attributed to the existence of a high proportion of labile verbs borrowed from French, García García and Ingham show once again that one needs to take a nuanced approach to the data, because this conclusion does not apply to destroy- and psych verbs. In these cases, French-derived verbs do not exhibit higher rates of lability than native ones. The difference in the impact of language contact—they argue—depends on whether there already existed in English a significant number of native verbs of the relevant semantic quality to act as ‘bridgeheads’ for this syntactic pattern. As with Fulton’s contribution, the last case study associated with morphosyntactic exploration offers a broader context for the outcomes of the contacts between medieval English and other languages. Kari Kinn and George Walkden investigate whether the interaction between Norn, the variety of Old Norse spoken in the northern isles of Orkney and Shetland until the eighteenth century, and Older Scots, a language whose origin lies in northern Old English, had an impact on the syntactic structure of possessive constructions in Norn. The significance of the paper lies not only in the fact that they show that language contact might be responsible for accelerating the tendency to place Norn possessors in prenominal position but also in that, by treating Norn as a heritage language, they further demonstrate the benefits of applying concepts from modern linguistic enquiry to the study of medieval languages / varieties. The last four chapters, on the textual and codicological representations of medieval multilingualism, remind us that medieval texts and their material contexts are key tools for the analysis of the sociohistorical and cultural milieux where language contact took place, as well as the

10 

S. M. Pons-Sanz and L. Sylvester

understanding of the complexities involved in establishing speakers’ perceptions and uses of the various languages at their disposal. In the overview that opens this section, Ad Putter, Joanna Kopaczyk and Venetia Bridges take the multilingual miscellany (or Sammelhandschrift) London, British Library, MS Harley 2253 as their starting point to outline the different levels at which one can—and should—investigate the coexistence and interaction of different languages in medieval texts (for instance, in terms of language mixing, textual pragmatics and manuscript materiality, personal and generic choices, and the relationship between a text and its source(s)). Relying on the varied linguistic nature of the texts included in this manuscript, often cast aside in the past, they also problematise common views amongst scholars of medieval England regarding the use of English, Anglo-French and Latin as three symbiotic and harmonious means to represent monocultural ‘Englishness’. The case studies in the section zoom in on specific linguistic and sociocultural aspects surrounding the multilingual character of textual and cultural interactions. Like Putter, David Callander scrutinises the opportunities for linguistic transfer and creativity offered by translation in a careful study of the Middle English and Middle Welsh adaptations of the Latin Life of Winefride/Gwenfrewy by Robert of Shrewsbury. With this contribution, he broadens the scope for the analysis of Welsh-English linguistic contact beyond the common scholarly interest in the Welsh March. Keller and Seiler take the bilingual sermon Amore langueo as the basis of a study of the interaction between Latin and English. Their comparison of the sermon’s four medieval manuscripts reveals that code-­switching from Latin to English was not motivated by lexical need, but instead responded to discourse-pragmatic and rhetorical functions associated with the sermon’s message. Interestingly, they also note that the frequency and presentation of the vernacular material are in keeping with the purposes of the manuscripts where the text is recorded. Manuscript materiality is the central topic of the final chapter in the volume. Emily Reed analyses the ten extant manuscripts of the Manières de langage, a group of dialogues composed to teach French in fifteenth-­ century England, so as to determine what manuscript size and various textual pragmatic features can tell us about late medieval language-­ learning processes and the changing sociolinguistic context in which it happened.

1 Introduction 

11

References Anglo-Norman Dictionary. 1977–1992. Ed. by Louise W. Stone, T. B. W. Reid and William Rothwell. Rev. edn, A-E. 2005. Ed. by William Rothwell, Stewart Gregory and David Trotter. AND2 Online Edition. 2021. Ed. by David Trotter (–2015), Geert de Wilde (2015–). Aberystwyth: Aberystwyth University. https://anglo-­norman.net. Accessed 18 May 2023. Dance, Richard, Sara M. Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge, Cardiff and Sheffield. https://gersum.org. Accessed 18 May 2023. Davidson, Mary Catherine. 2005. Discourse Features of Code-Switching in Legal Reports in Late-Medieval England. In Opening Windows on Texts and Discourses of the Past, ed. Janne Skaffari, Matti Peikola, Ruth Carroll, Risto Hiltunen, and Brita Wårvik, 343–351. Amsterdam: John Benjamins. Hunt, Tony. 2000. Code-Switching in Medical Texts. In Multilingualism in Later Medieval Britain, ed. D.A. Trotter, 131–147. Cambridge: D. S. Brewer. ———. 2011. The Languages of Medieval England. In Mehrsprachigkeit im mittelalter, ed. Michael Baldzuhn and Christine Putzo, 50–68. Berlin: Walter de Gruyter. Ingham, Richard, and Louise Sylvester. 2017–. The Bilingual Thesaurus of Everyday Life in Medieval England. https://thesaurus.ac.uk/bth/. Accessed 18 May 2023. Ingham, Richard, Louise Sylvester, and Imogen Marcus. 2019. Penetration of French-Origin Lexis in Middle English Occupational Domains. In Historical Linguistics 2015: Selected papers from the 22nd International Conference on Historical Linguistics, Naples, 27–31 July 2015, ed. Michela Cennamo and Claudia Fabrizio, 460–478. Amsterdam: John Benjamins. Kay, Christian, Marc Alexander, Fraser Dallachy, Jane Roberts, Michael Samuels, and Irené Wotherspoon, eds. 2020. The Historical Thesaurus of English, 2nd edn, version 5.0. Glasgow: University of Glasgow. https://ht.ac.uk/. Accessed 18 May 2023. The Lexis of Cloth and Clothing Project. University of Manchester. http://lexisproject.arts.manchester.ac.uk/index.html. Accessed 18 May 2023. Meecham-Jones, Simon. 2008. Where was Wales: The Erasure of Wales in Medieval English Culture. In Authority and Subjugation in Writhing of Medieval Wales, ed. Ruth Kennedy and Simon Meecham-Jones, 27–55. New York: Palgrave Macmillan. Middle English Dictionary. 1952–2001. Ed. by Robert E.  Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran

12 

S. M. Pons-Sanz and L. Sylvester

and Paul Schaffner (Ann Arbor: University of Michigan Press), available at https://quod.lib.umich.edu/m/med/. Accessed 15 September 2022. Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie, and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). https://www.oed.com. Accessed 18 May 2023. Pahta, Paivi, Janne Skaffari, and Laura Wright, eds. 2018. Multilingual Practices in Language History: English and Beyond, Language Contact and Bilingualism, 15. Berlin: Mouton de Gruyter. Rothwell, William. 1975/76. The Role of French in Thirteenth-Century England. Bulletin of the John Rylands University Library 58: 445–466. ———. 1980. Lexical Borrowing in a Medieval Context. Bulletin of the John Rylands University Library of Manchester 63: 118–143. ———. 1991. The Missing Link in English Etymology: Anglo-French. Medium Aevum 60: 173–196. ———. 2000. Aspects of Lexical and Morphosyntactical Mixing in the Languages of Medieval England. In Multilingualism in Later Medieval Britain, ed. D. A. Trotter, 213–232. Cambridge: D. S. Brewer. ———. 2007. Synonymity and Semantic Variability in Medieval French and Middle English. Modern Language Review 102: 363–380. Schendl, Herbert. 2000. Linguistic Aspects of Code-Switching in Medieval English Texts. In Multilingualism in Later Medieval Britain, ed. D.A. Trotter, 77–92. Cambridge: D.S. Brewer. ———. 2002. Mixed-Language Texts as Data and Evidence in English Historical Linguistics. In Studies in the History of the English Language: A Millennial Perspective, ed. Donka Minkova and Robert Stockwell, 51–78. Berlin: Mouton de Gruyter. ———. 2012. Literacy, Multilingualism and Code-Switching in Early English Written Texts. In Language Mixing and Code-Switching in Writing: Approaches to Mixed-Language Written Discourse, ed. Mark Sebba, Shahrzad Mahootian and Carla Jonsson, 27–43. London: Routledge. ———. 2013. Code-Switching in Late Medieval Macaronic Sermons. In Multilingualism in Medieval Britain (c.1066–1520), ed. Judith F. Jefferson and Ad Putter, 153–169. Turnhout: Brepols. Schendl, Herbert, and Laura Wright, eds. 2011. Code-Switching in Early English. Berlin: Mouton de Gruyter.

1 Introduction 

13

Stein, Achim, and Carola Trips. 2014. Les phrases clivées en ancien français: Un modèle pour l’anglais? Revue de linguistique romane 78 (309): 33–55. Sylvester, Louise. 2018. A Semantic Field and Text-Type Approach to Late-­ Medieval Multilingualism. In Multilingual Practices in Language History: English and Beyond, ed. Paivi Pahta, Janne Skaffari, and Laura Wright, 77–96. Berlin: Mouton de Gruyter. ———. Forthcoming. Early Standardisation. In New Cambridge History of the English Language, I: Context, Contact and Development, ed. Laura Wright. Cambridge: Cambridge University Press. Trips, Carola. 2020. Impersonal and Reflexive Uses of Middle English Psych Verbs under Contact Influence with Old French. Linguistics Vanguard 6 (2): 1–14. Trips, Carola, and Achim Stein. 2019. Contact-Induced Changes in the Argument Structure of Middle English Verbs on the Model of Old French. Journal of Language Contact 1: 232–267. Trotter, David. 2011. Language Labels, Language Change, and Lexis. In Medieval Multilingualism: The Francophone World and Its Neighbours, ed. Christopher Kleinhenz and Keith Busby, 43–61. Turnhout: Brepols. ———. 2012. L’anglo-normand dans le Middle English Dictionary. In Ki bien voldreit raisun entendre: Mélanges en l'honneur du 70e anniversaire de Frankwalt Möhren, ed. Stephen Dörr and Thomas Städler, 323–337. Strasbourg: Éditions de Linguistique et de Philologie. Wright, Laura. 1992. Macaronic Writing in a London Archive, 1380-1480. In History of Englishes: New Methods and Interpretations in Historical Linguistics, ed. Matti Risannen, Ossi Ihalainen, Terttu Nevalainen, and Irma Taavitsainen, 762–770. Berlin: Mouton de Gruyter. ———. 1995. A Hypothesis on the Structure of Macaronic Business Writing. In Medieval Dialectology, ed.  Jacek Fisiak, 309–321. Berlin: Mouton de Gruyter. ———. 2003. Models of Language Mixing: Code-Switching versus Semicommunication in Medieval Latin and Middle English Accounts. In Language Contact in the History of English, ed. Dieter Kastovsky and Arthur Mettinger, 363–376. Frankfurt am Main: Peter Lang. ———. 2005. Medieval Mixed-Language Business Texts and the Rise of Standard English. In Opening Windows on Texts and Discourses of the Past, ed. Janne Skaffari, Matti Peikola, Ruth Carroll, Risto Hiltunen, and Brita Wårvik, 381–399. Amsterdam: John Benjamins.

14 

S. M. Pons-Sanz and L. Sylvester

———. 2011. On Variation in Medieval Mixed-Language Business Writing. In Code-Switching in Early English, ed. Herbert Schendl and Laura Wright, 191–218. Berlin: Mouton de Gruyter. ———. 2013. The Contact Origins of Standard English. In English as a Contact Language, ed. Daniel Schreier and Marianne Hundt, 58–74. Cambridge: Cambridge University Press. ———, ed. 2020. The Multilingual Origins of Standard English. Berlin: De Gruyter Mouton.

Part I Research Contexts

2 Contact Theory and the History of English Susan Fox, Anthony Grant, and Laura Wright

1 Introduction This overview chapter aims to shed light on how we can understand the outcomes of medieval multilingualism in a broader context and how we might best use theories of language contact and modern sociolinguistic research (variationist sociolinguistics, the study of dialect contact/levelling) to understand the general effects that multilingualism had on the development of medieval English. We start with an overview of some of S. Fox (*) University of Bern, Bern, Switzerland e-mail: [email protected] A. Grant Edge Hill University, Ormskirk, UK e-mail: [email protected] L. Wright University of Cambridge, Cambridge, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_2

17

18 

S. Fox et al.

the important theoretical approaches to language contact which have emerged since Weinreich’s (1953) seminal work on the topic. Much of this work has looked at historical evidence and therefore focuses heavily on written documents. We then move on to consider some of the contemporary approaches to language/dialect contact, which have focused on variation and change in spoken language within the sociohistorical context in which it has taken place. During the course of the chapter, we aim to show the importance of drawing on both historical and contemporary approaches, in order to broaden our knowledge of the effects of language contact.

2 Theoretical Approaches to Language Contact The literature on language contact is vast and builds on linguistic ideas which were widespread in the years before the rise of American Structuralism. Nonetheless, a number of papers or books have been of especial importance. Language contact (especially contact-induced linguistic change) is multifaceted—sociological, psychological, diachronic, structural and so on—and requires a multifaceted approach which is best gained when the theoretical insights from different studies of language contact are brought together and integrated without diluting the insights and impetus of any of them. Different approaches or standpoints—psycholinguistic, structural-­typological, historical or others—have led to the development of different theoretical models, which have different focuses. Even before the work of Weinreich (1953), we come across some ideas whose power has yet to be exhausted. The tripartite division in Bloomfield (1933) between the three kinds of borrowing, namely cultural, intimate and dialect borrowing, is a case in point. All three are relevant to the history of English, although until recently most language contact studies focused on the issue of cultural borrowing, while less attention was paid to those kinds of borrowing which involve the replacement of labels for pre-existing concepts (which would fall under intimate borrowing) and

2  Contact Theory and the History of English 

19

borrowing between dialects, which would include koineisation as a major manifestation of dialect borrowing. Outcomes are important, but they must not be confused with the processes which give rise to them. Most kinds of contact-induced change can be abundantly exemplified from the history of English, especially if one is prepared to allow a little laxity of interpretation, and what is more, most of them can be illustrated from our knowledge of the history of English as it was before the Early Modern period. We may add that much of the terminology is vintage and not always appropriate or well chosen.

2.1 Weinreich (1953) Weinreich’s 1953 book Languages in Contact: Findings and Problems was an instant locus classicus. Weinreich studied under André Martinet (who wrote a preface to Weinreich’s book) and learned much from his own father, Max Weinreich, who had proposed Yiddish as being a fusion language. Weinreich (1953) draws on numerous sources, notably his experiences of working on, in and with Yiddish, and his doctoral work conducted in a Romansh / Swiss German area of eastern Switzerland. Weinreich saw the bilingual speaker as being the engine and locus of language contact, and conceived his monograph (filtered as it is through the viewpoint of American structuralism) as being a contribution to the literature on bilingualism. It is probable therefore that many of the cases which Weinreich cites are instances from idiolects rather than being typical of behaviour in a given speech community. Weinreich describes contact-­induced change in terms of ‘(phonic, structural etc.) interference’, often exerted upon the second language of a bilingual by the impact of their first language. Language shift, whereby a speaker of one language adopts another language as their main or only language while jettisoning the first language, is also discussed, though not to the extent that interference is. Weinreich distinguishes between substratum (the language(s) influencing the features of a given language), superstratum (the language which replaced the substratum language) and adstratum (which accounts for further elements which a language acquires as a result of contact with other languages), which is noteworthy as many authors in the late

20 

S. Fox et al.

twentieth century used ‘superstratum’ as though it meant ‘adstratum’. The table provided on pp. 64–65 is especially useful, not least because it classifies these phenomena and, in addition to discussing the factors which promote contact-induced change, also recognises that there may be countervailing influences on each instance of interference which try to inhibit the adoption of a particular feature, and which need to be specified and recognised. The idea of the bilingual or multilingual individual as the locus of contact-induced change is also supported in Lehiste (1987), a work informed by the insights of someone who grew up in a multilingual (Estonian-German-Russian) environment.1

2.2 Language Contact Theory in the 1980s An important, and for a long-time overlooked, approach is that of van Coetsem’s two transfer types in phonology and further afield (summarised most comprehensively in van Coetsem 2000). Again, this is a psycholinguistically rooted approach centred on the stance of the bilingual speaker and the nature of the language in which they are dominant. The transfer types are borrowing and imposition. The term borrowing needs little introduction, but by imposition van Coetsem means the influence of a speaker’s psycholinguistically dominant language (which is not necessarily the first language which they acquired) upon a recipient language, which is the second language into which the speaker transfers features. Mufwene (2001) has taken an approach influenced and inspired by biology as a means of analysing the origins of creoles and indeed other languages, equating what he calls a feature pool (a set of structural or other features which some or all languages in a contact situation may exhibit) with the biological gene pool. This has been adopted by other creolists, such as Aboh (2009). Nonetheless, its adoption is not universal: Lang (2012) is a nuanced refutation of several of Mufwene’s ideas by an investigator and lexicographer of Portuguese-lexifier creoles, and illustrates the ways in which a biological reading of language change and  Weinreich (1953) was soon reviewed in Language by Haugen (1954), while Matthews (2006) examines the work from the perspective of half a century later.

1

2  Contact Theory and the History of English 

21

formation is often a poor fit as an explanation of how language contact and change eventuates. Hancock’s (1986) componentiality hypothesis, first applied to West African English pidgins and creoles, examines features in languages which arose in areas where a range of typologically and genealogically diverse languages were in use, and where what Norval Smith calls ‘younger languages’ (Grant and Smith 2019: 303) have therefore developed as a result of population intrusion and disruption. This theory has strong parallels with Mufwene’s feature pool model, which it seems to predate, and Hancock’s work on the demographic detail behind the hypothesis is admirably strong. Both Hancock’s and Mufwene’s hypotheses have resulted from their separate analyses of the genesis of creole languages. These considerations are important in our understanding of the development of Standard English and in the patterns of formation of new dialects. There never was a time when the full body of one-time speakers of English had all ceased transmitting the language to their children, so that the idea of ‘abrupt transmission’ does not apply to the history of Standard English. The work by Thomason and Kaufman (1988; hereafter T & K) is the foundation for much, if not most, language contact research. T & K approach the issue from solid backgrounds in historical linguistics and with experience of first-hand research in a wide range of languages; the genealogical range of their examples is wide. They distinguish language shift, which they see as the source of structural influences (morphosyntactic as well as phonological ones), from borrowing, which is a mechanism they understand as bringing about changes in the lexicon of a language (not least by the addition thereto of new lexemes). Shift from other languages to English has played a large part in the history of English. We cannot be certain that any speakers of British Latin remained alive in order to be able to shift to the continental Germanic languages brought to Britain in the fifth century. But speakers of other languages came to England, and their descendants presumably acquired English at various times before 1500. We can be sure that populations whose L1s were British Celtic, Norse varieties, Anglo-French, Dutch/Flemish/Low German (the Netherlandic continuum), and very probably Goidelic Celtic (and maybe Breton) shifted to contemporary varieties of English.

22 

S. Fox et al.

T & K also distinguish between abrupt transmission of a language system (which is how they typify the rise of creoles) and uninterrupted or normal transmission, as English, Russian, and so on are assumed to have experienced, as a language is passed with a small number of changes from one generation to the next. An important innovation of theirs is the five-­ stage borrowing scale, developed from their studies of over fifty cases of language contact, which starts with 1 (borrowing of non-basic and then of basic vocabulary) and which moves onto 5 (in which contact-induced change occurs at many possible levels of a language’s structure, from the availability of particular allophones to micro-changes in syntactic structures, and the borrowing of various kinds of rules). The scale is cumulative, so that a situation in which one language has been influenced by another to the level of Stage 3 implies that speakers of the influenced language have already gone through Stages 1 and 2. Furthermore one imagines that the number of instances of languages which reach borrowing scale 1 is many times greater than those which reach borrowing scale 5. Eight case studies are appended, the longest being one which looks at the history of Old and Middle English, including Middle English varieties which had absorbed more features from Norse than survived in later varieties of English. Both transfer of fabric and transfer of pattern (the borrowing of morphemes or patterns, respectively, from one language to another; Grant 2002)—which are by no means mutually exclusive— occur in these cases, especially with contact involving Norse or French. Loan translations or calques, such as the use of an interrogative pronoun which to mark relative clauses, have come into English from French and Latin (it was a minor choice in Norse) and have shaped English syntax considerably through redeployment of inherited English morphs. The T & K borrowing scale is best illustrated when its categories are employed with pairs of languages; we can therefore use it to recognise the differences which occur when English meets British Celtic (scale 1, a small number of loans, although the Celtic Hypothesis, as espoused for instance by Filppula, Klemola and Paulasto 2008, surmises that Old English was profoundly influenced by a British Celtic substrate; see also Walkden et al., this volume). The outcomes of this are different from the changes which occurred when speakers of English came upon Latin (the most prolonged encounter, for which the influence is pitched at

2  Contact Theory and the History of English 

23

borrowing scale 3). We can also see that borrowing from British Celtic is not of the same nature or extent as typified borrowing from French (borrowing scale 4), or Norse (borrowing scale 3; a pattern of influence characterised by the absorption of vocabulary from Norse of which the vast majority comprises items for referents which already had labels in Old English). We also note layers of borrowing from Dutch, Flemish and Low German (borrowing scale 1, though many such loans may have had equivalents in Old English), and from Greek (also borrowing scale 1). Borrowing from Greek has usually been mediated via Latin (and thereby often via French), so that this is an instance of indirect diffusion, because it has entered English via an intermediary language, whereas borrowing from Norse, for instance, is an example of direct diffusion as no third language served as an intermediary. The fact that few loans from Norse which are preserved in modern English are cultural borrowings insofar as they name items which Old English probably had not lexicalised can be explained by the consideration that the first Norse loans which were taken into Old English frequently referred to features of life, such as parts of warships or legal practices, which have since become obsolete. Other theoretical insights can assist a study of the contact history of English. For instance, we can mention Wright’s 1976 article on semicultismos, a Romance (and especially but not necessarily solely Spanish) equivalent of the Sanskrit ardhatatsamas, that is, words which occur in post-Sanskrit Indic languages which partly but not wholly reflect the phonology of Sanskrit (if they affected this completely, they would be tatsamas) rather than the sound changes which directly inherited words (or tadbhavas) exhibit. Examples of semicultismos in English are perforce borrowings, but English circumstance (mediated through French) and the place-name Trinidad (mediated through Spanish) are instances of this; in Romance languages they number in their dozens. Back-borrowings from Old English are rare in modern varieties apart from technical terms such as scramasax, but back-borrowings from Latin in the French loan component of Middle and Modern English are considerable: for instance, regal derives from archaic French regal, which itself has been taken from L regālis ‘kingly’. However, we should also note the observation in Lass (1997: 172) that warns against adopting a contact-based explanation for linguistic changes unless the evidence for this is very strong. There are

24 

S. Fox et al.

also many instances of non-lexical influence upon English, be it the use of wh- words to introduce relative clauses on the model of French or Latin, the borrowing of French and Latin items of derivational morphology and their application to Germanic stems (wash-able, behav-iour) or the replication of politeness formulae such as good morning on French models. In some cases, such as the remodelling (with some loanwords) of the English kinship system, both lexical and semantic patterns have been taken over, while it is likely that the uptake of phrasal verbs in English owes something to the influence of Norse.

2.3 In Search of Terminological Exactitude Much of the terminology of language contact has been inherited from Neogrammarian linguistic practice but, as the complexities of the field have become clearer, many new terms have been developed (abrupt transmission being one). The work of Heath (especially Heath 1978, 1984) is particularly enlightening, rooted as it is in cases of extreme morphemic transfer among highly endangered languages in multilingual settings. There are some further important terminological items which we can harvest from the language contact literature post-1980: for instance, Clark’s (1982) distinction between ‘necessary’ and ‘unnecessary’ borrowings, with unnecessary borrowings being those words for which there is already a referent in the recipient language, and which undergo relexification by being replaced, in whole or in part, by words from other speech traditions. Another important distinction is outlined and illustrated by Ross’s (1996) discussion on metatypy, the process by which microsystems of patterns in one language are copied into another language (not necessarily with simultaneous transfer of morphemic fabric, though this is permitted). Ross has also pioneered the useful and scientifically more correct term genealogical relationship rather than the traditional genetic relationship. A third concept which originated with a linguist based in Australasia is the macro-level concept of system-changing versus system-­preserving contact-induced linguistic changes (see the discussion in Aikhenvald 2010). All this and more feeds into the scenarios approach (see, for instance, Muysken 2017). Muysken’s language contact scenarios illustrate different

2  Contact Theory and the History of English 

25

kinds of bilingualism, including literate bilingualism (which is certainly germane to the Early Modern English’s inkhorn debate or controversy; Hejná and Walkden 2022: 151–153). Muysken mentions eight scenarios. Some of these are relevant to contact-induced change in English: shift as a result of L2 learning, borrowing, grammatical convergence under prolonged bilingualism, relexification and levelling, whereas the roles of attrition, code-switching (in the more restrictive modern sense rather than in the older sense of the mixing of codes) and metatypy either are less important or (more probably) are scenarios whose overall level of influence upon English can no longer be ascertained clearly.

2.4 On Contact Effects in English: Where It Takes Us The various theoretical approaches are informed by the source material from which their ideas derive. Differing language contact scenarios have differing outcomes, and they affect different parts of a language’s structure, and each scenario affects the structure of a language to a certain extent—and yet some scenarios play a greater role than others do in the particular history of a language (including cases where the language is not affected in the slightest). Comparison of cases of heavy borrowing is an important requirement for understanding the field but it is one which is not often carried out, and we can apply this to different stages of English too. A study which Grant (2011) conducted is one such example. Inspired by the categories investigated in Smith (1987), Grant compared the presence or absence of borrowings in almost fifty categories of function words or other morphological features which are encoded by separate morphemes in many creoles but not usually so in other languages (tenseaspect-mood markers, plural noun markers and the like). Fifteen languages were surveyed; all had borrowed heavily on the Swadesh lists (see Swadesh 1955) but otherwise most had few typological characteristics in common (though three were Indo-European and five were Austronesian). English was the language surveyed which out of the fifteen showed the smallest number of features which were borrowed, namely six, all of which are free morphemes (want, [in] order [to], though,

26 

S. Fox et al.

because, they-them-their and the verb form are).2 An additional two items were borrowed from Norse into various Northern and Midland forms of Middle English—they occur in the language of the Ormulum, for instance—but they did not survive into Present-Day Standard English (the forms are oc ‘but’, sum ‘relativiser’). Had history been different and had Tostig defeated Harold Godwinson at the Battle of Stamford Bridge in 1066, Present Day English would have been a lot more Norse (and probably a lot less French and Latin) in terms of its morphemes than it is. (An examination of these categories in Late West Saxon English shows a borrowing level of no transferred features—but then again the item on the longest Late West Saxon Swadesh list which is a borrowing is sealtian ‘to dance’, from Latin, a word which did not survive Old English). We can make these observations, although there never was a unified version of Middle English. Post-Early Modern English contact-induced change is largely confined to the importation of myriads of words from Latin, French, Greek and other languages, and to the reappropriation of some borrowed forms for extended purposes. An example of this is the use of the get passive, based on a Norse loan in English, which may have first appeared in the late eighteenth century but which is barely noted until the nineteenth century. This usage can hardly be seriously attributed to the Vikings; this and thousands of others are purely English-internal developments using long-borrowed material. We can summarise a fairly conventional view of the contact history of English while employing some of the relevant terms. English is genealogically related to other West Germanic languages, and was introduced to England around the fifth century, gradually supplanting the British Celtic language of the prior population, which underwent language shift and resulted in imposition of British Celtic features upon superstratal Old English, a substratum which other West Germanic languages did not experience. Subsequent generations see the normal transmission of English, which acquires adstratal features, especially borrowings, through direct diffusion from Latin and other languages, while subsequent periods of borrowing bring about further tranches of borrowings, originally  On the Norse origin of these personal pronouns and verbal form, see, however, also Cole and Pons-Sanz (this volume), and Dance, Pons-Sanz and Schorn (2019: s.v. ar). 2

2  Contact Theory and the History of English 

27

‘necessary’ (including some, such as most of those from Greek, which come into English via indirect diffusion) but latterly also a tranche of unnecessary borrowings which indicate that some varieties of Middle English undergo partial relexification towards Norse varieties and also French, though this process does not continue to the end, as recent work by Sylvester and her collaborators shows (e.g. Sylvester et al. 2022). Most of these changes are system-preserving changes; typological readjustment resulting in system-affecting changes also occur, although these tend to increase the percentage of cases where certain kinds of such structures are employed.

3 Sociolinguistic Approaches to Language Contact and Their Relevance to Understanding the Outcomes of Medieval Multilingualism We turn now to contemporary sociolinguistic theories of dialect/language contact with a view to considering their application to medieval contexts. First, we consider Trudgill’s (2004) theory of new-dialect formation, developed to account for new dialects arising in tabula rasa contexts from contact between speakers of different dialects of the same language or mutually intelligible languages. We then turn to the multilingual and multicultural context of London, where the concept of a ‘feature pool’ (Mufwene 2001) has been used to explain innovation and language change in recent years.

3.1 Trudgill (2004) Trudgill (2004) argues that new-dialect formation consists of six key processes. The first of these is ‘mixing’, the coming together of speakers of different dialects of the same language or of mutually intelligible languages. Next is ‘levelling’, the loss of demographically minority variants. In determining who accommodates to whom, and therefore which forms

28 

S. Fox et al.

are lost, and which are retained, demographic factors involving proportions of different dialect speakers are vital. The third process is ‘unmarking’, where degrees of linguistic markedness and regularity or simplicity may be involved, such that unmarked and more regular forms may survive even if they are not majority forms. The fourth key process is ‘interdialect development’, which accounts for forms that were not actually present in any of the dialects contributing to the mixture, but which arise out of interaction between them. Such forms are of three types: (1) forms which are simpler/more regular than any forms in the original dialect mixture; (2) intermediate forms, usually phonetically intermediate between two contributing forms; and 3) forms which are the result of hyperadaption (e.g. hypercorrection). The fifth process is ‘reallocation’. Occasionally, even after levelling, more than one variant may survive from the original mixture. Where this happens, reallocation will occur such that variants originally from different regional dialects will, in the new dialect, become social class variants, stylistic variants or, in the case of phonology, allophonic variants. The final process is ‘focusing’ (Trudgill 2004: 88), the process by means of which the new variety acquires norms and stability. The first five of these processes can collectively be referred to as koineisation. Koineisation plus ‘focusing’ (process six) constitute new-­ dialect formation. Trudgill (2004) distinguishes between three different chronological stages which roughly correspond to three successive generations of speakers. Stage I refers to the rudimentary levelling and interdialect development that takes place among first-generation migrants. In this stage, speakers of different dialects accommodate towards each other in face-to-­ face interaction. When speakers of different dialects come into contact with one another, they tend to reduce the number of stereotypical features of their dialects in order to increase intelligibility. In addition, if the contact is extensive both in terms of intensity and duration, speakers will start to adopt features of the dialects they hear around them. Rudimentary levelling can therefore be said to be a conscious process. Interdialect development may also occur during this stage, where new forms which were not present in the input dialects may arise as a result of misanalysis and partial accommodation. Stage II refers to the variability and apparent levelling that can occur among second-generation migrants.

2  Contact Theory and the History of English 

29

Second-generation migrants tend to select linguistic forms from different dialects to be part of their own idiolect, which results in extreme variability. At the same time, the total number of variants available to them from listening to first-generation migrants will be somewhat reduced as the second generation will not select those forms that appear too infrequently to be noticed, resulting in the subconscious process of apparent levelling. Eventually, this leads to Stage III among third-generation migrants, who develop a focused and socially stable variety by reducing the amount of variation to usually one variant per function. This means that more forms will be levelled out, although in some cases more than one variant will become part of the new variety through the process of reallocation. During this process, the two forms which previously existed in free variation will now become either socially or even linguistically conditioned, as in the redistribution of the words shirt and skirt during the Middle English period: shirt was the English form and skirt was the Old Norse form, but both had the meaning of ‘an undergarment for the upper part of the body’. In the fifteenth century, the Old Norse form was reallocated to mean ‘the lower part of a woman’s dress or gown’ (van Hattum 2015: 111). Trudgill (2004) argues that the selection of forms in Stage III is determined by the majority principle, whereby speakers ‘simply selected, in most cases, the variants which were most common’ (Trudgill 2004: 115). These three stages present a very uniform, linear approach to the formation of a new dialect but, as Trudgill (2004) himself notes, the reality is likely to be much more complicated. In the tabula rasa context of New Zealand that Trudgill (2004) discusses at length, immigration continued over a long period of time, and there would have been much overlapping of the different stages for a considerable time until focusing was complete. Trudgill (2004: 163–164) also discusses the notion of the ‘founder principle’ (Mufwene 1996) in that one might expect the linguistic founding population of an area to exert the most influence, in terms of the survival of their speech forms, on the forming dialect. The ‘founders’ of New Zealand English, however, were not those who first arrived in 1800 but the large number of newcomers arriving from 1840 onwards who overwhelmed, both demographically and linguistically, the earlier settlers.

30 

S. Fox et al.

3.2 Sociolinguistic Work on Language Contact in London Next, we move to London where, in the last half-century, there has been significant dialect/language contact due to shifts in both in- and out-­ migration, and where the traditional vernacular variety of Cockney has been displaced by what has become known as Multicultural London English. Of course, migration is nothing new to London. Established by the Romans, London has continually attracted in-migration since its very inception. Some of the main immigrant groups to London from outside Britain have been the Huguenots (who fled persecution in France in the latter part of the seventeenth century), a large Irish community throughout the eighteenth century and a large increase in the Jewish population in London towards the end of the nineteenth century (see Fox 2015 for a fuller discussion). There is also evidence of a Dutch-speaking community from the Low Countries living in London during the medieval period (Putter 2021). No doubt all these groups influenced the London vernacular to some extent. However, in the post-war era, London underwent an intense period of both in- and out-migration (see Fox 2015 for a fuller discussion). From the 1950s, large-scale immigration to London resulted in the first ‘non-­ white’ groups of people becoming established in London. These groups included large inflows of people from former British colonies such as Jamaica, Guyana, Trinidad, India and Pakistan (including what is now Bangladesh). Since the 1990s, London has attracted smaller immigrant groups from more diverse places outside of Europe such as Somalia, Nigeria and Turkey, as well as people from the newer European Union states, such as Poland and Lithuania. London now has one of the highest proportions of foreign-born residents across cities globally (Fox and Sharma 2017), and more than half of inner-London schoolchildren are known or believed to have a first language other than English (Department for Education 2015). Unsurprisingly, the result of such changes has had linguistic consequences; young people acquire language in multilingual and multicultural neighbourhoods, and this has the potential to lead to contact-induced language change (Cheshire et al. 2011; Fox 2015).

2  Contact Theory and the History of English 

31

In the 1980s, Hewitt (1990: 191–192) reported the use of a ‘multiracial vernacular’ to refer to a ‘de-ethnicised, racially mixed language’ being used among young black-and-white speakers in London, and Fox’s (2015) analysis of speech data collected in East London in 2001 points to the continued effects of language contact and the impact of non-UK varieties on the language of London. In 2006, the label Multicultural London English (MLE) was coined by the research team working on two large-­ scale sociolinguistic studies of multi-ethnic inner-London districts in East and North London between 20043 and 2010.4

C  haracteristics of MLE Some of the main characteristic innovations reported for MLE are raised, narrow diphthongs or (near-)monophthongs for vowels that in other varieties of English are diphthongs, most notably in the face and goat lexical sets (Kerswill et al. 2008a), contributing to a more syllable-timed speech rhythm overall (Torgersen and Szakay 2012). H-dropping in stressed contexts, that is [æmǝ], for hammer, [ænd] for hand and [ed] for head, a long-standing feature particularly common in Cockney and popular London speech, has been reversed and is now hardly in evidence among MLE speakers. They also use DH-stopping, the pronunciation of the initial sound in words such as this and that as [d], and initial /θ/ in certain words (thief, thing and indefinite -thing compounds such as anything) as [t]. The most innovative consonant feature is perhaps the backing of /k/ to uvular [q] before non-high stressed back vowels (the vowels in the strut, start, lot and thought lexical sets; Cheshire et  al. Forthcoming). There has also been simplification in the indefinite and definite article allomorphy system. Many young speakers in inner London now use a [ə] and the [ðə] before both consonant-initial words and vowelinitial words ([ə] pear, [ə] apple, [ðə] pear, [ðə] apple) and in line with 3  October 2004–September 2007 Linguistic Innovators: The English of Adolescents in London (Economic and Social Research Council, ref. RES 000-23-0680); researchers: Jenny Cheshire, Sue Fox, Paul Kerswill and Eivind Torgersen. 4  October 2007–September 2010 Multicultural London English: The Emergence, Acquisition and Diffusion of a New Variety (Economic and Social Research Council, ref. RES 062-23-0814); researchers: Jenny Cheshire, Sue Fox, Paul Kerswill, Arfaan Khan and Eivind Torgersen.

32 

S. Fox et al.

many English contact varieties around the world, the speakers insert a glottal stop to avoid V#V hiatus (see Fox 2015). Furthermore, there is levelling in the past forms of be, with innovative variable levelling to was across person, number and polarity (we was going out wasn’t we? You wasn’t right, was you?), led by young people with African and African Caribbean backgrounds but also favoured by those in multi-­ ethnic friendship groups, both with and without immigrant backgrounds (Cheshire and Fox 2009). A new pronoun, man, is used mainly with first person reference (e.g. I don’t care what my girl looks like…it’s her personality man’s looking at; Cheshire 2013; Hall 2020). Innovations in vocabulary include borrowings from Jamaican English, such as the address term bruv, ends (‘neighbourhood’), whagwan (‘what’s up’) and the plural morpheme -dem, used on specific words such as mandem (‘men’). In the quotative system, a new quotative this is + speaker has emerged (Kerswill et al. 2008b: 10; Fox 2012).

The Role of Language Contact in the Emergence of MLE In tracing the origins of MLE, Kerswill and Torgersen (2021) argue that, as the first major post-war immigrant group, African Caribbeans have exerted a significant influence on MLE, with new immigrant groups having to relate their own speech to the emerging youth language norms, ‘in many cases choosing to assimilate with them’ (Kerswill and Torgersen 2021: 268). This, then, would seem to be the workings of the founder principle in post-war London. It should be emphasised, however, that not all of the innovations reported for MLE can be traced to an African Caribbean origin (e.g. the morphosyntactic innovations, such as loss of indefinite and definite article allomorphy and quotative this is + speaker). It would seem that MLE has emerged as a result of language contact, arising from social and historical changes. Children growing up in the multi-ethnic and multilingual inner-city areas will usually acquire English at an early age, but the English that they are exposed to may include different postcolonial and Creole-influenced varieties (if English is spoken in the home) as well as a range of different learner varieties that will be influenced to various degrees by the first languages of the caregivers (Fox

2  Contact Theory and the History of English 

33

and Torgersen 2018: 210; Oxbury 2021: 101–110). In cases where the caregivers do not speak English at all, the children will acquire their English from their (multi-ethnic) peer groups at school or from their older siblings. In other words, they do not have access to a stable target model of local English but, rather, they are exposed to a rich ‘feature pool’ (Mufwene 2001) of linguistic forms that are influenced by the many languages, in addition to involving features taken from different English varieties from around the world, the many learner varieties in the community, as well as the local existing forms. The forms that they acquire, through a process of unguided second language acquisitions, can be modified into new structures and innovations (Cheshire et al. 2011).

Relevance for the Study of Historical Language Contact This overview of the emergence of MLE in a relatively short time span has perhaps allowed us greater insights into how language contact processes work, together with their resulting linguistic outcomes. The processes themselves, though, are not new. As stated previously, London has always attracted in- and out- migration but perhaps at a slower rate over earlier periods of history. Nevertheless, with our knowledge of sociolinguistic language studies in London, it may be possible to consider whether the same principles can be applied to those earlier periods. In fact, the concept of the ‘feature pool’ has already been used to explain developments leading to Standard British English. For example, Moreno Olalla (2020: 142) considered spellings for the same words in four late fifteenth-­ century herbals, written by different scribes in unknown places (the late fifteenth century is when standardisation is traditionally reported as having occurred, although recent research puts it far later; see Stenroos 2020). He found that ‘[t]he feature-pool of variant spellings remained wide during the second half of the fifteenth century, but individual scribes had narrowed their choices from that pool’. Earlier, scribes had spelled words differently more or less each time they wrote them, also differing from each other. But by the late fifteenth century this was changing: ‘it can be stated that collectively, scribes used no more than five different spellings per wordtype, a single form is the rule for a substantial number

34 

S. Fox et al.

of words, and many more employed just a single alternative form (mostly a variation in the usage of -e)’. Further, Moreno Olalla noticed that words standardised from back to front: While the four scribes differed to a minor degree as to the spelling the final part of a word, they diverged dramatically when it came to spelling stems. The variation between single and double letters as length diacritics, for instance, is striking, and so is the fact that, other than function words, there are comparatively few wordtypes which all scribes spelt in the exact same way. (Moreno Olalla 2020: 160–161)

In other words, the feature pool of spellings remained wide, but scribes began to choose fewer word-final morpheme variants. Stenroos (2020) also found that spelling variation continued on in the late fifteenth-­ century administrative documents transcribed in the Corpus of Middle English Local Documents, but that the feature pool changed: The overall amount of variation does not seem to be reduced substantially: it is the actual pool of forms taking part in the variation that changes over time, as one would expect of a variable, dynamic system. […] There seems to be no substantial change towards greater homogeneity: in the early sixteenth century, the usage remains extremely variable, and as older, regionally marked forms disappear, new forms take their place in the pool of variants. (Stenroos 2020: 66–67)

The feature pool concept is thus useful for distinguishing between individual scribes’ reduced variation versus a countrywide reduction of forms.

3.3 Applying Trudgill’s Theory to Medieval Contexts Trudgill (2004) draws on linguistic data from many different sources and provides numerous examples from Colonial Englishes of the Southern Hemisphere to argue that similarities between different geographically separated regions are due to the fact that they have resulted from ‘mixtures of similar dialects coming together in similar proportions and

2  Contact Theory and the History of English 

35

developing at similar times’ (Trudgill 2004: 165). Perhaps, though, there is scope here to examine whether these same processes might have been at work in historical contexts where speakers of closely related languages / dialects came into contact with each other. The first four of Trudgill’s key processes can be illustrated by the late medieval development of Standard English, factoring in the multilingual contexts of the times.5

M  ixing In Trudgill’s schema, this is the coming together of speakers of mutually intelligible dialects. With regard to the dialects of parts of northern and eastern Britain, there has been debate in relation to intelligibility between speakers of Old English and Old Norse; that is, between the already-­ settled people of continental West Germanic origin, and the more recently immigrated people from Scandinavia who spoke North Germanic variants.6 Countrywide, from the Norman Conquest of 1066 to about 1500, there is plentiful evidence of a mix of Medieval Latin, Anglo-­ French and Middle English. Medieval Latin (the language of written record used by the new Anglo-French administration, as opposed to the Old English of the previous administration) came to include nouns and verb-stems from its spoken variant, its daughter-language Anglo-French, and also from the mother tongue of the scribes, English. This mixed-­ language variety conventionalised as a system for the purposes of keeping accounts, inventories, personal journals and other memoranda and, like all languages in constant use, it developed new features over time.7 It was, so far as we know, unnamed—presumably it was just seen at the time as the modern way of writing business Latin. The system of a Medieval Latin integrated with local languages in noun and verb root slots was also used widely on the continent for business purposes.  Stage 5, reallocation, has just been illustrated; Stage 6, focusing, happened after the medieval period. 6  See Fernández Cuesta (Forthcoming) for a summary of the positions taken as to whether speakers of Old English and Old Norse would have understood each other. See also Keller (2020). 7  See Wright (2023) for an introduction to the role of mixed-language in the standardisation of English, and Wright (2020) for a more detailed description of mixed-language. See also Timofeeva (2022) for a recent reconstruction of the early stages of this mix. 5

36 

S. Fox et al.

Levelling In Trudgill’s schema, this is the loss of minority variants. The development of Old English to Middle English to Early Modern English was characterised by loss of case-endings, a shift from a synthetic to an analytic system. Mixed-language also lost case-endings, illustrated by the development from the late fourteenth century of the noun phrase in [Def Art le + caseless English/Anglo-French N], where, for example, ad rammam (‘to the ram’) took the Latin accusative case-ending -am, but ad le Raam (‘to the ram’) did not (both examples are taken from the same document in the 1420s; a ram was an engine for driving piles).8 Lexically, Sylvester’s (2018) and Tiddeman’s (2020) work show how lexis crossed language boundaries in a business setting, resulting ultimately in the loss of variation. Sylvester (2018) investigates use of the words fur (which occurred in all text-types studied) and the more minority variant pelure (which was almost absent from wills), where both terms meant the same thing in the medieval period—a generic term for the pelt of any ­animal— but only fur subsequently lasted. Tiddeman (2020) discusses sweetmeatimports in the fourteenth and fifteenth centuries zucre caffatyne, suchre candi, cotegante, sport de suchre pot, madrian, confection, confit, which were brought to England via Italian merchants and were recorded in AngloFrench and mixed-language texts, but the lexemes themselves originated further afield in Arabic and Persian. Only sugar candy and confection(ary) were to last in English. For a fuller discussion of trade vocabulary, see Sylvester and Tiddeman (this volume) and Fulton (this volume).

3.4 Unmarking In Trudgill’s schema, this is the survival of regular forms even if they are not majority forms. Unmarking can be illustrated from the supralocalisation  See Fernández Cuesta (Forthcoming) for a summary of the debate on whether Celtic or Old Norse caused the loss of inflexions; see Wright (2010) for a discussion of the Anglo-French definite article le, which was borrowed into written Medieval Latin, blocking a case-ending on a following non-­ Latin noun. The ‘to the ram’ examples are taken from Wright (2010: 133); see also Walkden et al. (this volume) for further discussion of inflectional morphology and agreement. 8

2  Contact Theory and the History of English 

37

period which preceded the standardisation of English (supralocalisation is where linguistic features spread out from their dialect of origin over space and time, but ultimately did not enter the subsequent standard dialect). Mixed-language writing underwent a shift of language ratios at the end of the fifteenth century so that what had previously been a Medieval Latin matrix containing Anglo-French and English shifted to an English matrix containing Anglo-French and Medieval Latin—the system that is now simply regarded as the English language (Wright 2023).9 This Anglo-Frenchinfused English contained regularities which spread far and wide, but which ultimately did not last and standardise. Examples include the cliticisation of definite article the before a following word beginning with a vowel on the Anglo-French l’ + vowel pattern (e.g. thestate for ‘the estate’), which was still in use in the seventeenth century, and the adoption of AngloFrench spelling-conventions -lx, -aun (e.g. ‘shambles’, ‘Canterbury’, both words of English etymology; Wright 2020: 527–528).

I nterdialect Development In Trudgill’s New Dialect Formation schema, this is the emergence of new forms not present in the input systems. New forms to emerge in supralocal Englishes as a result of the long use of the mixed-language system were the many Anglo-French coinages and semantic developments that are first attested in the late fourteenth century, including new hybrids made up of English roots + Anglo-French affixes, and vice versa. Durkin (2020) assessed the origins of multilingual vocabulary in the present-day British National Corpus. He found that there was a ‘huge spike’ of French / Latin vocabulary still in use that was first attested in English in the late fourteenth century, so that ‘[t]he high-frequency vocabulary of contemporary written English thus appears to show a much more marked impact from late Middle English borrowing from French and Latin than is shown by the lexicon as a whole’ (Durkin 2020: 348). The point is that these were Anglo-French words created on English soil  On mixed-language writing, see Putter et al. (this volume), as well as also Keller and Seiler (this volume). 9

38 

S. Fox et al.

after Anglo-French had ceased to be a mother tongue, or a taught language in childhood.10 This sounds paradoxical but it is not, because the mixed-language system continued until the end of the fifteenth century, and so these Anglo-French lexical developments belong to that system rather than a monolingual one: as that system fell out of use, its lexical developments were absorbed into English. Sylvester’s (2020) work shows that the prolonged use of three languages in this way resulted in the development of technolects, that is, where Latin, French and English words for the same thing acquired shades of semantic precision. She investigated words for the concept ‘shepherd’ in the Bilingual Thesaurus of Everyday Life in Medieval England, and found eleven synonyms, of which five derived from Anglo-French and six derived from English. Similarly, for the concept of ‘shepherd’s equipment: crook and shears’, she found eight terms, four of Anglo-­ French and four of English etymology; and for the concept ‘sheep-­ shearer’, she found six terms, four of Anglo-French and two of English etymology. Far from being excessive or redundant, she says ‘the evidence suggests that standardisation depended on a range of near-synonyms’ (2020: 372); that is, that the multilingual lexis caused sufficient number of synonyms at any one time for nuances of meaning to develop and attach to each: ‘What we witness in the Middle English period is the evolution of a variety with a precise vocabulary’ (2020: 377). ‘(V)ariation of expression within the same conceptual area, equat(ed) to having the lexical resources to describe things at the superordinate, basic and hyponymic levels’ (2020: 375). However, for obvious reasons, scholars of medieval multilingualism work with written evidence, unlike Trudgill’s investigations into speech. Unsurprisingly, textual evidence of the multilingual situation in Britain shows facets not entirely captured by Trudgill’s key processes, and Stenroos and Schipor (2020) have developed a five-­ step schema to include the physical dimensions and visual properties of manuscripts which indicate a shift in linguistic code—be it shift of text-­ type, discourse function, or anything else. They investigated the kind of civic, bureaucratic and legal texts that consist of formulaic, semi-­formulaic and freely composed elements, with specific conventions about page  On the teaching of French, see Ingham (2012) and Reed (this volume).

10

2  Contact Theory and the History of English 

39

layout, finding that all these elements could influence the language systems employed. They analysed their corpus of Middle English Local Documents according to the following categories (Stenroos and Schipor 2020: 208–213): 1 . Linguistic mixing types (discernible main language/mixed code) 2. Syntactic structure (intersentential/intrasentential) 3. Textual structure (interelemental/intraelemental) 4. Visual structure (interelemental/intraelemental) 5. Visual marking (up/down/unmarked/conflicting) 6. Information content (parallel/complementary) 7. Predictability (formulaic/customary/free) Linguistic mixing was found to be a given: if multilingual events are taken to include any appearance of a single word or name that seems to represent a language different from that of the preceding or surrounding portion of text, then the great majority of all texts in the material (and, indeed, probably in any historical corpus) may be said to include multilingual events. Virtually all the Latin texts of any length contain English personal or place names or terminology, and words of Latin and French origin abound in the English texts. At the same time, single-word events are extremely difficult to define and categorise, as it is often impossible to decide whether a single word represents a multilingual event or a borrowing, or simply a variant spelling. (Stenroos and Schipor 2020: 209)

Categories 2–4 refer to the ways in which two or more languages were integrated, from a macro level such as chapters, to smaller units such as paragraphs, sentences, clauses, phrases and words, to micro-switches at the morphemic level. Paratextual material could be fluid in medieval manuscripts (e.g. distinguishing the body of the text from the commentary on it, or an accretion of later commentaries glossing between the lines or in the margins), so Stenroos and Schipor refer simply to ‘textual elements’, together with ‘visual elements’, referring to such features as marginalia, summaries, footnotes, dates, signatures. Category 5 refers to paginal layout and features such as use of blank space, coloured ink, script

40 

S. Fox et al.

size and change of script, ‘emblems’ or abbreviation and suspension symbols and other conventionalised marks, and orientation: ‘top is generally more important than bottom, while the centre is more important than the margins’, so that headings (large) are marked ‘up’ and footnotes (small) are marked ‘down’ (2020: 211). Category 6 refers to the same information provided in parallel, such as when English texts have Latin summaries, or an oath is written in Anglo-French and then again in English, or Latin texts have Anglo-French marginalia. (It remained common, for example, for abbreviated dates and an abbreviated form of solutus est ‘paid’ to be written in Latin in commercial bills into the late eighteenth century.) Finally, the formulaic, customary and free category takes medieval conventions into account. An example of ‘formulaic’ is set phrases such as le Roy le veult ‘the King wishes it’; an example of ‘customary’ is the appearance of specific text-types in specific languages, such as the goods-inventory section of a will, which was customarily written in mixed-language (see Wright 2015); and multilingual events are classified as ‘free’ when their content and form are unpredictable from the context. An example of ‘free’ multilingual events is the back-and-forth type of switching found in the macaronic sermon text-type, as discussed by Schendl (2020). An early fifteenth-century example, taken from Schendl (2020: 152), is Set nunc, prodolor, perfectus amor is laid o watur, caritas fere extinguitur, iste ignis is almost out. Quere vbi vis infra villam ex extra, poteris blowe super vngues tuos for any hete of loue (‘But now, for shame, perfect love is laid to waste, charity is entirely extinguished, this fire is almost out. Seek where you wish within the village or outside, you might [as well] blow on your two fingers for any warmth of love’). Present-day theories of code-switching (or translanguaging as it is sometimes now labelled), usually based on spoken data, tend to focus on switch-points and seek commonalities across languages in order to discover universal constraints (see Schendl and Wright 2011 for a discussion). However, when considering historical data, Schendl and Wright make the point that in Britain, from the Norman Conquest of 1066 to the end of the fifteenth century (at least), it is difficult to find monolingual writing in texts of any length. Accordingly, Stenroos and Schipor’s schema leads us to expect not a single underlying motivation but a range of mixing behaviours according to factors such as the conventions of the text-type, the intended audience

2  Contact Theory and the History of English 

41

and the cost of manuscript production. Trudgill’s New Dialect Formation schema is thus helpful when considering medieval states of English, but Lehiste’s (1987) concept of the multilingual individual as the locus of contact-induced change is crucial, in reference here to the collective linguistic behaviour of the trilingual later Medieval Latin/Anglo-French/ Middle English-writing scribes.

4 Conclusion Describing the linguistic history of English involves assembling a jigsaw in which a third of the pieces are missing. We have attempted to show that much of our knowledge concerning language contact processes and their linguistic outcomes has been derived from the detailed study of historical documents, resulting in the development of important theories of language contact. The missing part of the jigsaw in this case is of course spoken language, and we cannot be sure how the language was used by the people who spoke it at a given time, nor how the processes of accommodation impacted on the eventual linguistic outcomes. Contemporary sociolinguistic research focusing specifically on language variation and change in spoken language can therefore add to our knowledge in this respect, and we have attempted to show here how theories that have been used to explain language change in English in recent years (both in contexts where the dialects in contact are mutually intelligible and in a multilingual context such as London) might also help to explain some of the linguistic outcomes during the development of English. Even with the additional insights of contemporary sociolinguistic research, though, we are not able to pinpoint the exact cause nor precise time that a change is initiated, and the actuation problem (perhaps the final piece of the jigsaw) remains as elusive as ever.

42 

S. Fox et al.

References Aboh, Enoch O. 2009. Competition and Selection: That’s All! In Complex Processes in New Languages, ed. Enoch O. Aboh and Norval Smith, 317–344. Amsterdam: John Benjamins. Aikhenvald, Alexandra Y. 2010. Language Contact in Amazonia. Oxford: Oxford University Press. Cheshire, Jenny. 2013. Grammaticalisation in Social Context: The Emergence of a New English Pronoun. Journal of Sociolinguistics 17: 608–633. Cheshire, Jenny, and Sue Fox. 2009. Was/Were Variation: A Perspective from London. Language Variation and Change 21: 1–38. 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: 151–196. Cheshire, Jenny, Sue Fox, Paul Kerswill, and Eivind Torgersen. Forthcoming. Multicultural London English. In The Wiley Blackwell Encyclopedia of World Englishes, ed. Kingsley Bolton. Chichester: John Wiley. Clark, Ross. 1982. ‘Necessary’ and ‘Unnecessary’ Borrowings. In Papers from the Third International Conference on Austronesian Linguistics, III: Accent on Variety, ed. Amran Halim, Lois Carrington, and S.A.  Wurm, 137–143. Pacific Linguistics. van Coetsem, Frans. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter. Dance, Richard, Sara M. Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge, Cardiff and Sheffield. https://gersum.org. Accessed 30 May 2023. Durkin, Philip. 2020. The Relationship of Borrowing from French and Latin in the Middle English Period with the Development of the Lexicon of Standard English: Some Observations and a Lot of Questions. In The Multilingual Origins of Standard English, ed. Laura Wright, 343–364. Berlin: De Gruyter Mouton. Fernández Cuesta, Julia. Forthcoming. Key Events in the History of Early English: Reassessing the Celtic and the Scandinavian Hypotheses. In The Cambridge History of the English Language, I, ed. Laura Wright. Cambridge: Cambridge University Press. Filppula, Markku, Juhani Klemola, and Heli Paulasto. 2008. English and Celtic in Contact. London: Routledge.

2  Contact Theory and the History of English 

43

Fox, Susan. 2012. Performed Narrative: The Pragmatic Function of This Is + Speaker and Other Quotatives in London Adolescent Speech. In Quotatives: Cross-Linguistic and Cross-Disciplinary Perspectives, ed. Ingrid van Alphen and Isabelle Buchstaller, 231–257. Amsterdam: John Benjamins. ———. 2015. The New Cockney: New Ethnicities and Adolescent Speech in the Traditional East End of London. Basingstoke: Palgrave Macmillan. Fox, Susan, and Devyani Sharma. 2017. The Language of London and Londoners. In Urban Sociolinguistics  – The City as a Linguistic Process and Experience, ed. Patrick Heinrich and Dick Smakman, 115–129. Abingdon: Routledge. Fox, Susan, and Eivind Torgersen. 2018. Language Change and Innovation in London: Multicultural London English. In Sociolinguistics in England, ed. Natalie Braber and Sandra Jansen, 189–213. London: Palgrave Macmillan. Grant, Anthony P. 2002. Fabric, Pattern, Shift and Diffusion: What Change in Oregon Penutian Languages Can Tell Historical Linguists. In Proceedings of the Meeting of the Hokan-Penutian Workshop, June 17–18, 2000, Report 11, Survey of California and Other Indian Languages, ed. Laura Buszard-Welcher and Leanne Hinton, 33–56. Berkeley: University of California at Berkeley. ———. 2011. Bound Morphology in English (and Beyond): Copy Or Cognate? In Copies Versus Cognates in Bound Morphology, ed. Lars Johanson and Éva Á. Csató, 99–122. Leiden: Brill. Grant, Anthony P., and Norval S.H. Smith. 2019. Mixed Languages, Younger Languages, and Contact-Induced Language Change. In Oxford Handbook of Language Contact, ed. Anthony P.  Grant, 303–327. Oxford: Oxford University Press. Hall, David. 2020. The Impersonal Gets Personal: A New Pronoun in Multicultural London English. Natural Language and Linguistic Theory 38: 117–150. Hancock, Ian F. 1986. The Domestic Hypothesis, Diffusion and Componentiality: An Account of Atlantic Anglophone Creole Origins. In Substrata versus Universals in Creole Genesis: Papers from the Amsterdam Creole Workshop, April 1985, ed. Pieter Muysken and Norval Smith, 71–102. Philadelphia: John Benjamins. van Hattum, Marije. 2015. ‘Queensland for Ever & Augus un ballybug go braugh’: The Expression of Identity in Nineteenth-Century Irish Emigrant Letters. In Language and Identity: Discourse in the World, ed. David Evans, 55–80. London: Bloomsbury.

44 

S. Fox et al.

Haugen, Einar. 1954. Review of Weinreich 1953. Language 30: 380–388. Heath, Jeffrey. 1978. Linguistic Diffusion in Arnhem Land. Canberra: Australian Institute of Aboriginal Studies. ———. 1984. Language Contact and Language Change. Annual Review of Anthropology 17: 367–384. Hejna, Mia, and George Walkden. 2022. The History of English. Leipzig: Language Science Press. Hewitt, Roger. 1990. Youth, Race and Language in Contemporary Britain: Deconstructing Ethnicity? In Childhood, Youth and Social Change: A Comparative Perspective, ed. Lynne Chisholm, Peter Büchner, Heinz-Herman Krüger, and Phillip Brown, 185–196. London: Falmer Press. Ingham, Richard. 2012. The Transmission of Anglo-Norman: Language History and Language Acquisition. Amsterdam: John Benjamins. Ingham, Richard, and Louise Sylvester. 2017–. The Bilingual Thesaurus of Everyday Life in Medieval England. https://thesaurus.ac.uk/bth/. Accessed 30 May 2023. Keller, Jonas. 2020. The Leipzig-Jakarta List as a Means to Test Old English / Old Norse Mutual Intelligibility. NOWELE: North-Western European Language Evolution 73: 252–275. Kerswill, Paul, and Eivind Torgersen. 2021. Tracing the Origins of an Urban Youth Vernacular: Founder Effects, Frequency, and Culture in the Emergence of Multicultural London English. In Advancing Socio-Grammatical Variation and Change, in Honour of Jenny Cheshire, ed. Karen Beaman, Isabelle Buchstaller, Sue Fox, and James Walker, 249–276. London: Routledge. Kerswill, Paul, Eivind Torgersen, and Susan Fox. 2008a. Reversing ‘Drift’: Innovation and Diffusion in the London Diphthong System. Language Variation and Change 20: 451–491. Kerswill, Paul, Jenny Cheshire, Sue Fox, and Eivind Torgersen. 2008b. Linguistic Innovators: The English of Adolescents in London. Final Report Submitted to the Economic and Social Research Council, February 2008, Full Research Report ESRC End of Award Report, RES-000-0680. Swindon: ESRC. Lang, Jürgen. 2012. Dangers of the Linguistic Feature Pool. PAPIA: Revista do Associação Brasileira de Estudos Crioulos e Similares, 1–16. https://www.academia.edu/33332586/Dangers_of_the_linguistic_feature_pool. Accessed 28 May 2023. Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Lehiste, Ilse. 1987. Lectures on Language Contact. Boston: MIT Press.

2  Contact Theory and the History of English 

45

Matthews, P.H. 2006. Re-Reading Uriel Weinreich’s Language in Contact. In Rethinking Languages in Contact: The Case of Italian, ed. Anna Laura Lepschy and Arturo Tosi, 1–11. London: Legenda. Moreno Olalla, David. 2020. Spelling Practices in Late Middle English Medical Prose: A Quantitative Analysis. In The Multilingual Origins of Standard English, ed. Laura Wright, 141–164. Berlin: De Gruyter Mouton. Mufwene, Salikoko. 1996. The Founder Principle in Creole Genesis. Diachronica 13: 83–134. ———. 2001. The Ecology of Language Contact. Cambridge: Cambridge University Press. Muysken, Pieter C. 2017. Using Scenarios in Language Contact Studies: Linguistic Borrowing into Coptic. In Greek Influence on Egyptian-Coptic: Contact-Induced Change in an Ancient African Language, ed. Eitan Grossman, Peter Dils, Tonio Sebastian Richter, and Wolfgang Schenkel, 3–16. Hamburg: Widmaier. Oxbury, Rosamund. 2021. Multicultural London English in Ealing: Sociophonetic and Discourse-Pragmatic Variation in the Speech of Children and Adolescents. Unpublished doctoral thesis, Queen Mary University of London. Putter, Ad. 2021. Materials for a Social History of the Dutch Language in Medieval Britain: Three Case Studies from Wales, Scotland, and England. Dutch Crossing 45: 97–111. Ross, Malcolm. 1996. Contact-Induced Change and the Comparative Method: Cases from Papua New Guinea. In The Comparative Method Reviewed: Regularity and Irregularity in Language Change, ed. Mark Durie and Malcolm Ross, 180–217. New York: Academic Press. Schendl, Herbert. 2020. Multilingual Texts as a Reflection of Code-Switching in Medieval England: Sermons and Beyond. Medieval Worlds 12: 148–168. Schendl, Herbert, and Laura Wright. 2011. Code-Switching in Early English: Historical Background and Methodological and Theoretical Issues. In Code-­ Switching in Early English, ed. Herbert Schendl and Laura Wright, 15–45. Berlin: Mouton. Smith, Norval S. H. 1987. The Genesis of the Creole Languages of Surinam. Unpublished doctoral thesis, University of Amsterdam. Stenroos, Merja. 2020. The ‘Vernacularisation’ and ‘Standardisation’ of Local Administrative Writing in Late and Post-Medieval England. In The Multilingual Origins of Standard English, ed. Laura Wright, 39–89. Berlin: De Gruyter Mouton.

46 

S. Fox et al.

Stenroos, Merja, and Delia Schipor. 2020. Multilingual Practices in Middle English Documents. In Records of Real People: Linguistic Variation in Middle English Local Documents, ed. Merja Stenroos and Kjetil Thengs, 204–225. Amsterdam: John Benjamins. Swadesh, Morris. 1955. Towards Greater Accuracy in Lexico-Statistic Dating. International Journal of American Linguistics 21: 121–137. Sylvester, Louise. 2018. A Semantic Field and Text-Type Approach to Late-­ Medieval Multilingualism. In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 77–96. Berlin: De Gruyter Mouton. ———. 2020. The Role of Multilingualism in the Emergence of a Technical Register in the Middle English Period. In The Multilingual Origins of Standard English, ed. Laura Wright, 365–380. Berlin: De Gruyter Mouton. Sylvester, Louise, Megan Tiddeman, and Richard Ingham. 2022. Lexical Borrowing in the Middle English Period: A Multi-Domain Analysis of Semantic Outcomes. English Language and Linguistics 26: 237–261. Thomason, Sarah G., and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Los Angeles, CA: University of California Press. Tiddeman, Megan. 2020. More Sugar and Spice: Revisiting Medieval Italian Influence on the Mercantile Lexis of England. In The Multilingual Origins of Standard English, ed. Laura Wright, 381–410. Berlin: De Gruyter Mouton. Timofeeva, Olga. 2022. Sociolinguistic Variation in Old English: Records of Communities and People. Amsterdam: Benjamins. Torgersen, Eivind, and Anita Szakay. 2012. An Investigation of Speech Rhythm in London English. Lingua 122: 822–840. Trudgill, Peter. 2004. New-Dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press. Weinreich, Uriel. 1953. Languages in Contact: Findings and Problems. New York: Mouton. Wright, Laura. 2010. A Pilot Study on the Singular Definite Articles le and la in Fifteenth-Century London Mixed-Language Business Writing. In The Anglo-­ Norman Language and Its Contexts, ed. Richard Ingham, 130–142. (York: York Medieval Press). ———. 2015. On Medieval Wills and the Rise of Written Monolingual English. In Approaches to Middle English: Contact, Variation and Change, ed. Javier Calle Martín and Juan Camilo Conde Silvestre, 35–54. Frankfurt am Main: Peter Lang.

2  Contact Theory and the History of English 

47

———. 2020. Rising Living Standards, the Demise of Anglo-Norman and Mixed-Language Writing, and Standard English. In The Multilingual Origins of Standard English, ed. Laura Wright, 515–532. Berlin: De Gruyter Mouton. ———. 2023. How Multilingualism Came to Be Ignored in the History of Standard English. In Multilingualism and History, ed. Aneta Pavlenko, 107–122. Cambridge: Cambridge University Press. Wright, Roger. 1976. Semicultismo. Archivum linguisticum 7: 13–28.

3 From Original Sources to Linguistic Analysis: Tools and Datasets for the Investigation of Multilingualism in Medieval English Carola Trips and Peter A. Stokes

1 Introduction Like seemingly all research today, research into multilingualism and medieval English has seen the considerable impact of digital tools and datasets. The very first work in what is now known as the digital humanities was devoted to linguistic analysis, lexicography and corpora of historical texts (e.g. Cameron et al. 1970), and this has since expanded into what can sometimes seem to be a bewildering array of different, often disconnected and sometimes overlapping projects. Given the vast amount

C. Trips (*) University of Mannheim, Mannheim, Germany e-mail: [email protected] P. A. Stokes École Pratique des Hautes Études—Université PSL, Paris, France e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_3

49

50 

C. Trips and P. A. Stokes

of work and the continuing rapid evolution of projects and technology, a chapter of this length cannot hope to address all existing tools and datasets, and nor would such completeness be useful as it would surely become obsolete even before it is printed. Nevertheless, certain tools and datasets have been stable for a relatively long time now and are central to research in the field, while others are rapidly achieving this state. Furthermore, the core principles of best practice have changed little over the years, even if today they are much more clearly defined and largely more closely followed than in the past. The details here are more relevant to those who produce this material than to those who use them, but some understanding is essential to all. Indeed, all researchers need to apply a critical understanding to their materials, whether those materials are original manuscripts, modern editions, printed articles, datasets, digital tools or images, and this in turn implies knowing the source of a dataset or tool, how and why it was produced, and what assumptions underlie it. The chapter that follows is presented along what might be a typical workflow, starting with the original source (assumed to be a manuscript book) and ending with the conclusions of a linguistic analysis. This workflow implies numerous steps, including first finding the manuscript sources, then obtaining the texts (perhaps via automatic means), building the corpus, adding linguistic tagging, analysing and presenting the results, all in the context of given research questions. Not all researchers will be involved in all steps—on the contrary, few probably are in practice—but again an awareness of each step and its implications for those that follow is essential, as we hope to show. The earlier steps of this workflow are presented in relatively general terms, highlighting just some of the material that is available but raising methodological issues along the way, while subsequent steps include more concrete examples of linguistic analysis and the results that can be obtained.

2 Accessing the Manuscript The majority of surviving linguistic evidence has been preserved in the form of manuscripts, and so the first step is therefore finding and gaining access to this evidence. The corpus of manuscript books containing Old English has long been established as represented by Ker’s Catalogue of

3  From Original Sources to Linguistic Analysis 

51

Manuscripts Containing Anglo-Saxon (1957) and supplements (Ker 1976; Blockley 1982), as well as Gneuss’s Handlist of Anglo-Saxon Manuscripts and addenda (2001, 2003, 2011), which he subsequently expanded with Lapdige as Anglo-Saxon Manuscripts (2014). Gneuss’s (2001) Handlist is largely available online with supplemental information via DigiPal: Digital Resource and Database of Palaeography, Manuscript Studies and Diplomatic. Although not listed in Ker’s Catalogue, another area of access to content for those studying multilingualism is that of charters and writs. English charters from the ninth through to the late eleventh centuries were normally bilingual, with the body of the text in Latin but the charter bounds typically in English. Although a very specific genre and with a relatively restricted range of vocabulary and grammatical structure, these nevertheless provide further evidence for Latin-English bilingualism. The corpus of pre-Conquest charters was catalogued and published as Anglo-Saxon Charters: An Annotated List and Bibliography (Sawyer 1968), and this has also been supplemented and published online as the Electronic Sawyer. The Catalogue of English Post-Conquest Vernacular Documents (Pelteret 1990) extends Sawyer’s volume chronologically. Cartulary copies are catalogued in the Medieval Cartularies of Great Britain and Ireland (Davis 2010), first published in 1958 and then revised by Breay, Harrison and Smith. One of the more thorough catalogues from the linguistic point of view is The Production and Use of English Manuscripts 1060–1220 (Da Rold et al. 2010), particularly its ‘Index of Manuscripts Containing French’, which includes cartularies. There are also increasing numbers of digital resources for finding manuscripts containing vernacular texts, although their value for multilingualism varies. As well as allowing one to view images of manuscript pages, many digitisation projects also allow searching by different criteria, including by language, and in principle this can be very helpful for identifying examples of multilingual contact.1 Most often this is in the form of ‘faceted’  Some relevant examples include the Cambridge Digital Library (Search > Advanced Search), the Digital Bodleian (Advanced Search) and the joint Bibliothèque Nationale de France–British Library project France Angleterre 700–1200 (Recherche avancée). This is also possible for Parker on the Web, but it is not easy to find: one way is to click on ‘Search’ while leaving the search box empty, then one may or may not see options to ‘Limit your search’ depending on the size of your browser window. If not, one can click not on the text ‘Limit your search’ but instead on the box on the far right at the same horizontal level, and then the options appear. For full details of websites and software cited in this chapter, see ‘Digital resources and websites’ in the bibliography below. 1

52 

C. Trips and P. A. Stokes

searching, where one can filter by different criteria often including language. In principle, one can search for manuscripts containing texts in different languages, such as both ‘Middle English’ and ‘Middle French’, for example. In practice, however, this needs to be treated with some caution, as different sites can give very different results. First, it is often unclear whether selecting two different languages is treated as ‘AND’ (requiring both languages to be listed) or ‘OR’ (requiring at least one of the two), and experimentation shows that sometimes different pages on the same site give different responses in this respect. In addition, practices in cataloguing vary from country to country, institution to institution and sometimes even from person to person. For instance, the list of languages provided for a manuscript may represent those in the main text, or may also include those in marginal notes and glosses; it may only include those which seem to be part of the original writing or it may include later additions; it will most likely not include writing on flyleaves, even though such scribbles can provide important evidence for linguists; and so on. Furthermore, the definition of linguistic categories such as ‘Middle English’ are notoriously slippery, and manuscript cataloguers cannot be expected to enter into the finer details of such debates. This is particularly the case in aggregation projects such as Biblissima, which (among other things) is harvesting and linking data from collections of European manuscripts from the eighth to the eighteenth century.2 The images themselves are available from the different repositories, and the metadata is being combined into a single dataset that can be cross-searched very easily. For most users, the most accessible form of this is Biblissima Collections, which provides a relatively user-friendly means of searching records of approximately 80,000 manuscripts from around twenty different repositories in Europe. This includes searching by language or languages, and so (for instance) ‘English’ and ‘Middle French’ return a number of results particularly from Corpus Christi College in Cambridge, since the contents depend on the repositories agreeing to provide metadata and to host the appropriate servers for the images, but it also depends on the necessary metadata being included in the catalogue  The Biblissima project has recently received a second major grant and so has started its next phase as Biblissima+. 2

3  From Original Sources to Linguistic Analysis 

53

in machine-readable form, and in that metadata being consistent across the collections. For instance, searching Biblissima Collections for both ‘English’ and ‘Middle French’ gives thirteen hits at the time of writing, and although this is undoubtedly a useful start, it seems very unlikely that only thirteen out of 80,000 manuscripts contain texts in these two languages, and so presumably for the other cases the languages were not catalogued at all, or were done so in a way that could not be harvested and linked in to the search framework. Such searches are therefore by no means comprehensive and must not be relied on for statistical analyses in any sense, but they can nevertheless provide a useful starting point for research. Work also continues on cataloguing and aggregation, and so we are likely to see increasingly useful results in the relatively near future. Although visits to libraries and archives remain essential, the availability and prevalence of digital facsimiles has increased enormously in recent years thanks to digital imaging and the Internet. A large proportion of the surviving corpus of manuscripts containing texts in Old English is now readily available online, and those for Middle English are also progressing rapidly, where ‘available’ here means digital images of each page and, usually, of the binding as well. Most major repositories (and many smaller ones) now have hundreds or even thousands of manuscripts available in this way, including most main collections in Cambridge and Oxford, as well as large numbers in London, Paris and elsewhere, meaning that coverage of manuscripts containing Old English is now very thorough, and increasingly also for Latin, Middle English and medieval French.3 One difficulty is that images are normally hosted by individual repositories, and so one typically needs to visit many different websites in order to find the necessary material. However, this is changing rapidly with the International Image Interoperability Framework (IIIF). IIIF is a framework and standard that is now widely used across cultural heritage institutions, whereby an image viewer (or any other software) can communicate directly with any number of different servers hosted by different repositories, in order to display (or harvest, or search, or analyse) their  See Footnote 1, as well as Gallica (Recherche avancée), and the James Catalogue of Western Manuscripts from the Wrenn Library in Cambridge (Browse Catalogue), among many others. 3

54 

C. Trips and P. A. Stokes

images. This means that one can use a single viewer software such as Mirador or the Universal Viewer to simultaneously view images from dozens of different libraries and archives in the same interface, including the British Library, Bodleian Library, most Cambridge collections, the Bibliothèque Nationale de France, the Biblioteca Apostolica Vaticana and many more. As well as making the life of scholars significantly easier, this direct communication between computer servers also opens up many new possibilities for the near future, since a sufficiently powerful computer can now access hundreds of thousands of images of manuscripts and documents from dozens of different libraries, as well as their accompanying metadata. These images and metadata can then be analysed automatically at scale, as demonstrated by the Biblissima Collections discussed above as well as in some further ways that will be mentioned below. In addition to this access to manuscript evidence of known texts, the fact that all the pages of the manuscripts are normally being digitised, and not only a selection, can sometimes bring unexpected benefits for studies of multilingualism. For instance, many medieval manuscripts contain notes and ‘scribbles’ on endleaves, often in vernacular languages that may not be represented in the main text, and these have often been overlooked, but now one can at least search for them relatively systematically by going through online images. These are not normally catalogued or, if they are, this is often done for the texts in one specific language, and so the multilingual context is lost. Indeed, access to manuscripts or images thereof gives access to texts, but the palaeography of manuscript books, charters and later cartulary copies also provides important evidence for linguists. On the one hand, understanding the manuscript context of words can change their interpretation, most obviously in the case of silent editorial normalisation or emendation in editions, but also in glosses, marginal notes and other cases (e.g. Parkes 1969: xxix–xxx; Jenkyns 1991: 383–388; Page 1992: 43–44; Robinson and Solopova 1993: 19; Stokes 2009: 52–56, 2020b: 50–52). Furthermore, knowing if two texts in different languages were written by the same person or not, or in the same context or not, can significantly change the interpretation of those works and their cultural context, as they indicate scribal competence in writing different languages and, thereby, indirect evidence for bilingualism. For instance, a number of twelfth-century cartularies

3  From Original Sources to Linguistic Analysis 

55

containing copies of pre-Conquest charters are consistently well written for the Latin parts, but very much less so for the vernacular, sometimes in terms of orthography but also in terms of the quality and regularity of the script, even though the scribe is clearly the same throughout. This strongly suggests that the scribes in question were accustomed to copying in Latin and (presumably) relatively familiar with the language, but the fact that they had very much more difficulty in copying English suggests that they were also probably less familiar with the language, at least in its written form (Stokes 2011: 32–41, and see Fig. 3.1). However, accessing this information in a systematic way is very difficult. Perhaps the closest to a complete list is the Conspectus of Scribal Hands Writing English, 700–1100 (Scragg 2012), which includes books, cartularies and single-sheet charters, but it only lists vernacular writing and does not indicate which scribes also wrote in Latin (or other languages). Additional information is available from DigiPal, where some scribes are catalogued as writing in both English and in Latin, but the chronological scope here is limited to the eleventh century and even this is patchy and far from systematic. Computer-assisted palaeographical analysis and particularly scribal identification have also been the subject of significant research over the last decade or so. In addition to digital tools for aiding manual palaeographical analysis like Archetype and Graphoskop,4 a significant effort has also gone into the automatic or semi-automatic identification of scribes from sample images of handwriting (the problem known as ‘writer identification’ in the language of computer scientists), and computers are now very effective at automatically identifying subtle changes in handwriting, although it remains the work of the palaeographer to decide if such changes are due to a change in scribe or to any number of other reasons (change of pen, change of ink, change of parchment, change of weather, change of exemplar, change of lighting when taking the digital image and so on). At the time of writing, the best known here is probably the Monk project, developed at the University of Groningen. Other approaches aim to automatically identify the style of script (Praegothica, Hybrida etc.: Stutzmann 2016; Kestemont et al. 2017), or the date or  Archetype is the software used for the DigiPal project cited above, as well as others including Exon Domesday, Models of Authority, and VisigothicPal, among others. 4

56 

C. Trips and P. A. Stokes

Fig. 3.1  London, British Library, MS Cotton Tiberius A VI, 99v (another example is Cambridge, Trinity College, MS O.2.41, p. 87, reproduced in The James Catalogue (online)). Note the first twelve lines of col. a and the last seven lines of col. b (in English), compared to the rest of the page in Latin. Image © British Library Board

3  From Original Sources to Linguistic Analysis 

57

other details. Such methods are still largely under development and are perhaps less immediately relevant to linguists, but they give an idea of the rapid developments in computational methods applied to manuscript studies, and suggest new tools that are likely to be more widely available very soon. Although bringing clear benefits, online facsimiles are of course limited as many scholars have already noted. For instance, the codicological structure of the book is normally lost entirely in facsimiles, whether digital or printed. However, the structure and ‘archaeology’ of the book can have significant implications for understanding the text, including for multilingualism, for instance because correctly recognising codicological units of production is essential for understanding which texts in a manuscript are likely to have been produced at the same time and / or place (cf. Andrist et al. 2013; Stokes and Noël 2019; Stokes 2020a, b). Comparatively little work has been done in ‘digital codicology’, but probably the most important tool is VisColl, which allows one to produce diagrams and visualisations of the codicological structure of books: these include ‘traditional’ quire diagrams but also ‘virtual disbindings’, where one can see the sheets of parchment laid out as if the book had been entirely disbound. Although not yet widely used, such tools for codicological visualisation are starting to appear, for instance, as part of editions of particularly complex manuscripts or in online viewers.5

3 From Manuscript to Transcription Seeing images of pages from manuscripts and charters is an important step, but linguistic study normally requires access to the text, and this can be achieved in different ways. Scholars have been producing digital texts for many decades now by transcribing manually, whether as ‘pure’ text or with markup such as Text Encoding Initiative (TEI) XML. For cases where texts are not already available, automatic transcription of texts from images is increasingly becoming a viable option. This process of  Two examples are Exon Domesday (see Labs > Codicological Visualisation and Labs > Collation Visualisations) and BiblioPhilly, where some digitised manuscripts come with interactive quire diagrams. 5

58 

C. Trips and P. A. Stokes

optical character recognition (OCR) or handwritten text recognition (HTR) applied to manuscripts can now produce very good results, in ideal cases with less than one error per 100 or 200 characters.6 A number of different software packages are now available, with varying degrees of maturity, usability and the degree to which they are free and open source, as well as their applicability to medieval manuscripts. Some of the more important for this context include Tesseract, Calamari, kraken / eScriptorium and Transkribus.7 In general, the automatic transcription of medieval manuscripts is significantly more challenging than that of modern print, and for this reason one cannot normally use commercial systems such as ABBYY Finereader or Omnipage even though the underlying process is much the same. Instead, one must normally work in several stages, each of which will benefit from human interaction. The first stage is to obtain images of the pages to be transcribed, and to send these to the OCR / HTR system, for instance by uploading them from a home computer or by sending the URL of an IIIF manifest for the manuscript or collection from which the images can be imported automatically; in this way the files are transferred directly from the repository to the OCR server without ever passing by the user’s computer. The next stage is normally that of layout detection, where the software analyses the images to determine regions of the page such as the margins, the main text block, illustrations and so on, as well as lines of text on the page. Once the lines of text have been identified, the transcription itself can take place. In some systems, users can distinguish between the main text, running headers, page numbers, marginalia and so on, allowing them to treat the different regions in different ways such as excluding the page numbers, analysing only the marginalia or the ‘main’ text, or whatever else is required. Indeed, it is important to note that the automatic transcription is normally conducted line-by-line, and so reconstructing the text from  The distinction between OCR and HTR is not clearly defined: it can refer to the source material (OCR for print and HTR for manuscript), or the underlying process (recognition character-by-­ character or recognition line-by-line), but in modern systems there is little practical distinction. For this reason, we refer to OCR / HTR throughout. 7  Kraken is an OCR / HTR software which runs from the command line, and eScriptorium is a web-­based graphical user interface that interacts with kraken. The two are therefore often used together and so will be referred to collectively as a single package unless only one of the two parts is intended. 6

3  From Original Sources to Linguistic Analysis 

59

individual lines can be a difficult task if the page has a complex layout or significant amounts of paratext, such as marginal commentaries and interlinear glosses. Another important point is that both the layout analysis and the transcription normally require some manual correction if one wants an error-­ free text. In modern systems such as kraken / eScriptorium and Transkribus, both the transcription and the layout analysis are performed using machine learning, and one area of active development for the near future is for the recomposition of the text from the lines also with machine learning. This means that one can retrain the system on corrected material in order to obtain better results, and so in practice one can proceed iteratively by processing some pages, correcting, retraining, processing a larger number of pages with fewer errors, correcting and so on. Machine learning also means that the system can adapt to the particular needs of specific corpora, including their page layout and script. However, this also means that one cannot start without example material (‘ground truth’ data) to train the machine, and this material must normally be prepared manually. Both eScriptorium and Transkribus therefore provide interfaces for one to manually indicate lines and regions on an image, and to import or manually type transcriptions. In general, it is impossible to predict the exact amount of example material that is needed for good results, since it depends on many factors including the consistency of the writing, the complexity of the layout, the number of lines of writing per page, the number of distinct signs in the writing (letters, punctuation, abbreviation signs and so on) and indeed on how one defines ‘good’. For most linguistic analysis, the starting assumption is normally that no errors are acceptable, but there are cases where even very inaccurate transcriptions can still be used or some large-scale or ‘distant’ analyses. For instance, even an error every line is more than sufficient to automatically identify the language of the text, or potentially to identify the text itself, and so it is now possible and relatively easy technically to harvest digital facsimiles of hundreds or even thousands of manuscripts via IIIF, transcribe the images automatically, automatically identify the language(s) of these text and thereby produce estimates of multilingualism on a scale that was previously impossible. Such estimates would of course require care in their interpretation, and so one could argue that the number of

60 

C. Trips and P. A. Stokes

errors is less important than the clear communication of the level of accuracy, the likely sources of error and the biases that these can introduce. This need for training data means that careful consideration is required before deciding if it is worthwhile to prepare the ground truth at all: if one has several hundred pages written by a single scribe, then it is almost certainly more efficient to use automatic transcription, but the opposite is likely true for a small corpus of single-sheet charters each written by a different scribe. A number of initiatives are currently underway to help, including using the computer itself to generate artificial (‘synthetic’) data for training and a software that can take an existing transcription (for instance, from another manuscript copy of the same text) and automatically align it with the new manuscript. Another important approach is for researchers to publish their training data and standards for transcription so that they can be reused by others: examples here include OCR-D and SegmOnto. Indeed, kraken / eScriptorium allows users to freely export and share the trained model itself which contains the computer’s learned ‘intelligence’: existing models can therefore be reused and retrained, for instance taking a model developed for Latin written in fourteenth-­century Gothic cursiva and retraining it for Middle English in the same script. In this way, significantly less specific training data is required, saving in human effort and also in the environmental impact caused by the significant amounts of energy required for training these systems. A final consideration here is that of being free and open source. Apart from Transkribus, one can freely download all the systems described here and run them on a home computer. However, as just noted, training the machine learning on a large number of pages requires very large amounts of computing power and so is not feasible on a normal home system. For this reason, Transkribus has a desktop version for the user interface, but the processing is run on an external server that is managed and controlled by the Transkribus team. eScriptorium is also normally used this way, communicating with an instance of kraken that is installed on a remote high-performance computer, although this is not obligatory and it can be run on a single computer if desired. This in turn leads to different costing models: with Transkribus, the users do not need to (and cannot) manage the servers themselves, but instead pay for the service: the payment is

3  From Original Sources to Linguistic Analysis 

61

both direct in terms of a sum for each page transcribed and indirect in that all ‘ground truth’ data that the users provide is then used to improve the models which are retained by Transkribus and used as a commercial offering to other customers. In kraken / eScriptorium, as well as Calamari and Tesseract, there is no direct payment, and the models remain entirely the property of those who produce them, to be shared or not as desired. However, the trade-off is that using these at scale requires setting up the necessary servers, or finding an existing set-up that is available for use.

4 From Transcription to (Annotated) Text The output of an OCR / HTR system will normally be a ‘raw’ transcription: that is, the machine will produce a sequence of characters according to what it ‘sees’ in the image and how it has been trained to convert from image to text. In other words, one can train an OCR / HTR to distinguish between u and v or not, to treat as a distinct sign or as the word et (or and, or ond), and so on. However, this is only possible for the machine if one can provide a direct single mapping from an image to exactly one representation in text, and if not, then other approaches are needed. For instance, an OCR / HTR system on its own cannot expand sometimes as et and sometimes as and depending on the linguistic context. Similarly, a significant difficulty for modern analysis is the treatment of punctuation and spaces in medieval manuscripts, since spacing rarely corresponds clearly and consistently with that required for modern analysis. Overcoming this requires tools that include linguistic models, particularly for vernacular languages which typically show significant variation in orthography, and these can be either incorporated into the OCR / HTR software or run as a separate stage. Here again different approaches are available, usually varying between semi-automatic tools which aim to speed the work of the human user, or those which attempt to do the work largely or fully automatically. Tools for the analysis and linguistic tagging of Latin are relatively advanced, with resulting annotated corpora, as shown particularly by the LiLa project, which aims to integrate tools and corpora, but medieval vernacular languages are significantly more challenging. Ad hoc tools and approaches developed for specific projects are

62 

C. Trips and P. A. Stokes

relatively widespread (for two examples, see Stokes and Pierazzo 2009: 226–234, and discussion below). Some example projects aiming to produce more generic tools include BOUDAMS, for automatic tokenisation of Latin and Old French (Clérice 2020) and Pyrrha, which is designed to aid manual correction of automatically lemmatised and POS-tagged texts, particularly though not specifically Old French, and this in turn is built on PIE, a generic framework which can be developed to apply machine learning techniques to a very wide range of linguistic (and other) tasks. However, as this discussion suggests, these tools and approaches are largely language specific and are particularly difficult to apply to multilingual texts. Indeed, even OCR / HTR methods are sensitive to language: although the ones described here do not necessarily have an explicit language model as such, nevertheless modern systems do not treat each character independently but work line-by-line, and thereby incorporate the probabilities of different letter combinations. These probabilities are of course language-dependent, and for this reason a model trained for the transcription of one language often performs very badly when applied to a different language, even when written by the same scribe and in the same script. For these reasons, a tool that could analyse an image of a manuscript page and automatically identify the language(s) of the writing therein is potentially extremely valuable, since the system could then use this to apply different trained models for OCR / HTR and for subsequent processing and linguistic annotation. Such a tool would also be invaluable for studying multilingualism, since a sufficiently powerful computer could then scan images of hundreds of thousands of manuscript pages (all accessed via IIIF) in order to find multilingual content. Such a tool directly applicable to medieval manuscripts does not yet exist to our knowledge, but in principle it is well within the possibilities of existing technology.

5 From Annotated Text to Using Linguistically Annotated Corpora As noted in the introduction, the next step in the workflow includes more concrete examples of linguistic analysis and results that can be obtained by using methods and tools described for the previous steps.

3  From Original Sources to Linguistic Analysis 

63

More precisely, this section discusses collections of annotated texts, that is, linguistically annotated corpora, and how they can be used to investigate (morpho)syntactic effects of multilingualism in medieval Britain. We will also consider new developments in the field of corpus linguistics that enhance these corpus-based methods. It is well-known that there are a number of corpora available that reflect this time period, for example A Linguistic Atlas of Early Middle English (LAEME; Laing 2013–) and the electronic version of the Linguistic Atlas of Late Mediaeval English (eLALME; Benskin et al. 2013–) (for an overview, see the Corpus Resource Data Base, CoRD). LAEME is well suited to investigate (morpho)syntactic effects as it includes linguistic annotation in the form of tagged texts with lexico-grammatical tags. eLALME displays dialectal variation on the basis of 1000 medieval texts and includes Linguistic Profiles, maps and scribal languages. In their investigation of the h-type and þ-type pronominal system in the seven manuscripts of the Middle English poem known as La estorie del evangelie, Cole and Pons-Sanz (this volume) use these two resources to examine dialectal variation as well as linguistic, stylistic factors and scribal behaviour. Sometimes, however, linguistically annotated corpora are not available and researchers have to resort to a corpus of electronic texts alone. In their study of Norn as a historically attested heritage language spoken on the Orkney and Shetland Islands, Kinn and Walkden (this volume) examine Norn texts (charters and one ballad from 1300 to 1800) taken from the Diplomatarium Norvegicum and manually extract all instances of types of possessives. To identify contact effects, they adapt a method from psycholinguistics and create a baseline corpus of texts from contemporary Norwegian. By combining a philological / traditional approach with an experimental / statistical approach, they develop a new way to identify contact-induced change in historical linguistics. Apart from the corpora mentioned above, the most well-known linguistically annotated corpora are the Penn Parsed Corpora of Historical English, which have been available for historical stages of English for quite some time now. All texts in all corpora come in three formats (simple text, tagged text and parsed text) and adhere to the same annotation system, which can be queried with the programme CorpusSearch 2

64 

C. Trips and P. A. Stokes

(Randall 2010). In this way, the development of some aspect of English can be examined across corpora in a very comfortable way, that is, without learning a new annotation format and query language per corpus, which is, unfortunately, still often the case. By means of these corpora, a number of studies investigating contact-induced (morpho)syntactic change in medieval English were conducted (e.g. Kroch and Taylor 1997; Trips 2002; Taylor 2008; Haeberli 2018). Without a doubt, the availability of these corpora has transformed the way quantitative and qualitative studies can be conducted systematically, and they have also improved the accuracy of results. However, looking at the text samples of the corpora relevant to our field of interest, especially the Penn-Parsed Corpus of Middle English 2 (PPCME2; Kroch and Taylor 2000), one will quickly realise that they are all monolingual. In this respect, these electronic corpora do not represent a multilingual situation, although we have known for a long time that such a situation exists thanks to investigations using philological methods applied to the plethora of different text types (cf. above; Trotter 2000; Jefferson and Putter 2013; Wright 2018).8 So the question arises how we can use these corpora to investigate multilingualism and its effects on the development of English at all. One way to deal with this matter is to enrich these corpora with etymological information. We will briefly introduce the resources and tools that were created in the project Borrowing of Argument Structure in Contact Situations: The Case of Medieval English under French influence9 (BASICS) to investigate the copying of argument structure of lexical verbs in the medieval contact situation between Old French (Anglo-French) and Middle English. In addition to syntactic queries, the objectives of the project required a way to conduct lemma-based and semantic queries and to make a distinction between French-based and non-French-based verbs.

 These corpora do not reflect multilingualism primarily for technical reasons, for instance because paragraphs written in Latin that are part of the original texts were not included. However, traces are left in the form of foreign language sequences that are tagged FW (foreign word). 9  This project was funded by the Deutsche Forschungsgemeinschaft, DFG, research grant TRI555/6-1, TRI555/6-2 Sachbeihilfe. For further information, see https://tinyurl.com/dfgbasics/ [accessed 29 September 2021]. 8

3  From Original Sources to Linguistic Analysis 

65

6 Lemmatisation Since the PPCME2 (as well as all the other Penn corpora) comes without (verb)lemmatisation, this information had to be inserted. The first step was to identify the verb origin of all verbs found in the corpus. Collaborating with the Oxford English Dictionary (OED), the project was provided with a list of 2026 English verbs copied from French between 1066 and 1500. For French-based verbs, the attribute e (etymology) was defined as french, for non-French-based verbs, it was defined as nonfrench. The lemmatisation of verbs in the PPCME2 relied on a list of form-­ lemma correspondences which was obtained by extracting all verbs from the corpus and by manually assigning them according to the lemmata found in the Middle English Dictionary (MED). (1)  (VB biginnen@l=biginnen@m=4470@e=nonfrench@) (2)  (VBP ordayne@l=ordeinen@m=30818@e=french@)

As shown in (1) and (2) with the parts marked in bold, these pieces of information were added to the original corpus and are indicated by the character @. The attribute l specifies the verb lemma, the attribute m the identification number as given in MED, and as noted above, e the etymology of the verbs10 (for further details, see Percillier 2018). There are a number of other resources and tools that the BASICS Toolkit offers, and we will briefly introduce some of them here as it is directly linked to the lemmatisation of the corpus and can also be used to examine language contact between Old French (Anglo-French) and Middle English.11 Researchers interested in studying verb classes in Middle English will find it very hard to identify them since they are not readily available in any lexical resource or database. This is why the tools Verb Classes and Reverse Lookup were developed as they provide a way to define them.  Although the verb-lemmatised version of the PPCME2 cannot be provided for licensing reasons, the lemmatiser script is available for download on the BASICS website so that everyone who owns a copy of the corpus can perform the lemmatisation himself / herself. 11  The video tutorial on the BASICS Toolkit website provides an introduction to all resources and tools. 10

66 

C. Trips and P. A. Stokes

First, a list for the verb class under scrutiny based on Levin’s (1993) verb classes for Present-Day English can be extracted and exported. Second, based on this list, MED can be queried for verb entries in whose definition the Present-Day English verbs occur, thus providing a list of near-­ synonyms. A search for a verb class or a lemma part of this class, for example amuse, with the tool Verb Classes will produce a list of all verbs that are part of the class, according to Levin (class 31.1; Levin 1993:189). This list can be downloaded and used as input for the tool Reverse Lookup. Uploading this list the tool then matches the verb entries of the Levin list with the occurrences of these verbs in the definitions of verbs in the MED. Again, the result is a list that can be downloaded either in HTML or as a CSV table shown in Table 3.1. Apart from the MED lemma for the verbs, the MED ID, the corresponding verb in Present-Day English (OED), and the OED ID, the etymological information ((non)French-based) that was also added to the lemmatised PPCME2 is also given. The result of the process of Reverse Lookup shown in Table 3.1 is a good approximation of the class of amuse-­ type verbs in Middle English that clearly eases manual verification but does by no means replace it (for details, see Percillier 2018). This automated approach is similar to the one García García and Ingham (this volume) apply to investigate the lability of psych and destroy verbs in Middle English. However, they use different lexical resources and built-in tools where available (for Old English the Bosworth and Toller online Dictionary, the Dictionary of Old English, and for Middle English the ‘Definition and notes’ search tool of the MED to identify all verbs in whose definitions ‘destroy’ occurs). Table 3.1  Sample from Reserve Lookup MED-lemma

MED-ID

OED

OED-ID

French-based

comforten disporten gamen gladen

8533 12042 18152 18680

comfort disport game glad

36891 55102 76469 78624

TRUE TRUE FALSE FALSE

Amuse-type Verbs, Levin’s class 31.1; Levin (1993)

3  From Original Sources to Linguistic Analysis 

67

Overall, adding verb-lemmatisation and etymological information to the PPCME2 as well as creating the BASICS Toolkit enhanced methods to work on contact-induced change in Middle English (and Anglo-­ French). Using these resources, therefore, can offer further insights into medieval multilingualism in Britain (cf. Schauwecker 2019; Percillier 2019, 2020; Trips 2020). To demonstrate this, we will briefly discuss two studies. The first study is Percillier (2022), which examines contact-induced changes in Middle English on the model of Anglo-Norman (Old French) as well as the rise of Anglo-French (A-Fr). He applies Schneider’s (2003, 2007) Dynamic Model, which was originally developed to account for the rise of (post) colonial varieties of English. Percillier identifies Anglo-Norman as the settler strand (STL) and traces its development through the five stages assumed in this model. In the third stage of this model, the endonormative stabilisation, which spans from the thirteenth to the end of the fourteenth century, Anglo-French emerged as an independent variety of French coming to serve official functions whereas Anglo-Norman ceased to exist as a native language variety. Percillier sees Middle English as the indigenous strand (IDG), as its linguistic features, such as instances of copying and innovations,12 may have become nativised, meaning that they were no longer perceived as being foreign. This is why Percillier assumes that Schneider’s model can be applied to Middle English despite the fact that the IDG originally lies out of the scope of the model. Investigating verbs prefixed with a-, en-, es-, prepositional secondary predicates (PSP) and genitive / of-objects, Percillier seeks to explain the dynamics of this contact situation and thus to show that the Dynamic Model can be applied to other (non-postcolonial) contexts. The verbal prefixes a-, en-, es-enter Middle English by way of copies from Anglo-Norman (e.g. avengen ‘to take or get revenge’, enointen ‘to smear or daub’, eschaufen ‘to make warm’). To explore whether they were nativised, Percillier assumes that they have become productive in Middle English building new formations with native verbs. This is why Percillier calculates the potential productivity P (following Baayen 1992, 1993) for  Since copying stresses the non-identicality of copied material, Percillier opts for Johanson’s (2002) term for borrowing. 12

68 

C. Trips and P. A. Stokes

all three prefixes in hybrid formations. He shows that only en-occurs with native verb bases (e.g. enthrallen ‘to deprive of privileges’, enhongen ‘to hang someone’), whereas a- and es-occur in pre-existing formations. Thus, only en- has become productive building new formations with native bases,13 whereas the other two prefixes are used in formations copied from Anglo-Norman. Concerning (e.g. to take someone for a fool), Percillier found parallels in Anglo-Norman and Middle English. For example, ME chalengen, chaungen and prouen occurred with the prepositions as, for, into and to, respectively, just like their Anglo-Norman models (chalongier come, changier en, prover a / pour). The question is whether entire PSPs were copied from French or whether the copied French verbs are used with the English PSP markers. To answer this question, following Perek and Hilpert (2017), Percillier used a distributional semantic model to investigate possible semantic differences between PSPs with French-based and non-French-­ based (native) verbs. Percillier generated period vectors for all PSPs with French / non-French verbs with the four markers under scrutiny. The semantic information of the period vectors is used to examine semantic differences between the PSPs within a given period. By comparing the period vectors of the various PSPs within a given period via a hierarchical agglomerative clustering algorithm, these differences can be identified. Changes in the clusters from one period to another suggest changes in the semantic distinctions among the PSPs. Percillier focused on the sub-­ periods of M3 (1350–1420) and M4 (1420–1500) in the corpus, where all markers under scrutiny occur with French-based verbs. The dendrograms given in Fig. 3.2 show a clear semantic divide between the PSP with French-based and non-French based verbs in M3. This is marked with the box with a solid line, which contains only non-­ French verbs, and the box with a dashed line, which contains only French verbs. Percillier interprets this result as indicating an entire set of PSPs copied from French coexisting with the non-French set. This split is less clear in M4 because both French-based and non-French-based verbs occur with the marker as. Percillier takes this result to indicate a development towards nativisation of linguistic structures copied from  For a study of the copying of -able in Middle English, see Trips and Stein (2008).

13

3  From Original Sources to Linguistic Analysis 

69

Fig. 3.2  Dendrograms of PSPs in M3 and M4 (Percillier 2022: 41)

Anglo-Norman, which may then stretch into the Early Modern English period. The third linguistic feature that Percillier investigated are genitive / of-­ objects in Middle English. In Old English, partitive and causative objects were marked by the genitive and became increasingly marked by an of-­ phrase in Middle English. Some scholars (Mustanoja, 1960: 397–399; Visser, 2002: 360–361; Allen 2003: 8, 2008: Chapter 4; Myers 2009) have brought into play French influence, that is, French verbs taking de

70 

C. Trips and P. A. Stokes

Fig. 3.3  Types of of-objects per ME sub-period (Percillier 2022: 44)

that were copied to Middle English and gradually replaced the Old English genitive. Percillier notes that this assumption seems to be highly unlikely given that the of-phrase marking partitive and causative objects already existed at the onset of the Middle English period. Interestingly, of-objects that are neither causative nor partitive mainly occur with French verbs from M3 onwards, as Fig. 3.3 shows (Percillier labelled this type ‘other’). According to Percillier, the group with the ‘other’ objects comprised French-based and non-French-based verbs denoting, for example ‘ask’ or ‘stop’: (3) and preyde hym of paciense and prayed him of  patience ‘And asked him for patience’ (PPCME2, CMWYCSER,309.1472; Percillier 2022: 44)

3  From Original Sources to Linguistic Analysis 

71

(4) And, yyf þu  wylt Lord, þat  I sese of  wepyng, I prey þe and  if  you want lord  that I cease of weeping I pray you take me owt of þis world take me out of this world ‘And Lord, if you want me to stop weeping, I ask you to take me out of this world’ (PPCME2, CMKEMPE,142.3284; Percillier 2022: 45)

Overall, Percillier observed variation and competition between of and other markers (for, from in the cases at hand) in the second half of the Middle English period. Strikingly, non-French-based verbs behaved differently from French-based verbs: whereas the former set displayed the coexistence of competing forms, the latter set seem to have exclusively occurred with the of-phrase replicated from the French de-phrase. Gradually the of-phrase was replaced by the rival marker that supplanted the of-phrase of native verbs after the ME period (sixteenth century). Percillier concludes that Schneider’s Dynamic Model can be applied to the medieval contact situation because it can explain the emergence of the contact variety of Anglo-French (cf. Ingham 2012; Schauwecker 2019, 2022). Further, by extending the model to include the IDG language, it serves well to explain the structural properties and changes of Middle English. Moreover, it accounts for the nativisation and productivity of some properties (en-prefixation, PSPs) and the ephemeral occurrence of others (of-phrases replicated from de-phrases). In this way the model comprehensively accounts for the dynamics of this contact situation. The second study, by Percillier and Trips (2020), has two aims: first, it describes the extension of verb-lemmatisation and coding etymology to other Middle English corpora; second, it examines the rise of reflexives in Middle English and the potential role Old French might have played. It is well known to users of the PPCME2 that this corpus contains a data gap in the sub-period of M2 (1250–1350) (Truswell et al. 2018, 2019). Percillier and Trips (2020) note that, in comparison to the other sub-periods of the corpus, M2 contains only a fourth of the word counts, and in terms of diatopic variation, only texts from the southeast of England. This is problematic for any study that seeks to investigate the

72 

C. Trips and P. A. Stokes

Fig. 3.4  Word counts by ME sub-period for PPCME2, PLAEME and PCMEP (Percillier and Trips 2020)

diachronic development of Middle English by means of corpus-based analysis. Extending the database of linguistically annotated corpora of Middle English by adding The Parsed Corpus of Middle English Poetry (PCMEP; Zimmermann 2018) and A Parsed Linguistic Atlas of Early Middle English (PLAEME; Truswell et al. 2018) fills this gap in terms of diachrony and diatopy. This is illustrated by Figs. 3.4 and 3.5 from Percillier and Trips’s study. All three corpora have the same format now: Penn annotation, verb-­ lemmatisation and coding of verbs according to etymology (French / non-­French). Percillier and Trips (2020) demonstrate the benefit of using these annotated corpora collectively by means of a study of the rise of reflexivity in Middle English and possible influence from Old French. At the outset of their study, Percillier and Trips (2020) state that languages mark reflexivity in different ways. In a very general sense, a predicate can be called ‘reflexive’ whenever two of its arguments are coreferent, that is, refer to the same person. Present-Day English uses the set of pronouns with -self inflecting for person, number and gender to mark reflexive situations (see (5a–c)), also adnominally (see (5d)).

3  From Original Sources to Linguistic Analysis 

73

Fig. 3.5  Additions from PLAEME and PCMEP, text size shown in log10 (Percillier and Trips 2020)

(5)  a. I was muttering to myself.    b. You were muttering to yourself.    c. She was muttering to herself.     d. Ioanna herself wanted to become president.

This, however, is not the system that existed in previous times. The Old English system lacked reflexive pronouns, and instead personal pronouns had two functions: marking disjoint reference (like her, him, etc.) and coreference (like herself, himself, etc.; cf. Keenan 2009; König and Siemund 2000). Percillier and Trips (2020) give the` following examples to illustrate the use of personal pronouns in reflexive function: (6)  a. ða   gegyrede heo   hy   mid  hærenre tunecan    then dressed  shei.nom heri.acc with of-hair  tunic     ‘then she dressed herself in a tunic of hair’    (Mart 190 c. 875; Keenan 2009: 20)

74 

C. Trips and P. A. Stokes

b. forðæm  hi     him    ondrædað ða  frecenesse ðe  hi  ne  gesioð   because theyi.nom themi.dat fear      the danger    that  they neg see     ‘because they fear the danger that they do not see’    (CP 433 c. 880; Keenan 2009: 21)

Concerning intensifiers in Old English (self / seolf / sylf), the authors note that they were used in adnominal position (see (7a) ) but could also be combined with a personal pronoun in object position marking coreference (see (7b)): (7)  a. and hwæt Crist self tæhte  and his apostolas on þære niwan    and what Christ  self  taught  and  his  apostles   on  the  new    gecyðnisse    testament    “and what Christ himself and his apostles taught on the new Testament”     ([EGenPref 37]; König and Siemund 2000: 45)    b. se   Hælende sealde         hine       sylfne         for us    the lordi.nom.sg.m  gave himi.acc.sg.m self.acc.sg.m  for us     ‘The Saviour gave himself for us’     ([ELet 4 1129]; König and Siemund 2000: 45)

In Middle English, self occurred more and more often in compound forms (herself, himself, etc.) replacing the old monomorphemic forms (e.g. Crist self). This development then led to a situation where two alternative ways to mark reflexivity were in competition. Following Peitsara (1997: 280), Percillier and Trips (2020) call them the ‘simple strategy’ (use of personal pronouns) and the ‘self-strategy’ (use of self-compounds). In the course of the Middle English period, the paradigm of the reflexive self-forms developed, disambiguating the reflexive use of verbs (see van Gelderen 2000; Keenan 2009). König and Siemund (2000: 46) state, ‘[w]hether prosodic factors, influence from Romance (cf. elle-même), analogy or some other factor was responsible for this development is one of the many unsolved puzzles of the history of English’. Other authors clearly attribute this

3  From Original Sources to Linguistic Analysis 

75

development to French influence. For example, Einenkel (1916: §50), Mustanoja (1960: 502–503) and Visser (1963: 328) assume that the selfcompounds are calques of French expressions like (soi)-même or are built in analogy to French reflexive verbs which used first and second person pronouns to mark co-reference and had an invariable third person reflexive marker se just like in Modern French as the authors note: (8)  a. Je me réveille     b. Tu te réveilles.     c. Ioanna se réveille.     ‘I / you / Ioanna wakes up’.

Einenkel (1916: §138) gives the following example for French influence: These riottours … were set him in a tavern like Modern French ils se sont assis, where the personal pronoun him is used reflexively on the model of French se. Similarly, Peitsara (1997: 287) comments on one of her results of her case study based on the Helsinki Corpus (Rissanen et al. 1991): ‘[i]t could be supposed that the peak of frequency in ME3 might have to do with the grammaticalisation of the reflexive construction in English but, interestingly enough, it also coincides with the period of most profuse introduction of French’. Her results concerning the development of the ‘simple strategy’ and ‘self-strategy’ are shown in Fig. 3.6. As valuable as Peitsara’s work may be, Percillier and Trips (2020) stress that it is based on the Helsinki Corpus, which means that her results suffer from the same problem as those gained using the PPCME2, that is, the empirical gap in M2. On the basis of the collection of the verb-­ lemmatised Middle English corpora available, now Percillier and Trips (2020) are confident to gain more valid results than previously possible. They pursue the following research questions: (1) How often does the ‘simple strategy’ occur with native verbs and verbs copied from Old French? (2) How often does the ‘self-strategy’ occur with native verbs and verbs copied from Old French? To answer these questions, Percillier and

76 

C. Trips and P. A. Stokes

Fig. 3.6  Reflexive strategies in Middle English and Early Modern English per 1000 words (Peitsara 1997: 289)

Trips (2020) included both argument and non-argument reflexives,14 and compound and non-compound forms in their queries, as shown by the following examples:15 (9)  a. I wash me.     b. I fear me.       c. I wash myself.    d. I wash me self.

Figure 3.7 compares the results for the ‘simple strategy’ and the ‘self-­ strategy’ in the PPCME2 and the combination of all annotated Middle English corpora.

 This distinction is also called ‘reflexive’ and ‘inherent reflexive’ (cf. Steinbach 2002; see also Trips 2020). 15  The query for argument reflexives using the ‘simple strategy’ is limited to subject pronouns. For details about the limitations of querying these patterns, see Percillier and Trips (2020). 14

3  From Original Sources to Linguistic Analysis 

77

Fig. 3.7  Normalised frequencies of ‘simple’ and ‘self’ reflexive marking strategies in the PPCME2 only (left), and the PCMEP/PLAEME/PPCME2 combined (right) (Percillier and Trips 2020)

What is immediately striking, according to Percillier and Trips (2020), are discrepancies in the results of the PPCME2 alone with the M2 gap (left) and the combination of the corpora (right). In the PPCME2, the development of the ‘self-strategy’ seems to disappear in M2 only to surge in M3 (compare Peitsara’s curve above). The ‘simple strategy’ shows a steep rise in M2 only to drastically drop in M3. Overall, it seems that we have two changes in opposite directions. The picture that we gain from the results of the combined data is quite different: here, we see a steady increase of the ‘self-strategy’ and less drastic decrease of the ‘simple strategy’. On the whole, both strategies are stable for M1 and M2 and display considerable changes in M3 and M4. Percillier and Trips (2020) assume that this development is more probable than the one shown for the PPCME2 alone. The authors also examined French influence on the development of reflexive marking in Middle English, which is why their queries distinguished French-based and non-French-based verbs. But since the former type of verb entered the English language only from M2 onwards in increasing numbers, observed frequencies of the French-based verbs used

78 

C. Trips and P. A. Stokes

Fig. 3.8  Ratios of reflexive strategy use per 1000 verbs of the same etymological group in the combined data from the PCMEP/PLAEME/PPCME2 (Percillier and Trips 2020)

reflexively may reflect changes in their overall frequencies rather than changing characteristics of reflexive use with these verbs. This is why Percillier and Trips (2020) give the results of the reflexive use as ratios per 1000 verbs per etymological group, as shown in Fig. 3.8. The orange and blue lines show the development of the simple strategy and self-strategy, respectively. The dashed lines show the development of the French-based verbs, and the bold lines show the development of the non-French-based verbs. What becomes clear is that there are several differences between the group of French-based verbs and non-French-based verbs: for the former group, there is a steady increase of the ‘self-strategy’ from M1 to M3 and a decrease of the ‘simple strategy’ across all sub-­ periods (dashed lines). In contrast, the group of the non-French-based verbs exhibits a slight increase of the ‘self-strategy’ and a considerable decrease of the ‘simple strategy’ from M2 to M3. Furthermore, the French-based verbs display a higher ratio of verbs used reflexively than the non-French-based verbs. Percillier and Trips (2020) also took grammatical person (first, second, third person) into account; the results are given in Fig. 3.9.

3  From Original Sources to Linguistic Analysis 

79

Fig. 3.9  Ratios of reflexive strategy use per 1000 verbs of the same etymological group in the combined corpora accounting for grammatical person (Percillier and Trips 2020)

As in Fig. 3.8, the orange and blue lines show the development of the simple strategy and self-strategy, respectively. The dashed lines show the development of the French-based verbs, and the bold lines show the development of the non-French-based verbs. It is evident that changes in the way reflexivity is marked occur first with the third person for both verbs groups, French-based and non-French-based. But there are clear differences between the groups: French-based verbs show prominent changes in M2 when they start to enter English in larger numbers with the ‘self-strategy’ strongly increasing and the ‘simple strategy’ strongly decreasing. In contrast, non-French-based verbs display stagnation from M1 to M2 for both strategies (as already established above), and a less drastic increase of the ‘self-strategy’ and decrease of the ‘simple strategy’. How could the prominent role that the third person plays be explained? Percillier and Trips (2020) propose that it could be due to the fact that Old French possessed an unambiguous way to formally mark reflexivity (se or soi / elle / eux même(s)), something that the Middle English system completely lacked. This must have posed problems to speakers / writers, especially with regard to translating these markers (although they are

80 

C. Trips and P. A. Stokes

basically replicable in English as him / her / it / them self / selves). To come closer to a satisfying explanation, Perciller and Trips (2020) are of the opinion that it is desirable to gain further insights into language contact through translation. One step in this direction will be to add metadata of the texts of all Middle English corpora and to determine which texts are based on French originals. This would be crucial to examine translation effects. This information can also be used for a multifactorial statistical analysis where factors like text type can be taken into account as well. Overall, the study by Perciller and Trips (2020) clearly shows the benefit of verb-lemmatisation and the coding of etymology of all annotated corpora of Middle English presently available. Without it the authors could not have revealed the patterns that they found and the role the French verbs played in the development of reflexivity in Middle English. Very recently, the BASICS team enriched the three Middle English corpora mentioned above (PPCME2, PLAEME, PCMEP) with metadata, which is shown in the Table 3.2 for a part of the PPCME2. The following pieces of information were added to the corpora on the basis of the information given by the creators of the corpora: the period of composition, the period the manuscript was written (M1 = 1150–1250, M2 = 1250–1350, M3 = 1350–1420, M4 = 1420–1500), the date, the base of the text (either French or English), the dialectal area, the genre, the number of (non-)French-based verbs it contains, the overall word count of the text, and the longitude and latitude of the place where the text was written. By means of these metadata, studies can now be conducted exploring, for example, whether texts translated from French show more reflexive markers than texts where French influence can be excluded. Depending on the method one uses, instances of metadata can either be filtered out in spreadsheets, or tables in CSV-format can directly be imported into statistical programmes like R, where the metadata can be treated like variables.

M1

CMREYNAR

CMSAWLES

M3

M4

CMPURVEY

CMWYCSER

M3

CMPOLYCH

M1

M3

CMPETERB

CMVICES1

M1

CMMANDEV

M4

M3

CMMALORY

CMSIEGE

M4

CMAYENBI

Ayenbite of Inwyt Malory’s Morte Darthur Mandeville’s Travels Peterborough Chronicle John of Trevisa’s Polychronicon Purvey’s General Prologue to the Bible Caxton’s History of Reynard the Fox Sawles Warde

The Siege of Jerusalem Vices and Virtues English Wycliffite Sermons

M2

TextID

Title

Period composition

M3

M1

M4

M1

M4

M3

M3

M1

M3

M4

M2 1470

1340

1481

1388

1387

a1225 1225 (c1200) c1400 1400

c1225 1225 (?1200) c1500 1500

1481

c1388

a1387

en

en

en

en

en

en

en

en

fr

fr

fr

Religious Treatise Romance

Genre

Bible

West Midlands West Midlands East Midlands East Midlands

Religious Treatise Sermon

Romance

Homily

East Fiction Midlands

Southern

West Midlands East Travelogue Midlands East History Midlands Southern History

Kentish

Ms date simplif. Base Dialect

?a1425 1425 (c1400) c1150 1150

a1470

1340

Period manuscript Date

Table 3.2  Enrichment of the PPCME2, PLAEME, PCMEP with metadata

1458

69

149

21

189

1051

1010

4

810

1278

449

FBVs

5021

3037

884

495

963

3018

3763

733

4004

6096

4135

Latitude

52.5

0.1330202 51.755882 56,168 −1.008333

28,061

52.5

7757 −1.833333

52.18756

39,836 −0.142659

52.367426

51.683936

46,326 −2.449662

4304 −2.696323

52.574272

7333 −0.273244

52.98

51.917761

51,715 −0.273244

8799 −0.75

52.466352

1.1196628 51.31521

Longitude

60,297 −1.434

45,945

Word nFBVs count

3  From Original Sources to Linguistic Analysis 

81

82 

C. Trips and P. A. Stokes

7 Conclusion The discussion presented here shows the wide range and large number of tools and corpora that are available for researchers. This landscape is always changing, with projects constantly coming and going, and so we emphasise again that this chapter is by no means complete, and that any presentation of the state of the art will rapidly become out of date. At the same time, it is also striking how long-lived some of these examples are, particularly in this age of short-term funding and planned technological obsolescence. Work on digital dictionaries and corpora in medieval languages began in the 1940s, with numerous active projects by the 1960s (e.g. Cameron et al. 1970), some of which are still running and remain accessible today. This longevity stands as testament to the vision of the projects’ founders, but also to the value of complying to standards and best practice from the start, even when these may slow progress in the short term, and even though standards can often seem opposed to research and innovation almost by definition. Producing ad hoc solutions is always tempting and can indeed lead to faster results in the short term, but this is almost always a false economy in practice since the results are much more difficult to sustain and so unlikely to survive in the longer term (Stokes 2009; for further discussion of sustainability, see especially Smithies et al. 2019). This question of sustainability is also closely linked to that of interoperability and interchange, that is, the possibility of using existing tools and datasets created by others, combining different datasets, and creating viable workflows from different tools. Here again, best practice in standards and documentation are essential, as is allowing for the import and export of data in different standard formats. Ideally, if this is done, then one can chain together different tools to create a pipeline which can process a given corpus (or combination of corpora) with the need for very little technical expertise. In practice, however, it is often necessary to write custom code to transform data from one format to another, or one idiosyncratic implementation of a standard to another. Indeed, it is significant that the single largest category of lesson in the Programming Historian website is ‘Transformation’, suggesting that the need for programming skills to transform data from

3  From Original Sources to Linguistic Analysis 

83

one format to another is noticeably greater than that of other categories such as ‘Analyse’, ‘Present’, ‘Acquire’ and ‘Sustain’.16 This is even more the case for multilingualism, since the corpora and even the tools tend to be developed for one linguistic context, and so combining different datasets becomes essential. In this respect, arguably one of the more important developments in recent years has been the increasing availability of and attention to standards and interchange formats, such as TEI (now in its thirty-fifth year) and IIIF, but also increasing numbers of tools that work with them. This is further enhanced by the increasing emphasis on the openness and availability of data, particularly those funded by research bodies, with more and more guidelines on best practice, repositories for long-term storage, and requirements for documentation, openness, data management plans and compliance with standards (Pierazzo and Stokes 2022). This said, it remains also the fact that the same funding bodies are generally less and less open to supporting projects for the long term, tending for the most part to award relatively short grants of around five years at most. There are some exceptions, such as the European research infrastructure projects that include CLARIN, a digital infrastructure for language resources for research, but these are necessarily few and are usually not able or sufficient to sustain the development of the corpora themselves. Nevertheless, as the second part of this chapter shows very clearly, we already have a significant body of material that has already enabled important research with concrete results. Granted much more remains to be done, particularly for work on multilingualism, but the route ahead is clear and the level and quality of activity remains very promising indeed.

References Allen, Cynthia. 2003. Deflexion and the Development of the Genitive in English. English Language and Linguistics 7: 1–28. ———. 2008. Genitives in Early English: Typology and Evidence. Oxford: Oxford University Press.  The Programming Historian > ‘Lessons’. At the time of writing, the categories have thirty-three, twenty, twenty-one, twelve and two lessons respectively in English; other languages have different numbers of lessons but follow the same broad pattern. 16

84 

C. Trips and P. A. Stokes

Andrist, Patrick, Paul Canart, and Marilena Maniaci. 2013. La syntaxe du codex: Essai de codicologie structurale, Bibliologia, 36. Turnhout: Brepols. Baayen, R. Harald. 1992. Quantitative Aspects of Morphological Productivity. In Yearbook of Morphology 1991, ed. Geert Booij and Jaap van Marle, 109–149. Dordrecht: Kluwer Academic Publishers. ———. 1993. On Frequency, Transparency, and Productivity. In Yearbook of Morphology 1992, ed. Geert Booij and Jaap van Marle, 181–208. Dordrecht: Kluwer Academic Publishers. Blockley, Marion. 1982. Further Addenda and Corrigenda to N.R. Ker’s Catalogue. Notes and Queries 228 (n.s. 29): 1–3. Cameron, Angus, Roberta Frank, and John Leyerle, eds. 1970. Computers and Old English Concordances. Toronto: University of Toronto Press. Clérice, Thibault. 2020. Evaluating Deep Learning Methods for Word Segmentation of Scripta Continua Texts in Old French and Latin. Journal of Data Mining and Digital Humanities [Online] 5581. https://doi. org/10.46298/jdmdh.5581. Da Rold, Orietta, Takako Kato, Mary Swan, and Elaine Treharne, eds. 2010. The Production and Use of English Manuscripts 1060 to 1220. Leicester. https:// www.le.ac.uk/english/em1060to1220/index.html. Davis, Godfrey R. C. 2010. Medieval Cartularies of Great Britain and Ireland, revised by Claire Breay, Julian Harrison, and David M. Smith. London: British Library. Einenkel, Eugen. 1916. Geschichte der englischen Sprache, II: Historische Syntax. Strasbourg: Trübner. van Gelderen, Elly. 2000. A History of the English Reflexive Pronouns: Person, Self, Interpretability, Linguistics Today, 39. Amsterdam: John Benjamins. Gneuss, Helmut. 2001. Handlist of Anglo-Saxon Manuscripts: A List of Manuscripts and Manuscript Fragments Written or Owned in England up to 1100, Medieval and Renaissance Texts and Studies, 241. Tempe, AZ: Arizona Center for Medieval and Renaissance Studies. ———. 2003. Addenda and Corrigenda to the Handlist of Anglo-Saxon Manuscripts. Anglo-Saxon England 32: 293–305. ———. 2011. Second Addenda and Corrigenda to the Handlist of Anglo-­ Saxon Manuscripts. Anglo-Saxon England 40: 293–306. Gneuss, Helmut, and Malcolm Lapidge. 2014. Anglo-Saxon Manuscripts: A Bibliographical Handlist of Manuscripts and Manuscript Fragments Written Or Owned in England up to 1100, Toronto Anglo-Saxon Series, 15. Toronto: University of Toronto Press.

3  From Original Sources to Linguistic Analysis 

85

Haeberli, Eric. 2018. Syntactic Effects of Contact in Translations: Evidence from Object Pronoun Placement in Middle English. English Language and Linguistics 22: 301–321. Ingham, Richard. 2012. The Transmission of Anglo-Norman: Language History and Language Acquisition. Amsterdam: John Benjamins. Jefferson, Judith A., and Ad Putter, eds. 2013. Multilingualism in Medieval Britain (c. 1066-1520): Sources and Analysis. Turnhout: Brepols. Jenkyns, Joy. 1991. The Toronto Dictionary of Old English Resources: A User’s View. Review of English Studies 42 (167): 380–416. Johanson, Lars. 2002. Contact-Induced Change in a Code-Copying Framework. In Language Change: The Interplay of Internal, External and Extra-Linguistic Factors, ed. Mari C. Jones and Edith Esch, 285–313. Berlin: de Gruyter. Keenan, Edward L. 2009. Linguistic Theory and the Historical Creation of English Reflexives. In Historical Syntax and Linguistic Theory, ed. Paola Crisma and Giuseppe Longobardi, 17–40. Oxford: Oxford University Press. Ker, Neil R. 1957. Catalogue of Manuscripts Containing Anglo-Saxon. Oxford: Clarendon Press. ———. 1976. A Supplement to Catalogue of Manuscripts Containing Anglo-­ Saxon. Anglo-Saxon England 5: 121–131. Kestemont, Mike, Vincent Christlein, and Dominique Stutzmann. 2017. Artificial Paleography: Computational Approaches to Identifying Script Types in Medieval Manuscripts. Speculum 92: 86–109. König, Ekkehard, and Pieter Siemund. 2000. The Development of Complex Reflexives and Intensifiers in English. Diachronica 17: 39–84. Kroch, Anthony, and Ann Taylor. 1997. Verb Movement in Old and Middle English: Dialect Variation and Language Contact. In Parameters of Morphosyntactic Change, ed. A. van. Kemenade and N. Vincent, 297–325. Cambridge: Cambridge University Press. Levin, Beth. 1993. English Verb Classes and Alternations: A Preliminary Investigation. Chicago: University of Chicago Press. Mustanoja, Tauno F. 1960. A Middle English Syntax, Part I: Parts of Speech, Mémoires de la Société Néophilologique de Helsinki, 23. Helsinki: Société Néophilologique. Myers, Sara Mae. 2009. The Evolution of the Genitive Noun Phrase in Early Middle English. Unpublished MPhil thesis, University of Glasgow. Page, Raymond I. 1992. On the Feasibility of a Corpus of Anglo-Saxon Glosses: The View from the Library. In Anglo-Saxon Glossography: Papers Read at the International Conference Held in the Koninklijke Academie voor Wetenschappen Letteren en Schone Kunsten van België, Brussels, 8 and 9 September 1986, ed. René Derolez, 77–96. Brussels: Paleis der Academiën.

86 

C. Trips and P. A. Stokes

Parkes, Malcolm B. 1969. English Cursive Bookhands 1250–1500. Oxford: Clarendon Press. Peitsara, Kirsti. 1997. The Development of Reflexive Strategies in English. In Grammaticalization at Work: Studies of Long-Term Developments in English, ed. Matti Rissanen, Merja Kytö, and Kirsi Heikkonen, 277–370. Berlin: Mouton de Gruyter. Pelteret, David A.E. 1990. Catalogue of English Post-Conquest Vernacular Documents. Woodbridge: Boydell. Percillier, Michael. 2018. A Toolkit for Lemmatising, Analysing and Visualising Middle English Data. In Proceedings of the Second Workshop on Corpus-Based Research in the Humanities CRH-2, ed. Andrew U. Frank, Christine Ivanovic, Francesco Mambrini, Marco Passarotti, and Caroline I. Sporleder, 153–160. Vienna: TU Wien. ———. 2019. Dynamic Modelling of Medieval Language Contact: The Case of Anglo-Norman and Middle English. In Diachrone Migrationslinguistik: Mehrsprachigkeit in historischen Sprachkontaktsituationen, ed. R. Schöntag and S. Massicot, 79–99. Berlin: Peter Lang. ———. 2020. Allostructions, Homostructions or a Constructional Family?: Changes in the Network of Secondary Predicate Constructions in Middle English. In Nodes and Networks in Diachronic Construction Grammar, ed. L. Sommerer and E. Smirnova, 214–242. Amsterdam: John Benjamins. https://doi.org/10.1075/cal.27.06per ———. 2022. Adapting the Dynamic Model to Historical Linguistics: Case Studies on the Middle English and Anglo-Norman Contact Situation. In English Historical Linguistics: Historical English in Contact, ed. Bettelou Los, Chris Cummins, Lisa Gotthard, Alpo Honkapohja, and Benjamin Molineaux, 5–33. Amsterdam: John Benjamins. Percillier, Michael, and Carola Trips. 2020. Lemmatising Verbs in Middle English Corpora: The Benefit of Enriching the Penn-Helsinki Parsed Corpus of Middle English 2 (PPCME2), the Parsed Corpus of Middle English Poetry (PCMEP), and A Parsed Linguistic Atlas of Early Middle English (PLAEME). In Proceedings of LREC2020, 7170–7178. https://www.aclweb.org/anthology/2020.lrec-­1.886. Accessed 29 September 2021. Perek, Florent, and Martin Hilpert. 2017. A Distributional Semantic Approach to the Periodization of Change in the Productivity of Constructions. International Journal of Corpus Linguistics 22: 490–520.

3  From Original Sources to Linguistic Analysis 

87

Pierazzo, Elena, and Peter A. Stokes. 2022. Old Books, New Books and Digital Publishing. In The Bloomsbury Handbook to the Digital Humanities, ed. James O’Sullivan, 233–44. London: Bloomsbury Academic. Robinson, Peter, and Elizabeth Solopova. 1993. Guidelines for Transcription of the Manuscripts of the Wife of Bath’s Prologue. In The Canterbury Tales Project: Occasional Papers, ed. Norman F. Blake and Peter Robinson, 19–52. Oxford: Oxford University Computing Services, Office for Humanities Communication. Sawyer, Peter H., ed. 1968. Anglo-Saxon Charters: An Annotated List and Bibliography, Royal Historical Society Guides and Handbooks, 8. London: Royal Historical Society. Schauwecker, Yela. 2019. Le faus françeis d’Angleterre en tant que langue seconde? Quelques phénomènes syntaxiques indicatifs: The faus franceis d’Angleterre as an L2? – Some Distinctive Syntactic Features. Revue des Langues Romanes 123: 45–68. ———. 2022. Anglo-Französisch als Zweitsprache: Eine quantitative Untersuchung von Sprachkontakteinflüssen anhand der Lexikalisierung von Bewegungsereignissen. Unpublished Habilitation thesis, University of Stuttgart. Schneider, E. W. 2003. The Dynamics of New Englishes: From Identity Construction to Dialect Birth. Language 79: 233–281. ———. 2007. Postcolonial English: Varieties around the World. Cambridge University Press. https://doi.org/10.1017/CBO9780511618901 Scragg, Donald G. 2012. A Conspectus of Scribal Hands Writing English 960–1100, Publications of the Manchester Centre for Anglo-Saxon Studies, 11. Cambridge: D. S. Brewer. Smithies, James, Carina Westling, Anna-Maria Sichani, Pamela Mellen, and Arianna Ciula. 2019. Managing 100 Digital Humanities Projects: Digital Scholarship and Archiving in King’s Digital Lab. Digital Humanities Quarterly 13/1. http://www.digitalhumanities.org/dhq/vol/13/1/000411/000411. html. Accessed 14 December 2021. Steinbach, Markus. 2002. Middle Voice: A Comparative Study in the Syntax-­ Semantics Interface of German, Linguistics Today, 50. Amsterdam: John Benjamins. Stokes, Peter A. 2009. The Digital Dictionary. Florilegium 26: 37–69. ———. 2011. The Problem of Grade in Post-Conquest Vernacular Minuscule. New Medieval Literatures 13: 23–47. ———. 2020a. Cambridge, Corpus Christi College, 367 Part II: A Study in (Digital) Codicology. In Medieval Manuscripts in the Digital Age, ed. Benjamin

88 

C. Trips and P. A. Stokes

Albritton, Georgia Henley, and Elaine Treharne, 64–73. Abingdon: Routledge. ———. 2020b. Palaeography, Codicology and Stemmatollogy. In Handbook of Stemmatology, ed. Philipp Roelli, 46–56. Berlin: De Gruyter. Stokes, Peter A., and Geoffroy Noël. 2019. Exon Domesday: Méthodes numériques appliquées à la codicologie pour l’étude d’un manuscrit anglo-normand. Tabularia [En ligne]. https://doi.org/10.4000/tabularia.4118. Stokes, Peter A., and Elena Pierazzo. 2009. Encoding the Language of Landscape: XML and Databases at the Service of Anglo-Saxon Lexicography. In Perspectives on Lexicography in Italy and Europe, ed. Silvia Bruti, Roberta Cella, and Marina F. Albert, 203–239. Newcastle upon Tyne: Cambridge Scholars. Stutzmann, Dominique. 2016. Clustering of Medieval Scripts through Computer Image Analysis: Towards an Evaluation Protocol. Digital Medievalist [On line] 10. https://doi.org/10.16995/dm.61. Taylor, Ann. 2008. Contact Effects of Translation: Distinguishing Two Kinds of Influence in Old English. Language Variation and Change 20: 341–365. Trips, Carola. 2002. From OV to VO in Early Middle English. Amsterdam: John Benjamins. ———. 2020. Impersonal and Reflexive Uses of Middle English Psych Verbs under Contact Influence with Old French. Linguistics Vanguard 6 (Special Issue on Language Contact, ed. N. Lavidas and A. Bergs). https://doi. org/10.1515/lingvan-­2019-­0016. Trips, Carola, and Achim, Stein. 2008. Was French -able Borrowable? A Diachronic Study of Word-Formation Processes Due to Language Contact. In English Historical Linguistics 2006, II: Lexical and Semantic Change, ed. Richard Dury and Maurizio Gotti, 217–41. Amsterdam/Philadelphia: John Benjamins. Trotter, David, ed. 2000. Multilingualism in Later Medieval Britain. Cambridge: D.S. Brewer. Truswell, Robert, Rhona Alcorn, James Donaldson, and Joel Wallenberg. 2019. A Parsed Linguistic Atlas of Early Middle English. In Historical Dialectology in the Digital Age, ed. Rhona Alcorn, Johanna Kopaczyk, Bettelou Los, and Benjamin Molineaux, 19–38. Edinburgh: Edinburgh University Press. Visser, Fredericus T. 1963. An Historical Syntax of the English Language. Leiden: Brill. ———. 2002. An Historical Syntax of the English Language, 4th edn. Leiden: Brill.

3  From Original Sources to Linguistic Analysis 

89

Digital Resources and Websites (Last Verified 28 February 2022) Archetype, Version 2.8. 2022. London: King’s College. https://github.com/kcl-­ ddh/digipal. BASICS Toolkit, Release 5. 2021. Created by Michael Percillier. University of Mannheim. http://basics-­toolkit.spdns.org/. Benskin, Michael, Margaret Laing, Vasilios Karaiskos, and Keith Williamson. 2013. An Electronic Version of A Linguistic Atlas of Late Mediaeval English. Edinburgh: University of Edinburgh. http://www.lel.ed.ac.uk/ihd/elalme/ elalme.html. BiblioPhilly: An Interface for the Bibliotheca Philadephiensis Project, Philadelphia Area Consortium of Special Collections Libraries (PACSCL). Philadelphia: University of Pennsylvania. https://bibliophilly.library.upenn.edu. Biblissima IIIF Collections. Paris: Campus Condorcet. https://iiif.biblissima.fr/ collections/. Biblissima: Observatoire du patrimoine écrit du Moyen Âge et de la Renaissance. Paris: Campus Condorcet. https://projet.biblissima.fr. Bosworth, Joseph. 2014. An Anglo-Saxon Dictionary Online, Ed. by Thomas Northcote Toller, Christ Sean, and Ondřej Tichy. Prague: Faculty of Arts, Charles University. https://bosworthtoller.com/. Boudams: Le Boucher d’Amsterdam, tokenizer, Clean Release. 2019. Thibault Clérice and Jean-Baptiste Camps. https://github.com/PonteIneptique/ boudams. Calamari: Line-Based ATR Engine, Version 2.1.4. 2021. https://github.com/ Calamari-­OCR/calamari. Cambridge Digital Library. University of Cambridge. https://cudl.lib.cam.ac.uk/. CLARIN: The Research Infrastructure for Language as Social and Cultural Data. https://www.clarin.eu. CoRD: Corpus Resource Data Base. Helsinki: Research Unit for Variation, Contacts and Change in English. https://varieng.helsinki.fi/CoRD/. DigiPal: Digital Resource and Database of Palaeography, Manuscript Studies and Diplomatic. 2011–14. London: King’s College. http://digipal.eu. eScriptorium, Version 0.10.4. 2022. Paris: École Pratique des Hautes Études – Université Paris Sciences et Lettres. https://gitlab.com/scripta/escriptorium/. Exon: The Domesday Survey of South-West England. 2018. Ed. by Peter A. Stokes, Studies in Domesday, gen ed. Julia Crick. London: King’s College. http:// www.exondomesday.ac.uk.

90 

C. Trips and P. A. Stokes

France Angleterre 700–1200. Paris and London: Bibliothèque Nationale de France and the British Library. https://manuscrits-­france-­angleterre.org/. Gallica. Paris: Bibliothèque Nationale de France. https://gallica.bnf.fr/. Graphoskop: Aide à l’expertise paléographique, Maria Gurrado. Paris: Institut de Recherche et d’Histoire des Textes. https://www.irht.cnrs.fr/index.php/fr/ ressources/sites-­web-­outils-­corpus/graphoskop. Kraken, Version 4.0. 2022. Benjamin Kiessling. Paris: École pratique des Hautes Études – Université PSL. https://github.com/mittagessen/kraken. Kroch, Anthony, and A. Taylor. 2000. The Penn-Helsinki Parsed Corpus of Middle English, 2nd edn, Release 3. Philadelphia: University of Pennsylvania. https:// www.ling.upenn.edu/hist-­corpora/PPCME2-­RELEASE-­4/index.html. Laing, Margaret. 2013–. A Linguistic Atlas of Early Middle English, 1150–1325, Version 3.2. Edinburgh: University of Edinburgh. http://www.lel.ed.ac.uk/ ihd/laeme2/laeme2_framesZ.html. LiLa: Linking Latin; Building a Knowledge Base of Linguistic Resources for Latin. https://lila-­erc.eu/. Middle English Dictionary. 1952–2001. Ed. by Robert E. Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner. Ann Arbor: University of Michigan Press. https://quod.lib. umich.edu/m/med/. Mirador: Fully Featured IIIF Viewer. https://projectmirador.org. Models of Authority: Scottish Charters and the Emergence of Government. 2014–2017. London: King’s College. https://www.modelsofauthority.ac.uk/. OCR-D: DFG-Funded Initiative for Optical Character Recognition Development. https://ocr-­d.de. Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner (1989); 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Parker on the Web: Manuscripts in the Parker Library at Corpus Christi College, Cambridge, Version 2.1. Stanford: Stanford University Libraries. https:// parker.stanford.edu/. Programming Historian. 2022. https://programminghistorian.org. Pyrrha: A Language-Independant Post-Correction Ap for POS and Lemmatization, Version 3.0.0. 2021. Paris: École nationale des Chartes – Université Paris Science et Lettres. https://github.com/hipster-­philology/pyrrha.

3  From Original Sources to Linguistic Analysis 

91

Randall, Beth. 2010. CorpusSearch, Version 2.003.00. https://sourceforge.net/ projects/corpussearch/files/. Rissanen, Matti, et al. 1991. The Helsinki Corpus of English Texts. Helsinki: Department of Modern Languages, University of Helsinki. SegmOnto. https://github.com/SegmOnto/. Tesseract Open Source OCR Engine, Version 5.0.1. 2022. https://github.com/ tesseract-­ocr/tesseract. The Digital Bodleian. University of Oxford. https://digital.bodleian.ox.ac.uk/. The International Image Interoperability Framework. https://iiif.io. The James Catalogue of Western Manuscripts [Online Version]. Cambridge: Trinity College. https://mss-­cat.trin.cam.ac.uk/. The Universal Viewer. https://universalviewer.io. Transkribus. 2021. READ Coop SCE. https://readcoop.eu/transkribus/. Truswell, Robert, Rhona Alcorn, James Donaldson, and Joel Wallenberg. 2018. A Parsed Linguistic Atlas of Early Middle English. Edinburgh: University of Edinburgh. http://www.amc.lel.ed.ac.uk/amc-­projects-­hub/ project/p-­laeme-­a-­parsed-­linguistic-­atlas-­of-­early-­middle-­english/. VisigothicPal: ViGOTHIC: Towards a Typology of Visigothic Script: The Beatus British Library Add. ms. 11695 and Its Potential for Dating and Localising Visigothic Script Manuscripts. 2015–2017. http://visigothicpal.com. Zimmermann, Richad. 2018. The Parsed Corpus of Middle English Poetry. http:// pcmep.net/.

Part II Medieval Multilingualism and Lexical Change

4 Contact-Induced Lexical Effects in Medieval English Richard Dance, Philip Durkin, Carole Hough, and Heather Pagan

1 Introduction This overview chapter explores the roles of word studies in the investigation of multilingualism in medieval Britain. Drawing on a variety of research methods and resources, including insights from onomastics, etymology and lexicography, it introduces approaches to the collection, classification and analysis of lexical material. It surveys the complex

R. Dance University of Cambridge, Cambridge, UK e-mail: [email protected] P. Durkin (*) Oxford English Dictionary, Oxford, UK e-mail: [email protected] C. Hough University of Glasgow, Glasgow, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_4

95

96 

R. Dance et al.

evidence that word forms and meanings provide for contact, borrowing and diffusion in the medieval period, and discusses problems ranging from identification to periodisation and integration.

2 Place-Names Place-names occupy a unique position within language, in that they originate as meaningful expressions but develop over time into semantically bleached identifiers. They tend to be readily taken over by incoming speech communities, partly because their referents have not previously been encountered and therefore have no pre-existing identifiers, and partly because they can be used as labels without knowledge of semantic meaning. They thus offer both challenges and opportunities for the investigation of medieval multilingualism. Opportunities, because the toponymicon represents a palimpsest of historical languages; challenges, because placenames emerge and develop over a period of time, during which the different linguistic strands may become entangled and are difficult to unpick. Most place-names originate in spoken language, and may circulate orally for centuries before being written down. Again, this offers both challenges and opportunities. Opportunities, because the toponymicon provides evidence for prehistoric languages, including a demotic register mostly absent from written records; challenges, because it may be difficult to reconstruct the original form from later spellings. Among the most ancient names are those of major rivers, whose shared prehistoric roots in Britain and continental Europe point to a hydronymic system predating the splitting of Indo-European into daughter language families and the movement of speakers of these languages from the continent to Britain. One such root is *ad(u) / *adro ‘water-course, stream’, which forms rivernames in Austria (Adra, Attel), Germany (Eder), Italy (Adda), Latvia

H. Pagan University of Westminster, London, UK e-mail: [email protected]

4  Contact-Induced Lexical Effects in Medieval English 

97

(Adula), Poland (Oder, Odla) and possibly Scotland (Adder).1 The etymology of the Adder is uncertain because the earliest spellings date from the period of Northumbrian settlement in southern Scotland and are also consistent with a derivation from Old English (OE) ēdre ‘river’.2 The river-name must pre-date its earliest record (fluuius Edre, c.950 x 1050) as it forms the first (specific) element of two early settlement-­ names. Edrington ‘farm associated with the Adder’ and Edrom ‘settlement on the Adder’ both contain second (generic) elements associated mainly with the seventh and eighth centuries.3 What is uncertain is whether the rivername was created by the first wave of Northumbrian incomers in the seventh century, or taken over by them from the previous inhabitants. The two possibilities may not be mutually exclusive. Place-­names are particularly prone to folk-etymology, where an unfamiliar term is re-shaped by another speech community. Just as the river-name has now become homonymous with the unrelated word adder, perhaps reflecting a conceptualisation of the river as a long, powerful creature slithering through the landscape, so too the incoming Northumbrian speakers may have associated a pre-existing river-name with OE ēdre ‘river’. Recorded spellings and further discussion of the Adder and other names mentioned in this section can be found in the Berwickshire Place-­ Name Resource (BPNR), an online survey of all 1,224 place-names on the Ordnance Survey Landranger map series (1:50,000) for the historical county of Berwickshire in the Scottish Borders. Like the names used in the Welsh Marches discussed by Callander (this volume), they reflect competing influences across the present-day political divide. As well as the prehistoric language also found in other ancient river-names such as the Ale Water, Eden Water, and possibly Armet Water, six historical languages are represented: Brittonic, French, Gaelic, Old English, Old Norse and Scots. Some may have been spoken alongside each other, while others belong to different time periods, but again they can be difficult to untangle. French and Old Norse each accounts for a single place-name,  See, for instance, Nicolaisen (2001: 235–246, esp. 236–239). Separate branches of the river are now known as the Blackadder Water and Whiteadder Water. 2  The same word also means ‘vein’, reflecting the landscape is a body metaphor found in both the lexicon and onomasticon (see Hough 2016). 3  OE -ingtūn ‘farm associated with’; OE hām ‘settlement’. 1

98 

R. Dance et al.

neither of which reflects the presence of native speakers. Belville is a nineteenth-­century pun on the Scots name of nearby Belchester, reinterpreting the specific element as French belle ‘beautiful’, while Corsbie ‘village with a cross’ appears to have been created by analogy with the same name in areas of Scandinavian settlement elsewhere in Scotland and England. By contrast, Old English and its daughter language Scots are represented by 100 and 295 elements, respectively. Many appear in multiple place-names, but there is some overlap. The BPNR Element Glossary lists nineteen place-names under OE tūn ‘farm, settlement’ and eighteen under the Scots reflex toun. However, Manderston, Oxton and Preston appear in both lists, as their date of formation is ambiguous between the Northumbrian and Scots-speaking periods. First recorded in 1367 (Mandredestonam), c.1190 (Vlchilestun) and 1274 x 1275 (Prestun), their creation may be contemporary with these dates or many centuries earlier. Similarly, four of the five place-names listed under OE hlāw ‘tumulus, rounded hill’ are also listed under the Scots reflex law, together with fifty-­ one others. Since most place-names develop over time, with an appellatival description gradually becoming fossilised as a name, the process of creation may even have spanned the notional division between Northumbrian Old English and Older Scots. As discussed elsewhere in this volume, periodisation is a problematic concept. As regards Oxton, the specific element is an Anglo-Scandinavian personal name current in southern Scotland during the late twelfth and early thirteenth centuries, which may support a date of formation around that time. A further complexity is that although the personal name is linguistically from Old Norse (Úlfkell), this need not imply the presence of Scandinavian speakers, since names from disparate origins could be—and still are—quickly integrated into the common namestock. Indeed, Berwickshire was chosen for the survey partly because it would allow the Old English stratum to be studied relatively free from the Norse influence which, as Versloot (this volume) demonstrates, makes it so difficult to unpick West Germanic from North Germanic in areas of Scandinavian settlement, a topic explored further in Sect. 3. The issue of language attribution becomes more acute where place-­ names antedate other forms of evidence. According to BPNR, Fogo is first recorded 1139 x 1152 (foghou), so that the specific element is

4  Contact-Induced Lexical Effects in Medieval English 

99

ambiguous between OE *fogga and Scots fog, both referring to long grass left uncut for winter grazing, while the generic is ambiguous between OE hōh and Scots heuch, both referring to a steep hill. Whereas *fogga is unattested in English outside place-names until c.1400 (Oxford English Dictionary, OED, s.v. fog, n.1), fog is first attested in Scots about a century later (Dictionary of the Older Scottish Tongue, s.v. fog, n.). Clearly, Fogo provides earlier dating evidence for one of these dictionary entries, but which? The place-name may also throw light on the etymology of the term. OED describes it as ‘probably a borrowing from early Scandinavian’, citing Norwegian fogg ‘long-stalked, weak, scattered grass or heather, typically growing on wet ground’. The occurrence in an early place-name outside the areas of Scandinavian settlement raises the possibility that, as Versloot argues for other terms explained by OED as Norse loanwords, there may be a native West Germanic origin. In this case, though, there are no known West Germanic cognates, and the ulterior etymology of the word is opaque.4 In other instances, place-names provide unique evidence for terms or uses of terms. BPNR attributes Bemersyde to OE bēmere ‘trumpeter’, used in a transferred sense ‘bittern’ known from English place-names although unrecorded in other sources. As the earliest spelling (Bemersyd) dates from 1183 x 1212, it is fully possible that Bemersyde was created during the Older Scots period. If so, it would appear that only the transferred sense entered Older Scots, since the major dictionaries of Scots (Dictionaries of the Scots Language, DSL; and Concise Scots Dictionary, CSD) record the root beme ‘trumpet’ but not the agent noun in either its literal or transferred sense. Indeed, there is no evidence that the transferred sense was ever current outside the toponymicon. In recent years, there has been a growing recognition that the onomasticon is a separate register of language, overlapping with but not identical to the lexicon (see, e.g., Hough 2010, Cavill 2017). How should this and other specialised registers be treated by lexicographers? This is another question to which there is no straightforward answer, but which is explored further in contributions to this volume (e.g. Sylvester and Tiddeman’s chapter).

4

 In the Gersum project’s database (s.v. fogge), it is classified as a Type D1c.

100 

R. Dance et al.

As well as issues relating to diachronic strata, place-names illustrate some of the rich complexities of interaction between languages spoken alongside each other by neighbouring speech communities. Some of these interactions manifest in loanwords. For instance, Longformacus contains Gaelic longphort ‘encampment or hut, dwelling’, recently suggested to have originated as a loanword from OE lang-port ‘long market-­ place’ (Taylor with Márkus 2006–2012, v, 430–431). Although Old English is much more common in Berwickshire than Gaelic, a Gaelic etymology is supported by the characteristic element order of generic preceding specific, rather than the Germanic element order of specific preceding generic. Also suggestive—though not conclusive, as explained earlier—is that the specific is a personal name of Celtic origin. More complex is the development of Scots *carr ‘(coastal) rock’, found in a number of Scottish place-names, including four in BPNR: East Carr, Ebb Carrs, Heathery Carr and West Carr. None is recorded before the eighteenth century, and all have Scots specifics, so a Scots origin is beyond reasonable doubt. Again, however, the term has no entry in DSL or CSD. Its ancestor, OE carr ‘rock’, is believed to be a loanword from Brittonic (Parsons and Styles 2000: s.v. carr), and only the late date of the place-names, together with their Scots specifics, rules out a derivation from Brittonic itself. Notably, all four place-names refer to coastal rocks, although both the Brittonic and Old English terms have the more general meaning ‘stone, rock’ (James 2023: ii, s.v. *carr; Dictionary of Old English, DOE, s.v. carr). The inference is that semantic shift has taken place. As Gelling and Cole (2014) have shown, comparison of landscape features named from the same term can reveal subtle nuances in meaning.5 Place-­ name evidence thus challenges synonymy within the semantic field of topography, as does material presented by Sylvester and Tiddeman (this volume) in connection with the language of trade. Simplex place-names, containing a generic only, can be particularly problematic. With no structural or linguistic evidence available from a specific, it may be impossible to differentiate between alternative linguistic origins. OE dūn ‘hill’ is cognate with Gaelic dùn ‘(fortified) hill’, so  For instance, they demonstrate that each of the Old English hill-terms discussed here—hlāw, hōh and dūn—denotes a different type of hill. 5

4  Contact-Induced Lexical Effects in Medieval English 

101

either is possible as the etymon of Duns. If from Gaelic, the English plural -s, consistently recorded in historical spellings from c.1150 (Duns) onwards is unexpected, but it would also be unexpected with the feminine noun OE dūn. The main—and perhaps circular—argument against a Gaelic origin is the rarity of the language in this area, with a maximum of ten Gaelic elements in seven place-names. A third alternative offered by BPNR is a derivation from Brittonic dīnas ‘fort, refuge, stronghold’. While less persuasive in terms of the stem vowel, this would have the advantage of accounting for the final -s. Another simplex place-name potentially from Gaelic is Ross, referring to the topographical feature now known as Ross Point. However, the meaning ‘headland’ is shared by Gaelic ros and the cognate Brittonic *rōs. Moreover, the place-name is not recorded until 1615 (Ros), and similar coastal names in other parts of south-east Scotland have led to the suggestion that the term was borrowed into Scots. BPNR gives all three language options, and notes a further occurrence of Ross on the coast of north-east England. A reasonable case could be made that the term was borrowed, not into Scots or English, but as an isogloss common to southern Scots and northern Middle English. Should it be added to dictionaries of Scots or English, or both? OED (s.v. ross, n.3) has an entry for a later borrowing of the same Brittonic term, in its Welsh reflex rhos ‘a marsh, morass’, recorded as a rare regional Herefordshire term from the nineteenth century. This is clearly a much later borrowing of the term, following semantic shift, than the one found in coastal place-names. Would it be appropriate to add the early loan to the same entry, or to create a separate one? As with other questions raised in this section, the issues are multi-faceted, but reflect the intrinsic fluidity of linguistic boundaries both within and across languages.

3 Lexical Borrowings The effects of borrowing are different in some important respects from the situation we have seen with place-names, since in this case the features that are transferred (which may include word-forms, meanings and structures) are adopted by and hence become embedded in the

102 

R. Dance et al.

vocabulary system of a single, recipient language.6 But by the close of the Middle Ages, the number and variety of the source languages involved in the historical make-up of the English lexicon are no less striking than they are for the place-name record, and many of the same considerations apply to both. When their speakers first arrived in Britain, the varieties of Germanic that would become English were already far from homogeneous etymologically. The identification of originally non-Germanic material at this time-depth is difficult, but most linguists are confident that they can detect early continental imports from the Celtic languages (e.g. OE byrne ‘mail-shirt’, īren ‘iron’) and from Latin (e.g. OE cēap ‘purchase, payment’, citel ‘kettle’, pund ‘pound’). English shares many further Latin loans with the other West Germanic languages, including several ultimately from Greek (e.g. OE engel ‘angel’, dēofol ‘devil’, prēost ‘priest’), and at least one (OE cyrice ‘church’) apparently direct from Greek. And some have perhaps travelled from still further afield: notice OE pæð ‘path’ (etymologically problematic, but often derived from an Iranian language).7 But the amount of loan traffic into English, and the range of source languages involved, expanded dramatically in the insular setting during the Middle Ages. Borrowing is, of course, a very frequent outcome of the contact between all the languages spoken and written in medieval Britain, in all the diverse situations in which that contact came about. A variety of contexts for thinking about this subject are described in more depth elsewhere in this volume (see, e.g., the papers by Callander, Fulton, Keller and Seiler, and Reed). But our focus in this section will be on borrowings into English—and even confining ourselves to items about whose etymologies most authorities would agree, the number of borrowed words found in Old and Middle English texts is enormous. To take a fairly random selection of loans from the best represented source languages, our haul includes words for concepts as diverse as: OE brocc ‘badger’, bin ‘manger’ (Brittonic), drȳ ‘magician’, clugge ‘bell’ (Irish), ME genou ‘mouth’ (Welsh), gulle ‘gull’ (perhaps Cornish); OE abbod ‘abbot’, camp ‘battle’,  For introductory accounts of contact-induced change in the lexicon, and further references, see, for example, Haspelmath (2009) and Winford (2010). 7  See the entries in OED; for discussion and further examples, see Miller (2012: 22–23, 62–74) and Durkin (2014: 69–75, 158–161). 6

4  Contact-Induced Lexical Effects in Medieval English 

103

lopystre ‘lobster’, mancus ‘monetary unit (30 pence)’, weall ‘wall’, beside semantic loans like tunge ‘language’, and loan-translations like dǣlnimend ‘participle’ (all Latin); OE prūd ‘proud’, ME chaunge ‘change’, cours ‘course’, joinen ‘to join’, mesūre ‘measure’, paien ‘to please, pay’, pēple ‘people’ (varieties of French, including Anglo-French); OE ceallian ‘to call’, fēolaga ‘fellow’, lagu ‘law’, ME casten ‘to cast’, dīen ‘to die’, eg(ge) ‘egg’, il(le) ‘ill’, lōs ‘loose’, skī(e) ‘sky, cloud’ (the early Scandinavian languages (‘Old Norse’)); and ME dekke ‘deck’, grōt ‘groat’, luk ‘luck’, peg(g)e ‘peg’, rōver(e) ‘pirate, rover’ (Dutch or Low German).8 Most of the examples in this list are frequently cited in histories of English vocabulary, and are relatively secure etymologically. In practice, however, the degree of difficulty we face in identifying lexical material of non-native origin varies massively depending upon the source language concerned. Much of the time, for instance, it is a reasonably straightforward matter to pick out the loans from Latin or French in medieval English—as indeed it must have been for contemporary Romance speakers or Latin users to identify them. All the French borrowings cited above show characteristic French sound changes, which have befallen obvious Latin etyma. There are nonetheless more problematic instances, including some where Germanic word-elements have probably been borrowed into Romance, and then come (back) into English. Notice for example ME hālen ‘to draw, rise, come, go, pass’. Despite some attempts to derive this verb from Old Norse, the Scandinavian verb represented by Old Icelandic (OIc) hala ‘to haul, pull’ is now usually explained as a later loan from Low German, and the Middle English word is better accounted for as a borrowing of Anglo-French haler ‘to haul, hoist; (reflex.) climb, ascend’, itself ultimately from Germanic (probably via a Frankish *halōn) (Dance 2019: ii, 74). The most difficult scenario for the etymologist is when the language varieties in immediate contact are very similar genetically, and when our knowledge of their vocabularies before and during the contact period is more or less limited. The classic case involving English, which we will focus on here, is its encounter with the early  Middle English forms follow the head-words in the Middle English Dictionary (MED). For detailed discussions of borrowings from each of these language groups see Durkin (2014: 76–95, 99–170, 223–297, 171–221, 354–360), and also Miller (2012: 19–20, 44–46, 74–88, 148–191, 91–147); and on Old Norse and Dutch influence see, respectively, Versloot and Putter (this volume). 8

104 

R. Dance et al.

Scandinavian languages; although very much the same issues pertain to detecting borrowing between any pair of closely related languages (and on the Dutch influence on English see notably Putter, this volume). Early English and the Scandinavian languages came into close contact in Viking Age Britain, and varieties of the latter were spoken in parts of England from the ninth-century Viking settlements perhaps up until c.1100 in some parts of the country (Parsons 2001). The languages concerned were closely related descendants of their North-West Germanic parent, overlapping considerably in phonology, morphosyntax and lexis, and probably mutually intelligible enough for practical, face-to-face communication (see esp. Townend 2002: 181–185). In addition, we are faced with the relative dearth of witnesses to the varieties involved in the right times and places (with the North and East of England being sparsely represented in the earliest English writings). As a result, when it comes to word-forms and features (including developments in meaning) first attested in post-contact English, telling apart those which are genuinely the results of Scandinavian influence from those which simply continue unrecorded (or convergent) endogenous developments is inevitably extremely challenging. The following brief outline of the principles draws on the analysis in Dance (2019), the findings of which were applied in the etymological typology used in the Gersum project’s database of more than 900 items from late Middle English alliterative verse, and we present from that corpus some less familiar examples of Middle English words plausibly or possibly showing Scandinavian input.9 The ideal situation is when there are regular and predictable formal differences between Old English and early Scandinavian phonology and morphology, and when the characteristically Scandinavian feature is what is found in post-contact English (Type A). Thus, for example, we know that Proto-Germanic (PGmc) */ai/ regularly became Viking-Age Norse (VAN) /ɑi/, /ɛi/, rather than OE /ɑ:/, as in ME haylse ‘to greet’ (cf. OIc heilsa, OE hālsian; A1*c), wayth ‘hunting’ (cf. OIc veiðr, OE wāð;  Further information in each case may be found by consulting Dance (2019), and Dance et al. (2019); for each lexical item referred to below, the full Gersum etymological category label is given in brackets. On the etymological identification of Old Norse input, see also notably Björkman (1900–1902), Durkin (2014: 190–219) and Pons-Sanz (2013: 25–122; 2015). 9

4  Contact-Induced Lexical Effects in Medieval English 

105

A1*bc).10 Similarly, PGmc */sk/ remained in VAN but palatalized (in most positions) in Old English, giving ME, PDE /ʃ/; hence ME scowtes ‘jutting rocks’ (cf. OIc skúti; A1bc), skere ‘pure’ (cf. OIc skærr; A1) (neither Scandinavian word happens to have attested Old English cognates, but, because the phonological differences are systematic, this is immaterial).11 When formal criteria of this kind are not available, all possible cases of Scandinavian input are in principle more contentious. Sometimes we can be reasonably confident of its presence, as for instance in words like ME ille ‘ill’ (cf. OIc illr; B1c), or heþen ‘hence, away’ (cf. OIc heðan; B1c) or kay ‘left(-handed)’ (cf. Dan. dial. kei; B1c), whose Germanic roots are not (unambiguously) attested in early Old English (Type B). Even when the same Germanic root can be identified in English before Anglo-Scandinavian contact (Type C), most authorities agree that borrowing is probable for items like ME bonk ‘bank, hillside’ (cf. ODan. banke; C1abc), deʒe ‘to die’ (cf. OIc deyja; C1a), mynne ‘less’ (cf. OIc minni; C1ac), mon ‘must; will, shall’ (cf. OIc munu; C3c), when the alternative is to suppose a surviving but previously unrecorded English cognate for a very frequently occurring idea. But hundreds of other much less secure and often avowedly speculative suggestions have appeared in historical dictionaries and in other etymological scholarship since the nineteenth century, including many where the supposedly Scandinavian form or sense or usage could plausibly have developed independently in English; thus, for example, (Type B) ME wharred ‘whirred’ (cf. Dan. hvirre, Swed. (dial.) hvirra; BB1c), (Type C) dreped ‘slain, killed’ (cf. OE drepan ‘to strike’, OIc drepa ‘to strike; kill’; CCC3c). The circumstantial evidence of a word’s distribution (whether in the Germanic family as a whole, and / or in regional usage in English or Scots in the earliest period of its attestation) can sometimes help us to assess the probability of the proposed Scandinavian input. Notice for instance ille (whose existence) and mon (whose usage as a future auxiliary) are highly characteristic of  As in Dance (2019), we generally cite early Scandinavian words via their Old Icelandic reflexes, noticing the forms and meanings attested in other Scandinavian languages (or in earlier, reconstructed stages) where the differences are significant. 11  Careful attention must nonetheless be paid to all known endogenous variants and their possible developments, and the evidence for each word must be assessed on its own merits; for an important example, see Cole and Pons-Sanz (this volume). 10

106 

R. Dance et al.

the Scandinavian branch of Germanic, and which are most frequent in the North and East of England early in their recorded histories (and this is still so for mon). But this kind of additional information must be treated with caution (Dance 2019: i, 49–63). Inter alia, most etymologists do not consider the existence of a cognate in another West Germanic language sufficient basis on its own for assuming an unattested Old English etymon instead of a Scandinavian one.12 Given the inherent difficulties of the evidence, and the scope for dispute that it provides, it is important to recognise that there is no such thing as a ‘definitive’ list of Scandinavian borrowings in English upon which all etymological authorities can or will agree. This is especially obvious for those words where no single, unambiguous form-source can be agreed (Type D). Scandinavian input is often posited in cases like ME big ‘strong, big’ (cf. OIc byggva ‘to settle, inhabit’, as in the ME verb biggen, or Norw. dial. bugge ‘rich, wealthy, powerful’?; see OED, s.v. big, adj. and adv.; and Dance 2019: ii, 258–261; D1c), but it is more or less unsatisfactory as an explanation of their origins, and sometimes we must accept that a word’s history will continue to be obscure unless new evidence comes to light. In some such difficult cases, several distinct source languages could have been involved, and it is hard to choose between them. Examples include ME stayned ‘coloured’ (PDE stain; D1), which may show input from an Old Norse verb represented by OIc steina ‘to stain, colour, paint’ (itself of obscure etymology) or / and early Fr. desteindre ‘to remove colour’; and ME loupe ‘loop’ (DD1c), which perhaps derives in part from a Scandinavian noun (cf. OIc hlaup ‘leap, jump’), or a Celtic one (cf. Ir. lúb ‘loop, ring, circle’, Scottish Gaelic lùb ‘bend, winding; loop, noose’), or plausibly some combination of the two, and for that matter the stem of MDu. lūpen ‘to lie in wait’ (as in PDE loop-­ hole) may also have played a role (Dance 2019: ii, 321–322, 290–292). Examples like these are, of course, nicely emblematic of the challenges posed to etymologists by the known presence in the medieval English lexicon of material of so many different origins, an embarrassment of riches that reflects in turn the many-layered, complex multilingualism in  Pace Versloot (this volume), whose discussion nonetheless illustrates very well how much room for doubt and controversy remains. 12

4  Contact-Induced Lexical Effects in Medieval English 

107

medieval Britain and Europe. Even when the ulterior etymology of a word is clear, its proximate sources, its stopping-off points en route to medieval English, may sometimes be several, and it can be difficult to tell how many intermediary languages were actually involved. Likely examples of such colourful chains of transmission include: ME caple ‘horse’ (FCC5b) and cros ‘cross’ (FC5b), both best explained as Latin words (L caballus, crux) which underwent characteristically Irish sound changes before entering English (cf. Ir. capall, cros), but which may very well have reached Britain via (Hiberno-)Norse (cf. OIc kapall, kross) and / or in the latter case through Anglo-French (cf. the variant forms crosce, crosse); and ME tulk (FC2c), a word best known as a (northern) Middle English synonym for ‘man’ in alliterative verse, but which derives ultimately from a Slavonic or Lithuanian word meaning ‘interpreter, spokesman’, and which probably entered English via Scandinavian (cf. OIc tulkr ‘interpreter, spokesman; broker’), which may itself have acquired it through Dutch or Low German (cf. Du. tolk, MLG tolk, tollik ‘translator’; Dance 2019: ii, 181–182 and 245–247).

4 Challenges Presented by the Documentary Record If we are interested in lexical aspects of language contact, ascertaining the earliest date of currency of a word seems an essential first step in evaluating the likelihood that it results from a particular contact situation. An obvious first step is to consult historical dictionaries for the earliest attestations they record. However, it is important that the concepts ‘earliest surviving attestation’ and ‘earliest period of use’ are not conflated too unreflectingly. The dating of much of our medieval textual evidence is itself problematic, and today many historical dictionaries give their readers prominent reminders of this; for example, the DOE attaches no dates to citations in the body of its entries, the MED employs double-dating, with manuscript date followed by presumed date of composition, for most literary sources, while the new edition of the OED employs only a broad banding of ‘early Old English’, ‘Old English’, and ‘late Old English’

108 

R. Dance et al.

for the period up to 1150, while adopting double-dating similar to MED’s for most Middle English material from non-documentary sources. An important next step is to assemble as full a picture as we can of a word’s subsequent history in the borrowing language. Is there a gap before we find further examples? If so, what approaches can we apply to estimate whether the gap is accidental, or whether we in fact have separate episodes of borrowing, rather than a continuous history of use? If we find a large number of early attestations, is this evidence for rapid spread, or for multiple instances of borrowing by different individuals (not unlikely in a close contact situation), or was the word perhaps borrowed earlier and spread more slowly, and the appearance of rapid spread is an accident of our surviving evidence? If there is polysemy, does this show innovation in the borrowing language, or does it reflect further semantic borrowing from the donor, or indeed merger of multiple borrowings from the same source? (Complex histories of this sort are probably shown by some of the commonest words in modern English, such as person, nature, and use, reflecting multiple inputs from both Latin and Anglo-French during the Middle English period.) Broadening the perspective beyond the individual lexical item, we can investigate whether it has near or full synonyms, and if so what its relationship appears to have been with these, and whether it appears to have become (at any time, and in any speech community) the ‘usual’ word realising a particular meaning. Additionally, we can examine where the word sits in a semantic hierarchy, what its relationship is with any hypernym or co-hyponyms, or whether it has hyponyms of its own (some of the exciting potential of such an approach being demonstrated by Sylvester and Tiddeman, this volume). Given appropriate tools (such as, for English, the Historical Thesaurus of English; Kay et al. 2020) we can investigate the distribution of loans from a particular language across a range of semantic fields, and thus gain valuable perspectives on how a particular loan appears to fit into a wider picture. English in the medieval period is particularly rich in cases illustrative of each of these scenarios. It also confronts us repeatedly with instances which point up the limitations of our surviving evidence for the medieval stages of even this well-documented and intensely studied language. This problem extends even beyond the limits of the recorded history of

4  Contact-Induced Lexical Effects in Medieval English 

109

English: As noted in Sect. 3, Old English shows a considerable number of borrowings from Latin (although some of these may have been of very restricted currency); while various different chronological layers of borrowing can be distinguished on formal grounds through the application of techniques of historical philology, there remain many areas of uncertainty, and we cannot always be certain whether Latin loanwords in Old English were borrowed after the period of Anglo-Saxon settlement in Britain or result from earlier borrowing on the continent, being already part of the incomers’ vocabulary. Borrowings from early Scandinavian present probably the outstanding instance in the recorded history of English where dates of first attestation must be treated with extreme caution, informed by our knowledge of the contact situation. Analysis of the dates of first attestation of words included in the OED which are currently identified as either certainly or probably borrowed from early Scandinavian shows a clear peak in the fourteenth century, and a tail that extends as far as the nineteenth century (Durkin Forthcoming). Nonetheless, we can be confident that we are observing the lexical effects of a Norse / English contact situation which has as its very latest chronological limit c.1100. The explanation for this enormous chronological disparity lies chiefly in the poor representation in the documentary record of the geographical areas where Scandinavian settlement was densest. Sociolinguistic factors probably also play a part, especially if our point of comparison is medieval borrowing into English from Latin and Anglo-French, which was almost certainly densest in precisely those sections of society best reflected in our surviving Middle English evidence. Our earliest evidence for a Scandinavian loan is frequently not only far removed chronologically from the period of borrowing, but also at the far end of a long process of intralinguistic spread, from the original bilingual speech community to broader use. We can often see echoes of this spread in the comparative density of occurrence in different regional varieties, but here our richest data comes from dialect surveys of the nineteenth century and later, perhaps more than a millennium after borrowing occurred. As noted in Sect. 3, identification of Scandinavian loans requires extremely close linguistic and philological analysis, and this is also often indicative of date of borrowing. Thus, while English sprent ‘to move suddenly, quickly, or energetically’ is first attested in the

110 

R. Dance et al.

fourteenth century and the related sprint is first securely attested only in the nineteenth century, both are most easily explained by borrowing of the early Scandinavian strong verb *sprinta and the related (originally causative) weak verb *sprenta in the period of Scandinavian settlement: Subsequent changes affecting both the stem vowel and the following consonant cluster would make the hypothesis of later borrowing from Scandinavian languages much more difficult. However, it is undeniable that historical knowledge that there was a suitable contact situation before 1100 is also a major stimulus pointing us towards certain solutions. Interestingly, taking up a theme from Sect. 2, in this instance there is also a very realistic possibility that we have much earlier direct evidence for currency of closely related Scandinavian word forms in England in the river name Spritt, Sprit, also Sprett, Spret, which is attested in Cumbria from the twelfth century. Among other languages considered in this volume, borrowing into English from Dutch (examined by Putter) presents particular difficulties, as it could have occurred at almost any point in the medieval period (trading contacts and probably at least some immigration being found throughout the period), and the close relationship between these two West Germanic languages gives us relatively few linguistic characteristics by which a borrowing can be distinguished from a cognate. Therefore, however incomplete our data may be, close scrutiny of all of the surviving attestations is likely to give our best clues as to date and context of borrowing. Thus, pack, a key term in the wool trade, first appears in the early thirteenth century, and its Dutch counterpart pac only a couple of decades earlier, giving a clear indication at least of when the word first came to prominence; in the absence of a secure further etymology, what remains slightly in doubt is the direction of borrowing. Our dictionaries show very few loanwords from Welsh. In terms of the long-term history of the general vocabulary of English, there can be little doubt that this overall picture is broadly true (see the overview in Durkin 2014: 76–95). But, as the contributions by Fulton and Callander in this volume remind us, these two vernacular languages were in contact throughout the medieval period, in a wide variety of different contexts and settings. If we had more access to, and intensive study of, appropriate documents, there seems little doubt that evidence of at least localised

4  Contact-Induced Lexical Effects in Medieval English 

111

borrowing from Welsh would be much more abundant. It is possible that the research question might then shift somewhat from why so little was borrowed, to why so little showed the necessary intralinguistic spread to gain lasting currency in the general vocabulary. By contrast, in the case of borrowing from Anglo-French and Latin, most of our surviving written sources come from individuals who were either deeply involved in or closely connected with the very communities—literate, and professionally involved with writing—who were at the centre of the functional trilingualism of later medieval England. We are therefore confronted with a problem that is rather uncommon in the study of lexical change in the medieval period: Even though we have a very extensive documentary base for studying lexical borrowing from Anglo-French and Latin, there is a danger that it over-represents the extent of lexical influence on the vocabulary as a whole. Thus, the English verb carry in its modern basic meaning appears to show the result of semantic broadening of a borrowing from Anglo-French carrier, which normally refers more narrowly to transportation, especially of goods, especially in a cart or wagon (it is a derivative of car, meaning ‘cart or wagon’, its continental French counterpart being charrier). The word is first attested in English in the mid-fourteenth century, and appears to have become very frequent very quickly: in addition to being attested in a wide variety of sources, we find for instance that in the works of Chaucer attestations of carry even in its broadened sense are outnumbered by equivalent uses of native bear by only around 2:1 (see Durkin 2018). This is a striking example of just how receptive English could be to borrowings from Anglo-French in a crucial period of rebalancing of functions towards English in the developing functional trilingualism of later medieval England, and just as striking an example of how rich and potentially rewarding our documentation for this particular contact situation is. However, there is also perhaps a danger of being seduced somewhat by the very richness of this documentation. It is certainly possible that English carry existed rather earlier than our earliest attestations: It is notable that L carriare is found in British sources from the thirteenth century onwards alongside the expected form L carricare, hence a potential donor form had already been long in use in both of the other languages in daily use in the making of business records. It is probably also wise to bear in

112 

R. Dance et al.

mind Chaucer’s own professional closeness to those engaged in the making of business records: while a poet is unlikely to be consciously employing in non-technical use a word that will be unfamiliar to many of his readers, it is nonetheless possible that carry might look somewhat less established in English by around 1400 if our evidence were not so heavily weighted towards the literate communities who were closest to the functional trilingualism from which the word almost certainly results. It is very likely that similar factors affect the impression that our surviving documents give us of the extent of the early adoption and spread of many of the Middle English loans from Anglo-French and / or Latin which have most profoundly influenced the core vocabulary of later English (as person, nature, use, or their ilk), since registers that are socially high, as well as the vocabulary of intellectual life, administration, law, or the church, in which such loans were frequent, are all over-represented in our surviving documentation. Conversely, very early examples, which are outliers from other evidence, should also often be approached with caution. Trotter (2003) examines some strikingly early evidence for borrowings from Anglo-­ French shown in the early thirteenth century in the Ancrene Wisse, such as departunge ‘departing’ recorded earlier than the parent verb depart; this could show us indirectly that depart had also been borrowed by this date, but, as Trotter observes, we could instead have evidence here for the confident multilingualism of a particular linguistic community, freely creating hybrid formations with English suffixes on Anglo-French word stems, without the need for prior borrowing of the simplex. Given the complexity of the multilingual contact situation, it is not surprising that earlier evidence for the existence of a word is often provided by one or more of the other languages of medieval Britain. Ashdowne (2020) points to examples from Latin such as grasmannus ‘cottar’ (a1150) or undermannus ‘subordinate’ (1276) implying earlier currency of ME grasman and underman (otherwise first attested 1327 and 1372, respectively). In other instances, the evidence is much less clear-­ cut; haggis is first attested c.1400 in English, hagis in the thirteenth century in Anglo-French. If, as often assumed (but somewhat problematically), the word ultimately shows a French origin, then the relative dating does not present any difficulties. If, on other hand, the word is ultimately of

4  Contact-Induced Lexical Effects in Medieval English 

113

English origin (as often argued), then the Anglo-French evidence must show a loan from English, and hence demonstrates that the English word must have existed earlier. In cases such as this, dates of first attestation may be among the factors considered in weighing one etymology against another, but they may be open to multiple interpretations. Additionally, as explored further in Sect. 5, some areas of the vocabulary are much better represented in multilingual documents than in monolingual ones, and particularly in Latin-matrix business records, in which it is often impossible to tell whether vernacular items belong to Anglo-French or English (or to both). Lexical data from sources of this type is often the most revealing about the nature of the contact situation, but also presents the greatest challenges of interpretation.

5 Language Borrowing and Language Borders Determining the multilingual etymological origins of a word is not the only challenge facing the historical lexicographer. The multilingual text and its incorporation in the dictionary present unique challenges, first, in deciding whether the writer is engaging in an act of code-switching or is using a loanword; and second, in determining the borders where one language ends, and another begins. These issues are particularly relevant when considering administrative texts from later medieval England, in their mixing of Latin, Anglo-French and Middle English. This section will examine some of the recent strategies used by the Dictionary of Medieval Latin from British Sources (DMLBS), the Anglo-Norman Dictionary (AND) and the OED in their efforts to engage with the latest scholarship in medieval multilingualism. Recent work in linguistics has focused on distinguishing between code-switching and loanwords in multilingual texts, that is, differentiating one-off switches from one language to another by a single user versus the use of words borrowed from another language and deemed to be part of the lexis of the target language, though this distinction is not without controversy (see the competing theories by Poplack 2018 and

114 

R. Dance et al.

Myers-Scotton 1989; also Ingham, Sylvester and Marcus 2020 for discussion of this issue in relation to Middle English). Lexicographers have engaged in similar work, though the criteria for determining loanwords and code-switches (and thus inclusion or exclusion in a dictionary) has not always been made explicit. Furthermore, editorial policies towards the inclusion of loanwords have varied over time and between dictionaries. These differing methodologies reflect differing language ideologies; the OED has tended to be the most open to inclusion of borrowed words, as Queiroz de Barros (2018) points out, referencing Murray’s reflections in the first edition on the presence of both foreignisms and loanwords as headwords. By contrast, the first edition of the AND often silently excluded words deemed to be code-switches to Middle English rather than loanwords, possibly in response to assertions that contact with English had resulted in Anglo-French becoming ‘bad French’ (Ingham 2010 summarises the traditional criticisms towards the language). The borders between borrowing and code-switching are often blurred, relying on morphological, syntactical or phonological changes, or frequency of use to distinguish the two, but these are not always useful in distinguishing contact between Anglo-French, Middle English and Latin. Morphological differences between the languages are indeterminate; Middle English makes use of Latin and Anglo-French simplex words to create compounds that never existed in the source languages (see the discussion of gardeviance below); Anglo-French scientific treatises frequently contain nouns with Latin plural endings (e.g. rochys or rokis as a plural of roche ‘roach’). Latin matrix texts can use syntax to flag the use of the vernacular through the use of the article le, as discussed further below, though either vernacular can follow the article. In texts containing multilingual glosses, it can be impossible to determine the language of the gloss when borrowing has led to both having orthographically identical forms (see Pagan and Seiler 2019 for a discussion of Middle English and Anglo-French glosses). In addition, glosses to texts can seemingly misidentify the language of the gloss—leaving the lexicographer to question whether the medieval glossator or the modern linguist knows best which language is being used. Frequency of use is also problematic as a criterion for distinguishing loanwords from code-switches when dealing with

4  Contact-Induced Lexical Effects in Medieval English 

115

medieval documentation with many words poorly attested outside multilingual contexts. Generally, each of the dictionaries has moved towards a more inclusive editorial policy towards loanwords though each presents the effects of language contact in a different manner. The DMLBS indicates borrowing through the brief etymological information provided in square brackets after the headword: waga 2 [AN guage, wage; cf. ME waue, waghe]. Italics here, as in citation evidence, indicate vernacular words. Loanwords are generally (though not systematically) indicated in the AND through the application of a language tag, indicating the source language of the loan. The application of the tags is currently under revision as at times these tags have been used to indicate the ultimate etymological source instead, for example Greek words borrowed through Latin tagged as Greek. In contrast, the OED contains thorough etymological information about every loanword; however, the presentation of early attestations in languages other than English is in a process of revision. Such citations were previously presented between square brackets in the quotation evidence (as in the MED), but are now placed in the etymological commentary alongside any discussion about the linguistic classification of these citations. Lexicographers need to be clear about criteria for inclusion or exclusion of forms; studies on one-word code-switching and borrowing have relied on dictionary attestations to help make the distinction, suggesting editorial policy can have a profound effect on this type of work (Queiroz De Barros 2018). The incorporation of multilingual texts in these three dictionaries has also changed considerably, though there is not yet any clear consensus on how to best incorporate multilingual material, or indeed, which dictionary should include it. Each of these dictionaries is, by definition, monolingual; they present a snapshot of one language at a moment in time (Wright 2013). However, the traditional separate lexicographical treatment of the languages can obscure their frequent co-existence in documentary evidence and imply a linguistic separation at odds with the reality of the average medieval scribe in Britain. While modern linguists may consider Latin, English and French as distinct languages, this distinction may be less clear during the medieval period (Trotter 2009).

116 

R. Dance et al.

One of the effects of monolingual lexicography in the face of multilingual texts is that a word may be found solely in the ‘wrong’ language dictionary, whether this is as evidence for earlier use, as discussed above, or as the sole attested use of a word. For example, the AND has a single attested use of the word kernemen taken from the roll of the king’s council in Ireland in 1392–1393: .c. hommes appellés kernemen. The context, introducing the word with appellés, suggests that the scribe considers the word a borrowing, but no equivalent form is found in Middle English. The OED attests the use kern (‘a light-armed Irish foot-soldier’; a borrowing from Irish ceithern) in the same sense but there is no attested use of the compound kern + men outside this Anglo-French text. In this way, the AND conserves an otherwise unattested Middle English word. More concerning is the exclusion of words from one dictionary for being of the ‘wrong’ language, without subsequent uptake by the dictionary in the ‘correct’ language. In the case of Latin texts, this might mean that words deemed to be vernacular may be rejected by the DMLBS but remain unknown to the OED or AND as they do not form part of their standard corpus. For example, the following citation from 1431 in the Household Account Book of Richard Beauchamp [Warwickshire County Record Office CR1618/W19/5] (fol. xcv) reads: Item, solutum pro factura et emendacione ij gardevyands ...(‘For the making and mending of 2 gardeviands ...’). The word gardevyand (‘safe for meat’) is omitted from the DMLBS as being a vernacular word; both the morphology and phonology of the word suggest it is a French borrowing. The citation is not included in the AND as the account book does not appear in its corpus, though the dictionary does include an entry for gardevyand. The single attestation illustrating the word in the AND is in fact taken from the OED entry for gardeviance, n., in a citation from the Paston Letters from 1459. The language of that citation is difficult to determine, neither clearly Anglo-French or Middle English: Item, j gardevyaunt. While the word appears to be French, a compound of garde and viande, these two attestations are linguistically ambiguous. The MED entry for garde-viant, n., contains further evidence, with thirteen citations antedating the earliest OED citation, though most of these suggest the use of a vernacular term in a Latin context: une pare de gardevyandes ... una cofer vocata gardeviance ... unum gardevyant. But which vernacular? There is no

4  Contact-Induced Lexical Effects in Medieval English 

117

attested use of the term in an Anglo-French context and equally no attested continental variant, though similar compounds such as garderobe, gardemangier, gardelessive, meaning ‘places to keep things safe’, are well attested. The exclusion of the word from the DMLBS implies it should not be considered part of the lexis of Latin, but rather, reflects a moment of code-switching. As the word is etymologically Anglo-French, the word is given an entry in the AND, though it is not clear that it was ever a word in Anglo-French. The OED includes the ambiguous citations as early evidence for use of the word in Middle English, though it is unclear if the word should be considered a borrowing or a Middle English creation. This is a frequent pattern in multilingual sources, particularly in administrative documents, where the earliest attested use of a vernacular term is often in a Latin matrix (Durkin 2014). Recording these attestations remains lexicographically challenging, particularly in cases such as gardevyand, where they are the sole evidence of use. Work by Wright (2013 and 2017) and Roig-Marín (2018) suggests that the use of the vernacular in these Latin administrative texts represents a separate phenomenon from borrowing or code-switching: a fused or mixed-language used by multilingual clerks in specific contexts, such as account books, port books and other types of business records. Research on these types of texts suggests clerks made use of abbreviations to exploit the resulting linguistic ambiguity, removing clear indications of Latin inflection, though they also made use of the French definite article le to mark the use of a vernacular term. The use of a mixed-language for administration documents, though limited to a specific context, demonstrates how fluid the border between these three languages can be in this period. This movement between the languages is not confined to business texts; medieval scribes could take advantage of the fluidity between the vernaculars, switching between them to choose the precise term, with different vernaculars populating different levels of technical specificity, as Sylvester and Tiddeman’s work suggests. Glossators equally exploited the fuzzy borders between the vernaculars, choosing words which could be glossed as either vernacular, suggesting our modern divisions between the languages are anachronistic.

118 

R. Dance et al.

Thus, historical lexicography needs to reflect the fluidity found in the texts it is trying to record with projects cooperating to share multilingual data and include relevant citations in more than one dictionary. The dictionaries need to demonstrate the wide lexical range available to the medieval scribe, spanning multiple languages and making use of them all.

References [AND =] Anglo-Norman Dictionary. 1977–1992. Ed. by Louise W.  Stone, T. B. W. Reid and William Rothwell. Rev. edn, A-E. 2005. Ed. by William Rothwell, Stewart Gregory and David Trotter. AND2 Online Edition. Ed. by David Trotter (–2015) and Geert de Wilde (2015–). Aberystwyth: Aberystwyth University. https://anglo-­norman.net. Accessed 26 May 2023. Ashdowne, Richard. 2020. Mannus makyth man(n)? Latin as an Indirect Source for English Lexical History. In The Multilingual Origins of Standard English, ed. Laura Wright, 411–441. Berlin: De Gruyter. [BPNR =] Berwickshire Place-Name Resource. https://berwickshire-­placenames. glasgow.ac.uk/. Accessed 25 May 2023. Björkman, Erik. 1900–1902. Scandinavian Loan-Words in Middle English, Studien zur englischen Philologie, 7 and 11. 2 vols. Halle: Max Niemeyer. Cavill, Paul. 2017. A Vocabulary of the Everyday. In The Daily Lives of the Anglo-­ Saxons, ed. Carole Biggam, Carole Hough, and Daria Izdebska, 45–61. Tempe, AZ: Arizona Center for Medieval and Renaissance Studies. [CSD =] Concise Scots Dictionary. 2017. 2nd edn. Edinburgh: Edinburgh University Press. Dance, Richard. 2019. Words Derived from Old Norse in Sir Gawain and the Green Knight: An Etymological Survey, Publications of the Philological Society, 50. 2 vols. Chichester: Wiley-Blackwell. Dance, Richard, Sara Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge, Cardiff and Sheffield. https://www.gersum.org. Accessed 26 May 2023. A Dictionary of the Older Scottish Tongue. 1931–2002. Ed. by William Craigie, A. J. Aitken, James Stevenson, and Marace Dareau; published online as part of Dictionaries of the Scots Language. [DSL =] Dictionaries of the Scots Language, bringing together A Dictionary of the Older Scottish Tongue and The Scottish National Dictionary. https://dsl. ac.uk/. Accessed 15 July 2021.

4  Contact-Induced Lexical Effects in Medieval English 

119

[DMLBS =] Dictionary of Medieval Latin from British Sources. 2018. Ed. by Richard Ashdowne, David Howlett, and Ronald Latham. Oxford: Oxford University Press. http://clt.brepolis.net/dmlbs/. Accessed 26 May 2023. [DOE =] Dictionary of Old English: A to I online. 2018. Ed. by Angus Cameron, Ashley Crandell Amos, and Antonette DiPaolo Healey. Toronto: Dictionary of Old English. https://www.doe.utoronto.ca/pages/index.html. Accessed 26 May 2023. Durkin, Philip. 2014. Borrowed Words: A History of Loanwords in English. Oxford: Oxford University Press. ———. 2018. Exploring the Penetration of Loanwords in the Core Vocabulary of Middle English: Carry as a Test Case. English Language and Linguistics 22: 265–282. ———. Forthcoming. Norse Borrowings in the OED: A Fresh Examination. In The Legacy of Medieval Scandinavian Encounters with England and the Insular World, ed. Richard Dance, Sara M.  Pons-Sanz, and Brittany Schorn. Turnhout: Brepols. Gelling, Margaret, and Ann Cole. 2014. The Landscape of Place-Names, rev. edn. Donington: Shaun Tyas. Haspelmath, Martin. 2009. Lexical Borrowing: Concepts and Issues. In Loanwords in the World’s Languages: A Comparative Handbook, ed. Martin Haspelmath and Uri Tadmoor, 35–54. Berlin: De Gruyter Mouton. Hough, Carole. 2010. Toponymicon and Lexicon in North-West Europe: ‘Ever-­ Changing Connection’, E.  C. Quiggin Memorial Lectures, 12. Cambridge: University of Cambridge, Department of Anglo-Saxon, Norse and Celtic. ———. 2016. The Metaphorical Landscape. In Mapping English Metaphor Through Time, ed. Wendy Anderson, Ellen Bramwell, and Carole Hough, 13–31. Oxford: Oxford University Press. Ingham, Richard. 2010. Later Anglo-Norman as a Contact Variety of French? In The Anglo-Norman Language and Its Contexts, ed. Richard Ingham, 8–25. Woodbridge: York Medieval Press. Ingham, Richard, Louise Sylvester, and Imogen Marcus. 2020. Lone Other-­ Language Items in Later Medieval Texts. Journal of Historical Sociolinguistics 7: 179–205. James, Alan. 2023. Brittonic Language in the Old North, rev. edn. 3 vols. https:// spns.org.uk/resources/bliton. Accessed 23 May 2023. Kay, Christian, Marc Alexander, Fraser Dallachy, Jane Roberts, Michael Samuels, and Irené Wotherspoon, eds. 2020. The Historical Thesaurus of English, 2nd edn, version 5.0. Glasgow: University of Glasgow. https://ht.ac.uk/. Accessed 26 May 2023.

120 

R. Dance et al.

[MED =] Middle English Dictionary. 1952–2001. Ed. by Robert E.  Lewis. Online edition in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner. Ann Arbor: University of Michigan Press. https://quod.lib.umich.edu/m/med/. Accessed 26 May 2023. Miller, D. Gary. 2012. External Influences on English, from Its Beginnings to the Renaissance. Oxford: Oxford University Press. Myers-Scotton, Carol. 1989. Codeswitching with English: Types of Switching, Types of Communities. World Englishes 8: 333–346. Nicolaisen, W. F. H. 2001. Scottish Place-Names: Their Study and Significance, rev. edn. Edinburgh: John Donald. [OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 26 May 2023. Pagan, Heather, and Annina Seiler. 2019. Multilingual Annotations in Ælfric’s Glossary in London, British Library, Cotton Faustina A.x: A Commented Edition. Early Middle English 1 (2): 13–64. Parsons, David N. 2001. How Long Did the Scandinavian Language Survive in England? Again. In Vikings and the Danelaw: Select Papers from the Proceedings of the Thirteenth Viking Congress, Nottingham and York, 21–30 August 1997, ed. James Graham-Campbell, Richard Hall, Judith Jesch, and David N. Parsons, 299–312. Oxford: Oxbow. Parsons, David N., and Tania Styles. 2000. The Vocabulary of English Place-­ Names (BRACE-CÆSTER). Nottingham: Centre for English Name-Studies. Pons-Sanz, Sara. 2013. The Lexical Effects of Anglo-Scandinavian Linguistic Contact on Old English, Studies in the Early Middle Ages, 1. Turnhout: Brepols. ———. 2015. Identifying and Dating Norse-Derived Terms in Medieval English: Approaches and Problems. In Early Germanic Languages in Contact, ed. John Ole Askedal and Hans Frede Nielsen, 203–221. Amsterdam: John Benjamins. Poplack, Shana. 2018. Borrowing: Loanwords in the Speech Community and in the Grammar. Oxford: Oxford University Press. Queiroz de Barros, Rita. 2018. Twentieth-Century Romance Loans: Code-­ Switching in the Oxford English Dictionary? In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 61–76. Berlin: De Gruyter.

4  Contact-Induced Lexical Effects in Medieval English 

121

Roig-Marín, Amanda. 2018. When the Vernaculars (Anglo-Norman and Middle English) and Medieval Latin Fuse into a Functional Variety: Evidence from the Administrative Realm. Studia Neophilologica 90: 176–194. Taylor, Simon, with Gilbert Márkus. 2006–2012. The Place-Names of Fife. 5 vols. Donington: Shaun Tyas. 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. Trotter, David. 2003. The Anglo-French Lexis of Ancrene Wisse: A Re-Evaluation. In A Companion to Ancrene Wisse, ed. Yoko Wada, 83–101. Cambridge: D. S. Brewer. ———. 2009. Stuffed Latin: Vernacular Evidence in Latin Documents. In Language and Culture in Medieval Britain: The French of England c.1100– c.1500, ed. Jocelyn Wogan-Browne, Carolyn Collette, Maryanne Kowaleski, Linne Mooney, Ad Putter, and David Trotter, 153–163. Woodbridge: York Medieval Press. Winford, Donald. 2010. Contact and Borrowing. In The Handbook of Language Contact, ed. Raymond Hickey, 170–187. Oxford: Wiley-Blackwell. Wright, Laura. 2013. On Historical Language Dictionaries and Language Boundaries. In Evur happie & glorious, ffor I hafe at will grete riches, ed. Liliana Sikorska, and Marcin Krygier, 11–26. Frankfurt am Main: Peter Lang. ———. 2017. On Non-Integrated Vocabulary in the Mixed-Language Accounts of St Paul’s Cathedral, 1315–1405. In Latin in Medieval Britain, ed. Richard Ashdowne and Carolinne White, 272–298. Oxford: Oxford University Press.

5 The West Germanic Heritage of Yorkshire English Arjen Versloot

1 Introduction: The Abundance of Norse Loanwords in English In every academic and popular description of the history of English, one can read about the impact of Old Norse on English vocabulary as a heritage of the Viking invasions and subsequent settlements in the ninth and tenth centuries. The estimated numbers of loanwords are often fairly high, in particular for Middle English—although of a different scale in total than the French and Latin lexical impact. For instance, in the introduction to the Gersum database, Dance et al. (2019) note that ‘[…] there are about 2,000 Norse-derived terms recorded in medieval English texts.

A. Versloot (*) University of Amsterdam, Amsterdam, The Netherlands Fryske Akademy, Leeuwarden, Netherlands University of Groningen, Groningen, Netherlands e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_5

123

124 

A. Versloot

Of them, about 700 are still in use in Standard English, although many more can be found in dialects from areas such as the East Midlands, Yorkshire, Lancashire, and Cheshire’. Kastovsky (2006: 223) mentions ‘several thousand’ Norse loanwords in Middle English ‘[…] of which between 400–900 […] have survived in standard English, a further 600 or more in the dialects’. Viereck et al. (2002: 95) identify 1154 words alone with word initial sk- in Joseph Wright’s dialect dictionary (1905), all considered to be of Scandinavian origin on the basis of this phonological criterion (see Björkman 1900–1902: 119; Gersum database: Key to Phonological and Morphological Markers). The current version of the Oxford English Dictionary (OED) contains almost 1000 items with some form of reference to Old Norse (p.c. Philip Durkin). The influx of Old Norse vocabulary is particularly associated not only with the northern part of England (Kolb 1965) but also with Scotland (Corbett et al. 2003: 6); these are regions where Norse speakers were particularly present in the ninth and tenth centuries (Haywood 1995: 79). The earliest lexical loans from Norse (around hundred) appear already in Old English and are mostly confined to new concepts in specific semantic fields (Dance 2012: 1731–1732). However, non-technical Norse-derived vocabulary is already found in the tenth-century Northumbrian glosses (Pons-Sanz 2013, 2016: 307). It is frequently stressed that the Norse loanwords attested since the Middle English period include ‘many everyday words […]. Especially remarkable is the fact that function words were also borrowed […]’ (Kastovsky 2006: 249). Among these everyday and function words, there are many that seem to have replaced existing English words, such as both (cf. OIc báð-, OE bā) and call (cf. OIc kalla) replacing OE hrōpan.1 Pyles and Algeo (1982: 299) signal the concept of ‘semantic contamination’ where words such as bread, bloom and earl acquired their meaning from Old Norse.

 In the Gersum database, both and call are assigned to category C, which ‘indicates that the root is known in early Old English (or there is an unambiguous form-source in a third language), but some aspect of form, sense or usage suggests Scandinavian influence’. The latest online edition of the OED acknowledges that the origin of the word call (v.) is disputable and for both it mentions: ‘Partly (i) formed within English, by compounding. Partly (ii) a borrowing from early Scandinavian’ [all accessed 22 January 2023]. 1

5  The West Germanic Heritage of Yorkshire English 

125

Among these examples, there are words with well-attested West Germanic parallels, which, for that reason, may be inherited words in English after all (see Sect. 2). As a consequence, the fairly generous assignment of English words to Old Norse origin is a practice that has called for revision, as was made explicit by Dance (2012: 1731): ‘[o]ne sometimes senses that the attribution of a Norse derivation has as much to do with the enthusiasm of the scholar applying the labels as it does anything else […]’ (also Dance 2012: 1728; Dance et al. 2019).2 The word guest may serve as an example here. It is identified as an Old Norse loanword in Klein’s (1966) etymological dictionary, and so is it interpreted also in Hoad (1986), who summarises the etymologies from the OED. The 1989 edition of the OED acknowledges that guest is ‘usually explained as due to the influence of ON gest-r’, but the current 2000 edition describes the origin of the word as ‘Common Germanic’ (and thus an inherited word in English), hinting at the Old Norse origin as an option in the etymological literature (OED, s.v. guest; category A1* ‘the strongest case for Old Norse input’ in the Gersum classification, Dance et al. 2019: s.v. gest).3 An instance frequently quoted in the older literature as an Old Norse loan is bread, which has been refuted by Pons-Sanz (2017; category CCC3a in the Gersum classification, Dance et al. 2019: s.v. bred). This adjusted interpretation, namely, as a ‘word inherited from Germanic’ is also found in the OED (s.v. bread, n.). The necessity of revision of the set of likely Old Norse loanwords has recently gained momentum with an article of Cole (2018), who posited a native English origin for what may be called the crown-witnesses of intensive English-Norse language contact (see also Cole and Pons-Sanz, this volume): the pronouns of the third-person plural, they, them, their (Dance 2012: 1736; category A1*c in the Gersum classification, Dance et al. 2019: s.v. þay).  An example of what can be called an urban legend is the story found in Bragg (2004: 28): ‘A young soldier [from South Cumbria] went to Iceland […]. In Iceland […] he used words from his home dialect and made himself understood. Within a week or two he was conversant with the Icelanders. Old Norse was that deeply bitten into the Old North’. 3  At various places in the text, reference will be made to the classifications of loan word status in the database of the Gersum project (Dance et al. 2019). There are four main categories with multiple subtypes, allowing for 230 different code combinations. The explanation of all the labels can be found at https://www.gersum.org/about/explanation_of_summary_categories# [accessed 28 May 2023]. Relevant definitions are quoted in the text. 2

126 

A. Versloot

The aim of the present paper is to re-evaluate a number of words which have commonly been ascribed to the influence of Old Norse and to provide a more adequate estimation of the number of these loanwords in English. The first part of this study (Sect. 2) will discuss four factors that may have been underappreciated in the evaluation of potential Norse loanwords in English and whose application will lead to a reinterpretation of some words currently labelled as Norse loanwords as ‘inherited West Germanic’. These criteria will be applied to a set of presumed Norse loanwords in the Yorkshire dialect, which is commonly recognised as a variety with strong Norse influence (Kolb 1965). This case study analyses a set of 112 words that are identified as (potential) loans from Old Norse in Kellett’s (1994) Dictionary of Yorkshire Dialect, Tradition and Folklore (henceforth YD). A more detailed description of the material for the case study and its outcomes are presented in the Sects. 3 and 4.

2 The Overestimation of the Norse Component in (Yorkshire) English There are four major factors that may have contributed to the overestimation of the Norse component in English, including Yorkshire English. They are systematically discussed and illustrated with examples in this section. Unless otherwise stated, the examples are words that are identified as Norse loanwords in Kellett’s dictionary (YD) (1994). These insights are consistently applied in the case study of the YD in Sect. 4.

2.1 The Relation with Frisian In the first place, it is important to bear in mind that various English words traditionally labelled as Old Norse loanwords have proper West Germanic cognates which may have escaped the attention of etymologists. Of particular relevance here is the evidence offered by Frisian, the closest kin of English in the Germanic family (see e.g. Siebs 1889; Nielsen

5  The West Germanic Heritage of Yorkshire English 

127

1985). This relationship not only concerns the phonological and morphological developments but is also widely attested in the lexicon (Löfstedt 1963–1969; Århammar 1969: 18–19; Munske 1973). It is quite remarkable that the word ‘Frisian’ returns zero hits when searched for in the first volume of Björkman (1900–1902).4 The commonalities in lexicon of northern England and Scotland that result from the historical relations between English and Frisian were very early observed by Minssen, who studied the East Frisian dialect of Saterland in the nineteenth century: ‘Auch die gewöhnliche Volkssprache in Schottland […] bietet eine Analogie mit der der Saterländer [Friesen] dar, die beweis’t, wie weit das angelsächsisch-friesische Element im Norden Großbrittaniens vorgedrungen und in die Volkssprache übergegangen ist’ (Minssen 1854: 155).5 Frisian consists of four main branches, including West, East and two sub-branches of North Frisian. West Frisian is currently spoken by c.400,000 people in the Dutch province Fryslân, while East Frisian is confined to a group of c.1000 speakers in the German municipality of Saterland. Linguistic differences between West and East Frisian are already evident in the thirteenth-century Old Frisian sources. North Frisian has c.4000 speakers in the German ‘Kreis’ Nordfriesland, a region that until 1864 was part of Denmark. Archaeological and linguistic evidence indicates that a first group of Frisian speakers settled down on the North Frisian Islands in the seventh and eighth centuries, whereas a second group migrated to the adjacent coastal areas on the invitation of the Danish king in the eleventh century. All the branches of Frisian have been relatively high-contact varieties over the centuries. North Frisian has been in close contact with Danish for many centuries, and since the late Middle Ages with Low and later High German as well. East Frisian dialects show the impact of  See http://archive.org/details/scandinavianloan00bjuoft [accessed 28 May 2023].  ‘The common vernacular in Scotland […] offers an analogy with that of the Saterland Frisians as well, which shows how far the Anglo-Saxon-Frisian element has penetrated into the north of Great Britain and has passed into the vernacular’ (my translation). 4 5

128 

A. Versloot

long-­standing contact with Low German, whereas West Frisian has been in contact with Dutch already since the fifteenth century. Despite the Viking raids on the Frisian coasts (from the eighth to the eleventh century) and short-term Viking overlordship over parts of the medieval West and East Frisian regions, no substantial Viking settlement in the region has been mentioned in the historical sources or otherwise reconstructed. In contrast to the English Danelaw region, with its massive presence of Old Norse place names, no such patterns can be found in the medieval West and East Frisian coastal regions. Apart from a handful of older runic inscriptions (Kaiser 2021), the earliest attestations of Frisian date to the twelfth century. Despite this relatively late attestation date, the breadth of linguistic variation present until today allows for a remarkable reconstruction depth, using the historical-­comparative method. As a consequence, a surprisingly large part of the pan-Frisian vocabulary—including many words not attested in the Old Frisian attestations consisting nearly only of legal texts—has been preserved, if not in one dialect, then in another (Århammar 1989). Significantly, there are no indications of relevant language contact between northern England and the Frisian speaking area after the early Middle Ages that could have ‘contaminated’ the interpretation of the Frisian material in the context of this study.6 A modern dictionary of Old Frisian has only been published in 2008 (Hofmann and Popkema 2008), although the first Old Frisian dictionary from 1840 was remarkably reliable (von Richthofen 1840). Modern Frisian consists of multiple dialects, each with its own often complicated phonological history (Siebs 1889, 1901), differing from each other as strongly as, for example, Dutch and German, due to their largely separate

 See Munske (2001) for a comprehensive coverage of many aspects of linguistic and sociocultural history of Frisian. 6

5  The West Germanic Heritage of Yorkshire English 

129

and isolated developments. The lexicographical description of many of these dialects has only become available in relatively recent years.7 The (non-)existence of cognates in Frisian and other West Germanic languages is one of the important criteria in the identification of a potential Norse origin of English words when phonological or morphological criteria are inconclusive. This applies in particular to the words belonging to category B and C in the Gersum database: ‘Type B indicates lexemes formed on Germanic roots which are not attested early enough in Old English’. The absence of an early attestation in English is an argument to consider an external origin. Sub-type B1 comprises words whose ‘[…] root is only otherwise known in North Germanic’, which suggests that the word was not West Germanic in origin and not part of the Old English lexicon. For words belonging to B2, ‘the root is attested in at least one other Germanic language […]’, but the implicit interpretation seems to be that the lack of attestation in Old English indicates that the word was possibly no longer part of the initial Old English vocabulary at the time of language contact with Scandinavian and hence most likely (re-) introduced into English. However, for many low-frequency words, the absence from the Old English corpus is a less-relevant fact (Dance 2012: 1731). Despite the relatively large size of the attested Old English vocabulary, we have to realise that many more lexemes and their geographical variants are ultimately unknown. Therefore, given the close historical ties  One can mention the following dictionaries: West Frisian (WFri.): main dialect: van der Veen and de Boer (1984–2011); Hindeloopen: Blom and Dyk (2019); Schiermonnikoog: Spenter (1968) and Visser and Dyk (2002); East Frisian (EFri.): Saterland (Sat.): Fort (2015); Wangerooge (Wang.): Ehrentraut (1849, 1854); North Frisian (NFri.): Föhr and Amrum (FA): Sjölin (2002); Sylt: Möller (1916) and Kellner (2006); Bökingharde: Sjölin et al. (1988); Wiedingharde: Jensen et al. (1994); Halligen and Nordergoesharde: Löfstedt (1928, 1931). Cited examples from varieties of (Old) Frisian that appear as entries in these dictionaries will be quoted without further reference, unless specific remarks have been made or words are found under different entries. A proper pan-Frisian etymological dictionary does not exist. Spenter (1968) and Löfstedt (1928, 1931) are lexicological studies organised on the basis of historical phonology and ipse facto providing form-etymologies with mostly one-word German translations. Boutkan and Siebinga (2005) offer etymologies of the vocabulary of one Old Frisian manuscript (Oldenburg, Niedersächsisches Staatsarchiv, Bestand 24–1, Ab. Nr. 1, a.k.a. First Rüstring MS; Buma 1961), and Buma (1949) offers etymological references in the index on his text edition of the Old Frisian Brokmerbref, or Law of Brokmerland. Wider-ranging etymological studies are presented in Sjölin (2006) and Faltings (2010). Old and Modern West Frisian cognates to Dutch words are systematically included in the Dutch etymological dictionary by Philippa et al. (2003) . 7

130 

A. Versloot

between Frisian and English, I interpret the circumstance where a Proto-­ West Germanic word is attested as an inherited word in Frisian as a strong indication that a similar lexeme existed in the Old English period, even if it is not attested in its three million token corpus. ‘Type C indicates that the root is known in early Old English (or there is an unambiguous form-source in a third language), but some aspect of form, sense or usage suggests Scandinavian influence’ (Dance et al. 2019). The identification of other, hitherto unconsidered West Germanic parallels may provide evidence that the form, sense or usage found in English did exist in other West Germanic languages, which would relativise the unique English-North Germanic parallelism of such aspects. While Frisian is the most likely candidate to find so-far unidentified etymological parallels in West Germanic, other West Germanic languages, such as Dutch and Low German, may also contain hitherto unknown etymological parallels with (Northern) English words. By way of example, such West Germanic cognates seem to have been underappreciated or missed in the interpretation of the origin of Yorkshire English (YE) durn ‘doorpost’, to rive ‘to tear’ and to bensel ‘to beat, trash’. For durn, no West Germanic cognate is mentioned in the OED (s.v. durn), and yet it may be related to EFri. (Wangerooge) durn ‘door’. In the case of rive ‘to tear’, the OED (s.v. rive, v.1) mentions the Old Frisian cognate ūtrīva ‘to tear out’, but still labels the word as ‘[a] borrowing from early Scandinavian’, possibly because of its distribution in northern English and Scots. A fairly complicated example of a word that eventually seems best explained with a parallel from Low German is YE bensel ‘to beat, trash’. The OED (s.v. bensel, v.) marks it as a dialectal form, with a first attestation in 1673. Its origin is discussed in the OED’s entry for bensel (n.) with a first attestation in 1522, meaning ‘bending, tension, spring (of mental faculties); strong bent or determination; impetus (of a body in motion)’. The OED interprets the noun as ‘[a] borrowing from early Scandinavian. Etymon: Norse benzla’, meaning ‘bending, bent, tension’. The Modern Icelandic […] verb bensla means ‘to wrap around sth’. The Modern Icelandic word has been attested only since the sixteenth century and is related to Danish (Dan.) bændsel ‘a wrap’, which is a borrowing from Middle Low German bendsel (Ordbog over det danske Sprog, s.v. bændsel). This cannot be the source of YE bensel for reasons of semantics. Old Icelandic (OIc) benzl

5  The West Germanic Heritage of Yorkshire English 

131

means ‘bent state of a bow’ (Zoëga 1975: s.v. benzl), which is semantically close to the meaning of the noun bensel in the OED: ‘bending, tension’. In this meaning, it seems related to English to bend. The OED (s.v. bend, v. 2.a) considers the meaning ‘to constrain a bow with the string […]’, to be of Old English origin. The meaning, is however, still quite remote from YE bensel ‘to beat, trash’. Eventually, the best semantic fit is found in East Frisian Low German: bensel ‘stick, club’ or ‘bunch of twigs’ and a verb benseln ‘to hit with a bunch of twigs’ (Böning et al. 1998: s.v. Bensel). This word is semantically connected with the earlier mentioned MLG bendsel. Saterfrisian bänselje ‘to chase away, throw out’ can be connected to YE bensel as well. In this myriad of forms and meanings, the best formal and semantic parallel is found in West Germanic, in this case Low German and East Frisian (Sat.). Given the late attestations dates in English, the word may after all also be a much later loan from Low German in the latter meaning.

2.2 Lexical Support and Loanwords Language contact can strengthen the use and preservation of mutual cognates in either language (see also Sect. 2.5), but such selective preservation should not be taken for lexical borrowing. Lexical loan relations between languages can be found in many expressions, including loan formations and loan translations. All these processes add a new lexical element to a language ‘A’, bearing traces of a lexeme from language ‘B’. Already Björkman (1900–1902: 8–9) points to two more relations between the lexicon of two languages: ‘Many words, common to both languages, but differing somewhat in sense, must have adopted the sense of the other language. And many words which were becoming or had already become obsolete in one language may have been recalled to life by the influence of the other’. The first part of this quotation describes the concept of a semantic calque or semantic borrowing. The second process is even more subtle and can be called lexical support (‘lexikalische Stützung’, Århammar 1966): a lexeme is inherited and its meaning may even be constant, but the mere fact that the word has been retained is the result of contact with another language where such an etymon was (even

132 

A. Versloot

more) common. The Gersum database includes such instances under Type C5. The root of knife is attested in all North and West Germanic languages. It has specialised meanings in High German (HG; Kneif ‘cobbler’s knife’), Dutch (knijf ‘clasp-knife’) and West Frisian (knyft, kniif ‘clasp-knife’), but it is the generic term in North Frisian, English and the North Germanic languages. The East Frisian dialects preserve the word sax / soaks, which seems to have been the generic term in Old Frisian. West Frisian opted for mes, in line with Du. mes ‘knife’. The West Frisian form, only attested in late Old Frisian, can be derived from PGmc *matisahsa- and is not necessarily a loan from Dutch, given the appearance of the cognate in OE meteseax ‘meat-knife, dagger’ (Philippa et al. 2003: s.v. mes) and HG Messer ‘knife’. Considering the wide variety of words for knives attested in Old Frisian (and probably not only there), with seven different base words and multiple compounds, one may claim that the selection of English knife and North Frisian knif as the generic term from all these synonyms was a contact-phenomenon (OIc knífr, Dan. kniv), but it cannot be considered a loanword and hardly a semantic calque.8 Another example where lexical support might have played a role is YE creeaked, Standard English crooked ‘curved’. The word can easily be associated with North Germanic cognates such as OIc krókr ‘hook, curve’. It is ‘[a]pparently a borrowing from early Scandinavian’ (OED, s.v. crook, n. and adj.; Gersum database type BB1b). The root is not attested in Old Frisian either. Various modern Frisian dialects, however, show that this word stem must have existed in Frisian as well. NFri. (Sylt) kruk ‘bowed’ might still be an instance of borrowing from Danish, but the EFri. (Sat.) krouke ‘scythe handle’ confirms the existence of a word *krōk (noun / adj.) ‘curve(d)’ in Old Frisian and thus in North Sea West Germanic. Its survival in English and North Frisian is in my view a candidate for lexical support by Norse, but there is no need for identifying it as a loanword.

 Knife is labelled as BB2a in the Gersum database, indicating that West Germanic parallels exist, but the word is attested relatively late in English. 8

5  The West Germanic Heritage of Yorkshire English 

133

2.3 Northern English as a Peripheral Variety The previously described phenomenon of lexical support still includes some form of language contact. But even that is not always needed as an explanation for the presence of certain words in northern English / Yorkshire. Such common retentions do not need to imply any specific language contact situation. The resulting geographical distribution was in a way also noticed by Björkman (1900–1902: 9): It is a fact that, after the West-Saxon period, numerous words appear in English, which are not found in Old English, but are of a distinctly English stamp and cannot have been introduced from Scandinavian. It is therefore possible that many of the words, considered as Scandinavian, did actually belong to the vocabulary of the [northern, peripheral] dialects not represented by any literary monuments of an earlier date.

Northern England and Scotland are geographically and demographically peripheral regions on a European scale, and many examples from dialectology show that peripheral regions tend to preserve features and items, including lexemes, that were lost in the geographical centre (Viereck et al. 2002: 95).9 Yorkshire and Scotland share this peripheral status within the Germanic language family with regions such as Friesland, Scandinavia and the Alps in contrast to southern England, the rest of the Netherlands and central Germany.10 Such accidental commonalities among (northern) English, Frisian and the Scandinavian languages should not be too easily taken for expressions of the previously mentioned lexical support (Århammar 1966: 310). Whereas intense language contact with North Germanic is conceivable for North Frisian and (Yorkshire) English, a similar contact situation is not attested for West and East Frisian, although multiple contacts between Scandinavians (‘Vikings’) and Frisians took place (IJssennagger  It would be wrong to look at peripheral regions as being only archaic. Dahl (2015a: 185–186) illustrates that such regions are different from the centre, with both common archaisms and innovations. 10  For the European dimension, see Haspelmath (2001); for a combined geographical-­chronological comparison of English-Scottish varieties, see McMahon and Maguire (2012: 152–157). 9

134 

A. Versloot

2012). Still, as a result of the peripheral dialect-geographical position and due to the mere geographical vicinity, Frisian and northern English varieties sometimes show lexical similarities with North Germanic. A good example is the word for ‘child’: WFri. bern, NFri. (FA) bjarn and so on (cf. NGmc barn, northern English / Scots barn, bairn vs. Standard English child and Du. / HG Kind). Reflexes of PGmc *barna- are also found in Gothic and Old High German (Kroonen 2013: s.v. *barna-). As the word was preserved in West and East Frisian without any significant ‘Viking’ contacts, there is no reason to assume any expansion or even lexical support from North Germanic in English. Another example from the set of Yorkshire English words under consideration in this chapter is teem ‘to pour’, according to OED (s.v. teem, v.2) ‘[a] borrowing from early Scandinavian’. North Germanic parallels are at hand here, such as the Modern Icelandic verb tæma ‘to empty, vacate’ or the Danish adjective tom ‘empty’. The stem is not found in modern West Germanic languages, except for some North Frisian dialects; the dialect of Sylt attests to tem ‘to poor’, labelled as a borrowing from Danish in Möller (1916: s.v. tem). The West Germanic languages, however, show a wider spread of the root, including OFri. tēma ‘to let (water) flow’ and Old Saxon (OS) tōmian ‘to release’. Its presence in several West Germanic languages indicates that we might be dealing with a peripheral relic.

2.4 The ‘Velar’ Argument The occurrence of /g/ and /k/ in palatalisation environments has been considered a very important criterion for the identification of Norse loanwords in English (e.g. Dance 2012: 1729). The lack of palatalisation of *g, *(s)k is mentioned in 132 items (14%) in the Gersum database as an argument in favour of their status as an Old Norse loanword. Nonetheless, in his recent publication, Laker (2021) has relativised this important phonological criterion, showing that this phenomenon is at most phonologically influenced by contact with speakers of Norse, but cannot be used to identify words as Norse loanwords on the individual lexical level solely by this criterion.

5  The West Germanic Heritage of Yorkshire English 

135

The traditionally postulated contrast between palatalising and assibilating English and non-palatalising North Germanic is, in my view, problematic. A superficial comparison of, for example, church versus Ic kirkja, Dan. kirke, suggesting such a contrast, can be misleading. The phonetic reality of North Germanic languages is more complex, with each language showing a different palatalised variant of /k/: Icelandic [kj], Faroese [tʃ], Norwegian and Swedish [ç]; various forms of palatalisation, [kj], [tj] and [ç], can also be found in traditional Danish dialects in words such as ‘church’.11 The comparative method results in a reconstruction of an allophonic contrast between palatal and non-palatal realisations in Old Norse on the basis of the widespread expressions of velar palatalisation in the modern varieties. Palatalisation is not always mentioned in the phonological descriptions in the handbooks, for example Heusler (1967: 53), Haugen (1984: 197) and Schulte (2018: 41–43). For Old Norse (from c.1050), Haugen assumes a common tendency towards palatalisation of velar plosives before front vowels, which finds written expression from the thirteenth century. The use of instead of before front vowels indicates that this palatalisation had not led to assibilation or affrication at this early stage (Haugen 1984: 228, 247–248, 265). Bandle (2012: 80–81) considers the tendency towards palatalisation to be a pan-North Germanic phenomenon, incidentally spelled out already in the earliest manuscripts. The Icelandic pattern with palatal allophones [kj] and [gj] is considered to represent the most archaic stage. The eleventh century mentioned by Haugen does not entirely reach back to the ninth and tenth centuries, but an earlier date for the palatalisation in North Germanic seems plausible for such a pan-North Germanic phenomenon, stretching from Sweden to Iceland (Haugen 1984: 339–342), the latter inhabited since the late ninth century. As an areal feature, palatalisation was equally present in the geographically adjacent English, Frisian and Saxon (Old Saxon) languages. The loss of palatalisation can therefore be considered a typically northern English and Scots phenomenon rather than a characteristic of Old Norse.

11

 See https://dialekt.ku.dk/dialektkort/#map=10 [accessed 28 May 2023].

136 

A. Versloot

An important addition made by Laker (2021: 107–110) is the interpretation that words with [tʃ] in northern dialects can be ascribed to later lexical expansion from the south. Accordingly, in the north, words with /k/ may represent the inherited forms and words with /tʃ/ are possibly due to lexical expansion from southern varieties. This is the exact opposite of the earlier interpretations, where words with /tʃ/, also in northern varieties of English, are taken as inherited and words with /k/ as (Norse) loanwords. More research is needed to establish the exact conditioning of the lexical distribution of /k/ and /tʃ/, but it seems a reasonable assumption that palatalisation was genuine in every Old English dialect, whereas assibilation to [tʃ] was only realised in combination with strong palatal triggers in the north, being rather a default in the south. The northern realisation [kj] was subsequently reinterpreted as /k/ in many words at a later stage, and this reinterpretation was applied to both inherited English words as well as Norse loanwords with [kj]. An example illustrating the problematic interpretation of the origin of words containing a velar is English gate. In the meaning ‘road’, the word is commonly interpreted as a Norse loanword (OED, s.v. gate, n.2: ‘A borrowing from early Scandinavian’). The Gersum database (Dance et  al. 2019: s.v. gate) labels the word with the code B2abc and mentions the absence of palatalisation of */g/ as a potential diagnostic, even if not entirely reliable. However, the etymon of gate has to be separated from the one underlying OE geat ‘gate, door’, OFri. jet ‘hole, opening’. English gate (OIc gata, HG Gasse) goes back to PGmc *gatwōn- f. ‘road, alley’, whereas OE geat comes from PGmc *gata- n. ‘hole’ (Kroonen 2013). Palatalisation of the g- in gate is not expected as the (semi-)vowels in the second syllable, *-wō-, blocked fronting of the *a in all forms, possibly with the exception of the nom.sg.12 The stem vowel of a nom.sg. form *gætwe13 was probably levelled out by the /a/ from the other paradigm

 It is commonly considered to be the result of fronting and subsequent retraction (Ringe and Taylor 2014: 203). 13  Compare Early Old English (Épinal Glossary, Pheifer 1974) with -wæ in such words like quiquae ‘quitch grass’, sualuuae ‘swallow’, gearuuae ‘yarn’. 12

5  The West Germanic Heritage of Yorkshire English 

137

forms (cf. Hogg and Fulk 2011: 129; Ringe and Taylor 2014: 191–192).14 The modern English word gate continues the phonology of *gatwōn-, but the semantics of *gata-. The YD mentions the word yat ‘gate’, which is the formal continuation of OE geat < *gata-. Both the OED and Gersum database refer to the High German cognate Gasse ‘street’, but this is apparently too remote to be considered of any relevance for the interpretation of the English attestation. On a closer inspection, the word appears also in North Frisian dialects: jaat in the dialect of Föhr and Amrum, where it actually shows a palatalised consonant (Sjölin 2002, s.v. jaat), but gaat in the neighbouring dialect of Sylt (Möller 1916: s.v. gaat); in both dialects, it refers to a countryside road. Norse influence could have played a role here, but such an influence is unlikely for Middle Low German (Schiller and Lübben 1875: s.v. gate). The interpretation of gate offered in this chapter serves as an illustration for all four arguments against Scandinavian origin of loanwords, mentioned in Sects. 2.1, 2.2, 2.3, and 2.4: (a) The lack of palatalisation is not necessarily decisive for words from northern English and Scots as a criterion for Norse origin (Sect. 2.4) and is actually not relevant for the word gate; (b) Even if the lexeme were attested only in High German and northern English / Scots, it would not necessarily exclude the possibility that these are relics from a wider West Germanic distribution (Sect. 2.3); (c) ….which is in this case confirmed by the identification of the word in North Frisian and Middle Low German (Sect. 2.1); (d) The retention of gate in the meaning ‘road, street’ in northern English and Scots may have been facilitated by bilingual EnglishNorse speakers, who knew the cognate gata in the Norse language (Sect. 2.2).  It is unclear whether /g/ was palatalised before /æ/ that developed from fronted PGmc */a/ in Anglian dialects, as illustrated by a parallel example with /k/ in the northern dialect form caff ‘chaff’ (see Laker 2021: 108 for caff and 89–90 for a discussion of /æ/). Laker considers the lack of palatalisation in caff as part of a broader trend, rather than a specific development of */k/ before /æ/. Lack of fronting before /æ/ would also leave northern yat ‘gate’ unexplained. 14

138 

A. Versloot

2.5 The Acquisition of a Bilingual Lexicon Sections 2.1, 2.2, 2.3, and 2.4 have presented four reasons why some (Yorkshire) English words, often considered to be of Scandinavian origin, may in fact be native / inherited. In the first place, the lesser-studied West Germanic varieties, in particular Frisian, may, in my opinion, provide additional evidence for the words’ attestation in West Germanic. Without conclusive formal criteria—the case for non-palatalised velars being weakened—these words should rather be considered inherited English words, whose northern English distribution may be the result of reinforcement (i.e. lexical support) by their use in Old Norse spoken in the Danelaw, or the mere manifestation of a coincidental retention in peripheral Germanic vernaculars, such as northern English, Frisian and various Scandinavian languages. In the thought-provoking approach taken by Emonds and Faarlund, the Middle English lexicon is considered from the perspective of an individual Norse speaker, where every Norse-English cognate could be felt as ‘Norse’. This allows the identification of two-thirds of the lexicon as ‘Norse’ (Emonds and Faarlund 2014: 54–55) from the perspective of the Norse speaker, leaving aside the question of the matrix or intended target language of the speaker as well as the actual etymological origins of the words (for a critical discussion, see Pons-Sanz 2015; Bech and Walkden 2016: 68–71; cf. Dahl 2015b). A schematised overview over the various etymological interpretations is given in Table 5.1. Table 5.1  Evaluation of lexical items in English with respect to the question of a potential Old Norse origin Etymological status?

E & F Gersum This chapter

1. Non-cognates 2. Cognates + ON-phono- / morphological issues I 3. Cognates + ON-phonological issues II, e.g. velars 4. Full cognates + ON-semantics

E N

E A,C1,2,4 N

N

C2

E (Sect. 2.4)

N

C3,5

5. Full cognates + diachronic or geographical issues 6. Full cognates (no issues)

N

B

E / N (Sect. 2.2) E / (N) (Sects. 2.1, 2.3) E (Sect. 2.5)

N

E & F = Emonds and Faarlund, E = English, N = Norse

5  The West Germanic Heritage of Yorkshire English 

139

The Gersum classification applies a series of categories with decreasing likelihood of borrowing, with high for A and D for low likelihoods (with further nuances). This chapter is more critical of some of the phonological criteria (Sect. 2.4) and assigns more weight to other West Germanic parallels (Sect. 2.1), irrespective of diachronic or geographical issues of distribution in Old and Middle English, adding the lexical support (Sect. 2.2) as a major factor. For words with semantic issues, I prefer to separate borrowing of lexical material from contact-induced semantic shifts. While my approach maximises the inherited English component and reduces the Norse component in the etymologies, I agree with Emonds and Faarlund (2014) that, from the perspective of the Norse speaker learning English, all groups in Table 5.1 but (1) could be perceived as Norse or be adjusted to the Norse L1 (4). In addition, even without any borrowing or substantial semantic shift, two related languages spoken by bilingual speakers can converge. There is a tendency among bilingual speakers to opt for shared lexical items (cognate forms) and contents (shared meanings) in a situation where they are confronted with multiple variants. This can be observed in the word choice of present-day bilingual Frisian-Dutch children in the Dutch province of Fryslân, where (West) Frisian is spoken as the first language by approximately half the population (Klinkenberg et  al. 2018). Children with L1-Frisian use more Frisian-Dutch cognates in their active production than could be anticipated on the basis of the proportion of such cognates in their receptive Frisian vocabulary (based on data from Dijkstra 2013: 85–88). This attitude leads to convergence of the lexicon of the two languages, without borrowing in the strict sense. We could imagine a similar convergence of the lexicon in a bilingual Old English-Old Norse situation. Bosma et al. (2019) provide experimental psycholinguistic evidence for the ability of children to identify regular sound correspondences between two related languages. All word pairs in the two languages that comply with such simple conversion rules will be part of the preferred overlapping lexicon in the aforementioned sense of preference for cognates among bilingual

140 

A. Versloot

language learners.15 As a result, bilingual speakers have a large set of lexical items at their disposal in a myriad of formal and semantic relations, with eventually loose relations with the words’ diachronic etymological identities (compare Townend 2002: 182–183).

3 Data and Methodology of the Yorkshire Case Study In the second part of this chapter, I consider a set of just over hundred words that are described as (potential) loans from Old Norse in Kellett’s (1994) Yorkshire Dictionary of Dialect, Tradition and Folklore (YD). The traditional dialects of Yorkshire belong to the Northern English dialects. Yorkshire is part of the historical Danelaw, the region that was under Danish control from 866–954, with York as its main city (Haywood 1995: 70). This region is known for its large number of Norse loanwords (Kolb 1965; Viereck et al. 2002: 94, Map C). The YD is an idioticon, based on an extensive list of sources, including Wright’s (1905) English Dialect Dictionary and the OED. The introduction suggests that ‘[m]ost of the etymology is supported by the authority of the books in the bibliography’ (Kellett 1994: xiv), but no detailed references are provided. This chapter takes the perception of Norse influence on the English language as represented in the YD as a starting point of the case study. The words marked with ‘ON’ and ‘? ON’ in this book are considered to be of Scandinavian origin by a wider tradition of— partly gradually outdated—scholarly work. It was not possible to trace the individual sources used for the identification of individual etymologies. Moreover, the etymology of other words in YD (approx. 4700 in  One such correspondence rule in the Frisian-Dutch context is WFri. ân [ɔ̃ːn] = Du. and [ant], such as in WFri. lân, sân, brân = Du. land, zand, brand ‘land, sand, fire’. As a consequence, the Frisian form of the Dutch verb branden ‘to burn’ will be brâne. This is indeed the preferred form among most present-day speakers of Frisian (Goeman et al.: s.v. branden). However, the historical form of the verb showed metathesis of /r/: PGmc *brannjan > OFri. barna, hence WFri. baarne. The earliest instances of brâne appear in the early eighteenth century, while baarne is still in use in the late twentieth century, in particular in the peripheral northern region. 15

5  The West Germanic Heritage of Yorkshire English 

141

total), which may or may not be ascribed to the influence of Old Norse, was not further traced / analysed. The etymologies of claimed Norse loanwords in the Yorkshire English dialect (YE) will be scrutinised and re-evaluated in light of the issues raised in Sect. 2, with special attention to potential parallels in Frisian, Dutch and other West Germanic varieties. A systematic comparison is made with the etymologies provided in the current online version of the OED (2000–), which is more restrictive in the assignment of words to the category of Old Norse borrowings than earlier versions. The evidence presented in Sect. 4 allows us to conclude that the Norse character of the English language is still overestimated, for the reasons discussed in Sect. 2. Words were labelled as being of North Germanic (NGmc) or West Germanic (WGmc) origin, sometimes with some level of doubt (NGmc? 3x, WGmc? 3x). In the counting for Sect. 4, these six items were counted with the unambiguous instances. Five words were either entirely ambiguous or were fully acceptable as West Germanic in their formal origin but showed a clear North Germanic semantic profile. Such words have been labelled WGmc-NGmc (cf. Gersum classification C3) and in the analysis counted as ‘NGmc’. Three words could not be interpreted. Among the words with West Germanic origin both in form and general meaning, there are seven for which the geographical distribution in combination with details of their semantics points towards lexical support as a likely source for their preservation in Yorkshire English, but such an interpretation can only be conjectural (cf. Gersum classification C5).

4 Comparison of YD, OED and This Study The labelling of 112 entries in the YD as Norse loanwords was the starting point of this case study. All words have also been checked in the current online version of the OED (2000–).16 For twelve words, no satisfactory etymology was found, either in the OED or in this study. Figure  5.1 shows the interpretations for the remaining overlapping  I would like to thank research – MA – student Merel Luberti (University of Amsterdam) for her assistance in the analysis of the OED references to the items in this study. 16

142 

A. Versloot

Fig. 5.1  Etymological classification of the words identified as Norse loanwords in the YD. (The abbreviations should be read as follows: NGmc-NGmc = both this study and OED consider the word to be a Norse loanword; WGmc-WGmc = both this study and OED consider the word to be an inherited, West Germanic word; WGmc-NGmc = this study considers the word to be an inherited, West Germanic word, whereas it is labelled as a Norse loanword in the OED)

hundred etymologically interpretable words. Detailed etymological information and interpretations can be found in Appendices 1–3. Figure 5.1 shows consensus between the OED and this study in seventy-­one instances: forty-seven words are considered to be of Old Norse origin, while twenty-four are understood to be words of West Germanic / Old English origin. In the remaining twenty-nine cases, this study prefers a West Germanic etymology for the words, considering the issues described in Sect. 2. If the interpretations in the YD are representative of the generous interpretations as Norse loanwords found in earlier scholarship and handbooks, this study suggests that about half of them deserve a critical review. When we consider only those words that are

5  The West Germanic Heritage of Yorkshire English 

143

presented as Norse loanwords in the current online version of the OED, almost 40% of them may come out as being inherited, West Germanic words after all. A critical revision of presumably Old Norse loanwords in English could thus lead to a substantial reduction of positively identified loanwords. For seven out of the twenty-nine instances of ‘new’ West Germanic words, lexical support of Norse is not unlikely. However, one needs more data about the spread of cognates in earlier stages of English and a comparison with, for example, the North Frisian situation in order to make more conclusive statements about the likelihood of such an interaction between English and Norse. The label lexical support can never be a categorical interpretation, only a probability. As an additional step of the analysis, the list of words from the YD study was compared with corresponding lemmas in the Gersum database. Thirty-two words could be identified in the database. Eight of these words are considered to be of West Germanic origin in the OED, assigned to the C or D-categories in the Gersum database, six labelled with CC or CCC, where repetition of the consonant marks a decreasing likelihood of an Old Norse origin. For the remaining twenty-four words identified in the Gersum database, there is little agreement between the interpretations suggested in the present analysis and the Gersum classification: there is no significant correlation between the Gersum main classifications (A, B, C, D) and the interpretations West versus North Germanic in this analysis.

5 Conclusion A careful analysis of West Germanic languages, in particular but not only Frisian, provides evidence for a likely Old English / West Germanic origin of many words commonly considered to be of Norse origin, not only in older scholarship but also in the current version of the OED. An extension of the data underlying the etymological interpretation of English lexemes is therefore a fruitful enterprise. The number of Norse loanwords in Present-Day English, including dialects such as the ones in Yorkshire,

144 

A. Versloot

seems to have been overestimated in older scholarship but, even with the more conservative estimations nowadays, further scrutiny of all candidates seems desirable. At the end of the day, even after closer scrutiny, a stock of English words will remain that are positively of Old Norse origin. Together with the instances of lexical support, semantic adjustments and through the selection of shared cognates by bilingual speakers (Sect. 2.5), they bear witness to Norse-English language contact in the Danelaw during the Middle Ages.

Appendices Key to Appendices 1, 2, 3: Yorkshire:    English:    POS:      APV:      Gersum:    OED entry:  OED:    

words from YD (Kellett 1994) Standard English translation or description part of speech: V = verb; N = noun; A = adjective / adverb etymological interpretation by this author classification in the Gersum database (Dance et al. 2019) corresponding entry in the OED (2000–) etymological interpretation by the OED (2000–)

Etymological labels used: NGmc:       word of North Germanic (Old Norse) origin and thus borrowed WGmc:      word of West Germanic origin and thus inherited ?:          some level of doubt NGmc / WGmc: decisive North Germanic input, often in the semantics, but inherited language material or word formation processes may play a role as well

5  The West Germanic Heritage of Yorkshire English 

145

Appendix 1: APV and OED: North Germanic Words Yorkshire

English

POS APV

Gersum OED entry

OED

addle

to earn

V

C3c

NGmc

NGmc

addle

arval bread

funeral cake

N

NGmc

arval

NGmc

axle-tooth

molar

N

NGmc

axle-tooth

NGmc

brandrith, -ree

moveable iron frame

N

NGmc?

brandreth

NGmc

carr

marsh(y woodland) N

NGmc

carr, n.2

NGmc

cleg

horsefly

N

NGmc

cleg

NGmc

cletch

family of young (e.g. children, chickens)

N

NGmc

cletch

NGmc

deg

to sprinkle

V

NGmc

dill

to soothe, dull (pain)

V

NGmc

dolop

lump of something N soft

NGmc

ettle

to intend, aim, attempt

V

NGmc

flags

paving or floorstones

N

NGmc

flit

to move house

V

NGmc

garth

N small, grassed enclosure adjoining a house…

NGmc?

ghyll, gill

deep and wooded ravine

N

NGmc

gimmer

young female sheep

N

gloppened, astonished glottened

A1bc

deg

NGmc

dill, v.2

NGmc

dollop, n.

NGmc

ettle

NGmc

flag, n.2

NGmc

flit, v.

NGmc

garth

NGmc

gill, n.4

NGmc

NGmc

gimmer, n.2

NGmc

A

NGmc / WGmc

gloppen, v. NGmc gowk

NGmc

grain, n.2

NGmc

hag, n.3

NGmc

happen, v.

NGmc

gowk

cuckoo

N

NGmc

grain

prong of a fork, branch of a tree

N

NGmc

hagg

division of a wood; N …to be felled

NGmc

happen

to have something happen to one

NGmc / WGmc

V

CCC5c

A1*c

C1

A1bc

A1b

CC1c / CC3

(continued)

146 

A. Versloot

(continued) Yorkshire

English

POS APV

keld, kell

well

N

Gersum OED entry

NGmc

keld, n.2

OED NGmc

kilp

pot-hook

N

NGmc

kilp

NGmc

kittlin

kittin

N

NGmc

kitling

NGmc

laik

to play

V

NGmc

lake, v.1

NGmc

laithe

barn

N

NGmc

lathe, n.2

NGmc

lam

to strike hard

V

NGmc / WGmc

lam

NGmc

lop

flea

N

NGmc?

lop, n.2

NGmc

lug

something (such as a handle) that projects like an ear

N

NGmc

lug, n.1

NGmc

mense

decency, neatness

N

NGmc

C2 / CC1 mense, n.1 NGmc

mig, muck

muck, manure

N

NGmc

B1

muck, n.1

nieve

fist

N

NGmc

B2c

nieve

NGmc

poke

sack, bag

N

NGmc

poke, n.1

NGmc?

A1*bc

NGmc

seeaves

rushes

N

NGmc

seave

NGmc

seg

small metal stud… in a shoe

N

NGmc

seg, n.3

NGmc

skrike

to shriek

V

NGmc

skrike

NGmc

slack

depression in the ground

N

NGmc

slack, n.1

NGmc

snod

smooth, sleek, short

A

NGmc

snod

NGmc

stee

ladder

N

NGmc / WGmc

sty, n.2

NGmc

BB2a

CC1c / CC3c

steg

gander

N

NGmc

steg

NGmc

stithy

anvil (aambeeld)

N

NGmc

stithy

NGmc

stour, stower

rung of a ladder

N

NGmc

stower, n.1 NGmc

swarf, swarth

grit worn from grindstones

N

NGmc / WGmc

swarf, n.2

NGmc?

tyke

dog

N

NGmc

tyke

NGmc

wapentake

division of a shire

N

NGmc

wapentake NGmc

whinny

gorze, furze [plant] N

NGmc

whinny, n.2

NGmc

yawd

horse of inferior breeding

NGmc

yaud

NGmc

N

147

5  The West Germanic Heritage of Yorkshire English 

Appendix 2: APV and OED: West Germanic Words Yorkshire

English

POS APV

arse

posterior

N

WGmc

Gersum OED entry arse

WGmc

band

string, rope

N

WGmc

band, n.1

WGmc

barf

hill, esp. long and low

N

WGmc CCC5a

barrow, n.1

WGmc

barn, bairn

child

N

WGmc DD2

bairn

WGmc

bleck

thick and dirty greas N

WGmc

bleck n.

WGmc

boose

division in cowshed

N

WGmc CCC1

boose

WGmc

brig

bridge

N

WGmc

brig, n.1 / bridge

WGmc

clap

to apply quickly, esp. V the hand, slap

WGmc

clap, n.1

WGmc

dee

to die

V

WGmc C1a

die, v.1

WGmc

groop

drain in a cowshed

N

WGmc

groop

WGmc

handsel

money given to strike a bargain,…

N

WGmc CC4

handsel

WGmc

hoss

horse

N

WGmc

hoss

WGmc

ice shoggles

icicles

N

WGmc

icicle

WGmc

kittle

to tickle

V

WGmc

kittle, v.1

WGmc?

laverock

skylark

N

WGmc

lark / laverock

WGmc

lig(g)

to lie, to lay

V

WGmc CC2

lig/lie, v.1

WGmc

ling

long, slender sea fish N

WGmc

ling, n.1

WGmc

mickle

much, greater

A

WGmc CC2c

mickle

WGmc

nang-nail, anger-nail

in growing toenail

N

WGmc

agnail

WGmc

reckon

to pretend, think, consider

V

WGmc

reckon

WGmc

scuttle

basket for holding meal, etc.

N

WGmc

scuttle, n.1 WGmc

strang

strong

A

WGmc

strong

WGmc

yacker

acre

N

WGmc

acre

WGmc

yule

Christmas

N

WGmc CC2c

yule

WGmc

OED

148 

A. Versloot

Appendix 3: APV: West Germanic vs. OED: North Germanic For sources, see the main text, particular fn. 6. Additional etymological sources are as follows: EB =     online version of Philippa et al. (2003) GTB =    portal to diachronic Dutch dictionaries, Old, Middle and (early) Modern Dutch: https://gtb.ivdnt.org/search/ [accessed 28 May 2023] IOB =    Ásgeir Blöndal Magnússon (1989) LAE =    Orton et al. (1978) Linguistic labels (exhaustive): Dial.   dialectal EFri.   East Frisian HG     High German ME     Middle English NFri.    North Frisian OE    Old English OFri.   Old Frisian ON     Old Norse PWGmc  Proto-West Germanic Sat.    East Frisian from the Saterland region Sy.     North Frisian from the island Sylt Wang.   East Frisian from the island Wangerooge (extinct) WFri.    West Frisian WGmc   West Germanic YE       Yorkshire English

5  The West Germanic Heritage of Yorkshire English 

149

Yorkshire

English

OED POS Gersum entry

beck

stream

N

beck

WGmc?; OE bece, bec, bæc ‘brook’; the variant with single /k/ is dominant in West Germanic, but double-spelled consonants, indicating a closed syllable, are also found in ODu. (GTB, s.v. beki) and in many modern North German place names in -beck (Haverbeck, etc.)

bensel

to beat, thrash

V

bensel, n.

WGmc; both meaning (‘to beat’) and form are found in Oldenburg Low German (Böning et al. 1998: s.v. bensel).

blaeberry

bilberry

N

blaeberry

WGmc: YE (/ble:/) can possibly be derived from OE blǣ, rather than Norse (OIc blá); berry is WGmc anyway, cf. HG Beere.

cam

bank, slope, ridge

N

cam, n.2

WGmc; both Du. kam and Scots came have both meanings of ‘comb’ and ‘ridge’. Cf. for the phonological development YE lam ‘lamb’ (cf. kuəm ‘comb’) (Wright 1892: § 281).

crake

crow

N

crake, n.

WGmc; NFri. (Sy.) kreek < OFri. *krek < *PWGmc *krak-, also various German dialects: krack, krak(e). http://woerterbuchnetz.de

creeaked

crooked

A

crook, n. and adj.

WGmc; EFri. (Sat.) krouke ‘scythe handle; NFri. (Sy.) kruk ‘bowed’

durn

doorpost, gatepost

N

durn

WGmc: EFri. (Wang.) durn ‘door’.

eldin

kindling, firewood

N

elding, n.1

WGmc; NFri. (Sy.) jöl’ing ‘kindling’; jöl ‘fire’ = OS eld (Tiefenbach 2010: s.v. eld); a morphological parallel with the ‘fire’-lexeme is found in EFri. (Sat.) fjúrenge, HG Feuerung. Both the formation and the lexeme eld- can be found in West Germanic. Lexical support from Norse (OIc elding) may have played a role here.

BB1b

Class APV

(continued)

150 

A. Versloot

(continued) OED POS Gersum entry

Yorkshire

English

flaik, fleeak

hurdle, railings

gain

quick A (way), near

gat

got (past tense)

gate

N

Class APV

flake, n.1

WGmc; The original root vowel is *a in open syllable (IOB, s.v. fleki), cf. Du. vlaak ‘hurdle’. YD flaik [e:] and fleeak [iə] can be derived from OE *a in open syllable (LAE, map Ph60). YE fleik (Wright 1892: § 87) rather points towards *fleke.

A1c

gain, adj.

WGmc; PGmc *gagin (EB) developed into OE gæġn > ME gein > YE geən (Wright 1892: § 84). The lack of palatalisation before *æ is not conclusive.

V

A1*c

get, v.

WGmc; the strong verb PGmc *getana- is amply attested in all Germanic languages (see OED), e.g. OS bigetan ‘to seize’ (Tiefenbach 2010). The past tense vowel a is the regular quality in the strong class 5 (OE æ, Northern English a). Apparently, YE generalised the vowel from the singular here.

street, way

N

B2abc

gate, n.2

WGmc; see the chapter text, end of Sect. 2.4.

gilt

young sow

N

gilt, n.1

WGmc; cf. Du. (dial.) gelte ‘young sow’. The lack of palatalisation before i is not conclusive (for parallel contexts, see Laker 2021: 109).

ing

meadow

N

ing, n.

WGmc; the word is very frequent in Dutch and Low German as a field name (Berkel and Samplonius 2018: s.v. eng), where it mostly refers to cultivated acres. The semantic shift to ‘meadow’ may have been inspired by Norse, but could just as well be a reaction to a different physical environment.

(continued)

5  The West Germanic Heritage of Yorkshire English 

151

(continued) OED POS Gersum entry

Yorkshire

English

kist

large box, chest

N

kist, n.1

WGmc; loanword from L cista, widely attested in West Germanic (EB, s.v. kist). For the initial k- compare gilt.

lisk

groin

N

lisk

WGmc; cf, WFri. ljisk. The YE form shows shortening of the vowel, which is also not found in Scandinavian (OIc ljóski) (EB, s.v. lies). Word final -sk can be retained in YE (Wright 1892: § 312.6).

mawk

maggot

N

mawk, n.1

WGmc: the OED is undecided between OIc maðkr and a local diminutive form from PGmc *maþōn-. Similar diminutives are also attested in Frisian: WFri. maits, maik; EFri. (Wang.) maðuuk. The YE /ɔ:/ < *au suggests syncope of /ð/ in the sequence -aðu- rather than vocalisation in -aðk-, cf. the Wang. form.

nay

no

A

nay, adv.

WGmc; English no is the regular (southern) continuation of OE nā, with ā < PGmc *ai. YE is the normal continuation of ME ai, ei and could very well represent the Norse adverb (OIc nei; Wright 1892: § 84). However, a special development in final position, joining the development of lengthened short a would give YE /eə/ (spelling Wright). A similar development is assumed for they (Cole 2018: 191–200). Lexical support is not unlikely here.

A1*

Class APV

(continued)

152 

A. Versloot

(continued) Yorkshire

English

OED POS Gersum entry

rive

to tear

V

B2a

rive, v.1

WGmc; with OFri. ūtrīva, there is a cognate in a closely related WGmc language, so there is no reason not to consider it a North / North Sea Germanic word, with a peripheral distribution in YE.

sca(u)r

cliff or rocky outcrop

N

A1abc

scar, n.1

WGmc; the word is derived from a root PGmc *sker- ‘to cut’. WFri. skar, Du. schaar designate plots of land in the sense of share. The meaning is probably influenced by Norse (cf. ing).

seeat

seat

N

A1*

seat

WGmc; the word has West Germanic parallels (see OED), to which OFri. sēte ‘farmstead’ can be added.

skep

WGmc; WFri. skeppe, Du. schep ‘spade, shovel’ (cf. EB). IOB considers a Low German origin of the Icelandic word.

skep, skip basket, N coal bucket

Class APV

skitters

diarrhoea

N

skitte, n.1 WGmc; the OED explicitly mentions West Germanic cognates as potential sources. Given the pan-Germanic attestations, a local origin seems obvious, when the ‘velar’ argument is considered not conclusive.

skive

to split, pare, leather or hide

V

skive, v.1

WGmc; OFri. attests to skīved ‘divided’. IOB considers the verb to be derived from the noun skífa, which it says to be a loanword from Low German.

(continued)

5  The West Germanic Heritage of Yorkshire English 

153

(continued) OED POS Gersum entry

Yorkshire

English

stang

pole, shaft, N stake

storken

to set, to V stiffen as it cools

storken

stoup

post, gatepost

N

stoop, n.1 WGmc; archaic Du. stolp(e) ‘pole, post’ (GTB, s.v. stolp), also Middle Low German (IOB, s.v. stólpi).

tang

projecting part of knife

N

tang, n.1

WGmc; pan-Germanic word, e.g. Du. tang, HG Zange ‘pliers, tongs’. The meaning ‘protruding tip of land’, nowadays common in Icelandic (IOB), is also found in Old Dutch (GTB, s.v. tanga). Some semantic interference or reinforcement from Norse is not unlikely.

teem

to pour

V

teem, v.2

WGmc; compare OFri. tēma ‘to let (water) flow’; see the discussion in the text at the end of Sect. 2.3.

CC1abc

Class APV

stang, n.1 WGmc; the OED mentions various West Germanic cognates, to which one may add WFri. stange ‘pole’. YE preserves pan-Germanic lexicon here. WGmc; the verb has a pan-­ Germanic spread (IOB, s.v. storkna). The suffix -en was productive in English itself, cf. lighten that does not have a cognate in Scandinavian (OED, s.v. -en, suffix5).

References Århammar, Nils. 1966. Nordische Lehnwörter und lexikalische Stützung im Nordfriesischen. In Nordfriesisches Jahrbuch,  II, 302–316. Bredstedt: Nordfriisk Instituut. Århammar, Nils R. 1969. Die Amringer Sprache: Die Amringer Literatur, mit e. Anh. über d. Amringer Pflanzen- u. Vogelnamen/Nils Århammar,  2nd edn. Itzehoe-Münsterdorf: Hansen & Hansen.

154 

A. Versloot

———. 1989. Zur Rekonstruktion des altfriesischen Lexikons mit Hilfe der neufriesischen Dialekte. In Philologia Frisica 1988, 94–128. Leeuwarden: Fryske Akademy. Bandle, Oskar. 2012. Die Gliederung des Nordgermanischen, Reprint der Erstauflage (1973) mit einer Einführung von Kurt Braunmüller, Beiträge zur Nordischen Philologie, 47. Tübingen & Basel: A. Francke Verlag. Bech, Kristin, and George Walkden. 2016. English is (Still) a West Germanic Language. Nordic Journal of Linguistics. 39: 65–100. Berkel, G. van, and Kees Samplonius. 2018. Nederlandse plaatsnamen verklaard. Rotterdam: Mijnbestseller. Björkman, Erik. 1900–1902. Scandinavian Loan-Words in Middle English, Studien zur englischen Philologie, 7 and 11. 2 vols. Halle: Max Niemeyer. Blom, Gosse, and Siebren Dyk. 2019. Graet Hylper Wordebook. Leeuwarden: Afûk. Böning, Hermann, Jürgen Lange, Hans-Joachim Mews, and Dieter Stellmacher. 1998. Plattdeutsches Wörterbuch für das Oldenburger Land. Oldenburg: Isensee. Bosma, Evelyn, Elma Blom, Eric Hoekstra, and Arjen Versloot. 2019. A Longitudinal Study on the Gradual Cognate Facilitation Effect in Bilingual Children’s Frisian Receptive Vocabulary. International Journal of Bilingual Education and Bilingualism 22: 371–385. Boutkan, Dirk, and Sjoerd Michiel Siebinga. 2005. Old Frisian Etymological Dictionary, Leiden Indo-European Etymological Dictionary Series, 1. Leiden: Brill. Bragg, Melvyn. 2004. The Adventure of English: The Biography of a Language, 2nd edn. London: Hodder & Stoughton. Buma, Wybren Jan. 1949. Die Brokmer Rechtshandschriften. The Hague: M. Nijhoff. ———. 1961. De eerste Riustringer codex. The Hague: M. Nijhoff. Cole, Marcelle. 2018. A Native Origin for Present-Day English They, Their, Them. Diachronica 35: 165–209. Corbett, John, J. Derrick McClure, and Jane Stuart-Smith. 2003. The Edinburgh Companion to Scots. Edinburgh: Edinburgh University Press. Dahl, Östen. 2015a. Grammaticalization in the North: Noun Phrase Morphosyntax in Scandinavian Vernaculars. Berlin: Language Science Press. ———. 2015b. Linguistic Mysteries around the North Sea. In From Clerks to Corpora: Essays on the English Language Yesterday and Today, ed. Gunnel Melchers, Britt Erman, and Philip Shaw, 17–34. Stockholm: Stockholm University Press. Dance, Richard. 2012. English in Contact: Norse. In Historical Linguistics of English: An International Handbook, ed. Alexander Bergs and Laurel Brinton, II, 1724–1737. Berlin: de Gruyter.

5  The West Germanic Heritage of Yorkshire English 

155

Dance, Richard, Sara M. Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge, Cardiff and Sheffield. https://www.gersum.org/. Accessed 28 May 2023. Dijkstra, Jelske. 2013. Growing up with Frisian and Dutch. Leeuwarden: Fryske Akademy. Ehrentraut, Heinrich Georg. 1849. Friesisches Archiv: Eine Zeitschrift für friesische Geschichte und Sprache, I. Oldenburg: Rudolf Schwarz. ———. 1854. Friesisches Archiv: Beiträge zur Geschichte der Friesen und ihrer Sprache, auch der Grafschaften Oldenburg und Delmenhorst,  II.  Oldenburg: Rudolf Schwarz. Emonds, Joseph E., and Jan Terje Faarlund. 2014. English: The Language of the Vikings. Olomouc: Palacký University. Faltings, Volkert F. 2010. Etymologisches Worterbuch der friesischen Adjektiva. Berlin: De Gruyter. Fort, Marron Curtis. 2015. Saterfriesisches Wörterbuch mit einer phonologischen und grammatischen Übersicht. Hamburg: H. Buske. Goeman, Tom, Johan Taeldeman, and Piet van Reenen, eds. MAND-Database. http://www.meertens.knaw.nl/mand/database/. Accessed 28 May 2021. Haspelmath, Martin. 2001. The European Linguistic Area: Standard Average European. In Language Typology and Language Universals: An International Handbook, ed. Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible, II, 1492–1510. Berlin: de Gruyter. Haugen, Einar. 1984. Die skandinavischen Sprachen: Eine Einführung in ihre Geschichte, trans. Magnús Pétursson. Hamburg: H. Buske. Haywood, John. 1995. The Penguin Historical Atlas of the Vikings. London: Penguin Books. Heusler, Andreas. 1967. Altisländisches Elementarbuch,  7th edn. Heidelberg: C. Winter. Hoad, T.F. 1986. The Concise Oxford Dictionary of English Etymology. Oxford: Clarendon Press. Hofmann, Dietrich, and Anne Tjerk Popkema. 2008. Altfriesisches Handwörterbuch. Heidelberg: C. Winter. Hogg, Richard M., and Robert D. Fulk. 2011. A Grammar of Old English, II: Morphology. Chichester: Wiley-Blackwell. IJssennagger, Nelleke. 2012. Vriend, vazal en vijand? Contacten tussen Friezen en Scandinaviërs in de Vikingtijd: Teksten en objecten. De Vrije Fries 92: 37–59. Jensen, Peter, Adeline Petersen, and Bo Sjölin. 1994. Freesk uurdebuk = Friesisches Wörterbuch: Wörterbuch der Wiedingharder Mundart. Neumünster: Wachholtz.

156 

A. Versloot

Kaiser, Livia. 2021. Runes across the North Sea from the Migration Period and Beyond: An Annotated Edition of the Old Frisian Runic Corpus.  Berlin: De Gruyter. Kastovsky, Dieter. 2006. Vocabulary. In A History of the English Language, ed. Richard Hogg and David Denison, 199–270. Cambridge: Cambridge University Press. Kellett, Arnold. 1994. Dictionary of Yorkshire Dialect, Tradition and Folklore. Otley: Smith Settle. Kellner, Birgit. 2006. Sölring Uurterbok: Wörterbuch der sylterfriesischen Sprache: Deutsch-Sölring, Sölring-Deutsch. Sylt: Söl’ring Foriining. Klein, Ernest. 1966. A Comprehensive Etymological Dictionary of the English Language. Amsterdam: Elsevier. Klinkenberg, Edwin, Reitze J. Jonkman, and Nika Stefan. 2018. Taal yn Fryslân: De folgjende generaasje. Leeuwarden: Fryske Akademy. Kolb, Eduard. 1965. Skandinavisches in den nordenglischen Dialekten. Anglia: Zeitschrift für englische Philologie 83: 127–153. Kroonen, Guus. 2013. Etymological Dictionary of Proto-Germanic. Leiden: Brill. Laker, Stephen. 2021. Palatalization and Assibilation of /k/ in English and Scottish Place-Names. NOWELE 74: 80–115.  Löfstedt, Ernst. 1928. Die nordfriesische Mundart des Dorfes Ockholm und der Halligen. Lund: Gleerup. ———. 1931. Nordfriesische Dialektstudien. Lund: Gleerup. ———. 1963–1965, 1966, 1967, 1969. Beiträge zur nordseegermanischen und nordseegermanisch-nordischen Lexikographie. Niederdeutsche Mitteilungen, 19–21: 281–345; 22: 39–64; 23: 11–61; 25: 25–45. Magnússon, Ásgeir Blöndal. 1989. Íslensk orðsifjabók. Reykjavik: Orðabók Háskólans. McMahon, April M.S., and Warren Maguire. 2012. Quantitative Historical Dialectology. In Analysing Older English, ed. David Denison, Ricardo Bermúdez-Otero, Chris McCully, and Emma Moore, 140–158. Cambridge: Cambridge University Press. Minssen, Johann Friedrich. 1854. Mittheilungen aus dem Saterlande. In Friesisches Archiv: Beiträge zur Geschichte der Friesen und ihrer Sprache, auch der Grafschaften Oldenburg und Delmenhorst, ed. Heinrich Georg Ehrentraut, II, 135–227. Oldenburg: Rudolf Schwarz. Möller, Boy P. 1916. Söl’ring Uurterbok: Wörterbuch der Sylter Mundart. Hamburg: Meissner. Munske, Horst Haider. 1973. Der germanische Rechtswortschatz im Bereich der Missetaten: Philologische und sprachgeographische Untersuchungen. Berlin: de Gruyter.

5  The West Germanic Heritage of Yorkshire English 

157

———, ed. 2001. Handbuch des Friesischen / Handbook of Frisian Studies. Tübingen: Max Niemeyer. Nielsen, Hans Frede. 1985. Old English and the Continental Germanic Languages: A Survey of Morphological and Phonological Interrelations, Innsbrucker Beiträge zur Sprachwissenschaft, 33, 2nd edn. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. [OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 28 May 2023. Ordbog over det danske Sprog (Danske Sprog- og Litteraturselskab). https://ordnet.dk/ods. Accessed 28 May 2023. Orton, Harold, Stewart Sanderson, and John David Allison Widdowson. 1978. The Linguistic Atlas of England. London: Croom Helm. Pheifer, Joseph John Michael Donovan. 1974. Old English Glosses in the Épinal-­ Erfurt Glossary. Oxford: Clarendon Press. Philippa, Marlies, F.  Debrabandere, Arend Quak, Tanneke Schoonheim, and Nicoline van der Sijs. 2003. Etymologisch Woordenboek van het Nederlands, digital and extended version. Amsterdam: Amsterdam University Press. http://www.etymologiebank.nl/. Accessed 28 May 2023. Pons-Sanz, Sara M. 2013. The Lexical Effects of Anglo-Scandinavian Linguistic Contact on Old English, Studies in the Early Middle Ages, 1. Turnhout: Brepols. ———. 2015. Review of English: The Language of the Vikings, by Joseph Emonds and Jan Terje Faarlund (Olomouc: Palacký University, 2014). Apollon. http:// www.apollon.uio.no/bokanmeldelser/2015/faarlund_2015.html. Accessed 28 May 2023. ———. 2016. A Study of Aldred’s Multiple Glosses to the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 301–328. Berlin: De Gruyter. ———. 2017. Reassessing the Semantic History of OE brēad/ME brēd. English Language and Linguistics 20: 47–67. Pyles, Thomas, and John Algeo. 1982. The Origins and Development of the English Language, 3rd edn. New York: Harcourt Brace Jovanovich. von Richthofen, Karl. 1840. Altfriesisches Wörterbuch. Göttingen: Dieterich. http://archive.org/details/altfriesischesw00richuoft. Accessed 28 May 2023. Ringe, Don, and Ann Taylor. 2014. The Development of Old English. Oxford: Oxford University Press.

158 

A. Versloot

Schiller, Karl, and August Lübben. 1875. Mittelniederdeutsches Wörterbuch. Bremen: Kühtmann. https://drw.hadw-­bw.de/drw-­cgi/zeige?index=siglen& term=schiller-­luebben&firstterm=schiller-­luebben+i. Accessed 28 May 2023. Schulte, Michael. 2018. Urnordisch: Eine Einführung, Wiener Studien zur Skandinavistik, 26. Vienna: Praesens Verlag. Siebs, Theodor. 1889. Zur Geschichte der englisch-friesischen Sprache. Halle: Max Niemeyer. http://www.archive.org/details/zurgeschiehtede00siebgoog. Accessed 28 May 2023. ———. 1901. Geschichte der friesischen Sprache. In Grundriss der germanischen Philologie, ed. Hermann Paul, II, 1152–1464, 2nd edn. Strassburg: Trübner. Sjölin, Bo. 2002. Fering-Öömrang Wurdenbuk = Wörterbuch der friesischen Mundart von Föhr und Amrum. Neumünster: Wachholtz. ———. 2006. Etymologisches Handwörterbuch des Festlandsnordfriesischen. Kiel: Fach Friesische Philologie, CAU. Sjölin, Bo, Alastair G.H. Walker, and Ommo Wilts. 1988. Friesisches Wörterbuch: Wörterbuch der Mooringer Mundart = Frasch uurdebök. Neumünster: Wachholtz. Spenter, Arne. 1968. Der Vokalismus der akzentuierten Silben in der Schiermonnikooger Mundart: Eine geschichtliche Studie des autochthonen westfriesischen Inseldialekts. Copenhagen: Munksgaard. Tiefenbach, Heinrich. 2010. Altsächsisches Handwörterbuch = A Concise Old Saxon Dictionary. Berlin: De Gruyter. Townend, Matthew. 2002. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English. Turnhout: Brepols. Veen, K. F. van der, and R. de Boer. 1984–2011. Wurdboek fan de Fryske taal = Woordenboek der Friese taal. Leeuwarden: Fryske Akademy. https://gtb.ivdnt. org/search/. Accessed 28 May 2023. Viereck, Wolfgang, Karin Viereck, and Heinrich Ramisch. 2002. Dtv-Atlas englische Sprache. Munich: Deutscher Tachenbuch Verlag. Visser, Willem, and Siebren Dyk. 2002. Eilander Wezzenbúek: Woordenboek van het Schiermonnikoogs. Leeuwarden: Fryske Akademy. Wright, Joseph. 1892. A Grammar of the Dialect of Windhill, in the West Riding of Yorkshire. London: Paul, Trench, Trübner & Co. ———. 1905. The English Dialect Dictionary. London: Henry Frowde. Zoëga, Geir T. 1975. A Concise Dictionary of Old Icelandic. Toronto: University of Toronto Press.

6 Reframing the Interaction between Native Terms and Loanwords: Some Data from Occupational Domains in Middle English Louise Sylvester and Megan Tiddeman

1 Introduction This case study is concerned with the relationship between native terms and loanwords in the later medieval period. This is the period in which, compared to that of Old English, the vocabulary of English is marked by extensive lexical borrowing from French, one of the main languages of record following the Norman Conquest. The evidence from the Oxford English Dictionary (OED) and the Middle English Dictionary (MED) is that 44%–48% of headword entries are borrowings from French and/or Latin (it is not always possible to distinguish between the two sources) (Durkin 2014: 256–257). This influx of foreign-language lexical items has been described as going beyond the conventional understanding of word borrowing. It has been characterised as a ‘partial relexification of L. Sylvester (*) • M. Tiddeman University of Westminster, London, UK e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_6

159

160 

L. Sylvester and M. Tiddeman

English from French and Latin sources’ (Schendl 2000: 78) and as a ‘lexical transfer’ (Trotter 2012: 1789). Borrowing from Latin also continues, and the input from Old Norse, which must have occurred much earlier in the spoken language, becomes visible in the textual record. Middle English emerges in a period of multilingualism in Britain, and in histories of English and research on the lexis the relationship between the native terms and lexical items borrowed from other languages has generally been understood as one of competition. We see this conceptualisation in early scholarship such as Rynell (1948), and in more recent work such as Akimoto (2008), in which he argues that ‘[t]he verbs of the wanting type, e.g. desire, hope, want and wish have developed in competition with each other. Hope and wish were already in existence in the OE period, and desire and want came in during the ME period: desire from Old French, want from Old Norse respectively’ (2008: 117). Akimoto notes Rynell’s focus on native nimen and the Norse-derived take, and Samuels’s (1972) discussion of forms in competition, in particular the example of the native throw and the Norse-derived cast. He also observes that lexical replacement and semantic shift have been discussed in relation to semantic fields. Here we may note Käsmann’s (1961) study of the Middle English religious lexis up to 1350, in which he considers the obsolescence and replacement of the Old English vocabulary in this area. Käsmann presents a list of possible scenarios for lexical items relating to religion in the Middle English period including obsolescence and replacement by loanwords (1961: 34–37) and is particularly interested in the neglected story of the survival of native terms (1961: 19–21). Käsmann’s work provides the foundation for Timofeeva’s (2018a) examination of this particular issue. Her conception of the relationship between native terms and loanwords is also informed by the competition model. She observes that ‘such core lexemes as gospel or Easter compete with evangelium and pasque and eventually survive, while dozens of other old religious terms are replaced with newer loans […]. Just how this competition is resolved in early Middle English is the focus of the present paper’ (2018a: 225) . In another paper on pastoral care and lexical innovation in the thirteenth century, Timofeeva notes that ‘there is a lot of variation and competition between Middle English (ME) terms within the individual subdomains […] and between English and French terms’ (2018b: 58–59).

6  The Interaction between Native Terms and Loanwords 

161

It has been suggested that the expansion of the lexicon through lexical borrowing creates a need for systemic regulation. Samuels observes that ‘[i]f, for extralinguistic reasons such as cultural borrowing or foreign conquest, two exact synonyms exist for a time in the spoken chain, either one of them will become less and less selected and eventually discarded, or a difference in meaning, connotation, nuance or register will arise to distinguish them’ (1972: 65; italics in the original). He suggests that, where there is genuine free variation, it is probable that the processes he mentions have already begun, but we have not yet noticed. In histories of English we find well-known examples of semantic shift such as the narrowing of the meaning of the term meat from ‘food’ to ‘animal flesh eaten as food’ and the broadening of the sense of bird from ‘young bird or young of a bird’ to ‘bird’. These examples (with a few others) appear in a range of scholarship and textbooks including Ullmann (1962: 228–231), McMahon (1994: 178–179), Smith (1996: 120), Trask (1996: 42), Schendl (2001: 30) and Kay and Allan (2015: 75–76). In the literature, such shifts are often accounted for as having been triggered by the competition created by word borrowing.1 Discussing variation in the lexicon, Smith observes, ‘[m]ore important are occasions when a borrowed word and a synonymous or near-syonymous native word […] become differentiated in meaning within the borrowing language’ (1996: 125). In this chapter we examine the lexical field of Trade, tracing the interactions between the native and borrowed items in that domain. This is a lexical field relating to everyday life in medieval Britain and thus contrasts with the more elite domains, such as religion and church, that have generally been the focus of scholarship on French borrowing in the medieval period. We now know that prestige was not the only factor in the borrowing of French terms after the Norman Conquest. The lexical data assembled in the Bilingual Thesaurus of Everyday Life in Medieval England (BTh; Ingham and Sylvester 2017–) shows that French vocabulary was  It is noticeable that even when all the terms involved are native, the idea of competition persists; for example: ‘OE hound referred to any dog and fugol to any bird, but their meanings narrowed in competition with other words such as dog and bird’ (Kay and Allan 2015: 32). The OED suggests that the sense ‘[a]ny feathered vertebrate animal’ that is the same as bird n. 2 was still in use in the nineteenth century, while the sense ‘“barn-door fowl”, a domestic cock or hen’ is first attested a.1586. See OED (s.v. fowl, n., senses 1 and 3). Hound is cited in the sense ‘[a] dog, generally’ up to the mid-nineteenth century and the hunting dog sense is first attested in c.1200. See OED (s.v. hound, n.1, senses 1 and 2). 1

162 

L. Sylvester and M. Tiddeman

borrowed into English at the lower end of the socio-economic scale. We find it not just in domains such as the church, the law and government, but also in the records of occupations such as farming and building. Following that, we report on research undertaken as part of the Technical Language and Semantic Shift in Middle English project.2 The results of our analyses suggest that there are far fewer instances of relexification, and far more of native terms and loanwords continuing to co-exist as synonyms during the Middle English period than might have been expected. Furthermore, proportions of outcomes involving narrowing and broadening were low in the Middle English period, regardless of language of origin. These findings support the idea that this period sees the early beginnings of standardisation. This process ultimately involves the achievement of minimal variation of form (Haugen 1966: 931), and this is evidently true for some elements of language, such as spelling. It has been suggested, however, that standardisation of the lexis requires sets of synonyms, which can then be differentiated by register or region (Sylvester 2020: 365–366). We suggest that key to this development was the enrichment of vocabulary that happened via the lexical borrowing of the Middle English period. This is noticeable in some of the literary genres, for example the poetry of the Alliterative Revival. This poetic form requires a large stock of alliterating terms. Lester notes, for example, that ‘man, warrior’ might be expressed by the general terms kni3t, lorde, mon but also by the words burne, freke, gome, hathel, lede, renk, schalk, segge, tulke, or wy3e (1996: 101–102). Turville-Petre cites this and other examples, and reinforces the point in his observation that ‘this use of a wide vocabulary goes far beyond the practical necessities of filling the line and becomes an aspect of a sophisticated style’ (2018: 3), further noting that the demands of the form mean that ‘great weight is placed upon the ‘content’ words of the line: the nouns, adjectives, verbs and adverbs’ (2018: 4). We must emphasise here that differentiation by register is one of the ways that terms are retained in the language and competition resolved so that neither needs to undergo sematic shift or drop out of use. As noted above, Samuels includes distinction of register as one of the ways in which  The principal investigator on this project was Louise Sylvester, with Richard Ingham as co-­ investigator. The research associates were first Harry Parkin, then Megan Tiddeman. We are grateful to the Leverhulme Trust for funding this project across 2017–2020. 2

6  The Interaction between Native Terms and Loanwords 

163

synonymy is avoided in the lexicon (1972: 65). In this period, therefore, which saw a huge influx of loanwords into English, the model of a competition between the native lexis and the incoming loanwords seems limiting and even misleading; rather, we might wish to speak of continued co-existence of native terms and borrowed items.

2 Introduction to the Semantic Hierarchy The data we are using to investigate the relationship between native terms and borrowed vocabulary comes from the Technical Language and Semantic Shift in Middle English project. The project dataset comprises 5276 Middle English words and 2307 senses, which have been organised into a semantic hierarchy based on the system of classification created for the Historical Thesaurus of English (HTE; Kay et al. 2020).3 The project began with the Middle English vocabulary of the seven occupational domains that were collected for the BTh project: Building, Domestic activities, Farming, Food preparation, Manufacture, Trade, and Travel by water.4 We added two more domains (Medicine and Hunting) to achieve a fuller picture of medieval English society. The dataset provides a variety of opportunities for semantic analysis, including a focus on lexicalisation (number of co-hyponyms per sense); specificity (a word’s position in the hierarchy); polysemy; narrowing, broadening and sense extension; word borrowing and obsolescence. As the title suggests, the original focus of the Technical Language and Semantic Shift project was the vocabulary items with the greatest specificity of meaning, which we equated with technicality (see Wright 1995). As noted above, the idea that semantic shifts include broadening as well as  The principal investigator of the project was Louise Sylvester, with Richard Ingham as co-­ investigator. The research associates were first Harry Parkin, then Megan Tiddeman. We are grateful to the Leverhulme Trust for funding this three-year project between 2018 and 2020. We would like to express our thanks to Dr Harry Parkin, who created the hierarchy during the first half of the project. 4  The Bilingual Thesaurus project was led by Richard Ingham with Louise Sylvester as co-­investigator (later principal investigator). Our thanks are due to the Leverhulme Trust for funding for the project between 2013 and 2016. The Bilingual Thesaurus is available online hosted by the University of Glasgow (Ingham and Sylvester 2017–). 3

164 

L. Sylvester and M. Tiddeman

narrowing of sense—though the former is less common (Kay and Allan 2015: 77)—is prevalent in the histories of English. In order to examine these shifts in meaning, it was evident that the project needed to include the lexical items at the higher levels of the semantic hierarchy. Vocabulary at the contiguous higher levels taken from the HTE was therefore added to the dataset. This means that it is possible to track semantic shifts such as narrowing (which sends a lexical item down the semantic hierarchy) and broadening (which sends a term up to the more general levels). Like the contributors and editors of the HTE classification, we made use of componential analysis to classify the vocabulary: where the dictionary definitions add a component of meaning, we go down a level in the semantic hierarchy. Each sub-section of the hierarchy is divided into levels of technicality (or specificity) known as category levels (CLs), with superordinate senses found at CL0 and hyponymic senses found below at CL1 through to CL4. For example, the hierarchy extract below begins with vocabulary under the superordinate term Merchant. The example below offers an extract from the semantic hierarchy from the semantic domain Trade.5 1. Merchant                  [CL0] marchaunt c1225–1450+ Old French;Anglo-French mercer c1230–1450+ Old French chafferer a1382 Old English Group/body of merchants                [CL1] gild(e) 1114–1450+ Old English;Old Norse hanse a1135–1450+ Old French compaignie 1389–1450+ Old French livere a1422–1450+ Old French .Specific group/body of merchants            [CL2] merceri(e) c1375–1450+ Old French  The notation used in the project’s hierarchy is as follows: Two languages separated only by a semi-­ colon denote multiple origins of a term, or a case where it is not possible to determine from which language the term entered English, for example tapistere c1440–19th century, Latin;Old French. Two languages separated by a hyphen denote the origins of each part of a hyphenated compound term, for example web-lome 1338–PDE Old English-Old English. Two languages separated by a space denote the origins of each element in a non-hyphenated compound, for example flex monger 1297–1307 Old English Old English. 5

6  The Interaction between Native Terms and Loanwords 

165

.Member of a specific group/body of merchants        [CL3] lombard c1390–1450+ Old French esterling 1422–1450+ Old English hanser 1442–1449 Old French Foreign merchant                    [CL1] alien 1450+ Latin;Old Fr ench Trader/dealer in textiles/clothing/yarns           [CL1] mercer c.1230–1450+ Old French cloth-monger 1272–1311 Old English-Old English clother 1286–1450+ Old English draper 1350–1450+ Old French teler c.1400 Anglo-French yarn-chopper 1429 Old English-uncertain lin-draper 1436–1450+ Old English-Anglo-French worsted man 1450+ Old English Old English Traders in flax                      [CL1] flex-man 1266–1381 Old English-Old English flex monger 1297–1307 Old English Old English Traders in wool                     [CL1] lainer 1292–1305 Old French wol-monger 1296–1450+ Old English-Old English wol-man 1390–1450+ Old English .Female dealer in textiles/clothing/yarns          [CL2] flex womman 1324–1450+ Old English Old English threde-womman 1349 Old English-Old English silk-womman 1368–1450+ Old English-Old English flex wif 1378–1450+ Old English Old English .Female dealer in purple cloth              [CL3] purpuresse a1384–c1425 Latin purpurere c.1425 Latin

3 Focus on the Vocabulary Relating to Trade This domain was included in the BTh dataset because of its outward-­ facing nature (as opposed to, e.g., farming or domestic activities). As mentioned, however, lexis borrowed from French occurs in all the domains for which vocabulary was collected.

166 

L. Sylvester and M. Tiddeman

We can see at a glance that in the Middle English period this section of the semantic domain of Trade is made up of a mixture of native terms and loanwords. The group of superordinate (most general) terms includes one word derived from Old English, although this form is not attested in Old English. The MED offers the sense ‘[a] trader, merchant’ with a citation from the Wycliffite Bible.6 The OED provides the same citation and derives the term from the verb chaffer, first recorded in the Middle English text Ayenbite of Inwyt and in use until the eighteenth century with the sense ‘to mingle, interchange, exchange’ or possibly the nineteenth century with the sense ‘to talk much and idly’.7 The term is putatively traced back to Old English through the noun chaffer, thought to be derived from cheapfare, chapfare from OE *cēapfaru, derived from cēap ‘bargain, sale’ combined with faru ‘faring, going’. This compound is not attested in Old English, but there is a cognate Old Norse term kaupfǫr with the sense ‘trading journey’. Dates are difficult to be sure about for the Middle English period; it is likely that terms entered the spoken language before they were recorded in texts, and they may have had different levels of formality and so fulfilled different functions. Nevertheless, the written record suggests that the loanwords preceded the native term in use in the sense ‘merchant’.8 This may be an issue that has to do with the nature of the evidence that survives, since at CL1, the earliest word for a group or body of merchants is gild, a term which was evidently present in Old English.9 The focus of this chapter is on Middle English, and so we are taking attestation in Old English to stand for native vocabulary as lexemes from the OE period that survive into the ME period seem likely to have been naturalised in English even if they did not begin as native terms, and came ultimately from Latin, Old Norse, and so on. The other  See MED (s.v. chafferer, n.).  OED (s.vv. chafferer, n.; chaffer, v.; and chaffer, n.1, senses 1, 5 and 6). Note that OED, OED2 and OED3 refer to the successive editions of the Oxford English Dictionary. The online edition, which we have used, is the most up to date, and it alerts readers to the edition from which each entry derives, including the dates when entries were compiled. 8  This is partly an artefact of the way that the lexical items were classified, for example chapman and monger appear in the data but are classified under Society < Trade and finance < Trader while mercer, chafferer and marchaunt appear below Society < Trade and finance < Trader < Merchant. 9  See MED (s.v. gild(e, n.) and OED (s.v. guild | gild, n.). 6 7

6  The Interaction between Native Terms and Loanwords 

167

terms expressing this sense, hanse,10 compaignie11 and livere,12 are all borrowings from French, and the first of these is contemporary with the earliest English term. The OED tells us that the merchant guild is an institution not found in England before the Conquest, and that on the continent the name and thing were older’. The terms at CL2 and CL3 are all borrowings, and were evidently the names of specific groups of merchants.13 Of the terms for traders/dealers in clothing and yarn, the earliest attested is the French borrowing mercer. The native origin terms clothmonger and clother appear to be derived from terms recorded in Old English but not recorded in these forms and senses.14 The terms at CL2 for a female dealer in textiles/clothing/yarns are all native formations. Some of the elements of the compounds are found at other places in the same subcategory of terms, for example the terms under the sense ‘traders in flax’, at CL1, flex-man and flex monger.15 We can note that the semantic hierarchic arrangement reveals that there are varying levels of lexicalisation for the different senses. There is only one term for a foreign merchant (alien), but there are eight co-­ hyponyms with the sense ‘trader/dealer in textiles/clothing/yarns’ (mercer, cloth-monger, clother, draper, teler, yarn-chopper, lin-draper and worsted man). This is a heading with a broad range of senses, as we would expect at the level of the semantic hierarchy dealing with the most general vocabulary, and this accounts for the large number of co-hyponyms listed under this sense. In functionalist approaches to vocabulary, density of lexicalisation is thought to play a role in semantic shift and in the obsolescence of lexical items: the multiplication of new meanings for a single form, if allowed to continue unchecked, would lead to the breakdown of language as a communicative system (Samuels 1972: 65). This idea  See MED (s.v. hanse) and OED (s.v. hanse, n., sense 1a).  See MED (s.v. compaignie, n., sense 1e) and OED3 (s.v. company, sense 5a). 12  See MED (s.v. livere, n., sense 4e). 13  See MED (s.v. mercerie, n., sense 1d) and OED3 (s.v. mercery, n., sense 2). See also MED (s.vv. lombard, n., sense 1b; and sterling, n., sense 1) and OED (lombard, n.1 and adj., sense 2). 14  See MED (s.v. clother, n.) and OED (s.v. clothier, n.). For cloth-monger, see MED (s.v. cloth, n., sense 8). 15  On flex-man, see MED (s.v. flex-man); the term is attested in OED (s.v. flax, n., compounds C1b), but only from 1509. On flex monger, see MED (s.v. flex, n., sense 2d). 10 11

168 

L. Sylvester and M. Tiddeman

underlies the conceptualisation of a relationship of competition between native terms and foreign loanwords, in its suggestion that there is a functional limit to the number of synonyms that are possible in the language at any one time. This extract from the semantic hierarchy highlights the way in which the vocabulary in the Middle English period does not show the pattern suggested in many accounts of this period. We might expect to find a few native terms at the beginning of the period and their replacement by loanwords, mainly from French or perhaps of French and/or Latin origin. We will now examine the lexical data from this semantic domain from a different angle, to see if the same picture of native terms and loanwords co-existing in the vocabulary in this period, rather than the former terms being shifted within its variational space or dropping out of the language upon the arrival of the latter vocabulary, is borne out.

3.1 The Categories Making up the Semantic Domain of Trade As expected, the bulk of the vocabulary is clustered at CLs 0, 1 and 2, with a dramatic falling-off at CL3 and only one term at CL4, sterling ‘the standard currency of England’. This distribution in general matches that found in the other domains. CL1 (i.e. the second most general category level in any given section of the hierarchy) is the most prolific, followed by CL0 and CL2 on either side. There is a very sharp decrease as we move down to the more technical levels of vocabulary. One of the aspects of interest to our project was how the various levels of the semantic hierarchy, from the most general to the most specific terminology, were populated in the Middle English period. We now know that prestige was not the only factor in the borrowing of French terms after the Norman Conquest. The lexical data assembled in the BTh shows that French is found in occupational domains associated with everyday life in medieval England such as Farming and Building. This finding underlines the point that the foreign loanwords served a purpose in the emergence of Middle English as a fully developed variety with a range of registers. As we have seen, there is considerable variation in the

6  The Interaction between Native Terms and Loanwords 

169

number of terms found under the senses at the different category levels, and each subcategory has a varying number of sub-groups, depending on how fine-grained the distinctions are in the linguistic expression of the different concepts making up the semantic domain (Table 6.2). Of the categories at CL0, the largest, Trade and finance contains thirteen nouns. There are two categories with ten to twelve co-hyponyms: Selling and Merchandise; ten categories have between five and nine co-­ hyponyms; and twenty categories have four or fewer, including four categories each containing only one term. As we might expect, there are far more categories at CL1. Many of them denote sellers of specific items: for example, Sellers of candles, Sellers of dishes/pots, Sellers of strings/laces, Sellers of headdresses, Sellers of hay and so on. There are also a number of categories describing specific goods as merchandise: for example, Skins as merchandise, Beeswax as merchandise, Wool as merchandise and so on, as well as categories denoting the goods by seller: for example, Goods sold by a haberdasher, Goods sold by a mercer, Goods sold by grocer/greengrocer and so on. The largest category, Money-dealer/banker, has nine co-hyponyms. There are two categories with eight: Trader/dealer in textiles/clothing/yarns and Italian coins; one category with six co-­ hyponyms: Coiner; and five with five co-hyponyms: To set a monetary value on, Trader/dealer in skins/furs, Itinerant/pedlar, Market-place and (A) silver coin. The remaining 121 categories at CL1 have four or fewer terms each. The third biggest category level in terms of numbers of categories is CL2. The categories at this level denote particular kinds of sellers (note that the dots designate categories further down the hierarchy): for example, .Seller of herbs, .Seller of garlic, .Seller of mustard, .Seller of condiments/sauces and particular kinds of coin: for example, .A coin with a cross on it, .A coin bearing the image of a knight and .A debased coin. The numbers of co-hyponyms collected under these senses are smaller still, as we might expect. There is one category with nine co-­ hyponyms, .Seller of specific type of cloth; all the remaining categories at this level contain either one, two, three or four terms. As expected, the picture at CL3 is similar, though with overall reduced numbers of terms (see Table  6.1): there is one category containing six

170 

L. Sylvester and M. Tiddeman

Table 6.1  Distribution of terms across the category levels in Trade Number of terms

CL0

CL1

CL2

CL3

CL4

Total

145

256

102

31

1

535

Table 6.2  Numbers of categories in each of the category levels in Trade Number of subcategories

CL0

CL1

CL2

CL3

CL4

Total

33

130

55

17

1

236

Table 6.3  Numbers of terms per category at each of the category levels of Trade Numbers of co-hyponyms per category

CL0

CL1

CL2

CL3

CL4

13 10, 11 or 12 5–9 2, 3 or 4 1 Total

1 2 10 16 4 33

0 0 9 60 61 130

0 0 1 22 32 55

0 0 1 9 7 17

0 0 0 0 1 1

co-­hyponyms, .Market-place for sale of meat, and the rest of the categories each contain one, two or three terms. Finally, there is only one category at the most specific level of technicality, CL4: …Standard currency of England and it contains only one term (Table 6.3).

3.2 Vocabulary at Different Category Levels of the Semantic Domain Trade For the purposes of this study, we need to exclude all the subcategories that have only one item in them. We may assume that these denote concepts that were not heavily lexicalised in English in the medieval period, or perhaps that the processes mentioned by Samuels of obsolescence or differentiation have already taken place. First, we turn to a closer examination of the largest category in the semantic domain Trade, Trade and finance. As noted above, this subcategory contains thirteen nouns: Trade and finance n. [CL0] chepinge a1225 Old English

6  The Interaction between Native Terms and Loanwords 

171

chaffare c.1230–1450+ Old English marchaundise a1300–1450+ Old French;Anglo-French chep c.1325 Old English market 1340–1450+ Old English;Old Norse;Latin;Old French ned(e)-doinge 1382 Old English chapmanhede c.1390–1450+ Old English marchaundi(e) c.1390–1450+ Old French corserie ?c.1430 Latin cos(s)eri c.1440 Anglo-French marchaundrie a1450 Old French chapmanri 1450+ Old English merceri(e) 1450+ Old French

We can classify the terms in this subcategory in a way that allows us to focus on thinking about the effect of lexical borrowing, in particular the impact of the extensive borrowing from French in the Middle English period. Table 6.4 separates the native origin lexis in the category from the French loanwords, and these are further divided off from the other-­ language borrowings. Table 6.4  Terms in the category Trade and finance arranged according to language of origin OE

French origin

chepinge a1225 Old English

marchaundise market 1340–1450+ Old a1300–1450+ Old English;Old Norse; French;Anglo-French Latin;Old French marchaundi(e) corserie ?c.1430 Latin c1390–1450+ Old French

chaffare c.1230–1450+ Old English chep c.1325 Old cos(s)eri c1440 English Anglo-French ned(e)-doinge 1382 marchaundrie a1450 Old Old English French chapmanhede merceri(e) 1450+ Old c1390–1450+ Old French English chapmanri 1450+ Old English

Other-language origin

172 

L. Sylvester and M. Tiddeman

What is of interest here is the dates of usage. Of the six terms for Trade and finance attested in Old English, four (chaffare, chapmanhede, chapmanri and market) are still in use in 1400, and all continue in use in this sense beyond the Middle English period. Chaffare is still in use in the seventeenth century;16 chapmanhede is still attested in this sense at the end of the sixteenth century;17 chapmanri is recorded in use in this sense at the end of the eighteenth century;18 and market is thought to have been borrowed into Germanic before its occurrence in Old English. It is ‘late and uncommon’ in Old English and ‘not recorded before the twelfth century’, but it is recorded as being in use in the ‘trade’ sense until the 1960s.19 It is not generally the case for this category that the native terms dropped out of use, or underwent semantic shift as a result of the incoming loanwords. Of the five terms of French origin, two, marchaundise and marchaundi(e), are in use in the fourteenth century and beyond, and one of them, marchaundise, appears to have been borrowed before 1300.20 Three of the French loanwords continue in use beyond 1450: marchaundise is in use in the ‘trade’ sense until the nineteenth century;21 marchaundi(e) is recorded in the ‘trade’ sense until almost the end of the sixteenth century;22 and merceri(e) is attested until the seventeenth century.23 Even when the term narrows in meaning, such as merceri(e) narrowing from ‘trade, commerce’ to ‘textile trade’, the original sense remains in use for hundreds of years before any semantic shift is complete. These dates of usage equally suggest that it is not the case that the borrowed terms were short-term loans, coming into the language and quickly

 See OED (s.v. chaffer, n.1, sense 1a). The final citation for term in the related sense ‘chaffering, bargaining, haggling as to price’ is dated 1878. 17  See OED (s.v. chapmanhood | chapmanhead, n.) for citations; for the definition see OED (s.v. chapmanship, n.). 18  See OED (s.v. chapmanry, n.). 19  See OED (s.v. market, n., etymology and sense 5a). 20  See MED (s.v. marchaundīse, n., sense 1a). 21  See OED (s.v. merchandise, n., sense 1). 22  See OED (s.v. merchandy, n., sense 1a). 23  See OED (s.v. mercery, n., sense 2). 16

6  The Interaction between Native Terms and Loanwords 

173

dropping out of use as the usage dates of the native co-hyponyms in the category and the competition metaphor would lead us to expect. There is one term in this category that is borrowed from Latin, the lexical item corserie. In terms of the textual evidence, the term sits slightly oddly in this category. The only citation for it in both OED and MED from the ME period is from the Wycliffite Bible: ‘Alle doyng in þis mater is cursed corserie of symonye, ʒevynge þe sygne of holy ordris for temperal drit’ (the dictionaries use the same citation, though in the OED the quotation is more abbreviated).24 There is one further citation in the OED from the middle of the sixteenth century. While the term cannot be said, therefore, to have been only briefly present in the language, it does not seem as if it was ever the usual term for trading (‘buying and selling’ in the OED definition). This fits with research that suggests that loanwords are often borrowed in their non-basic senses (Allan 2014, 2015) and also suggests that it would be difficult to make a case that this Latin loanword was in competition with native terms expressing the idea of trade. What we see in this subcategory appears to be a wealth of terms, both native and foreign borrowings, expressing the idea of Trade and finance within the same period. We can now look at subcategories of differing sizes in this semantic domain to see if they have similar patterns. The terms for Money are also at CL0 in the semantic hierarchy. Below it is the subcategory Ready money/cash at CL1. Also at CL0 is the category Sum of money and below that, at CL1, Large sum of money, and below that, at CL2, .Coin as a type of small sum, .English monetary units and .Scottish monetary units. The category Money is much smaller than Trade and finance, consisting of five terms: Money n. [CL0] shat a1150–a1325 Old English peni a1225–1450+Old English god a1325–1450+Old English monei(e) a1325–1450+Old French pecuni(e) c1400–1450+Latin;Old French 24

 See OED (s.v. corserie, n.) and MED (s.v. corserīe, n.).

174 

L. Sylvester and M. Tiddeman

We can categorise the terms in this subcategory, as shown in Table 6.5, in order to discover if we see the same patterns. In the case of this subcategory, the term monei(e) does indeed oust the incumbent native terms shat, peni and god. We can note, however, that in two out of three cases, the native term did not fall out of use in this sense in the ME period: peni is still attested in the plural form with the sense ‘money’ until the seventeenth century (it is difficult to pin down exactly where the later depreciative use begins); and god becomes obsolete in this sense in the sixteenth century. The other French loanword amongst the co-hyponyms is pecuni. The citation record for this term is much more restricted, but it is attested until 1653 and then occurs in two citations from 1719 and 1832 as specifically Scottish in use. There are some very small subcategories, which consist only of borrowed terms, as in this example from a subcategory at CL1 (Table 6.6): The HT data does not include any terms under this sense for the Middle English period, though there are two terms for the sense ‘seller of herbs’ listed in use in the sixteenth century, croker 1577, derived from Latin crocus, and herb-man 1580–1598, a compound derived from Old French erbe and Old English man. At the same level (CL1), there is a subcategory which contains no borrowed terms (Table 6.7). Table 6.5  Terms in the subcategory Money arranged according to language of origin OE

French origin

shat a1150–a1325 Old monei(e) a1325–1450+ English Old French peni a1225–1450+ Old English god a1325–1450+ Old English

Other language origin pecuni(e) c1400–1450+ Latin; Old French

Table 6.6  Terms in the category Trader/dealer in herbs and spices arranged according to language of origin OE

French origin

Other-language origin

spicer 1311–1450+ Old French; Anglo-French

pigmentari(e) a1398 Latin

6  The Interaction between Native Terms and Loanwords 

175

Table 6.7  Terms in the category Trader/dealer in grain arranged according to language of origin French origin

OE

Other-language origin

corn mongere 1177–c.1450 Old English Old English ote mongere 1300–a1400 Old English Old English

Table 6.8  Terms in the subcategory .Seller of meat arranged according to language of origin Other-language origin

OE

French origin

flesh-mongere 1130–1450+ Old English-Old English ket mongere 1175–c.1275 Old Norse Old English lire mongere 1324–?c.1425 Old English-Old English

bocher c.1325–1450+ Old French

Table 6.9  Terms in the subcategory .Seller of pepper arranged according to language of origin OE

French origin

pepperer 1309 Old English

peverer 1294–1389 Anglo-French

Other-language origin

The numbers here are very small, however. It should also be noted that so far, we have only looked at examples from CL0 and CL1, that is the superordinate (most general) vocabulary. Most of the categories at CL2 seem to present a similar picture of a mixture of native terms and loanwords with the native and the borrowed terms in use at overlapping periods. Here is an example of a subcategory at CL2: .Seller of meat (Table 6.8). If we look at subcategories at CL3, the terms with the most specific meanings, we find a similar pattern. Here is an example of a CL3 subcategory. We are now two levels below the category Trader/dealer in herbs and spices. The particular spice is specified (Table 6.9):

176 

L. Sylvester and M. Tiddeman

Table 6.10  Proportion of all subcategories which contain one or more mixed-­ language compounds in Trade All Subcategories containing subcategories in one or more mixed-­ TRADE language compounds CL0 33 CL1 130 CL2 55 CL3 17 CL4 1

1 12 7 5 0

Percentage of subcategories containing mixed-language compounds (%) 3 9 13 29 0

Following examination of this one semantic domain, it appeared to be the case that lower down the semantic hierarchy there is more mixing of the languages as more compound terms appear which combine a native and borrowed element (such as monei(e)-maker, mal(le) mongere, lin-­ draper and chepinge-place). These mixed compounds express greater specificity of meaning as we would expect to find at the lowest and most technical level of the semantic hierarchy. We can see in Table 6.10 that, while the numbers involved are quite low, the proportion of mixed compounds does noticeably increase as we move down through the levels of technicality from CL0 to CL3 (although there is only one term at CL4, as previously noted, and it is not a compound).

4 The Relationship between Native Terms and Loanwords in Small Groups of Terms across the Whole Dataset In order to further test the idea that in the Middle English period loanwords did not tend to lead to the obsolescence or semantic change of native terms, but rather co-existed alongside them in the lexicon, we decided to examine a large number of small subgroups of terms across a wider range of data. Senses that have two co-hyponyms (pairs), three (trios) or four (quads) have been examined in data assembled for the Technical Language and Semantic Shift project, both in a pilot study looking at lexical pairs in four semantic domains and in an investigation of an

6  The Interaction between Native Terms and Loanwords 

177

Table 6.11  Number of word and senses for lexical pairs, trios and quads analysed, per domain Pairs Domain

Trios

Quads

All

Senses Words Senses Words Senses Words Senses Words

Building 54 Domestic Activities 39 Farming 86 Food Preparation 24 Hunting 38 Manufacture 71 Medicine 60 Trade 50 Travel by Water 31 Total 453

108 78 172 48 76 142 120 100 62 906

11 11 12 11 11 11 11 11 11 100

33 33 36 33 33 33 33 33 33 300

5 7 21 12 5 18 14 9 9 100

20 28 84 48 20 72 56 36 36 400

70 57 119 47 54 100 85 70 51 653

161 139 292 129 129 247 209 169 131 1606

extended dataset of a total of 1606 words, divided across 453 two-item senses, 100 three-item senses and 100 four-item senses (for fuller accounts of the pairs study and the pairs, trios and quads study see Sylvester et al. 2022, 2023). These pairs, trios and quads are taken from all nine of the project’s domains, as shown in Table 6.11. All suitable two-item subcategories (pairs) were examined, amounting to a total of 453 pairs (906 words).25 A sample of 100 three-item senses (trios) was collated, twelve from Farming (the largest domain in our dataset) and eleven from each of the other eight domains. The dataset only contains a small number of four-item senses (quads), as only 6% of categories in the hierarchy have four terms. For this reason, some domains (especially Farming and Manufacture) are more heavily represented than others in the sample of 100 quads. In all cases, senses with words with uncertain etymologies were discarded. Once the samples had been collated, each word in the pairs, trios and quads was tagged as either ‘native’ or ‘non-native’ based on their language(e)s of origin using the following system of categorisation: N = Native:

 Twenty-one pairs of terms were excluded because of lack of certainty about the etymology of one or both terms, or because they were hyponyms but not close synonyms. 25

178 

L. Sylvester and M. Tiddeman

• Single terms of native origin; • Compound terms where both elements are of native origin; • Words of non-native origin attested in Old English and therefore assumed to have been assimilated by the Middle English period. NN = Non-native: • Single terms of non-native origin first attested in the Middle English period; • Compound terms made up of two loanwords; • Compound terms made up of a loanword and a native term. Each pair, trio and quad was categorised on the basis of its composition; that is, whether it was composed of all native words or all loanwords or was a mixture of the two. The categories were as follows: (i) All N = all terms are native For example, the following lexical pair: ‘wire-maker/worker’ wir-drawer(e) 1368–PDE Old English-Old English wir-smith 1438–PDE Old English-Old English

(ii) All NN = all terms are loanwords For example, the following lexical trio: ‘cider’ sider c.1350–PDE Old French pommade c.1400 Old French pomis c.1450 Latin

(iii) Mixed N1  =  mixed group where the earliest attested term is native For example, the following lexical quad:

6  The Interaction between Native Terms and Loanwords 

179

‘animal flesh as food’ flesh OE–PDE Old English mete a1325–PDE Old English braun 1381–17th century Old French char a1450–15th century Old French

(iv) Mixed NN1 = mixed group where the earliest attested term is a loanword For example, the following lexical trio: ‘art of cooking’ curie a1387–16th century Old French cokerie a1393–PDE Old French kitchen(e) (c.1400) Old English

Once the dataset had been collated and categorised, we were able to compare long-term semantic outcomes across the lexical groups. As with the language labelling, the focus is on the word in the group—tagged as Term 1—which is attested first, using the dates given in the OED and MED. Term 2 is the second to be attested, with trios also having a Term 3 and quads, a Term 3 and Term 4. Outcomes track whether Term 1 is retained until PDE, whether it is replaced or shifts to another sense, following the arrival of incomers into the semantic space in the Middle English period, noting where Term 1 is a native term, and where it is a borrowed item. In order to track the effects of the lexical borrowing on the native vocabulary, a typology of possible outcomes was devised, specifically for this project (see Table 6.12). Therefore, from the examples given previously above, the trio under ‘Cider’ and the quad under ‘Animal food as flesh’ are both classed as Type 2 because the earliest term is not ultimately replaced by any incoming co-hyponyms which enter the semantic space in the ME period. However, the trio under ‘art of cooking’ is classed as Type 1 because the earliest term is replaced by an incomer, and the pair under ‘Wire-maker/worker’ is classed as Type 3 because both terms remain in use as synonyms until the PDE period.

180 

L. Sylvester and M. Tiddeman

Table 6.12  Outcome types Type Outcome 1

2

3

4

5

6

An existing term (Term 1) drops out of use in a particular sense (relevant to our domains) before the PDE period following the arrival of incoming terms (Terms 2–4) during the Middle English period. An existing term (Term 1) is joined by incoming terms (Terms 2–4) in a particular sense during the ME period, all or some of which then drop out of use before the Present-Day English period. An existing term (Term 1) is joined by incoming terms (Terms 2–4) in a particular sense during the Middle English period. All terms go on to exist as (near) synonyms until the Present-Day English period. An existing term (Term 1) is joined by incoming terms (Terms 2–4) in a particular sense during the Middle English period. One or all of the incoming terms then undergoes semantic change through narrowing, broadening or metonymy prior to 1500. An existing term (Term 1) is joined by incoming terms (Terms 2–4) in a particular sense during the Middle English period. Term 1 then goes on to undergo semantic change (through narrowing, broadening or metonymy) prior to 1500. An existing term (Term 1) is joined by incoming terms (Terms 2–4) in a particular sense during the Middle English period. All terms go on to undergo semantic change (through narrowing, broadening or metonymy) prior to 1500.

Outcomes of Types 1 and 5 are the ones that we would expect to be the most prevalent in the dataset based on the textbook accounts of semantic shift. Type 1 outcomes (an existing term drops out of use following the arrival of a term with the same sense) do indeed occur in our data, as we saw in the example above from Money, where the first attested term under that sense, the native shat, was replaced by the loanword monei(e) in the Middle English period. We found examples of this outcome in our data but they accounted for fewer than a quarter of the outcomes across the pairs, trios and quads. Type 5 outcomes (an existing term undergoes semantic change via narrowing, broadening or metonymy following the arrival of term or terms with the same sense) have the kinds of examples we have seen in the scholarly and textbook accounts. There were very few examples like these in our data, for example when besom (c.1000–?19th century) is joined by brom (1346–PDE), besom narrows to mean ‘broom made of specific material e.g. the plants, broom or birch’ (c.1400–19th century). Such examples are extremely rare in our data and account for only 2% of the outcomes overall.

6  The Interaction between Native Terms and Loanwords 

181

Our studies showed that the most common outcome was that incoming terms fail to replace the existing term regardless of the language(s) of origin of the component terms involved. Further, proportions of outcomes involving narrowing, broadening and metonymy are very low across all combinations of native and non-native terms. It is not the case that loanwords regularly oust native terms. In mixed pairs, there is no great difference in the rate of non-native terms replacing native ones, compared to the other way around. In essence, whilst failed replacement of Term 1 is the most common across all language combinations, replacement of Term 1 is the second most common outcome in mixed language groups, whereas long-term synonymy is the second most popular outcome in non-mixed language groups, be they all native terms or all loanwords. We may also note that Type 3 outcomes (where all terms remain in use until Present-Day English as long-term synonyms) are much more common overall when all words under a sense are of native origin: 34% for All N, compared to 25% for All NN, 15% for Mixed N1 and 12% for Mixed NN1. Type 1 (replacement of Term 1) is also the least likely outcome (17%) when all the words are native, compared to 22% for All NN, 26% for Mixed N1 and 35% for Mixed NN1. The drop-out rate in the Middle English period is 7% higher for native than for non-native Term 1s: forty-six out of eighty-two native Term 1s (or 56%) are obsolete by 1500, compared with thirty-six out of seventy-­ three (or 49%) of non-native Term 1s. Overall, 82 out of 155 Term 1s (or 53%) have become obsolete by the end of the 1400s. Crucially, the time period with the highest drop-out rate for all Term 1s, regardless of language of origin, is the end of the Middle Ages, from 1400 to 1499. These findings served to support the idea that obsolescence was not the immediate consequence of the lexical borrowing.

5 Conclusion The trend towards examining big data in lexical semantics is still current, and has brought valuable insights. The brief for this chapter, however, to provide a case study, offered the opportunity to focus in detail on one

182 

L. Sylvester and M. Tiddeman

semantic domain, that of Trade. Homing in one lexical field allowed us to compare the numbers of subcategories and sub-groups across the domain, and to investigate the varying amounts of vocabulary contained within those subcategories and sub-groups. We were able to examine the variations in the way in which lexical items were distributed at the different levels of the semantic hierarchy within the domain. All of this enabled us to gain a sense of how fine-grained the distinctions are in the lexicalisation of this concept. Having looked at the semantic domain as a whole, we were then able to focus on the lexical items within two subcategories at CL0 (the top of the semantic hierarchy), the large grouping Trade and finance and the slightly smaller subcategory Money. We also looked at the vocabulary from two subcategories at CL1: Trader/dealer in herbs and spices, Trader/dealer in grain; one sub-group at CL2: .Seller of meat; and one at CL3: .Seller of pepper. We investigated which terms within each lexical group are native and which are loanwords, and compared the dates of usage of the lexical items, many of which were in overlapping use in the Middle English period. The close focus on a subset of the data meant that the report we offered of our examination of a larger set terms, 1606 lexical items that are found in two-, three- and four-item groupings, is contextualised within a picture of the kinds of senses and vocabulary items contained in our dataset. Our analyses produced the key finding that replacement did occur but is rarer than expected, and wholesale relexification did not happen during the later medieval period. We can see that loanwords did not generally oust native terms; there is not much difference in replacement rates for non-native terms replacing native ones and vice versa. Across all our data, the most common outcome is that the first term in any group of two, three or four co-hyponyms is likely to remain in the language. We know that Middle English emerged from this period of borrowing as a language whose vocabulary is made up of both native Germanic terms and romance borrowings. The multilingual context of the period produced the availability of a set of variants for the expression of different levels of specificity within the same conceptual space, allowing for the emergence of different registers. The initial processes of standardisation of the lexicon that we see in Middle English do not seem to depend on setting

6  The Interaction between Native Terms and Loanwords 

183

boundaries determining which terms do and do not belong in the language. This idea goes against some of the most recent ideas that make use of Haugen’s (1983) work, such as a recent chapter by Andrew Linn et al., which discusses what Haugen calls lexication, described as the process of ‘standardizing the list of words used in the language (lexicon)’ (Linn et al. 2018: 26). We may note that Haugen’s view is that lexication is ‘the selection of an appropriate lexicon’ in which he includes ‘the assignment of styles and spheres of usage for the words of the language’ (Haugen 1983: 271). What we see in administrative and romance texts of the later Middle English period is the creation of something approaching a technolect in which terms attested in a variety of languages could be used because they would be readily understood and because they had accreted both precision, making them worth using, and universality, making them easily comprehensible (Sylvester 2018: 93–94). We can point to usefulness of the expansion of the lexicon for the poets of the Alliterative Revival in the fourteenth and fifteenth centuries. As noted above, Lester (1996: 101–102) points to the distinctiveness of the poetic register of the alliterative poetry of this period and observes that the requirements of the form created the need for a large stock of alliterating terms (see also Turville-Petre 2018: 3–5). Lester cites a number of concepts for which a range of terms might be necessary, such as verbs of movement, generalised complimentary adjectives and adverbs of prompt and speedy movement (Lester 1996: 102). This process may be traced as far back as the lexical borrowing from French in the thirteenth century. It has been suggested that the creation of a French-influenced high register of English, well before the end of the medieval period, was already underway (Ingham 2017). The availability of variation of expression within the same conceptual area equated to having the lexical resources to describe things at the superordinate, basic and hyponymic levels (Sylvester 2020: 375). This enrichment of vocabulary, and its affordances for a range of text types, ultimately allowed the standardisation of vocabulary, a move which, we suggest, would not have been possible if the relationship between native terms and loanwords was only one of competition for a place in the language.

184 

L. Sylvester and M. Tiddeman

References Akimoto, Minoji. 2008. Rivalry among the Verbs of Wanting. In English Historical Linguistics 2006: Selected Papers from the Fourteenth International Conference on English Historical Linguistics (ICEHL 14), Bergamo, 21–25 August 2006, ed. Maurizio Gotti, Marina Dossena, and Richard Dury, II, 117–138. Amsterdam: John Benjamins. Allan, Kathryn. 2014. An Inquest into Metaphor Death: Exploring the Loss of Literal Senses of Conceptual Metaphors. Cognitive Semiotics 5: 291–311. ———. 2015. Lost in Transmission? The Sense Development of Borrowed Metaphor. In Metaphor and Metonymy Through Time and Cultures, ed. J. E. Diaz Vera, 31–50. Berlin: Mouton de Gruyter. Durkin, Philip. 2014. Borrowed Words. Oxford: Oxford University Press. Haugen, Einar. 1966. Dialect, Language, Nation. American Anthropologist 68: 922–935. ———. 1983. The Implementation of Corpus Planning: Theory and Practice. In Progress in Language Planning, ed. Juan Cobarrubias and Joshua A. Fishman, 269–289. Berlin: Mouton. Ingham, Richard. 2017. Middle English Borrowing from French: Nouns and Verbs of Interpersonal Cognition in the Early South English Legendary. In The French of Medieval England: Essays in Honour of Jocelyn Wogan-Browne, ed. Thelma Fenster and Carolyn P.  Collette, 128–139. Woodbridge: Boydell & Brewer. Ingham, Richard, and Louise Sylvester. 2017–. The Bilingual Thesaurus of Everyday Life in Medieval England. https://thesaurus.ac.uk/bth/. Accessed 29 May 2023. Käsmann, Hans. 1961. Studien zum kirchlichen Wortschatz des Mittelenglischen 1100–1350. Tübingen: Niemeyer Max Verlag. Kay, Christian, and Kathryn Allan. 2015. English Historical Semantics. Edinburgh: Edinburgh University Press. Kay, Christian, Marc Alexander, Fraser Dallachy, Jane Roberts, Michael Samuels, and Irené Wotherspoon. eds. 2020. The Historical Thesaurus of English, 2nd edn, version 5.0. Glasgow: University of Glasgow. https://ht.ac.uk/. Accessed 29 May 2023. Lester, G.A. 1996. The Language of Old and Middle English Poetry. Basingstoke: Macmillan. Linn, Andrew, Guro Refsum Sanden, and Rebecca Piekkaria. 2018. Language Standardization in Sociolinguistics and International Business: Theory and

6  The Interaction between Native Terms and Loanwords 

185

Practice across the Table. In English in Business and Commerce: Interactions and Policies, ed. Tamah Sherman and Jiří Nekvap, 19–45. Berlin: De Gruyter Mouton. McMahon, April M.S. 1994. Understanding Language Change. Cambridge: Cambridge University Press. [MED =] Middle English Dictionary. 1952–2001. Ed. by Robert E.  Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner. Ann Arbor: University of Michigan Press. https://quod.lib.umich.edu/m/med/. Accessed 29 May 2023. [OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 29 May 2023. Rynell, Alarik. 1948. The Rivalry of Scandinavian and Native Synonyms in Middle English Especially Taken and Nimen. Lund: C. W. K. Gleerup. Samuels, Michael. 1972. Linguistic Evolution. Cambridge: Cambridge University Press. Schendl, Herbert. 2000. Linguistic Aspects of Code-Switching in Medieval English Texts. In Multilingualism in Later Medieval Britain, ed. David Trotter, 77–92. Cambridge: D. S. Brewer. ———. 2001. Historical Linguistics. Oxford: Oxford University Press. Smith, Jeremy J. 1996. An Historical Study of English: Form, Function, Change. London: Routledge. Sylvester, Louise. 2018. A Semantic Field and Text-Type Approach to Late-­ Medieval Multilingualism. In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 77–96. Berlin: de Gruyter Mouton. ———. 2020. The Role of Multilingualism in the Emergence of a Technical Register in the Middle English Period. In The Multilingual Origins of Standard English, ed. Laura Wright, 365–380. Berlin: De Gruyter. Sylvester, Louise, Megan Tiddeman, and Richard Ingham. 2022. Lexical Borrowing in the Middle English period: A Multi-Domain Analysis of Semantic Outcomes. English Language and Linguistics 26: 237–261. ———. 2023. Lexical Replacement, Retention and Borrowing in Middle English: A Case Study. In Language Contact and the History of English:

186 

L. Sylvester and M. Tiddeman

Processes and Effects on Specific Text-Types, ed. Gabriella Mazzon, 165–187. Berlin: Peter Lang. Timofeeva, Olga. 2018a. Survival and Loss of Old English Religious Vocabulary between 1150 and 1350. English Language and Linguistics 22: 225–247. ———. 2018b. ‘Mid ðare soðe luue ðe is icleped karite’: Pastoral Care and Lexical Innovation in the Thirteenth Century. Journal of the Spanish Society for Medieval English Language and Literature (SELIM) 23: 55–85. Trask, R.L. 1996. Historical Linguistics. London: Arnold. Trotter, David. 2012. Middle English in Contact: Middle English Creolization. In English Historical Linguistics, ed. Alexander Bergs and Laurel J. Brinton, II, 1781–1793. Berlin: De Gruyter Mouton. Turville-Petre, Thorlac. 2018. Description and Narrative in Middle English Alliterative Poetry. Liverpool: Liverpool University Press. Ullmann, Stephen. 1962. Semantics: An Introduction to the Science of Meaning. Oxford: Basil Blackwell. Wright, Laura. 1995. A Hypothesis on the Structure of Macaronic Business Writing. In Medieval Dialectology, ed. Jacek Fisiak, 309–321. Berlin: Mouton de Gruyter.

7 Cheapside in Wales: Multilingualism and Textiles in Medieval Welsh Poetry Helen Fulton

1 Introduction The ‘poetry of the gentry’ composed in Welsh mainly in the fourteenth and fifteenth centuries is a relatively untapped source of evidence for borrowings from French and Middle English into Welsh. While French loanwords are appearing in Welsh from at least the twelfth century, the influence of Middle English is somewhat later, with loanwords appearing in significant numbers during the fourteenth century and gaining pace in the fifteenth as cultural integration, especially on the March of Wales, takes hold.1 The lexis of the cloth trade, including the production and consumption of textiles and related commodities, is a particularly fruitful  There are earlier borrowings, from Old English into Welsh, dating from the ninth century; see Parry-Williams (1923: 22–47). This paper is an expanded and updated version of research delivered as the O’Donnell Lecture at the University of Oxford in May 2018. 1

H. Fulton (*) University of Bristol, Bristol, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_7

187

188 

H. Fulton

arena of investigation with regard to linguistic borrowings, and one that has not yet been fully explored (cf. Sylvester and Tiddeman, this volume). In this case study of French and English loanwords into Welsh relating to the cloth trade, I aim to show that the patterns of usage of such terms were closely related to the rise of an urban economy in Wales and its consequent exposure to English commercialisation and long-distance trade. Cloth production became a major industry in Wales, as in many parts of Britain, in the later Middle Ages and was entirely dependent on the agricultural economy in terms of sheep farming and wool production as the raw material of most cloths produced in Wales. References in Welsh to textiles and how they were used tell us not only about what kinds of commodities were available in Welsh markets but also about loanwords from English and French into Welsh and how these articulate the multilingual environments of Welsh towns in the later Middle Ages and the importance of long-distance trade as a sociocultural factor in linguistic developments. A related argument is that Welsh poetry composed in the fourteenth and fifteenth centuries is essentially an urban form, one which engages with the urban culture of Wales and which addresses audiences of gentry who relied on urban trade to deliver the consumer goods which marked their elite status. As the towns of Wales grew and flourished from about 1300, markets and fairs opened up to merchants travelling from major port cities such as Bristol and London, bringing with them a wide range of commodities imported from the continent and from the eastern trade routes. These new commodities, especially wines, spices, precious stones, and textiles, provided a means by which socio-economic status could be visibly reinforced. In Wales as in England, and elsewhere in Europe, the aspirational gentries of the late Middle Ages found that they could buy the kinds of commodities that were once available only to the highest nobility.2 Like the Nike trainers of the late twentieth century, the medieval cloth industry became a contributor to social mobility. The anxiety engendered by this class mobility, sometimes perceived as ‘the world turned upside down’, was one of the reasons why the English government  On the social and economic conditions that made this new wealth possible, especially in the wake of the Black Death, see Dyer (2002). 2

7  Cheapside in Wales 

189

of Edward III attempted to restrict the type of clothing that different social classes could wear by imposing a sumptuary law in 1363, a law that was largely ignored and then repealed in 1365 (Dyer 1989: 88). This process of social change, most marked during the fourteenth century, is visible through the changing references to textiles, especially luxury fabrics, in literary texts from Wales. Before 1300, such fabrics were unavailable to any but the highest echelon of society, the princes of the Welsh nobility. As fully imported and rare commodities, these fabrics were objects of wonder to poets and storytellers who used them as signifiers of aristocratic prestige and of the world of French romance, another import from beyond the boundaries of Wales. After 1300, as the urban economy grew and the cloth trade became more extensive, these textiles became available to a wider social group who used them to enhance their status as the new nobility of Wales after the fall of the princes. The earlier romance of rich fabrics and colours, unobtainable except by a tiny few, was largely replaced by a keen understanding of their value to an emergent nobility as a proxy for their distinctive social status.

2 Linguistic Evidence before 1300 Some early evidence for the importation of luxury textiles into Wales comes from the Marcher lordship of Glamorgan. In 1316, the Welsh nobleman Llywelyn Bren led a rebellion against the English crown which had recently taken temporary control of Glamorgan following the death of its lord, Gilbert de Clare, the Earl of Gloucester, at Bannockburn in 1314 (see Davies 2000: 388). Llywelyn paid the price for his treason. Along with ‘other felons slain and dead in the war and outlawed’, as recorded in an indenture of 1316, written in French in Cardiff, Llywelyn’s property, both land and movable goods, was appropriated by the Crown, and Llywelyn was subsequently executed in 1317 (Matthews 1898–1904: iv, 49–58). Among the lists of goods seized from Llywelyn are these items: j romanz de la rose. iij liures Galeys. iiij autres lyures. ‘1 Roman de la Rose, 3 Welsh books, 4 other books’.

190 

H. Fulton

This reference is the main reason why Llywelyn Bren is well-known among modern scholars: it shows that the Welsh gentry of south Wales were collecting books, in both Welsh and French, as part of their personal property. Llywelyn owned not just one book, but a library of books, at least the eight which are mentioned here. Speculating about the three Welsh books is interesting; considering that books in Welsh seem to have started to appear from the middle of the thirteenth century, it is likely that Llywelyn’s Welsh books were law tracts or possibly historical chronicles or religious texts. Most significant of all is the specific mention by name of the Roman de la Rose, the great thirteenth-century French romance, written in the form of an allegorical dream vision examining the philosophy of courtly love.3 Here is proof, if any were needed, that the Welsh nobility living in the Marches of Wales enjoyed French romances at least from around 1300 and, presumably, were able to read them in the original language. Llywelyn Bren may ultimately have rebelled against English rule, but he also acknowledged French language and literature, as used by the French-speaking kings and magnates of England and Wales, within his cultural identity. But there is more to Llywelyn Bren’s property list than books. This is the complete list of his property that was seized on his arrest: j. Aketon. j gaumbeyson. j peyre des quissens. j coleret de linge teille. v poz darein ij peyles darein j bacin. j vel chapel de fer, iij cheys galeys. j cofre oue chartres et munemenz. j romanz de la rose. iij liures Galeys. iiij autres lyures. x aneus dor. j fermail dor. j fermail dargent [...] j cote darme de bocram. j gaunbeyson vermail. j Aketon nyent p[a]rfet. ij napes, j vele sele. j sarge raie. viij quillers dargent. ‘1 cuirass, 1 riding-coat, 1 pair of cushions, 1 collar of slashed linen, 5 brazen pots, 2 brazen pails, 1 basin, 1 old breastplate of iron, 3 Welsh chairs, 1 coffer for charters and muniments, 1 Roman de la Rose, 3 Welsh books, 4 other books, ten gold rings, 1 golden clasp, 1 silver clasp [...] 1 coat-of-arms of buckram, 1 bright-red riding coat, one cuirass not perfect,

 On the significance of this evidence that at least one copy of the Roman de la Rose was circulating in Wales by the early fourteenth century, see Bromwich (1986). 3

7  Cheapside in Wales 

191

2 tablecloths, one old seal, one rayed serge, 8 silver spoons’ (Matthews 1898–1904: iv, 58).

This list indicates a man of wealth and property. A coat of arms, tablecloths, cushions, chairs, gold rings and silver spoons are all commodities which indicate the already thriving market economy associated with Cardiff and the other English towns in Wales, an economy in which the Welsh also participated, as far as they were able, as both producers and consumers. What is interesting is the number of references to the textiles from which these goods were made. There is a collar of slashed linen (linge teillé), that is, with decorative cuts; a coat-of-arms embroidered on buckram (bocram), a stiffened cotton or linen; and a rayed serge (sarge raié), a fine silk or woollen fabric that was striped or streaked, often used for hangings and bed covers, which is probably what is signified here. Like other finely woven linens, buckram is associated with the cloth manufactures of the Low Countries, places such as Ghent and Bruges, and these cloths were imported into England, usually through London but also through Bristol, and then sold at the provincial fairs. Throughout the fourteenth century, the cloth towns of Flanders were the market leaders; this is why Chaucer compared the Wife of Bath, a cloth-maker, to those of Ypres and Ghent as a form of high praise. We can assume then that Llywelyn Bren had either inherited or could afford to buy fully imported fabrics that would have been at the upper end of the price bracket.4 The Old French (OFr.) word sarge, borrowed into Middle English (ME) as sarge, is derived from the Latin (L) adjective sericum, meaning ‘made of silk’ (Middle English Dictionary, MED, s.v. sarge). The root of the word is Seres, the classical Greek and Latin word for China, indicating the origin of this luxurious imported fabric. The earliest reference to serge in English, according to the Oxford English Dictionary (OED) (s.v. serge) is in Geoffrey Chaucer’s Knight’s Tale of about 1385, where the city of Athens is arrayed for a great tournament:

 For a useful set of spreadsheets showing the prices of woollen and linen textiles in the fifteenth and sixteenth centuries, see Munro (n.d.). 4

192 

H. Fulton

Up goon the trompes and the melodye,    trumpets And to the lystes rit the compaignye,     lists [jousting arena] rode By ordinance, thurghout the citee large, Hanged with clooth of gold, and nat with sarge. (ll. 2565–2568; Benson 2008: 59)

The importance and high status of the tournament is signified by the expensive silk ‘cloth of gold’ that decorates the city, rather than the cheaper and more ordinary woollen fabric. Already by the late fourteenth century in England, serge had become a fairly standard fabric, made of wool rather than silk, and used for hangings, bed covers and so on (Benson 2008: 839). Its lowly status was apparent much earlier in France, and Chrétien de Troyes, composing in the 1170s, refers to it dismissively in his romance, Erec et Enide, when King Arthur is holding a sumptuous coronation ceremony: Molt fu li rois puissanz et larges: ne dona pas mantiax de sarges, ne de conins ne de brunetes, mes de samiz et d’erminetes, de veir antier et de dïapres, listez d’orfrois roides et aspres. (ll. 6605–6610; Roques 1981: 201) ‘The king was very powerful and generous: he did not give mantles made of serge or rabbit or dark-brown wool, but of samite and ermine, of whole miniver and mottled silk, bordered with orphrey [embroidery], stiff and rough’ (Kibler 1991: 119).

In England, serge remained an imported fabric, produced mainly in the Low Countries, and brought into England via the London merchants.5 Llywelyn Bren’s serge is described as ‘rayed’, meaning ‘striped’, where the yarn of the warp and the weft are of different colours, adding a further layer of production to the process and therefore additional expense.

 For examples of serge imported into London, see Sutton (2005: 151–156).

5

7  Cheapside in Wales 

193

There is one further indicator of Llywelyn’s wealth, which is his ‘bright-­ red riding coat’, coloured with vermail, or vermillion, which comes from the Italian vermilium (‘little worm’; Walton 1991: 334). This was one of the most expensive dyes to produce, since it came from crushed insects or ‘worms’ which produced the bright red colour. These insects were native to the eastern and southern Mediterranean and also parts of Asia Minor, so Llywelyn’s bright red coat was both exotic and expensive. Llywelyn’s material possessions therefore had a value over and above their practical purpose: what Marx called an exchange value, as well as a use value. The coat of arms, the collar, and the bed hangings became commodities because of the fabrics from which they were made; their value was woven into their material. The case of Llywelyn Bren provides us with a name and an identity— rather a shadowy one but certainly a real person—to whom we can attach specific consumer goods exemplifying the growing urban economy in medieval Wales. He stands at the cusp of a major social and economic change which announced itself partly through the greater availability of expensive imported commodities in the larger towns close to England such as Cardiff. Before about 1300, towns in Wales were small and mainly of Norman origin (Fulton 2012: 2). After 1066, the Normans had gradually colonised what became known as the March of Wales, that is, the border regions between Wales and England and also the coastal areas in the south and south-west of the country. Formed around the characteristic stone castles which secured the new Norman lands, urban settlements developed at key sites such as Rhuddlan, Brecon, Caerleon, Cardiff, Swansea and Carmarthen. The majority of the towns in Wales were populated by immigrants, mainly Flemish and English, and many of the larger towns received some kind of charter of liberties during the eleventh and twelfth centuries. This meant that they had the right to hold markets and fairs on particular days (markets were normally held weekly and fairs once or twice annually on specific feast days, such as that of the local patron saint; see Everitt 1990: 82). While the weekly markets catered mainly for staples, such as food, cattle, tools, leather goods, and local textiles, the annual fairs attracted traders from much further afield bringing goods from other parts of the

194 

H. Fulton

country, from England, and even from the continent. This was where the nobility, both Welsh and Normans, acquired the expensive commodities that marked their status. Going back to the case of Llywelyn Bren, he was arrested in Cardiff, which was already the largest town in Wales. Occupied by the Normans in the 1080s, Cardiff had been a borough town with its own market since the early twelfth century, and a number of fairs are recorded during that century. A fair held by Gilbert de Clare, Earl of Gloucester and lord of Glamorgan, was recorded in 1296 and two regular fairs, beginning on 23 June and 7 September, respectively, were recognised in the charter of liberties granted to Cardiff by Hugh, lord Despenser, in 1340 (Griffiths 1971: 346; Walker 1978: 123).6 There were many opportunities, then, for Llywelyn Bren to buy his consumer goods in Cardiff, where merchants sold goods brought in from the surrounding countryside and also imported through the long-distance trade routes. Wales did, of course, have its own cloth industry to provide basic materials to local consumers and to more distant markets. Wool and woollen cloth were among the largest exports in the medieval Welsh economy, with wool being exported via Bristol to the Low Countries to be finished and dyed before being re-imported into England. Flemish settlers in south Wales from the twelfth century had helped to develop the industry, and the earliest fulling mills were established by the thirteenth century. It has been estimated that there were about twenty-five fulling mills in Wales in the thirteenth century and a total of 202 mills in Wales and the March by 1547 (Jack 1980–1981: 447). Carmarthen, Cardiff, and Shrewsbury became staple towns in 1326, that is, they were central collection points which controlled the export of wool from Wales and the March to the continent (Lloyd 1977: 115). During the fourteenth and fifteenth centuries, the cloth trade expanded significantly in Wales, especially in the areas near the March, supported by the Marcher lords, some of the religious houses, and a number of entrepreneurial individuals. Before about 1300, however, luxury imported fabrics were a rarity in Wales, found only at the courts of the great princes and probably brought  See also Sharp and Stamp (1912: no. 371), which refers to a fair in the borough of Cynffig in Glamorgan during the reign of Edward I in 1263. 6

7  Cheapside in Wales 

195

there directly from London. One of the poets of the princes, Llywarch ap Llywelyn, who flourished in the second half of the twelfth century, sang a praise poem in 1175 to Dafydd ab Owain Gwynedd, briefly prince of the whole of north Wales, mentioning the splendour of his garments: Ei anaw a ddaw, a ddyfnad, I anant ei eni bu mad, Ei aur rhudd yn rhodd o’i geiniad A’i emys a’i amryw dillad; Ei bali porffor parth nâd—a wyrthia A’i werthfawr ysgarlad. (Jones and Jones 1991: no. 1, ll. 39–44) ‘His wealth is bestowed, it is his custom, on his poets, blessed was his birth, his red gold a gift to his singer, and his horses and his various garments; his purple silk brocade in exchange for a song, which causes amazement, and his valuable scarlet’.7

Not only does the king possess these amazing fabrics, the kind of things that are also listed in the wardrobe of the King of England, but he gives them as gifts to his poets as a mark of his royal generosity. The wonder inspired by the foreignness of these luxury fabrics in the twelfth and thirteenth centuries is translated into the language of romance in the prose tales of the Mabinogion, a collection of largely unrelated prose tales (apart from the ‘Four Branches’ which are linked together) composed at various times between about 1100 and 1300. In one of the three tales derived from the twelfth-century French poems of Chrétien de Troyes, the romance of Owein or ‘The Lady of the Fountain’, the hero Owain, one of Arthur’s knights, finds himself in a foreign land whose otherness is marked by expensive high-quality commodities. In the castle of the noblewoman whom he marries, the wooden panels are painted with gold patterns, the furniture is set with silver and gold, the dishes are all of silver or gold or ivory, and the towels and tableware are of the finest linen, or bliant, a borrowing either from ME bleaunt, bliaunt (‘cambric, lawn’; MED, s.v. bleaunt) or directly from Anglo-French (A-Fr.) bliaunt. When Owain sees a funeral procession from the window, he is amazed by  Translations are mine unless otherwise credited.

7

196 

H. Fulton

the barons who are all dressed in pali a seric a syndal (‘brocaded silk and damask and sendal’).8 When Owain goes to sleep, his bed has been made up with ysgarlat a gra a phali a syndal a bliant (‘scarlet cloth and ermine and brocaded silk and sendal and linen’).9 These are all luxury fabrics that were not made in Wales at this time— hence their symbolic otherness—and had to be imported from France or the Low Countries. The term pali (‘silk brocade’) occurs fairly often in the prose tales of the Mabinogion, including the ‘Four Branches’, which date from the early twelfth century. Signposting the wealth of the nobility, the word is borrowed from OFr. palie and suggests its value in linking the courts of the Welsh nobility with the prestige of those of France.10 Serig or sirig is the Welsh form of L sericum, meaning ‘serge’, which I have discussed above in relation to Llywelyn Bren and his list of goods. According to the OED (s.v. serge), ‘[t]he material originally designated by the name [serge] must have been silk, though there is no evidence of this in the early English (and apparently not in the Romance) uses of the word. But names of textile fabrics often come to be applied to materials cheaper and coarser than those which they originally designated’. I suggest that the reference to seric in the Welsh tale of Owein, which can be dated to the second half of the thirteenth century, must therefore be the earliest insular reference to serge as a silk fabric rather than as the more standard and cheaper woollen fabric, as it later became.11 Syndal, ‘sendal’, is likely to be a borrowing from OFr. cendal, though a ME form sindal is also attested as well as ME cendal, borrowed directly from French (MED, s.v. cendal).12 Found in medieval Latin as cendalum, the word occurs in most of the Romance languages and clearly refers to a relatively well-­ known fabric. The texture was probably some kind of silk, though the Greek cognate, sindon, refers to a fine linen, and this usage is found in  Thomson (1968: ll. 352–353); translation by Davies (2007: 125).  Thomson (1968: l. 335) and Davies (2007: 124). See also Williams (2012: 93–95). 10  The term is also attested in Old English, Middle English, Anglo-French, Older Scots and Cornish, indicating the fabric’s ubiquity throughout Britain (Lexis of Cloth and Clothing Project, s.v. pall). 11  The etymology of serge suggests an original association with silk, though by the Middle Ages it was a woollen cloth (LexP, s.vv. serge and sericum). 12  Compare Middle Cornish sendall, cendal (Geiriadur Prifysgol Cymru / Dictionary of the Welsh Language, s.v. syndal). See also LexP (s.v. sendal). Surridge (1984: 247) includes syndal as one of her list of Welsh words borrowed from French. 8 9

7  Cheapside in Wales 

197

Middle English from the fourteenth century. The word is attested in Welsh from the twelfth century (GPC, s.v. syndal) and seems to mean ‘silk’ rather than ‘linen’. Sendal could also be dyed in different colours, as we see in another of the tales from the Mabinogion collection, Breuddwyd Rhonabwy, ‘The Dream of Rhonabwy’, composed in the first half of the thirteenth century. When Arthur and Owain are playing their deadly game of gwyddbwyll (a boardgame similar to chess) and Owain’s ravens are butchering Arthur’s men, a knight comes riding up with his horse covered in sendal, partly bright red and partly bright yellow: Cwnsallt y varch o’r gorof vlaen idaw y vynyd yn syndal purgoch, ac o’r gorof y waeret yn syndal puruelyn.13 ‘His horse’s covering, from the front pommel of his saddle upwards, was of bright red sendal, and from the pommel downwards, of bright yellow sendal’.

Lists of fabrics and precious stones are a major stylistic feature of this particular tale, where every troop of knights and many individuals are distinguished by their colours, fabrics, fringes, jewels, and other ornaments. Pali, silk brocade, features fairly regularly, but these are the only references to sendal, suggesting its rarity and the brightness of its colours. The many references to colour in ‘The Dream of Rhonabwy’, reds and golds and greens and blues, is part of the tale’s atmosphere of the marvellous; compared to the sober browns and greys of homespun cloths, these are the colours of the otherworld. Returning to the list of fabrics which make up Owain’s bedclothes, we come to ysgarlad (‘scarlet’), the same fabric mentioned in tones of wonder by the poet Llywarch ap Llywelyn. The Welsh ysgarlad is a borrowing from OFr. or A-Fr. escarlate, and the word is known throughout the European vernaculars, both Romance and Germanic, including medieval Latin, but its origins are obscure.14 The word may be related to the verb  Richards (1948: 15, ll. 17–19); Davies (2007: 222).  The OED suggests an etymology from Persian saqalat, siqalat, suqlat, a kind of rich cloth (OED s.v. scarlet), but it may have its origins in Arabic (LexP s.v. scarlet) .

13 14

198 

H. Fulton

‘to shear’, since the main feature of the cloth was its triple shearing, creating a voluptuously soft finish. According to Munro, ‘scarlets were indisputably the most costly of all woollens, and usually the largest’; and he adds: ‘the world of the scarlet was for the few, the very few indeed: a world of popes, emperors, archbishops, princes — and also of Flemish civic aldermen’ (Munro 1983: 37 and 70). It is not surprising, then, that the wearing of scarlet was specifically forbidden to all but the wealthiest classes in the various sumptuary laws issued during the fourteenth century, laws that attempted, not very successfully, to arrest the spread of luxury fabrics down the social hierarchy and to resist the forces of consumerism (Harte 1976; Hunt 1996).15 Scarlet was originally produced in the three basic colours of white, that is, undyed; blue, woven from wool that had already been dyed indigo using woad, a plant dye; or medley, using a mixture of different wools, which could also be used to create stripes or ‘rays’ (where the warp yarn and weft yarn were of different colours). The blue-coloured cloths could be re-dyed to produce other colours such as browns, purples, violets and greens. However, from about the thirteenth century scarlet was dyed with the most expensive dye then available, which was an insect dye, or kermes, belonging to the same type as that which produced vermillion. Because the cost of the dye and the processing was so high, the cloth itself was one of the most expensive on the market. Although the English word from kermes is crimson, the colour of the cloth gradually became fused with the cloth itself, and with the dye that produced it, since ‘scarlet’ could also refer to the dye (LexP, s.v. scarlet). Throughout the Middle Ages, then, scarlet cloth was likely to be red, but not always. Going back to Owain and his bedclothes, we can see that they were made from some of the finest, softest, and most expensive fabrics then being produced. Unlike the ordinary homespun wools, none of these bedclothes would have scratched his skin or brought him out in a rash. The final word in the list is gra, a type of fur, which Davies has translated as ‘ermine’, that is, the grey-white fur of the stoat, and one of the most expensive furs available in Europe. The term also signified the colour  Some scholars have argued that the sumptuary laws were in fact a response to the fear that social class could not be simply ‘read off’ from clothing. See Sponsler (1992) and Sylvester (2017). 15

7  Cheapside in Wales 

199

‘grey’ as this was the general colour of the fur (LexP, s.v. grey). The word gra is borrowed from ME gra or grai / grei (‘a grey fur’).16 The MED (s.v. grei, n. 2) suggests it signified squirrel fur, but ME grei was also the word for a badger and clearly could mean any small animal with grey-coloured fur. As in the Welsh example from Owein, the concept of gra in English was associated with expensive luxury fabrics. In Homily 3 of the Northern Homily Cycle, for example, dating to about 1300, Christ emphasises the humility of John the Baptist by saying that es he nan of tha / That er cled in gren and gra (‘he is not one of those who are clothed in green and grey’), and he goes on to say In kinges houses…won thai / That er cled in gren and grai (‘In kings’ houses…dwell those who are clothed in green and grey’; Thompson 2008: ll. 197–202). The late twelfth-century homiletic poem, the Poema Morale, distinguishes between several different kinds of fur: Ne sal þar ben foh ne grai ne cunin ne ermine (‘nor shall there be fou [particoloured fur] nor grey nor rabbit nor ermine’; Morris 1873: l. 365). The latter two nouns were borrowed from French, and are also attested in Latin, Anglo-French, Older Scots and Middle English (LexP, s.vv. cony and ermine). The evidence from Welsh poetry and prose before 1300 is that imported luxury fabrics were available in Wales but only to the very wealthy. Their value and rarity made them objects of wonder, signifiers of a world of romance and fantasy far removed from everyday life, a world where conspicuous consumption was confined to a tiny minority. There is also a relatively high number of textile and clothing terms in Welsh that are likely to be borrowings from French or Anglo-French rather than English, such as bliant, pali, bwckran, cordwal (OFr. cordoan; A-Fr. cordewan, ‘cordwain’), swrcot (OFr. sorcot, surcot), syndal.17 Such borrowings are indicative not only of linguistic interactions between the Welsh and French languages in the centuries following the Norman settlement of the March of Wales but also of the prestige consumer goods, including  The most common ME form is grei, while gra suggests a northern form derived from Old Norse grá-r (LexP, s.v. grey). The Welsh gra may therefore have come into the language from the northern dialect of Middle English, possibly suggesting a specialised trade with the north of Britain. 17  Surridge (1984: 249) comments that, in her study of French loanwords in the Mabinogion, ‘it is clear that the largest bloc of words relates to the world of luxurious and fashionable clothing and ornament’. 16

200 

H. Fulton

fabrics, that became available to the highest levels of Welsh society, mainly the regional princes and their families.

3 The Rise of Consumerism in Wales After 1300, the economy of Wales changed dramatically, resulting in a new influx of luxury and imported goods and a new market sector who wanted to buy them. The cause of this change was the military conquest of north Wales by Edward I in 1282. The last independent prince of north Wales, Llywelyn ap Gruffudd, was killed and his territories were brought under the control of the English crown. From then on, Wales was divided into two major zones: the Marcher lordships in the east, south, and south-west; and the Crown lordships (or Principality) in the west and north, now under the direct control of the English king. When Llywelyn ap Gruffudd was killed, the entire structure of the Welsh princely aristocracy went down with him. Most of the men were killed or imprisoned, the women were sent to convents, the children were disinherited and their lands removed. Even apart from the massive social and cultural trauma of this change, which happened within a few years, there was a sudden change too for the poets of the princes, men who had lived at the royal courts and had been paid to sing the praises of the princes and their families. The year 1282 marked a rupture in Welsh cultural production, an abrupt halt to the kind of literature that had been produced because the aristocratic patrons who had once supported this literature were no longer there. Without a native aristocracy at the top of society, namely the princes who marked themselves out by wearing scarlet and purple, the social hierarchy became slightly flatter. A new class of Welsh gentry emerged to fill the vacuum, the uchelwyr, or high-born men, an aspirational class of wealthy landowners and administrators, many of whom worked for the Marcher lords or the English Crown and looked to the French-speaking English aristocracy to set the tone in terms of cultural practice. The aspirations of the uchelwyr were supported by the new economic situation in Wales. After 1282, the urban economy, already established under the Normans, went into overdrive. Immediately following the

7  Cheapside in Wales 

201

conquest, Edward I put into action a hugely ambitious project of town-­ building, bringing a reinvigorated and greatly expanded urban culture to the north and north-east of the country. The immense castles built by Edward to control the Welsh population—castles such as Flint, Conwy and Caernarfon—all had sizeable towns attached to them. Filled with English settlers looking to profit from trade, the new towns, like some of the old ones, were not welcoming to the Welsh, and yet a culture of urban trade began to infiltrate Welsh society, including its literary production. Alongside the new gentry, a new class of poets emerged in the early fourteenth century to serve this audience, poets who responded to the rise of urbanism in Wales, and indeed in England, by adapting their frame of reference to include urban culture and urban audiences. While the new Welsh gentry lived in estate houses or ‘courts’ where they could receive poets and musicians, replacing the courts of the Welsh princes lost after 1282, the towns offered an alternative economic location. Towns were now the public spaces where poets looked for business, found new patrons and audiences, and sold their work as commodities in the open market. The new commercialism was already making itself felt in the poetry of Dafydd ap Gwilym, who flourished in the middle decades of the fourteenth century. He sets some of his poems in the market towns of north Wales where he mixes with the merchants and tradesmen and conveys the bustle and noise of the fair. In one poem, he describes his passion for a merchant’s wife who is involved in the cloth trade: O cherais wraig mewn meigoel Wrth lyn y porthmonyn moel, Gwragennus esgus osgordd, Gwraig, rhyw benaig, Robin Nordd, Elen chwannog i olud, Fy anrhaith â’r lediaith lud, Brenhines, arglwyddes gwlân, Brethyndai bro eithindan, Dyn serchog oedd raid yno. Gwae hi nad myfi fai fo! Ni chymer hon, wiwdon wedd, Gerdd yn rhad, gwrdd anrhydedd.

202 

H. Fulton

Hawdd oedd gael, gafael gyfa’, Haws no dim, hosanau da. Ac os caf liw gwynnaf gwawn, O fedlai y’m gwnâi’n fodlawn. (Johnston et al. 2010: no. 120, ll. 13–28) ‘If I loved a woman half on trust over a drink with the bald merchant, a hunchback with an excuse for a retinue, the wife of a certain leader, Robin North, Elen eager for wealth, my treasure with the sticky foreign accent, the queen, the mistress of wool, of cloth-factories in the land of gorse-fire, a loving man was certainly needed. Bad luck for her that I wasn’t that man! Face like a beautiful wave, she won’t take a poem for free, steadfast her honour. It’s easy to get some good stockings, to take the lot, easier than anything. And if I can have her, colour of the whitest gossamer, she will make me content with medley’.

In a style that is characteristic of Dafydd’s poetry, he juxtaposes the language of courtly love with the mercantile imagery of urban trade. Robin North is very likely a real person, a burgess of Aberystwyth (Dafydd’s nearest market town) in the mid-fourteenth century whose name appears as Robert le Northern in a legal document of 1344, one of the many Englishmen who traded in the towns of Wales (Parry 1963: xvi). Robin and his wife are cloth-merchants, so the poet hopes to be rewarded for his poetry, not with cash, but in kind, in the form of good woollen stockings and medley, an inexpensive cloth spun from different coloured wools. The reference to Elen’s ‘sticky foreign accent’ is a comment on her Englishness, which makes her sound foreign to the Welsh-­ speaking majority of Aberystwyth. We can assume that both Robin and Elen would have learned some Welsh, in order to conduct their trade, but presumably not enough to remove their ‘foreignness’. It is in the fifteenth century that the effects of international trade and greater consumer choice become increasingly evident in Welsh poetry. References to textiles and a wide range of consumer goods, especially prestige commodities, proliferate among the court poets who sang to the wealthy gentry of the Marcher lordships, whose growing towns provided material status symbols for Welsh and English alike. Around the middle of the fifteenth century, the Welsh word sieb, borrowed from ME chep

7  Cheapside in Wales 

203

(which had the separate meanings of ‘price’ and ‘bargaining’; Ingham and Sylvester 2017–:, s.v. chep), starts to be used in Welsh poetry in the context of urban commerce, indicating that Welsh towns followed an English urban model of commodity markets. Although a ‘cheap’ or a ‘cheapside’ in English could mean a market in any town, its dominant reference came to be the area of Cheapside in the city of London, the largest marketplace in Britain, and in most cases, though not all, the use of sieb in Welsh poetry does literally refer to London’s Cheapside. Gutun Owain (fl. c.1460–1498), praising his patron Huw Conwy and his wife Elizabeth, compares the consumer goods at their court to the riches of London and St David’s: Plas main, fal Powls a Mynyw, / Elsabeth, ail i Sieb yw (‘Elizabeth’s fine mansion, like St Paul’s and St David’s, is a second Cheapside’; Bowen 1959: no. 13, ll. 23–24). An unknown love poet, addressing a salmon and asking it to visit his beloved by the lake, describes the girl as llinos hwy no llin o Sieb (‘a linnet taller than a length of linen from Cheapside’; Fulton 1996: no. 37, l. 26). Using sieb in a more generic sense, Dafydd ab Edmwnd (fl. c.1450–1497) imagines a kiss from a beautiful girl to be like the senses and flavours of the market: sipio hin siop ei hwyneb / swpera ar sawr siopau’r sieb (‘sipping a measure from the shop of her face, supping on the aroma of the market shops’; Rowlands 1976: no. 13, ll. 23–24).18 It is worth noting here that the Welsh hin is borrowed, via Middle English, from hin, a Hebrew word referring to a liquid measure. The Middle English examples all come from the Wycliffe Bible of the late fourteenth century (MED, s.v. hin), and following the translation of the Bible into English, the word is likely to have circulated in popular oral usage, first in England and then in Wales, around marketplaces where drinks were sold.19 The urban Welsh looked towards London as their commercial yardstick mainly because most of the imported goods that were sold in Welsh  Another poem by Dafydd ap Edmwnd describing the sensory delights of a girl’s kiss compares it to drinks and spices from the marketplace, many of them imported: osai pêr (‘sweet wine’), sinamwn (‘cinnamon’), siwgr gwyn (‘white sugar’), orains (‘orange’), pupur (‘pepper’); see Roberts (1914: no. 26). 19  Welsh hin is also used in William Salesbury’s Welsh translation of the Bible of 1588, the source of the earliest attested examples in the Geiriadur Prifysgol Cymru. But the example found in Dafydd ab Edmwnd’s poem proves that the word was in circulation in Wales around a century before the 1588 Bible, no doubt picked up from English usage. 18

204 

H. Fulton

towns and in other provincial markets came through London. This was especially true of textiles, a market that was dominated by the guild of Mercers in London. The London Mercers used middlemen called chapmen, local tradesmen in the provincial regions, to sell their goods around the country, but the goods had to have been sold to the chapmen in the London marketplace before they could transport them to other fairs and markets in other parts of the country (Sutton 2005: 212–216). The point of this was to keep control of the trade, keep it centred in London, and obscure the prices of buying and selling so that the final customers at the markets and fairs did not know the wholesale price. This meant that most of the commodities and all the imported textiles that came to Welsh towns in the fifteenth century had come through London, which became the de facto capital of Wales. There are two well-­ known poems to the town of Oswestry, one by Guto’r Glyn, who flourished from about 1435 to 1490, and one by Tudur Aled, who flourished in the last quarter of the fifteenth century and the first quarter of the sixteenth century. Both these poems compare Oswestry to London’s Cheapside (Fulton 1997). Oswestry had long been a key player in the urban economy of the March, a town that was right on the border of Wales and England and which, unlike many Welsh towns, welcomed both Welsh and English burgesses (Smith 1978). It was also an important cloth-making centre, supporting a local industry that comprised merchants as well as the wool producers and the craftsmen and women who made the finished garments. The poems to Oswestry are part of a genre of Welsh praise-poems to towns, which conform broadly to the rhetorical principles of the classical encomium urbis (Fulton 2006). Guto’r Glyn praises the town for its churches and scholars, its wine and its women, and he declares Ynddi mae marsiandi Siêb (‘in this place is the merchandise of Cheapside’), meaning that almost everything you could find in London is also for sale in Oswestry. He even calls it Llundain gwlad Owain a’i dir (‘London of Owain’s territory and land’; Williams and Williams 1939: no. 69, ll. 35 and 21), meaning that Oswestry was the capital of the region where the former rebel Owain Glyndŵ r had his estates at the end of the fourteenth century. Attacked by Owain and his forces in the early fifteenth century, Oswestry had restored itself to

7  Cheapside in Wales 

205

become the jewel in the crown of the Marcher lordship of the FitzAlans, the earls of Arundel. Tudur Aled celebrates Oswestry as a city of crefftwyr, llafurwyr (‘craftsmen and labourers’), whose markets are full of all the commodities that anyone could want: Cistiau da, ’n costio dierth, Cwmin, bocs, caem win heb werth; Siwgr, sarsned, ffelfed a phân, Siêb-Seid yn siopau sidan… ….Seigiau o garw gwresowgwyllt, Serch, gwin, a sew o iwrch gwyllt; Cwrw a siwgr caer wresowgwin, Cwnffets, pomgarnets, a gwin. (Jones 1926: no. 65, ll. 61–64 and 87–90) ‘Chests of goods, expensive for aliens, cumin, boxwood, we get wine without tax; sugar, sarsnet, velvet and fur, Cheapside here in the silk shops…Hot dishes of wild venison, love, wine, and a soup of wild deer; beer and sugar in a mulled-wine fortress, comfits, pomegranates, and wine’.

The sheer number of loanwords from English indicates not only the bilingualism of the March but also the exoticism of the commodities for sale, items like cumin and pomegranates which have come from the Middle East along the international trade routes, showing the extent to which Oswestry is a cosmopolitan city because it can boast of having such things for sale. The textiles mentioned here are three of the most expensive and all imported. This is probably the earliest reference in Welsh to sarsnet, a fine silk fabric made in Bruges, and the word first appears in English about a century earlier (LexP, s.v. sarsenet; MED s.v. sarsinet, with the earliest citation from a will of 1373). Velvet was also made from silk and imported from the continent, while pân, or fur, meaning ‘ermine’, was mainly imported from the Baltic and Prussia via the eastern European fairs (Sutton 2005: 259). Oswestry, then, not only produced its own woollen cloths but brought in the finest and most exotic fabrics to sell in its markets, which could

206 

H. Fulton

stand comparison with Cheapside in London. As these expensive commodities became more available in Welsh market towns, they became more visible and therefore more normalised among the wealthy classes of Wales. Poets often used textile names as descriptors of women, birds, and animals, like the anonymous fifteenth-century poet who described a blackbird wearing fine clothes all in black: Gwisgaist, enynnaist annerch, Gwisg ddu, nid er selu serch, Gwisg a ddanfones Iesu Is dail it o sae du, A dwbwl gwell no deuban, Mawr ei glod, o’r mwrrai glân; Sidan gapan am gopa Yn ddu rhoed yn ddiau’r ha’, Dwbled harddgled mewn rhedyn Blac-y-lir uwch glandir glyn. (Fulton 1996: no. 35, ll. 7–16) ‘You wore, you kindled a greeting, a black robe, not to watch for love, a robe which Jesus sent beneath leaves for you, of black serge, and a lining, better than homespun, great its praise, of pure murrey; a silk cap for a crest, all black, was surely given in the summer, a doublet, beautiful and warm in bracken, of black-a-lire, above a valley of fine land’.

Mwrrai, borrowed from ME murrei (originally from OFr. moré) was a mulberry-coloured woollen cloth, dark red or purple, often worn by churchmen or by people in mourning or even as a shroud. Black-a-lire, also an English borrowing, was a fine black woollen cloth originally made in Lire in Brabant. The overall effect is to personify the bird as someone very smartly dressed in black and dark purple, like a wealthy churchman or an administrator, or even a poet of the gentry. The Englishness of the description, like the goods listed in Tudur Aled’s poem to Oswestry, implies a distinctive Marcher culture, one where Welsh and English gentries shared similar values of status and hierarchy expressed in part through clothing and commodities. Despite the relative domestication of fine fabrics in Wales, their primary signifier remained as markers of wealth and status, the wearing of

7  Cheapside in Wales 

207

clothes that set one social class apart from others. As Burns has said, ‘[t]extiles stand at the nexus of the personal and the cultural’, while clothing has the ability ‘to generate public identity and create social status’ (Burns 2004: 1, 10). The social function of fabrics is manipulated skillfully by Lewys Glyn Cothi in his poem of the late fifteenth century celebrating the marriage of a Marcher gentleman, Robert Whitney, to Elis, the daughter of Tomas ap Rhosier Fychan, one of the Vaughans of Hergest, an old and distinguished Marcher family. This was a very suitable marriage between people of the same class and the same social circle, a marriage designed to reinforce the exclusiveness and prestige of the Marcher gentry as a single status group where Welsh and English joined together to protect their shared interests. Lewys describes the bride-to-be like this: Mae i’r Mastres Elis, mewn tŵ r caead, mwnai a thrysor, main a thrwsiad; hi a wisg Dduwllun ddamasg ddillad, siamled â felwed un ddyfaliad, ac a arwedd aur uwch grudd a iad, ac a wisg gerlont ac ysgarlad. (Johnston 1995: no. 123, ll. 17–22) ‘Mistress Elis, enclosed in her tower, has money and treasures, jewels and clothes; on Monday she will wear damask garments, camlet with velvet of some description, and she will display gold above her cheek and temple, and she will wear a garland and scarlet’.

There is a wonderful formality to this description, suggesting the dignity and beauty of the young woman as she goes to be married. But the strongest implication is that of wealth: her clothing and jewels are both the visible exterior signs of her value and the proxies for her social status. The poet leaves us in no doubt that these are what she brings to her husband: her money and her class.

208 

H. Fulton

4 Conclusion The growth of the urban economy in Wales after 1300 had a particular impact on the Welsh: on the one hand, they knew themselves to be economically disadvantaged by the English ownership and control of towns, but, on the other hand, they were able to benefit from the greater access to consumer goods. In the extensive body of Welsh court poetry composed in the fourteenth and fifteenth centuries, we find evidence of increasing engagement with the urban economy and with the status differentials conferred by particular consumer goods, such as jewellery, furnishings, and clothing. Luxury fabrics, once a source of wonder and suggestive of French romance, became a key consumer commodity which, along with praise poetry, supported the aspirations of the gentry to be the new nobility of Wales. What we also see clearly in the number of French and English loanwords into Welsh during this period, with many French terms likely to have been borrowed through the medium of English, is the increasing dominance of English speakers in the social and economic life of Wales, mainly due to their larger control of the urban economy on the March, where goods flowed into Wales. The examples of commodity terms, particularly textiles, borrowed from English into Welsh illustrate the larger trends towards anglicisation of Welsh and high levels of bilingualism, especially among Welsh speakers and especially on the March. They also illustrate the significance of the urban economy as a major conduit for linguistic influence, bringing English traders and English commodities into Welsh-speaking areas. Finally, the textiles themselves, and the clothing and furnishings made from them, functioned as key indicators of wealth and status for both Welsh and English gentries, aligning their interests as the dominant class group and eliding differences of language, culture and history. In that sense, Wales had become, by the fifteenth century, truly European, part of a ‘display culture, where luxury goods such as dress, jewels and arms not only signaled but also actually constituted political power, social bonds, and hierarchy itself ’ (Howell 2010: 5).

7  Cheapside in Wales 

209

References Benson, Larry D., ed. 2008. The Riverside Chaucer, 3rd edn. Oxford: Oxford University Press. Bowen, D.J., ed. 1959. Barddoniaeth yr Uchelwyr. Caerdydd: Gwasg Prifysgol Cymru. Bromwich, Rachel. 1986. Tradition and Innovation in the Poetry of Dafydd ap Gwilym. In Aspects of the Poetry of Dafydd ap Gwilym: Collected Papers, by Rachel Bromwich, 57–89. Cardiff: University of Wales Press. Burns, E. Jane. 2004. Why Textiles Make a Difference. In Medieval Fabrications: Dress, Textiles, Cloth Work, and Other Cultural Imaginings, ed. E. Jane Burns, 1–18. New York: Palgrave Macmillan. Davies, R.R. 2000. The Age of Conquest: Wales 1063–1415. Oxford: Oxford University Press. Davies, Sioned, trans. 2007. The Mabinogion. Oxford: Oxford University Press. Dyer, Christopher. 1989. Standards of Living in the Later Middle Ages: Social Change in England, c. 1200–1520. Cambridge: Cambridge University Press. ———. 2002. Making a Living in the Middle Ages: The People of Britain, 850–1520. New Haven, CT: Yale University Press. Everitt, Alan. 1990. The Marketing of Agricultural Produce, 1500–1640. In Agricultural Markets and Trade, 1500–1750, ed. John Chartres, 15–138. Cambridge: Cambridge University Press. Fulton, Helen, ed. and trans. 1996. Selections from the Dafydd ap Gwilym Apocrypha. Llandysul: Gomer Press. ———. 1997. Trading Places: Representations of Urban Culture in Medieval Welsh Poetry. Studia Celtica 31: 219–230. ———. 2006. The Encomium Urbis in Medieval Welsh Poetry. Proceedings of the Harvard Celtic Colloquium 26: 54–72. ———. 2012. Urban Culture in Medieval Wales. Cardiff: University of Wales Press. [GPC =] Geiriadur Prifysgol Cymru / Dictionary of the Welsh Language. https:// welsh-dictionary.ac.uk/gpc/gpc.html. Accessed 30 May 2023. Griffiths, Ralph A. 1971. The Medieval Boroughs of Glamorgan and Medieval Swansea. In Glamorgan County History, III: The Middle Ages, ed. T.B. Pugh, 333–360. Cardiff: University of Wales Press. Harte, N.B. 1976. State Control of Dress and Social Change in Pre-Industrial England. In Trade, Government and Economy in Pre-Industrial England: Essays Presented to F. J. Fisher, ed. D.C. Coleman and A.H. John, 132–165. London: Weidenfeld and Nicolson.

210 

H. Fulton

Howell, Martha C. 2010. Commerce before Capitalism in Europe, 1300–1600. Cambridge: Cambridge University Press. Hunt, Alan. 1996. Governance of the Consuming Passions: A History of Sumptuary Law. Basingstoke: Macmillan. Ingham, Richard, and Louise Sylvester. 2017–. The Bilingual Thesaurus of Everyday Life in Medieval England. https://thesaurus.ac.uk/bth/. Accessed 30 May 2023. Jack, R.  Ian. 1980–81. The Cloth Industry in Medieval Wales. Welsh History Review 10: 443–460. Johnston, Dafydd, ed. 1995. Gwaith Lewys Glyn Cothi. Caerdydd: Gwasg Prifysgol Cymru. Johnston, Dafydd, Huw Edwards, and Dylan Foster Evans, eds. 2010. Cerddi Dafydd ap Gwilym. Caerdydd: Gwasg Prifysgol Cymru. Jones, T.  Gwynn, ed. 1926. Gwaith Tudur Aled. 2 vols. Caerdydd: Gwasg Prifysgol Cymru. Jones, Elin M., and Nerys Ann Jones, eds. 1991. Gwaith Llywarch ap Llywelyn ‘Prydydd y Moch’. Caerdydd: Gwasg Prifysgol Cymru. Kibler, William W., trans. 1991. Chrétien de Troyes, Arthurian Romances. London: Penguin. Lloyd, T.H. 1977. The English Wool Trade in the Middle Ages. Cambridge: Cambridge University Press. The Lexis of Cloth and Clothing Project. University of Manchester. http://lexisproject.arts.manchester.ac.uk/index.html. Accessed 02 August 2021. Matthews, J. H., ed. 1898–1904. Cardiff Records. 5 vols. Cardiff: Corporation of Cardiff. [MED =] Middle English Dictionary. 1952–2001. Ed. by Robert E.  Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner (Ann Arbor: University of Michigan Press). https://quod.lib.umich.edu/m/med/. Accessed 27 July 2021. Morris, R., ed. 1873. Old English Homilies of the Twelfth Century, Early English Text Society, o.s. 53. London: Trübner. Munro, John H. 1983. The Medieval Scarlet and the Economics of Sartorial Splendour. In Cloth and Clothing in Medieval Europe: Essays in Memory of Professor E.  M. Carus-Wilson, ed. N.B.  Harte and K.G.  Ponting, 13–70. London: Heinemann. ———. n.d. Textile Production, Wages, and Prices. Medieval and Early Modern Data Bank. https://memdb.libraries.rutgers.edu/munro-prices-wages. Accessed 30 May 2023.

7  Cheapside in Wales 

211

[OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 27 May 2023. Parry, Thomas, ed. 1963. Gwaith Dafydd ap Gwilym, 2nd edn. Caerdydd: Gwasg Prifysgol Cymru. Parry-Williams, T.H. 1923. The English Element in Welsh: A Study of English Loan-Words in Welsh. London: Honourable Society of Cymmrodorion. Richards, Melville, ed. 1948. Breudwyt Ronabwy. Caerdydd: Gwasg Prifysgol Cymru. Roberts, Thomas, ed. 1914. Gwaith Dafydd ab Edmwnd. Bangor: Bangor Welsh Manuscripts Society. Roques, Mario, ed. 1981. Les Romans de Chrétien de Troyes, I: Erec et Enide. Paris: Honoré Champion. Rowlands, Eurys I., ed. 1976. Poems of the Cywyddwyr. Dublin: Dublin Institute for Advanced Studies. Sharp, J.E.E.S., and A.E. Stamp, eds. 1912. Calendar of Inquisitions Post Mortem, iii: Edward I. London: H.M. Stationery Office. Smith, Llinos B. 1978. Oswestry. In Boroughs of Mediaeval Wales, ed. R.A. Griffiths, 219–244. Cardiff: University of Wales Press. Sponsler, Claire. 1992. Narrating the Social Order: Medieval Clothing Laws. Clio 21: 265–283. Surridge, Marie. 1984. Words of Romance Origin in the Four Branches of the Mabinogi and Native Welsh Tales. Études Celtiques 21: 239–255. Sutton, Anne F. 2005. The Mercery of London. Aldershot: Ashgate. Sylvester, Louise. 2017. Dress, Fashion and Anti-Fashion in the Medieval Imagination. In Stasis in the Medieval West? Questioning Change and Continuity, ed. Michael Bintley, Martin Locker, Victoria Symons, and Mary Wellesley, 253–270. New York: Palgrave Macmillan. Thompson, Anne B., ed. 2008. The Northern Homily Cycle. Kalamazoo, MI: Medieval Institute Publications. Thomson, R.L., ed. 1968. Owein, or Chwedl Iarlles y Ffynnawn. Dublin: Dublin Institute for Advanced Studies. Walker, D.G. 1978. Cardiff. In Boroughs of Mediaeval Wales, ed. Ralph A. Griffiths, 103–128. Cardiff: University of Wales Press.

212 

H. Fulton

Walton, Penelope. 1991. Textiles. In English Medieval Industries: Craftsmen, Techniques, Products, ed. John Blair and Nigel Ramsay, 319–354. London: Hambledon Press. Williams, Patricia. 2012. Dress and Dignity in the Mabinogion. In Medieval Clothing and Textiles, VIII, ed. Robin Netherton and Gale R. Owen-Crocker, 83–113. Woodbridge: Boydell and Brewer. Williams, Ifor, and J. Ll Williams, eds. 1939. Gwaith Guto’r Glyn. Caerdydd: Gwasg Prifysgol Cymru.

8 Caxton’s Linguistic and Literary Multilingualism: English, French and Dutch in the History of Jason Ad Putter

1 Introduction William Caxton needs no introduction to anyone interested in the history of the language. Most textbooks on the subject accord him of a place of honour, and the stories they tell about him go something like this: Prior to the development of standard written English, there was much variation in writing. In the fifteenth century, William Caxton was keenly aware of this issue. Rather like the automobile industry, Caxton’s wish for some kind of standard had an economic dimension. Clearly, it would be very costly to print a different version of a book for every variety of English. So, Caxton needed to pick one variety of English that was widely ­understood and was socially valued (and thus would be taken seriously) […]. Caxton set up his printing press in Westminster in 1476, close to the government offices. His adoption of a London-based variety of English, including some

Ad. Putter (*) University of Bristol, Bristol, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_8

213

214 

Ad. Putter

features of the English of official circles, was the obvious choice (Culpeper 2015: 97).

The reason why a ‘London-based variety’ was the obvious choice for Caxton, Culpeper explains, is that a standard language needs to command respect and needs to be supralocal. Since London English had become the language of Chancery and was a hybrid Midland dialect mixing northern and southern features, it offered Caxton the best opportunities to make his books attractive and accessible to the widest possible market, so Culpeper and others argue (Nevalainen and van Ostade 2006, but see Wright 2020b for a different account of the origins of standard English). Short textbooks inevitably simplify (and nothing is more simplistic than the idea that you can pick your language), but they have the advantage of giving us the bare bones of received wisdom, and Culpeper’s account conveniently combines two widely shared ideas about Caxton that need more scrutiny. The first, which owes much to Blake’s debunking of Caxton’s reputation as a cultivated man of letters, is the assumption that Caxton was above all an entrepreneur: ‘[A]s a merchant Caxton could appreciate the position of the market, and he based his publishing policy upon that assessment […]. His livelihood depended upon his output […]. He was a businessman, not a scholar’ (Blake 1969a: 214). I call this an assumption because I know of no evidence that Caxton needed the income from his printing press to make ends meet. Caxton had enjoyed a hugely successful career in the cloth trade. The best part of this career he spent in the Low Countries, where in 1462 he rose to become Governor of the English Nation, that is, ‘chief executive’ of the company of Merchant Adventurers headquartered in Bruges (Driver 1997). An alternative theory to Blake’s is that by the 1470s he had become wealthy enough to indulge his interest in books and to take his chances with a new technology, the printing press, in the knowledge that any gains or losses he made would not be the making or breaking of him, though of course he preferred gains. Since this alternative theory cannot be evidenced either, I will speculate no further about Caxton’s personal motives as a printer. My topic is Caxton’s multilingualism as evidenced by his History of Jason, though I do in that context wish to highlight the

8  Caxton’s Linguistic and Literary Multilingualism 

215

impressive breadth of Caxton’s reading in a range of languages. Whatever profit margins Caxton wanted from his printing, his books certainly show him to be a scholar as well as businessman. The second assumption is that Caxton was concerned to implement a standard language for his books. This assumption owes nothing to Blake, who did not share it (Blake 1969a: 171–193), but has a much longer history. Wright (2020a) traces that history all the way back to the entry for English Language in the early-twentieth-century Encyclopaedia Britannica by James and Hilda Murray: Every reason conspired that this ‘English’ should be the midland dialect. It was the intermediate dialect, intelligible, as Trevisa has told us, to both extremes, even when these failed to be intelligible to each other … . By its exclusive patronage of the midland speech, it [Caxton’s press] raised it still higher above the sister dialects, and secured its abiding victory … . Modern English thus dates from Caxton (Murray and Murray 1910–1911: 594).

‘What is missing from previous discussions of standardisation’, Wright goes on to say, ‘is the fourteenth- and fifteenth-century multilingual background against which English began to be written’ (Wright 2020a: 29). It is the multilingual background to Caxton that I hope to explore and that explains some of the peculiarities of his language. In the study of Caxton’s English, there are special reasons for paying attention to this multilingual background. First, Caxton himself read and wrote in a variety of languages apart from English (Hsy 2013: 116–130): He translated from Latin, Dutch and French, and we have texts from him in English, French, Latin and Dutch (more about this Dutch text later). Second, Caxton spent around three decades on the Continent, mainly in Flanders (Hellinga and Hellinga 1976–1977; Goodman 2006), and so we should expect his English to have been influenced by the two vernacular languages, French and Dutch, that were spoken there. The only record of Caxton’s English language is that of the books he published, but that language shows interference from French and Dutch of all kinds (orthographical, grammatical, lexical) and is highly idiosyncratic rather than ‘standard’. I have argued this before with specific reference to Caxton’s earliest printed book in English, The Recuyell of the Historyes of Troye

216 

Ad. Putter

(Putter 2021). My aim here is to develop the case by examining the language of what was probably one of the last books that Caxton translated before his return from Flanders to England: The History of Jason. Of particular interest to me are Caxton’s lexical innovations, many of them overlooked in the Oxford English Dictionary (OED) and the Middle English Dictionary (MED). I shall begin by providing some necessary context to Jason and to Caxton’s multilingual credentials as reader, translator, printer, merchant and scholar. I shall then examine the influence of French and Dutch on the English language of Jason. The conclusion spells out the implications of my findings for editors of Caxton, for historical lexicographers, and for historians of the language who want a more nuanced assessment of Caxton’s place in the history of English.

2 The History of Jason and Caxton’s Multilingual Context Caxton’s Jason was printed in Westminster in 1477, just after Caxton’s return from the Low Countries. The book is his own translation of Raoul Lefèvre’s Histoire de Jason (c.1460), which tells the story of Jason’s conquest of the Golden Fleece and his eventful love life, including his stormy relationship with Medea. It is more than likely that the translation work was carried out in Flanders, just before Caxton returned to England. Caxton dedicated the work to Edward, Prince of Wales, son of King Edward IV, for reasons that shed light on the multilingual culture of this period. To present an English translation of a French original to King Edward IV would have been very inappropriate, Caxton writes in his Preface: not presumyng to presente it vnto his highnesse for asmoch as I doubte not his good grace hath it in Frensh, which he wel vnderstandeth, but not displesing his most noble grace I entende by his licence & congye [‘permission’] & by the supportacion [‘backing’] of our most redoubted liege lady, most exellent princesse the Quene to presente this sayde boke vnto the most fayr, and my moost redoubted yonge lorde, My lord,

8  Caxton’s Linguistic and Literary Multilingualism 

217

Prynce of Wales [...] to th’entent he may begynne to lerne rede Englissh (2).1

The literature of English royalty was still dominated by French books, and the holdings of Edward IV’s personal library, which includes a manuscript of Lefèvre’s second romance, Le receuil des histoires de Troyes, c.1464 (Kekewich 1971), amply justify Caxton’s sentiment that an English translation of Lefèvre’s would be out of place. I am not aware that Edward IV himself owned a French text of Lefèvre’s Histoire de Jason, but he might well have read it while he was in exile in Bruges and a guest of Louis de Gruuthuse. Gruuthuse’s sumptuous library (consisting largely of French books, but with some Latin and Dutch books too) inspired Edward IV’s own book collecting and certainly did contain a copy of Lefèvre’s Histoire de Jason, now Paris, Bibliothèque Nationale, MS fr. 331 (Smeyers 1999: 405). Caxton therefore reckoned that a translation of this romance in English would be better bestowed on Edward’s son, who was about seven at the time, to help him learn to read English. Whether the book was actually presented to the young prince or whether this was just a publisher’s ploy no-one really knows. Although Caxton’s Jason was printed in Westminster, its immediate context is Burgundian Flanders. Caxton had already translated Lefèvre’s Receuil, and this translation was probably printed around 1473 (Hellinga 2010: 48-50). An English translation of Lefèvre’s Histoire de Jason was a natural sequel. Caxton’s Recuyell was dedicated to Margaret of York, wife of Charles the Bold, Duke of Burgundy. Hellinga (2014: 307) has suggested that the History of Jason was also initially intended for her but that the childlessness of the couple made it awkward to publish the work, for it begins with a poignant reflection on the pain of noble couples without offspring: Anciently the kynges and princes of hye felicite were attendaunt [‘eagerly anticipating’] and awayted whan their seed shold bringe forth generacion 1  All citations are from the diplomatic edition by Munro (1913). I have added punctuation and glosses in square brackets, regularised capitalisation, and occasionally emended the text. Emendations are indicated by square brackets. References are to page numbers. I have included line numbers when notable words or phrases are in question.

218 

Ad. Putter

[offspring]. But whan so was that they myghte not come therto, what [whatever] prosperite they had, their lyf was trauersid in contynuell bewailing (4).

Publication was thus delayed until 1477, after Charles the Bold’s death. Around the same time as the English-language edition of Jason was published in Westminster, there appeared in Flanders the French-language printed edition that Caxton had been working on before his return to England. The biases of nationalism in historiography are such that Caxton is universally known as the ‘first English printer’, but his printing of Lefèvre’s Histoire de Jason, completed by a Flemish colleague, shows not only that Caxton worked across languages but also that the earliest books he produced were published under the imprint of Flemish publishers. The first printed book in English, Caxton’s Recuyell (Sommer 1894), was probably published in partnership with David Aubert in Ghent (or Collard Mansion in Bruges), and such collaboration was essential. To engage in the making and selling of books in a Flemish city one needed to be a guild member (Hellinga 2010: 40). Foreigners like Caxton could not set up a printing shop, and the official business of publishing a book in English thus had to be fronted by a local. Like Jason, the Receuil was also printed in both an English- and a French-language edition. Jason uses the same printing type as the French and English editions of the Recuyell, and so they were probably brought out by the same Flemish publisher (Hellinga 2010: 51). Caxton’s French Receuil is just as ground-­ breaking as his English version. Setting aside a fragment of a translation of Donatus’s Ars Minor (undated and unattributed, c.1470), printed in Holland, it is amongst the earliest French books to be printed. The first dated French-language book from France is the Legende doree, printed in Lyon by Guillaumme Leroy in 1476 (Colin and Hellinga 1973: 171). Caxton, in short, was a pioneer of French as well as English printing. The other languages that Caxton knew were Dutch and Latin. Since any formal education involved reading and writing Latin, Caxton would have learned Latin as a schoolboy, for in the prologue to Charles the Grete (printed in Westminster, 1485) he mentions that his parents sent him to school (Blake 1973: 68). He was involved in the early 1470s in the printing of Latin texts (De proprietatibus rerum and Gesta Romanorum) in

8  Caxton’s Linguistic and Literary Multilingualism 

219

association with Johan Veldener and personally translated from Latin the Life of St Roch (Finnegan 1952; Corsten 1975–1976). Caxton’s proficiency in Dutch is evident from his printed books. Basing himself on the Dutch Hystorie van Reynaert die vos (Hellinga 1952), he published in 1481 the hugely popular Reynard the Fox (Blake 1970) and also reworked a version of the Flemish–French conversation manual Livre des mestiers, originally composed in Bruges around 1340, into an English–French version which was published in 1483, and was edited by Bradley (1900) as The Dialogues in French and English. Less well-known is the fact that there survives a Dutch document written by Caxton. A discovery in the archives of Hoorn (North Holland) brought to light a copy of a letter, dated 1475, in impeccable Dutch, by Willem Kaxton (Bakker and Gerritsen 2004). The simple explanation for Caxton’s mastery of Dutch is that it was essential for his working life as a merchant in Flanders. While the medium of the Burgundian court and deluxe literary production was French, Dutch was the language of the people and of local government, and even the Burgundian dukes and duchesses (including Margaret of York) deemed it necessary to learn the language (Armstrong 1983: 189–212; Spufford 1993: 11). Rulings by the court of aldermen in Bruges were translated into French for the convenience of French merchants, but German- and English-speaking merchants were expected to cope with Dutch (Kowalewski 2017: 208). The privileges granted to foreign merchants by the rulers of the Dutch-speaking principalities were increasingly issued in Dutch. The Mercers’ Archives in London contains a manuscript anthology of some of these privileges, known as the Book of Privileges (Sutton and Visser-Fuchs 2011). Copied in London c.1485, it is quadrilingual (Latin, French and Dutch, with translations of each into Middle English). As Blake (1969b) showed, the Middle English translations of the Dutch privileges provide a close parallel to Caxton’s Dutch-­ inflected English in Reynard the Fox. What Blake did not know is that this anthology was probably based on a dossier produced for the negotiations of 1468 to end the Duke of Burgundy’s ban on the import of English cloth, and that the man in charge of these negotiations was none other than Caxton (Sutton and Visser-Fuchs 2011: 35). Caxton’s multilingualism is also apparent from the reading that went into his Jason. His immediate source was Lefèvre’s Histoire de Jason, in

220 

Ad. Putter

elegant French prose, which Caxton had read in a manuscript version (now lost) closely related to the surviving manuscript copy owned by Louis de Gruuthuse (Pinkernell 1971: 62–65). He was also impressively knowledgeable, however, about the larger literary landscape to which Lefèvre’s Histoire de Jason belonged, and shows this in his prologue and epilogue to the English translation. First of all, Caxton knew of the relation between Lefèvre’s Histoire de Jason and his other romance, the Receuil, which, as noted above, Caxton had already translated. Lefèvre had only briefly touched on Jason and the Golden Fleece in his Receuil, and the reason for this, Caxton explains in the prologue, is that Lefèvre had by this stage already written his Histoire de Jason: as to the historie of lason, towchyng the conqueste of the Golden Flese, myn auctor hath not sett in his boke but breuely [‘briefly’] and the cause is for asmoche as he hadde made before a boke of the hoole lyf of lason (1).

How Caxton knew that Lefèvre wrote Jason first and Receuil later is uncertain, but he appears to have been correct. Pinkernell (1971) figured out the chronology on the basis of internal evidence, and perhaps Caxton had done the same. Caxton also adduces other authorities for the history of Jason. From the world of books, he knew Dares Phrygius and of Guido delle Colonne, for he mentions in the prologue the historie of him whiche that Dares Frigius and Guido de Columpnys wrote in the begynnyng of their bokes (1). All this, including the detail that the Jason material can be found in the begynnyng of their bokes, is also correct: The story of Jason and Medea is told in the first three of the thirty-five books of Guido’s Historia Destructionis Troiae (Griffin 1936) and in the first two chapters of Dares Phrygius’ De Excidio Trojae Historia (Cornil 2012). More remarkable is Caxton’s knowledge of Boccaccio’s Genealogia Deorum Gentilium, which he mentions at the end of his translation. Lefèvre ended his story with the reconciliation of Jason and Medea, but Caxton felt obliged for the sake of completeness to supplement his source with a summary of relevant facts concerning Jason gleaned from Boccaccio, including Jason’s conquests in Asia, the fortunes of his two children, Thoas and Euneus, and so on:

8  Caxton’s Linguistic and Literary Multilingualism 

221

howe be it that myn auctor writeth that he hath founde nomore of th’istorie of lason, yet haue I founden & red in the boke that Bochace made of the Genelagie of Goddes in his .xiij. boke, that whan so was that lason & Medea were reconciled agayn togeder after that shee fled from Egeon that he […] after went into Asie, where he had victorie in many batailes and made so many conquestes with grete magnificence […] But what cam afterward of these two sones it is incertayn, this saith Bochace in the xiij. boke of the geneolagye of goddes (198–199).

The reference is exact enough to allow it to be easily checked, and again it is correct: Caxton got his information from Genealogia, Book XIII, chapters 26–27, which finish with Boccaccio’s statement that he does not know what became of Thoas and Euneus: Quid autem ex eis postea actum sit mihi incertum est (Romano 1951: 656). If Caxton was a merchant, not a scholar, you would not know it from these procedures: he checked Lefèvre’s history against other sources, notes where in these sources we need to look, summarises supplementary material based on Boccaccio, and judiciously ends with Boccaccio at the point the evidence runs out. What Caxton did not know from books he supplied from personal experience. As pictorial evidence of the story he mentions the paintings of Jason and Medea that Philip the Good had installed in a room of his castle at Hesdin (Campbell 2007), which he had personally seen (in whiche chambre I haue ben and seen the sayde historie so depeynted [‘depicted’]; 2) and he was obviously aware of the controversy about whether the Biblical Gideon or Jason should be regarded as the patron of the Order of the Golden Fleece. The issue with Jason was not that he was pagan but rather that he was thought to have betrayed Medea, and it was precisely to rehabilitate Jason’s reputation that Lefèvre wrote his Histoire de Jason. The debate occupied a number of contemporary writers (Bakelaar 1981), but Caxton apparently knew about it from hearsay, and refused to be drawn on it: somme persones afferme and saye that the sayd ordre hath taken his orygynal of the Flese of Gedeon, where in I will not dispute (2).

222 

Ad. Putter

3 The Influence of French and Dutch on the Language of Jason Caxton, then, was multilingual in the literature he read and produced and in the languages he juggled in his working life in Flanders. This multilingualism of life and letters sheds more light on his English than does the theory that he tried to standardise it. The greatest influence on the language of his Jason was exerted by the book he was translating. Caxton generally translated very closely. His aim, he says in the prologue, was to follow his auctor as nygh [‘closely’] as I can or may, not chaungyng the sentence, ne presumyng to adde ne mynusshe [‘reduce’] ony thing otherwyse than myne auctor hath made in Frensshe (1). Many of the lexical and semantic innovations he introduced were taken over from his source. As far as loanwords are concerned, the Oxford English Dictionary (hereafter OED) records 124 words that are first attested in Caxton’s Jason, almost all of them of French origin. A further 156 words, again mostly French, are recorded in OED with senses first found in Jason. An example of the kind of prose resulting from Caxton’s literal translation are Medea’s words to Jason, wherfor I pray you that ye wil haue regarde to myn offre abandonned (122/36), based on Lefèvre’s Si vous prie que vous n’ayés regard a mon offre habandonnee (197).2 OED gives abandonned as the first attestation of the adjective in the sense ‘forsaken, deserted’ (s.v. abandoned, adj., sense 2), but Caxton also follows Lefèvre by using it in the novel sense ‘wanton, unrestrained’ (Pinkernell 1971: 115). It thus belongs in OED under 1b, ‘chiefly attributive (without to). Uninhibited, unconstrained’, for which sense OED gives 1690 as the first attestation. Another example of direct influence from the source is and the damoiseau lason ... began thenne to foundre in teeris right habondantly (5), based on Et le damoisel Jason...commenca lors a fondre en larmes mout habondamment (fol. 3v). The earliest  Quotations from Lefèvre’s Histoire de Jason are from Pinkernell (1971). Because this edition is based on the autograph manuscript and Caxton used a manuscript version that contained elaborations not found in the autograph, I sometimes cite for purposes of comparison the text from Paris, Bibliothèque Nationale, MS fr. 12148, which contains many of these same elaborations. Quotations from the manuscript are referenced by folio number and are based on my own transcription of the digitised manuscript (with addition of accent égu), available at https://gallica.bnf.fr/ark:/12148/ btv1b90611849/f13.item.r=Histoire%20de%20Jason%20Lefrevre# [accessed 23 May 2023]. 2

8  Caxton’s Linguistic and Literary Multilingualism 

223

attestation of damoiseau and the earliest (of only two attestations) of founder in the sense of ‘to break down (in tears)’ come from this sentence in Caxton, and they clearly entered his language via his source. The number of loanwords and semantic loans resulting from Caxton’s literal translations is certainly underreported by OED. For instance, style in the sense ‘profession, métier’ is not in OED (s.v. style, n.). This sense was normal in Middle French (see Dictionnaire du Moyen Français, DMF, s.v. stile, sense B.2), but we find it also in Caxton’s Jason—he semed aboue all other to be right well vsid in the faites [‘feats’] of Armes, & that in alle his lyff he had don [non] other style but to iouste (7/30)—because he lifted it from his source (autre stille fors de iouster, fol. 6r). The use of the word devise in the obsolete sense of ‘conversation, discussion’ is in OED (s.v. device, n., sense 5), but with Caxton’s Blanchardyn & Eglantine (1490) as the first attestation. But here it is in Jason: The ladyes helde their deuises apart (8/7); cf. tindrent leur deuises apart (fol. 6r). Peisaunt, meaning ‘heavy, massive’ (31/3), and the related noun peysanteur (30/2) also predate the attestations in OED. The kind of English resulting from this close translation is peculiar, to say the least. It should have discouraged the assumption that ‘modern English dates from Caxton’, but it seems only to have confirmed the prejudice that he was driven by mercantile pragmatism. Translating a foreign language into idiomatic English is time-consuming, and it has been argued that Caxton, ever the businessman, took over words from his source for the sake of expedience. ‘His method’, according to Blake, ‘was to finish his translation as soon as possible […]. Since he was translating to keep his own presses in work, it is only natural to assume that there were many occasions when financial gain took precedence over literary quality’ (Blake 1969: 126). Expedience alone, however, will not explain the borrowings in Caxton’s prose. One complicating factor is that Caxton, like many of his contemporaries, regarded French as beautiful (Ailes and Putter 2014). In his prologue to Eneydos (c.1490), he praised the French diction of his source: the fayr and honest [‘worthy’] termes and wordes in Frenshe (Blake 1973: 79). Taking over French lexis was therefore not a matter of putting ‘financial gain’ before ‘literary quality’, but a way of enhancing that quality. The use of binomials with Gallicisms in the prose of Caxton and other writers

224 

Ad. Putter

of the period confirms that speed or convenience was not always what motivated Caxton. The context for this conspicuous feature in Caxton (Kubaschewski 2017) is the aureate prose style that Burnley (1986) called ‘curial prose’. Since Caxton lived and worked in Flanders, the internationalism of this mannerism is worth noting. Caxton would certainly have come across it in Flemish officialese, for Burgundian administrators embraced doublets—formeerde ende maecte (‘formed and made’), bereyt ende diligent (‘committed and diligent’), and so on—as a badge of honour, an ‘attractive means of staging professional identity’ (Schoenaers 2017: 21–22). In Caxton’s Jason, as in Schoenaers’s examples from Burgundian prose, the binomials often coordinate a French borrowing with a native Germanic word or a naturalised French loanword with a new one. Examples from Caxton’s Jason are with grete loenge and preysynges (34), for to enuahye [‘battle’] and fighte (36), their gayn & butin [‘booty’] (41), with litil noise and bruyt [‘sound’] (181). With one exception (cf. le gaaing et le butin, fol. 40v), the doublets cited above are not in the source but expand on a single term in the source (loange, 145; envahir, 146; a pou de bruit, 230). The theory that Caxton’s imported French words because he translated at speed misses the point that he often took the trouble of both adopting and translating them. A more fundamental complication is that the languages that influenced Caxton’s English were not just buried in the books he was translating but active in his mind. French and Dutch were the foreign languages he needed on a daily basis in Flanders, and it is reasonable to suppose that, after decades of using them, they were hardwired in his brain. That is why the prologue and epilogue to Jason show some of the same kinds of linguistic interference from French as does the rest of the work. Thus, the anomalous spellings that occur in Caxton’s translated text—for example, dishonneur (3/15–16), corresponding with deshonneur in Lefèvre (125); radeur (39/17), corresponding with radeur ‘force’ (148)3—also intrude in the prologue and epilogue where Caxton had no French source to contend with: here he writes foundeur (2/4) and serviteur (199/34). Neither spelling is recorded in OED or Middle English Dictionary  Caxton normally spells these words with the suffix -our: cf. dishonour (48/24, 48/31), radour (152/18). 3

8  Caxton’s Linguistic and Literary Multilingualism 

225

(hereafter MED), and neither is explicable as a result of hasty translation. Caxton spelled these words as one might expect them to be spelled in his other works,4 but French spellings occur in the books from his Flemish phase (Wiencke 1930: 1951–1954). His lexis in the prologue is also overshadowed by French. For instance, the formula by licence & congye & by the supportacion of our most redoubted liege lady, is French-­ dominated. It is by thinking of French that we can make sense of aparte of in this sayd boke is late [‘recently’] newe made, aparte of alle th’istories of the sayd Jason & the historie of him whiche that Dares Frigius & Guido de Columpnys wrote. The prepositional phrase is puzzling. It does not match anything I have found in English but corresponds with French à part de (‘separate from’; DMF, s.v. part, sense 5). Perhaps the issue here is not lazy translation but interference between languages that Caxton had fully internalised. When he was translating Jason, his source-language was French and his target-language English, and so Dutch should not come into it, but in fact it sometimes interfered with his English. First, there are some examples of spellings that appear to be influenced by Dutch. In Caxton Dutch-­ derived Reynard (Blake 1970), such spellings—for example, for as in moed (7/33); for as in hier (7/35); for as in vlycche (9/7)—are more obtrusive. Blake dismissed them as ‘superficial lapses’ (1963–1964: 532) made in haste: ‘Caxton knew how such words were normally spelt in English but in his haste took over the Dutch spellings without thinking’ (Blake 1970: liv). However, this does not explain why they occur sporadically not only in Recuyell (Putter 2021) but also in Jason, which, for instance, has hier (44/3) as well as the usual here; proeue (20/9) as well as the usual proue; vliese (86/10, 89/13, 108/36) as well as the usual flees / fleece; and chiese (44/34) as well as chese ‘to choose’. Such spellings also occur in earlier Middle English writings from Kent, where Caxton was born (Jordan 1974: 116), but since they are anomalous in fifteenth-century English it is relevant to note that these were the normal spellings in Middle Dutch.  Cf. founder (Eneydos 71/9, Culley and Furnivall 1890; and Book of Curtesye, l. 330, Furnivall 1867) and seruitour (Charles the Grete, 168/35, Herrtage 1881; and The Canterbury Tales, 167/3, available at https://www.bl.uk/treasures/caxton/homepage.html [accessed 23 May 2023]). 4

226 

Ad. Putter

There are also examples of semantic loans or ‘loanshifts’ from Dutch, that is, words which take on novel senses based on the semantic range of the equivalent word in Dutch. Such semantic contamination is a well-­ known feature of the language of bilingual speakers. For instance, humoroso means ‘capricious’ in Portuguese, but in American-Portuguese it acquired the meaning ‘humorous’ based on its English equivalent (Winford 2003: 43–44). It is therefore interesting to find various examples of Dutch semantic loans even when Caxton was working on a French–English translation. ME forthon meant ‘for this reason’, or ‘forthwith’, but in Caxton’s English it also means ‘from this time on’.5 This sense was influenced by Dutch voortaan, as is shown by Caxton’s Reynard, where Eme see now forthon that ye doo good werkis (28/6) goes back to Oem seit dat ghi nv voert an goede werken doet (Hystorie van Reynaert, 105/1230–1231). However, Caxton clearly did not need a Dutch source to produce this loan because he also uses it in the Recuyell (23/18) and in Jason: fro thenne forthon he named him his broder (7/10). Caxton’s use of how well as a concessive conjunction (‘although’) is also Dutch-inspired, as noted by OED (s.v. how, adv., sense IV.14.a). In Reynard, it can be traced back to the Dutch of his source (and how well that he had supposed [ende hoe wel hi meende, 151/1776] that he had made al faste [‘everything secure’], I was not so moche a fool, 37/ 26)—but when it occurs in Jason (howe well somme persones afferme and saye, 2/1) we must trace it back to the Dutch in Caxton’s head. The verb lap as Caxton uses it in Jason offers another example of a semantic loan. In yclad with clothes all to rente and lapped (188/8; [vestiments deschirés, 235]), the verb is not used in any recorded sense of ME lappen ‘to enfold’, but rather in the sense ‘to patch’, which is what the word means in Dutch.6 Some idioms in Jason also point to interference from Dutch. Where the French reads au point du jour (182), Caxton has in the crekyng of the daye (98/3). Munro emended to brekyng, but the authenticity of crekyng is confirmed by the same expression in Caxton’s Recuyell, where the first krekyng of the day (37/1) again translates au point du jour (Aeschbach 1987: 44). Caxton met the challenge of translating a French idiom by  OED (s.v. forthon, adv.) gives Caxton’s Game and Playe of the Chesse as the first attestation.  Neither MED (s.v. lappen) nor OED (s.v. lap, v.2) records this sense.

5 6

8  Caxton’s Linguistic and Literary Multilingualism 

227

adopting a Dutch one: het krieken van de dag is the Dutch equivalent of ‘the crack of dawn’ (Middelnederlandssch Woordenboek, MNW, s.v. crieken).7 OED only records the modern noun creaking, with the earliest attestation dated 1520. Dutch influence is probably also responsible for over dead: And yf thou be not ouer dede, that they wyll hold the so long alyue that I might yet ones [‘once’] speke with the (185/38). I have not come across this phrase in English but have often seen it in Middle Dutch, where it means ‘as good as dead’ (see MNW, s.v. doot). Because Dutch and English are such closely related languages, it is not always possible to know for sure whether we are dealing with a native formation or a borrowing. An interesting case in point is the word drunkard. OED (s.v. drunkard, n.) explains this as a native formation and gives 1530 as the first attestation: ‘J. Palsgrave Lesclarcissement 155 Yuroygne, a man droncarde; yuresse, a woman droncarde’. As the source is Palgrave’s French–English dictionary, it is worth pointing out that droncart existed in Middle French, where it had been borrowed from Dutch late in the fifteenth century (see DMF, s.v. dronquard). Although MNW omits it, dronckaert existed in Dutch by at least the middle of the fifteenth century,8 and in English it is actually in Caxton’s Jason that the word first appears: And folowed after the dronkardes [yvroignes, 130] that went relyng on alle sydes (10/5). Caxton may have encountered the word in the Dutch of his surroundings. In a couple of cases in Jason we could be dealing with native words or senses that Caxton reintroduced into English via Dutch. The verb overthink in I haue wel ouerthought these saide thinges (17/26) is a case in point. Translating French pensé (fol. 16r), the verb clearly means ‘to consider’ rather than ‘to regret’ (the normal Middle English sense: MED, s.v. overthinken) or ‘to think too much’ (the modern sense: OED, s.v. overthink, v.1). Caxton uses the verb as it is attested once in Old English (in  Historical Dutch dictionaries can be searched at https://gtb.ivdnt.org/search/ [accessed 23 May 2023]. Users should be warned that the historical Dutch dictionaries are patchy compared with OED and MED. 8  I have found it in Utrecht, University Library, MS Hattem C5 (s. xv med.), available at https:// www.dbnl.org/tekst/_hat001hatt03_01/_hat001hatt03_01_0003.php [accessed 23 May 2023], and in the Statutes of the Hospital of Bunderen (1473), transcribed at https://www.tenbunderen. be/transcriptie2.html [accessed 23 May 2023]. 7

228 

Ad. Putter

Bishop Waerferth's late ninth-century Old English translation of Gregory’s Dialogues). However, since Caxton appears to have been the only person after the ninth century to do so,9 overthought may well be calqued on the common Middle Dutch verb overdenken, which also meant ‘to consider’. The adjective tocoming (‘future’) presents a similar case, except that there is further circumstantial evidence that it was indeed reintroduced via Dutch. Caxton uses it in the preface with reference to the crown prince: my lord Prynce of Wales our tocomynge souerayne lorde (2/22). Again, the only precedent for tocoming as an attributive adjective comes from Old English (OED, s.v. tocoming, adj. and n.2), but in Middle Dutch toecomende was commonly used as an attributive adjective. Two slightly later usages in English confirm Dutch influence. One of these, not in OED, comes from the quadrilingual Book of Privileges I mentioned earlier. In the charter granted to the English merchants by the Lord of Bergen op Zoom, the Dutch phrase in toecomende tijden was translated as in tocomyng tyme (Sutton and Visser-Fuchs 2011: 263).10 The other attestation, which is in OED—‘?1575 tr. H. Niclaes Epistolæ ix. i. 191 “The to-coming Daye of the Reuealing of the Coming of Christ”’— comes from the printed prose of a Dutchman, Christopher Vittels, who translated the Dutch / Low German works of Hendrik Niclaes into English (Hamilton 2014). Finally, there are some grammatical peculiarities in the books from Caxton’s Flemish period that become comprehensible if we think of him as a trilingual English / French / Dutch speaker. Some of the possible grammatical symptoms of interference from Dutch in Jason—the frequent use of men as the indefinite pronoun (54/15, 57/10, 58/11, etc.),11 the archaic use of a modal infinitive followed by another infinitive, as in to mowe come to your goode grace (17/34) and to so wille lose your lyf (21/4)—also occur in Caxton’s Recuyell, and Reynard and the influence of  OED gives, apart from a later Caxton example, one further attestation from 1903, but the sense here just seems to be the modern one: ‘think too much about’. 10  The editors misconstrue the Middle English as into comyng tyme. 11  The indefinite pronoun men was gradually ousted (though not yet entirely replaced) by they in the fifteenth century, but it remains common in Caxton, who was apparently also the last writer to use me as the indefinite pronoun, probably under influence of Dutch me(n) (Mustanoja 1960: 219–222; Putter 2021: 217). 9

8  Caxton’s Linguistic and Literary Multilingualism 

229

Dutch on these usages has already been discussed elsewhere (De Reul 1901: ix, fn. 2; Blake 1991: 246; Putter 2021: 217 and 222–223). However, there are other cases in Jason that deserve our attention. Consider Caxton’s use of bleef as the past tense of Middle English bileven (‘to leave, remain’). This is a weak verb in Middle English and can be found as such in Caxton (e.g. Tantalus beleuyed [‘left’] not lightly his wache [‘vigilance’], Recuyell 134/29), yet surprisingly he also treats it as a strong verb in Jason (pret. sg. bleef 23/8, 12/18), and earlier in Recuyell (pret. pl. bleuen, 327/5). OED (s.v. belive, v.1) solves this enigma by putting these under a different verb, belive, from OE belīfan, which was strong; however, since that verb appears to have died out in Early Middle English, it seems more likely that Caxton’s usage was influenced by Middle Dutch, where the past tense of the equivalent verb (bliven) was strong (singular bleef, plural bleven). Dutch influence is confirmed by Caxton’s Reynard 16/15, ther ne bleef nether man ne wyf, translating daer en bleef wijf noch man (555; Hellinga 1952). A final example of possible grammatical interference from Dutch is Caxton’s use of fulcome as a transitive verb. The verb, rare in Middle English and beyond, was intransitive, as it had been in Old English, and meant ‘to arrive at a particular state, to come to full knowledge of something’. According to OED (s.v. fullcome, sense 2), only Caxton uses it transitively, in the sense ‘to achieve, accomplish’, once in The Dialogues in French and English (God late them theyr waye [‘journey’] well fulcome!), translated from a French-Flemish original, and once in Jason, for to fulcome [en parfaisant, fol. 14r] his emprise (20/30). In both instances, the transitive use of fulcome seems to accord less with English than with Dutch usage. Cf. Si volquamen hare saken (‘they accomplished their business’; cited by MNW, s.v. volcomen).

4 Conclusion and Implications for Future Research As I hope to have shown, the language of Caxton’s History of Jason shows signs of the kinds of linguistic interference that might be expected from someone who moves abroad and spends years in a country where

230 

Ad. Putter

different languages are spoken. In the case of Caxton, who worked in the Low Countries for some thirty years, the languages of everyday use in Flanders were Middle Dutch and Middle French. The exotic elements in his language are usually explained as the result of hasty translation from his sources. This certainly explains many of the French loans in his translations from French and many of the Dutch loans in his translations from Dutch, but it fails to explain why we find Dutch influence also in his translation of Jason and why we find French influence also in the prologue and epilogue. The explanation is that Caxton was profoundly trilingual and that the three vernaculars that were hardwired in his brain interfered with each other. Although the predominant view of Caxton is that he published to make money, Jason shows a different side: he was a scholar and a polyglot who went beyond his French source to include other information that he had read, seen, and heard. It is therefore unfortunate that his mercenary reputation has even come to define his place in the history of the English language: As a merchant he wanted to sell books which last a long time, linguistically speaking, and which can be read by as large a reading public as possible. In other words, he was seeking a relatively stable language variety that could serve a supraregional function to speakers of different dialects […] by the end of the fifteenth century economic motivations contributed significantly to earlier linguistic and political ones in the standardisation process of the language (Nevalainen and van Ostade 2006: 278).

The point that needs to be emphasised in the face of such textbook accounts is just how unstable the language of his earliest printed books was. One source of that instability is that Caxton’s compositors introduced variant spellings (Shute 2017); a second one is that Caxton’s own language changed over time (Wiencke 1930); but another important source was interference from other languages that Caxton carried in his mind. The influence of French in Jason is not surprising, since it was the language of his source, but Jason also shows the influence of Dutch. Editors of Caxton’s books will thus at times need to triangulate three languages, Middle English, Middle French, and Middle Dutch, as Caxton

8  Caxton’s Linguistic and Literary Multilingualism 

231

himself did. As we have seen, Munro was unprepared for the Dutch substrate in Caxton’s English and edited out the authentic Dutch idiom the crekyng of the daye (98/3). That was a long time ago, but modern editors of Caxton seem no better prepared for it. Judging by her punctuation and lack of glossing, Adams in her edition of The Game and Playe of the Chesse did not realise that how wel means ‘although’ rather than ‘how well’ when she printed: How wel that the lyon be the strengest beest. Yet somtyme a lityl byrde eteth hym (Adams 2009: 20). Here as elsewhere (cf. Adams 2009: 29), a subclause introduced by a Dutch-derived subordinating conjunction is presented as if it were as an independent clause introduced by an adverb. Moll (2013: 377) in his edition of Caxton’s Ovide Moralisé edits out the Dutch loanword bedwingen (‘to restrain’) in I oughte well thenne bedwynge [‘control’] myn herte, by replacing it with bend, despite the fact that the French source—Bien en doy mon cuer fleschir—militates against the emendation (fleschir here means ‘subdue’, humble’, not ‘bend’), and despite the fact that bedwynge also occurs in Caxton’s Dutch-derived Reynard, where Blake (1970: xxvi) recognised it for the Dutch borrowing it is. Historical lexicographers also need to be more alert to Caxton’s multilingualism. Caxton’s French borrowings in Jason have generally been well dealt with in OED and MED, though some earliest attestations in the text have not been spotted. But the Dutch influence on Caxton’s Jason has not been captured at all. The only lexical innovation in Jason that OED connects with Dutch is the word loop (They of Oliferne...ran vnto the bateillement & lowpes of the walles [aux creniaux, 135], 19/3), which OED gives as the first attestation of loop, n.2, sense 2 (‘an opening in the parapet of a fortification; an embrasure’) and relates to Middle Dutch lupen ‘to lie in wait, watch, peer’. The connection with Dutch seems much more tenuous in this case than in the ones discussed above. A revaluation of Caxton’s language and vocabulary that takes the full measure of his multilingualism is badly needed.12

 The research for this essay was funded by the Leverhulme Trust (RPG-2018-124), whose support is gratefully acknowledged. 12

232 

Ad. Putter

References Adams, Jenny, ed. 2009. The Game and the Playe of the Chesse, TEAMS. Kalamazoo, MI: Medieval Institute. Aeschbach, Marc. 1987. Raoul Lefèvre, Le Recoeil des Histoires de Troyes. Frankfurt: Peter Lang. Ailes, Marianne, and Ad Putter. 2014. French in Medieval England. In European Francophonie: The Social, Political and Cultural History of an International Prestige Language, ed. Vladislav Rjéoutski, Gesine Argent, and Derek Offord, 51–80. Bern: Peter Lang. Bakelaar, Bette Lou. 1981. Patron of the Golden Fleece — Jason or Gideon? Fifteenth Century Studies 4: 25–35. Bakker, F.J., and J. Gerritsen. 2004. Collecting Ships from Holland and Zeeland: A Caxton Letter Discovered. The Library 5: 3–11. Blake, Norman. 1969a. Caxton and His World. London: Deutsch. ———. 1969b. Some Low Dutch Loan-Words in Fifteenth-Century English. Notes and Queries 16: 251–253. ———, ed. 1970. The History of Reynard the Fox, Early English Text Society, o.s. 263. London: Oxford University Press. ———, ed. 1973. Caxton’s Own Prose. London: Deutsch. ———. 1991. William Caxton and English Literary Culture. London: Hambledon. Bradley, Henry. 1900. Dialogues in French and English, Early English Text Society, e.s. 79. London: Oxford University Press. Burnley, J.D. 1986. Curial Prose in England. Speculum 61: 593–614. Campbell, Caroline. 2007. Lorenzo Tornabuoni’s History of Jason and Medea Series. Renaissance Studies 21: 1–19. Colin, G., and W. Hellinga. 1973. De vijfhonderdste verjaring van de boekdrukkunst in de Nederlanden. Brussels: Koninklijke Bibliotheek. Cornil, Jonathan, ed. 2012. Dares Phrygius’ De Excidio Trojae Historia: Philological Commentary and Translation. Unpublished master's thesis, University of Ghent. https://libstore.ugent.be/fulltxt/RUG01/001/891/500/ RUG01-001891500_2012_0001_AC.pdf. Accessed 23 May 2023. Corsten, Severin. 1975–1976. Caxton in Cologne. Journal of the Printing Historical Society 11: 1–18. Culley, W.T, and F. J. Furnivall, eds. 1890. Caxton’s Eneydos, Early English Text Society, e.s. 57. London: Oxford University Press. Culpeper, Jonathan. 2015. History of English, 3rd edn. London: Routledge.

8  Caxton’s Linguistic and Literary Multilingualism 

233

[DMF =] Dictionnaire du Moyen Français (1330-1500), version 2015 (DMF 2015) (ATILF -CNRS & Université de Lorraine). htttp://www.atilf.fr/dmf/. Accessed 25 May 2023. De Reul, Paul. 1901. The Language of Caxton’s Reynard the Fox: A Study in Historical English Syntax. Ghent: Vuylsteke. Driver, Martha. 1997. Caxton, William. In The Encyclopedia of Medieval Literature in Britain, ed. Siân Echard, Robert Allen Rouse, Jaqueline A. Fay, Helen Fulton, and Geoff Rector, IV: 390–399. Oxford: Blackwell. Finnegan, Sister Mary Jeremy. 1952. Caxton’s Life of St. Rocke. Modern Language Notes 67: 313–317. Furnivall, F. J., ed. 1867. Caxton’s Book of Curtesye, Early English Text Society, e.s. 3 (London: Trübner). Goodman, Jennifer. 2006. Caxton’s Continent. In Caxton’s Trace: Studies in the History of English Printing, ed. William Kuskin, 101–123. Notre Dame, IN: Notre Dame University Press. Griffin, Nathaniel Edward, ed. 1936. Guido de Columnis, Historia Destructionis Troiae. Cambridge, MA: Medieval Academy. Hamilton, Alastair. 2014. Niclaes, Hendrik, Founder of the Family of Love. In Oxford Dictionary of National Biography. https://oxforddnb.com. Accessed 23 May 2023. Hellinga, Lotte. 2010. William Caxton and Early Printing in England. London: British Library. ———. 2014. Texts in Transit: Manuscript to Proof and Print in the Fifteenth Century. Leiden: Brill. Hellinga, Lotte, and Wytze G.  Hellinga. 1976–1977. Caxton in the Low Countries. Journal of the Printing Historical Society 11: 19–32. Hellinga, Wyzte G., ed. 1952. Die hystorie van Reynaert die vos. Zwolle: Tjeenk Willink. Herrtage, Sidney J. H., ed. 1881. Charles the Grete, Early English Text Society, e.s. 38. London: Trübner. Hsy, Jonathan. 2013. Trading Tongues: Merchants, Multilingualism, and Medieval Literature. Columbus, OH: Ohio State University Press. Jordan, Richard. 1974. Handbook of Middle English Grammar: Phonology, Trans. Eugene J. Crook. The Hague: Mouton. Kekewich, Margaret. 1971. Edward IV, William Caxton, and Literary Patronage in Yorkist England. The Modern Language Review 66: 481–487. Kowaleski, Maryanne. 2017. French Immigrants and the French Language. In The French of Medieval England: Essays in Honour of Jocelyn Wogan-Browne,

234 

Ad. Putter

ed. Thelma Fenster and Carolyn P.  Colette, 206–224. Cambridge: D.S. Brewer. Kubaschewski, Elisabeth. 2017. Binomials in Caxton’s Ovid (Book I). In Binomials in the History of English, ed. Joanna Kopaczyk and Hans Sauer, 141–158. Cambridge: Cambridge University Press. [MED =] Middle English Dictionary. 1952–2001. Ed. by Robert E.  Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner (Ann Arbor: University of Michigan Press). https://quod.lib.umich.edu/m/med/. Accessed 23 May 2023. [MNW =] Middelnederlandsch Woordenboek. 1885–1952. Ed. by E. Verwijs and J.  Verdam. 11 vols. The Hague: Martinus Nijhoff. htttps://gtb.ivdnt.org/. Accessed 23 May 2023. Moll, Richard. 2013. The Booke of Ovyde Named Methamorphose. Toronto: Pontifical Institute of Mediaeval Studies. Munro, John, ed. 1913. Caxton’s History of Jason, Early English Text Society, e.s. 111. London: Oxford University Press. Murray, James, and Murray, Hilda. 1910–1911. English Language. In Encyclopaedia Britannica, 11th edn, 29 vols, IX: 587–600. Cambridge: Cambridge University Press. Mustanoja, Tauno F. 1960. A Middle English Syntax, Part I: Parts of Speech, Mémoires de la Société Néophilologique de Helsinki, 23. Helsinki: Société Néophilologique. Nevalainen, Terttu, and Ingrid Tieken-Boon van Ostade. 2006. Standardisation. In A History of the English Language, ed. Richard Hogg and David Denison, 271–311. Cambridge: Cambridge University Press. [OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 23 May 2023. Pinkernell, Gert. 1971. L’Histoire de Jason: Ein Roman aus dem 15. Jahrhundert. Frankfurt: Peter Lang. Putter, Ad. 2021. Dutch, French and English in Caxton’s Recuyell of the Historyes of Troye. In Medieval Romance, Arthurian Literature: Essays in Honour of Elizabeth Archibald, ed. A.S.G. Edwards, 205–226. Cambridge: D.S. Brewer. Romano, Vincenzo, ed. 1951. Genealogie deorum gentilium. 2 vols. Bari: Laterza.

8  Caxton’s Linguistic and Literary Multilingualism 

235

Schoenaers, Dirk. 2017. ‘French: Middle Frenchified’: A Contact-Based Approach to Transculturation and Linguistic Change in Holland-Zeeland (1428/33-c.1500). In The Multilingual Muse: Transcultural Poetics in the Burgundian Netherlands, ed. Adrian Armstrong and Elsa Strietman, 12–41. Cambridge: Legenda. Shute, Rosie. 2017. A Quantitative Study of Spelling Variation in William Caxton’s Printed Texts. Unpublished doctoral thesis, University of Sheffield. Smeyers, Maurits. 1999. Flemish Miniatures: The Medieval World on Parchment. Leuven: Davidsfonds. Sommer, H.  Oskar, ed. 1894. The Recuyell of the Historyes of Troye.  2 vols. London: Nutt. Spufford, Peter. 1993. The Burgundian Netherlands. In Splendours of Flanders, ed. Alain Arnould and Jean Michel Massing, 1–11. Cambridge: Cambridge University Press. Sutton, Anne F., and Livia Visser-Fuchs, eds. 2011. The Book of Privileges of the Merchant Adventurers of England, 1296-1483. Oxford: Oxford University Press. Wiencke, Helmut. 1930. Die Sprache Caxtons. Leipzig: Tauchnitz. Winford, Donald. 2003. An Introduction to Contact Linguistics. Oxford: Blackwell. Wright, Laura. 2020a. A Critical Look at Previous Accounts of the Standardisation of English. In The Multilingual Origins of Standard English, ed. Laura Wright, 17–38. Berlin: De Gruyter Mouton. ———, ed. 2020b. The Multilingual Origins of Standard English, Topics in English Linguistics, 107. Berlin: De Gruyter.

Part III Medieval Multilingualism and Morphosyntactic Change

9 An Overview of Contact-Induced Morphosyntactic Changes in Early English George Walkden, Juhani Klemola, and Thomas Rainsford

1 Introduction This chapter aims to provide an overview of contact-induced morphosyntactic change in English before 1500. ‘Morphosyntactic’ is here understood in the broadest possible sense, encompassing both morphological and syntactic change as well as change in the inventory of functional words (e.g. pronouns, copulas and auxiliaries). G. Walkden (*) University of Konstanz, Konstanz, Germany e-mail: [email protected] J. Klemola Tampere University, Tampere, Finland e-mail: [email protected] T. Rainsford University of Stuttgart, Stuttgart, Germany e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_9

239

240 

G. Walkden et al.

The identification of morphosyntactic changes in the historical record brings with it challenges that are unlike those faced in the study of lexical borrowing. In particular, the regularity of sound change provides a relatively robust formal diagnostic for lexical transfer in many cases (see Dance et al. and Fox et al., this volume). For grammatical influence, we have no such diagnostic. To be sure, in some domains the formal similarities between two languages may be so strikingly idiosyncratic and robust as to be crying out for a contact explanation. But how can we tell whether (for instance) an increase in use of verb-object (VO) constituent order is due to contact with another language with VO order, or an independent parallel development? VO order is common, and the change from object-­ verb (OV) to VO is also attested in other languages that are outside obvious contact situations—yet this does not allow us to rule out contact influence.1 Historical linguists have grappled with this problem for some time (see e.g. Bowern 2008; Erschler 2009; Heine 2009). Very often, the case for a particular change being contact-induced is by necessity circumstantial. Returning to our example of OV to VO: if it can be established that contact influence from the VO language was most dominant in a particular geographical area, and texts from that area show an earlier and faster shift from OV to VO, then a plausible (though not watertight) case can be made that the change was at least partly contact-induced. In making arguments of this kind, extralinguistic evidence, such as archaeological findings or documentary evidence, often plays an important role. In arguing for or against contact-induced grammatical change, it is thus crucial to have a clear picture of who was where, when, and what they were doing—bearing in mind McIntosh’s (1994: 137) dictum that ‘what we mean by “languages in contact” is “users of language in contact” and to insist upon this is much more than a mere terminological quibble and has far from trivial consequences’. Moreover, contact linguistics over the last half-century has taken steps towards understanding the mechanisms and pathways of contact-induced change. Focusing in  That is, unless one subscribes to the methodological principle of always preferring ‘endogenous’ explanations whenever they are available (see Lass 1997). For discussion of this ‘if-in-doubt-dowithout’ mentality, see Farrar and Jones (2002). 1

9  Overview of Contact-Induced Morphosyntactic Changes 

241

particular on transfer of material from one language to another, van Coetsem (1988, 2000) has drawn a distinction between borrowing and imposition, which will be elaborated on in Sect. 1.1. And with an eye to the overall consequences of contact on a language’s typological profile, Trudgill (2011) has made the case that different socio- and psycholinguistic circumstances give rise to different types of change, either simplifying or complexifying the languages involved (see Sect. 1.2) . The rest of Sect. 1 of this chapter is devoted to introducing these ideas, which will then be used as a prism through which to evaluate and compare the different contact situations that English entered into during late antiquity and the medieval period. Here our focus is on three different languages (or clusters of closely related varieties), in approximate chronological order: Celtic in Sect. 2, Old Norse in Sect. 3 and Old French in Sect. 4. In a single chapter we cannot hope to deal with these contact situations comprehensively or catalogue all possible contact-induced morphosyntactic changes. Rather, we zoom in on a selection of changes that help us to characterise the nature of each of these contact situations. Over the last two centuries, attitudes to each of these contact situations have followed different paths. It has never been in serious doubt that Old Norse was a major influence on the structure of English, though the nature of that influence has been debated. There is a substantial tradition of research on Old French influence too, though for the most part such research has focused on the lexicon and on derivational morphology; in recent years, the question of syntactic influence has attracted renewed attention. By contrast, for many decades the question of Celtic influence on English languished in obscurity, only being broached by isolated voices (see van der Auwera and Genee 2002). Since the 1990s this book has been reopened, however, and today the topic of Celtic influence is the subject of lively discussion. We leave Latin influence out of consideration in this chapter, primarily because the nature of the contact situation was substantially different: alone among the languages that have had a major influence on English, by the time of the arrival of speakers of Old English in the fifth century, Latin was already a variety with no or very few native speakers, and probably less than 1% of speakers of Old English had any knowledge of Latin

242 

G. Walkden et al.

at all (Timofeeva 2010a).2 This is arguably not a scenario in which we would expect to find significant morphosyntactic influence (see in particular Sect. 1.1), and indeed such influence has not usually been argued for. To be sure, we see imitation of Latin syntactic constructions in Old English prose style, in particular with a handful of non-finite constructions. However, since this type of structure is overwhelmingly found in texts that are direct translations from Latin, the consensus is that on the whole this form of syntactic influence was not deep or long-lasting.3

1.1 Borrowing and Imposition In an attempt to provide a general theory of language contact, van Coetsem (1988, 2000) has proposed that cross-linguistic transfer comes in two types: borrowing and imposition. This simple dichotomy is psycholinguistic in nature and depends on the dominant language of the agent of transfer, which is typically their first language.4 Borrowing takes place under recipient-language agentivity: the person doing the borrowing is psycholinguistically dominant in the language that they are borrowing into, not the language that they are borrowing from. A present-day example would be a British native-English-speaking manga fan who likes to sprinkle Japanese words into their English usage. Imposition, meanwhile, takes place under source-language agentivity: the agent of imposition is psycholinguistically dominant in the language from which the linguistic feature in question originates. A present-day example would be a French student who has moved to England and transfers their articulatory norms into the English they speak.  Schrijver (2002, 2014) has recently made the case that British Latin survived in lowland Britain for much longer than normally assumed, and that this contact left behind traces in early English phonology. The consequences of this theory have yet to be explored in the morphosyntactic domain, however. 3  See Fischer (2013) and Fischer et al. (2017: 56–65) for an overview, and Timofeeva (2010b) for an in-depth study. 4  Whether dominance always correlates with order of acquisition is a matter of some debate. Van Coetsem (2000) and Winford (2003, 2005) propose that dominance can shift over time, for instance when a person moves abroad and spends the rest of their life immersed in a different language. For Lucas (2012, 2014), on the other hand, dominance is an immutable consequence of first-language status. 2

9  Overview of Contact-Induced Morphosyntactic Changes 

243

The distinction is important because the two types of transfer are likely to be associated with different types of linguistic features. Borrowing is much more likely to be agentive in the literal sense, that is, above the level of consciousness and as an intentional act; it then stands to reason that borrowing is more likely to involve aspects of the source language that are accessible to conscious reflection, prototypically open-class lexical items. By contrast, imposition is likely to be a subconscious, unintentional process, to which, if anything, those aspects of language that are more stable and less mutable in the individual are more prone. Since most morphosyntactic features fall into the latter category, we are led to expect that most, if not all, contact-induced morphosyntactic transfer will be in situations characterised primarily by imposition rather than borrowing. However, it is important to avoid circular argumentation here; see also Lucas (2012) for a defence of syntactic borrowing as a process of change. Either way, the socio- and psycholinguistic circumstances surrounding a language contact scenario will be crucial in evaluating what sort of changes are likely to have been contact-induced. This is also a consequence of Trudgill’s (2011) theory of sociolinguistic typology, to which we now turn.

1.2 Complexification and Simplification Trudgill (2011) is concerned with the effects of contact on linguistic complexity. Observing that sociolinguists have tended to emphasise simplification as a consequence of contact, and that typologists have tended to emphasise complexification, he proposes that both are correct, but that the outcome is situation dependent. In situations of long-term co-­ territorial language contact, additive complexification (i.e. transfer of linguistic properties without replacing existing ones) is predicted, whereas short-term contact involving extensive adult second-language acquisition and use is predicted to lead to structural simplification. Trudgill’s theory has been applied to the history of English (Trudgill 2011: 50–55; Warner 2017; see also Adamczyk, this volume),5 and ­provides another way to contrastively characterise the sociohistorical  These two authors reach drastically different conclusions on the respective roles of Old Norse and Celtic, as is discussed in Sect. 3.1. 5

244 

G. Walkden et al.

circumstances, and hence the likely types of changes, associated with each of our three contact situations: Do we see long-term balanced bilingualism, or a short-term, more asymmetric relationship between the languages involved? A note of caution is in order. Just like languages, contact situations can change: the relationship between present-day Celtic languages and English, for instance, is certainly not the same as it was in the fifth century. Contact situations can also vary locally, down to the level of the individual. As Dance (2012: 1727) puts it (in reference to Old Norse), ‘one should be wary of assuming that all the (putative) effects of this contact arose from a single type of encounter […] even if historical distance has effectively turned them into one cluster of phenomena’. We believe it is of value to characterise contact situations globally according to typologies like van Coetsem’s and Trudgill’s, assessing what kind of circumstances were most prominent, but at the same time nuance is needed: at a granular enough level, every contact situation can probably be associated with every type of process, albeit to different extents. We have tried to balance the drive to generalise with the responsibility to be sensitive to nuance of this kind.

2 Celtic The earliest contact influences on (Old) English are those deriving from British Celtic and (British) Latin, the two predominant languages spoken in Britain at the time of the Adventus Saxonum in the late fourth to mid-­ fifth century CE. The consensus view has been that, despite a close co-­ existence between speakers of British Celtic languages and English spanning for over a millennium, the influence of Celtic languages on the development of English has been practically non-existent. In recent years, however, the question of Celtic influence on English, and especially on English morphosyntax, has received increasing attention and is currently being reassessed. This section begins with a short survey of the contact situation in early medieval England, followed by a discussion of two plausible cases of Celtic contact influence: the double paradigm of the verb ‘to be’ in Old English and the Middle English and Older Scots comparative nor.

9  Overview of Contact-Induced Morphosyntactic Changes 

245

2.1 Contact Between British Celtic and English There had been contacts between the Germanic invaders and the indigenous people of Britain, that is, the British Celts, even before the arrival of Germanic settlers to Britain in the mid-fifth century. These contacts had not, however, led to the kind of large-scale invasions and settlements that followed in the aftermath of those led by Hengest and Horsa in 449 (see e.g. Sims-Williams 1983). These invasions were to bring almost the whole country under Germanic rule within the next couple of centuries. As Jackson (1953: 199) states, our main source of information regarding the Germanic invasions is the historical account by the British monk Gildas, who wrote his De Excidio et Conquestu Britanniae sometime in the first half of the sixth century. In addition to De Excidio, information about the Germanic invasions can be obtained from other important near-contemporary sources, the two Gallic Chronicles of 452 and 511 (see Higham 1992: 69). Another important, though significantly later, source is the Historia Ecclesiastica Gentis Anglorum from the early eighth century, authored by the Anglo-­ Saxon monk Beda Venerabilis (the Venerable Bede). Later still, this was followed by the famous Anglo-Saxon Chronicle, which was compiled by several authors working in different places at different times, with the earliest versions dating from the ninth century. Despite problems of dating the different waves of invasion exactly, the overall picture emerging from the mentioned sources is fairly clear: the first hostile encounters between the native Britons and the newcomers, that is, armies consisting of Angles, Saxons and Jutes, did not lead to permanent settlements by the latter except in some eastern parts of the country. It was not until the second half of the sixth century that the Anglo-Saxon kingdom of Wessex, formed in the first half of the sixth century by the Saxon chiefs Cerdic and Cynric, managed to expand its territory as far west as the River Severn, and further south, to the borders of Wiltshire and Somerset. This meant that the Britons of Wales became separated from the Britons of the south-west of Britain (Cornwall), leading eventually to the separation and division of the (Late) British dialects into Welsh and Cornish, respectively (for further discussion, see Jackson 1953: 203–206).

246 

G. Walkden et al.

In the north of Britain, the Anglo-Saxon conquest and settlements proceeded along major waterways such as the Trent and the Humber. According to Jackson (1953: 207), the northern and Midland settlements led to the establishment of two Anglian kingdoms, Lindsey and Mercia, in the seventh century. Under their king Penda (d. 655), Mercia conquered large areas both from their West Saxon cousins in the south and from the Welsh in the west. Jackson refers here to the often-expressed view according to which the Mercians also managed to reach the sea in the north and thus break the land connection between the Welsh and the Britons of the North. Jackson does not, however, find any solid evidence to substantiate this claim (Jackson 1953: 210–211). In any case, the Anglo-Saxon advances to the north proved to have significant consequences for the later development of the British Celtic languages, as it meant an areal separation of the Welsh and Cumbric dialects of Late British.6 The rapidly growing extent of the Anglo-Saxon settlements in the centuries following the Adventus raises the question of what exactly happened to the indigenous population of Britain. Up till fairly recently, history writing was dominated by a persistent myth about ‘ethnic cleansing’ of the indigenous British and Romano-British population, especially in the southern and eastern parts of the country (see e.g. White 1971; Myres 1986). This ‘Germanist’ view is no longer upheld in current research: the evidence does not support widespread massacre of the Romano-British population in either towns or countryside; widespread intermingling of the two cultures was more likely than sharp polarisation and conflict (see e.g. Laing and Laing 1990; Tristram 2002; Schrijver 2007; Laker 2008a; Trudgill 2010). There are differing estimates of the immigrant : native ratio in the first centuries after the Adventus Saxonum (see e.g. Laing and Laing 1990; Higham 1992; Härke 2003), but, despite these differences, it is evident that the Germanic immigrants formed only a relatively small proportion of the population of Britain. Therefore, instead of wholesale extermination of the Romano-British population, a process of acculturation, assimilation and language shift was a more likely outcome of the contact.  For a recent assessment of the historical accounts of Anglo-Saxon settlement, see also Carver (2019).

6

9  Overview of Contact-Induced Morphosyntactic Changes 

247

Support for the ‘acculturation theory’ can also be obtained from population-­genetic studies. For example, Capelli et  al.’s (2003) study shows that (i) no complete population replacement occurred anywhere in the British Isles; (ii) there was considerable continental introgression in the Central-Eastern part of England; and (iii) the data from southern England indicate significant continuity of the indigenous population. Further corroboration for the acculturation theory is offered by The People of the British Isles Project (see Leslie et al. 2015). Based on genome-­ wide data, the results of this project provide clear evidence of Saxon migration in the modern English data. As Leslie et al. (2015: 313) note, however, the DNA contributed by Saxon settlers in the fifth and sixth centuries ranges from as low as 10% to not higher than 40% in the data from central and southern England, thus ‘clearly excluding the possibility of long-term Saxon replacement’. This points towards the conclusion that intermarriage and acculturation rather than genocide must have taken place during the centuries following the Adventus Saxonum. The historical evidence discussed above indicates that in many parts of Britain conditions favourable to bilingualism existed for a considerable period of time after the first arrival of the Anglo-Saxons. Indeed, Jackson (1953: 245) considers it likely that, before the eventual language shift, there was a bilingual stage when the Britons were able to speak both Old English and British Celtic;7 on the other hand, the Anglo-Saxons probably had no particular need to learn the language of those whom they had conquered. In terms of van Coetsem (1988), the primary mechanism of transfer would thus have been imposition, where British Celtic-dominant bilinguals, forming the majority of the population, imposed some of their L1 features on their L2, English. The effects of imposition can be seen primarily in morphosyntax rather than lexicon, and this is also the case with the majority of plausible cases of language transfer from British Celtic.8

 Cf. also Warner (2017: 364–369), who considers it likely that bilingualism, including childhood bilingualism, was widespread among first-language speakers of British Celtic. 8  Van Coetsem (1988: 3 and 26) notes that phonology and grammar are areas where transfer typically takes place in source language imposition. Phonology, however, is not discussed in this chapter. 7

248 

G. Walkden et al.

2.2 Celtic Influence on English Morphosyntax Filppula et al. (2008: 30–117) identify eleven areas of English morphosyntax that show possible influence of Celtic in their development. These include features such as the internal versus external possessor construction, the double paradigm of the verb ‘to be’ in Old English, the Northern Subject Rule, do-periphrasis, the progressive aspect, the cleft construction, contact relatives, -self reflexives and intensifiers, and comparative nor. In the following, we will briefly discuss two Old and Middle English constructions widely considered as plausible cases of Celtic influence on English: the Old English double paradigm of the verb ‘to be’ and the Middle English and Older Scots comparative nor.

Double Paradigm of the Verb ‘To Be’ The Old English paradigm for the verb ‘to be’ distinguished between two meanings, habitual and ‘actual’ or future. The former was based on a reconstructed PIE *es-form, the latter on a reconstructed PIE *bheu-form of the verb ‘to be’. The first scholar to pay attention to the possible Celtic background of this distinction was Wolfgang Keller (see Keller 1925), who noted that the Old English forms based on the reconstructed root *bheu (so-called b-forms) and their meanings are closely paralleled by the corresponding Celtic and especially Brittonic forms. He further points out that, although partially similar parallels are found in other Germanic dialects, none of these have developed a full present-tense indicative paradigm for both roots with clearly distinct meanings. Keller (1925: 60) concludes that this feature was introduced into English by the early Britons trying to acquire English: ‘[D]ie altenglischen Formen und Funktionen der Wurzel *bheu, die den anderen germanischen Dialekten fremd sind, entstanden im Munde und im Denken von englisch sprechenden Briten’.9 Table 9.1 (based on Lutz 2009: 232) summarises the parallels between the Germanic and Brittonic paradigms:  ‘Old English forms and functions of the root *bheu, which are alien to the other Germanic dialects, arose in the mouths and minds of English-speaking Britons’. 9

9  Overview of Contact-Induced Morphosyntactic Changes 

249

Table 9.1  The double paradigm of present indicative of the verb ‘to be’ in West Saxon and Brittonic and the simple paradigms in the other Germanic languages West Saxon

Brittonic

‘habitual’ ‘actual’

‘habitual’ ‘actual’

bīo bist bið bīoð

byðaf byðy byð byðwn byðwch byðant

eom eart is sind(on)

wyf wyt yw ym ywch ynt

Old Saxon

Old High German

Old Norse

Gothic

bium bist is(t) sind(un)

bim bist ist birum birut sint

em est es erom eroþ ero

im is ist sijum sijuþ sind

The same parallelism was later noticed by Tolkien (1963), who considered it as one of his prime examples of linguistic contact between the two languages. He noted the distinction that both English and Welsh make between what he termed the ‘actual present’ and the ‘consuetudinal present’ / ‘future’. Each of these was expressed by a different set of forms, the latter relying on forms beginning with b- both in Old English and in Welsh. Tolkien also pointed out the uniqueness of the Old English system among Germanic languages. Besides the similarities in the forms and functions of the Old English and Welsh ‘be’ verbs, he noted the difficulty of explaining the short vowel in the Old English 3sg. form bið as a regular development from earlier Germanic, while there would be no such problem if the corresponding Welsh form bydd (from earlier *bið) was considered (Tolkien 1963: 30–32). In more recent research, Keller’s and Tolkien’s accounts have been taken up with renewed interest. Lutz (2009: 234) concludes that ‘[t]he twofold paradigm of “to be” represents the most obvious but not the only syntactic evidence for early Celtic substratum influence due to language shift by speakers of Celtic which was addressed by Keller’. Similar ideas have also been expressed on the Celticist side. Thus, Ahlqvist (2010) quotes Tolkien’s article at length and devotes a fair amount of space to a detailed comparison of the relevant verbal paradigms in Old English and early Welsh. He, too, comes to the conclusion that the parallelism between Old English and early Welsh with respect to the twofold paradigm of ‘to be’ must be due to an early contact situation, ‘based on both

250 

G. Walkden et al.

languages having forms both with and without b- in the paradigm of the verb “to be”, and these forms, moreover, having rather similar functions’ (Ahlqvist 2010: 54).10 Finally, Ahlqvist also considers the possibility of early influences from across the Irish Sea from Old Irish, following up on a proposal in that direction by Schrijver (2007). However, he notes certain differences between Irish, on the one hand, and Welsh and English, on the other, which make this scenario less likely than the one between Welsh and English (Ahlqvist 2010: 55–56). The strongest arguments speaking for an early contact between the English and Welsh paradigms for the verb ‘to be’ rest on the existence of very close formal and functional parallels and the uniqueness of Old English amongst Germanic languages with respect to this feature. Together, they make it very likely that the Old English distinction between the PIE *es- and PIE *bheu-forms of the verb ‘to be’ was imposed on the language by British Celtic-dominant bilingual Britons.

Comparative Nor The use of nor and its variants na, no, ne, nai, nag instead of than in comparative clauses, exemplified in (1) and (2), is well attested in English and Scots from the fourteenth century onwards, and this nor construction is also widely used in dialects, especially in the north of England. (1) Odere tythynges cannot I tell yow no thes for soothe but be here sey (?1438; Oxford English Dictionary, OED, s.v. no, conj.2) (2) Ye schall here myche more in thys pertys nor I can at Brytys (1479; OED, s.v. nor, conj.2).

 It is worth emphasising that some other Germanic languages also have forms beginning with b- in some parts of their paradigms of the verb ‘to be’, as Keller (1925) already noted. Thus, Schumacher (2007) argues for possible earlier continental contact between Celtic and West Germanic. While Ahlqvist considers this kind of contact quite possible (Ahlqvist 2010: 54), Lutz (2009: 237) prefers Keller’s original account, which rests on the idea of early substratal influence between Celtic and Old English. 10

9  Overview of Contact-Induced Morphosyntactic Changes 

251

There have been a number of attempts to find a source for the construction. Holthausen (1913: 339–340) argues that a construction of the type ‘He is older nor I’ simply represents the combination of the two propositions ‘he is older, and not I’ through the loss of what he calls a ‘syntactic pause’ and consequent shift of stress. Small (1924) explains the construction by phonetic reduction and consequent reanalysis, while Joly (1967) attempts to explain it through a Middle English reanalysis of the Old English comparative particle þonne. These explanations, however, are untenable (for details, see Filppula et al. 2008: 99–102 and Laker 2008b: 9–14), and the OED and the Dictionary of the Older Scottish Tongue both consider the origin of the nor construction to be obscure or uncertain. Filppula et al. (2008) and Laker (2008b) argue that a plausible origin for the negative comparative particle in English can be found in (Old and Middle) Welsh, where a formally and functionally very similar particle, na(c) / no(c), antedates the English comparative nor, as in (3) and (4) (from Evans 1976: 43):11 (3)  Ny byd gwaeth itt (ms. in) yno noc et y Arthur yn y llys    ‘It will not be worse for thee there than for Arthur in the court’   (c. 1050–1100, Kulhwch ac Olwen, WM 456. 28–29) (4)  iawnach yw idaw dy gynnhal nogyt y mi    ‘more fitting is it for him to maintain thee than for me’   (twelfth century, Pedeir Keinc y Mabinogi 26.28).

Although comparative nor is first attested in fourteenth-century Middle English dialects, Laker (2008b: 20–21) considers it likely that the putative transfer took place during the Old English period.12 To support his  Laker (2008b) also considers the possibility that the Old French (OFr.) ne explétif construction could have played a role in the development of the English comparative nor / negative comparative particle (NCP). He concludes, however, that ‘several formal linguistic divergences existing between the OFr. ne explétif construction and the NCP of Middle and Modern English dialects argue against French influence’ (Laker 2008b: 25). 12  Comparative nor is also attested in Irish English. Laker (2008b: 21–22) suggests that the Irish English NCP is a borrowing from colloquial British English or a loan translation of the corresponding Irish construction. 11

252 

G. Walkden et al.

argument, he (2008b: 21) refers to Tolkien (1963: 28): ‘The records of Old English are mainly learned or aristocratic; we have no transcripts of village-talk. For any glimpse of what was going on beneath the cultivated surface we have to wait until the Old English period of letters is over’.

Complexification or Simplification? Trudgill (2011: 50–55) argues that contact between speakers of Old English and Late British was the factor that triggered the process of simplification and consequent typological change from highly fusional and inflecting Old English to the much more isolating type of morphology in Middle English. However, as Warner (2017: 364) observes, ‘few of the eleven areas of development in grammar claimed to show the possible influence of Celtic on English listed in Filppula et al. (2008: 30–117) clearly involve simplification’, and constructions such as the double paradigm of the verb ‘to be’ clearly represent complexification instead. Indeed, as Warner (2017: 367–368) argues, the contact situation between speakers of Old English and British Celtic, with both childhood and adult bilingualism over an extensive period of time, was of the type where complexification rather than simplification is the expected outcome.

3 Old Norse One important feature sets contact with Old Norse apart from contact with French and Celtic:13 both are Germanic languages, and hence typologically and genealogically they are very close to one another. At the very least, certain cognate words and similar structures would have been immediately recognisable to speakers of the other language. Townend (2002) has made a compelling case that the two languages would have  In this chapter we use the term ‘Old Norse’ broadly, to refer to any and all North Germanic varieties spoken and written during the medieval period, rather than narrowly in the sense of Old West Nordic (as opposed to Old East Nordic). This latter distinction is not trivial, especially since the bulk of Scandinavian settlement in England was by speakers of Old East Nordic; however, the differences between the two varieties are unlikely to be relevant to any of the changes discussed in this section. 13

9  Overview of Contact-Induced Morphosyntactic Changes 

253

been mutually intelligible under receptive bilingualism (see Keller 2020 for lexical evidence in support of this claim). This section first provides an overview of the contact situation (Sect. 3.1), then discusses potential transfer of functional lexical items (Sect. 3.2), effects on inflectional morphology and agreement (Sect. 3.3), and syntactic effects (Sect. 3.4).

3.1 Contact Between English and Old Norse Chronologically, contact between English and Old Norse can be divided into four phases.14 Pons-Sanz (2013) names (1) the ‘hit-and-run’ phase, from the earliest Viking incursions in the eighth century to the middle of the ninth; (2) the ‘settlement’ phase, from the mid-ninth century to AD 1000, characterised by wholesale settlement of Old Norse speakers in the British Isles;15 and (3) the brief ‘conquest’ phase, from 1000 to 1042, ending with the cessation of direct rule over Britain by Cnut and his sons. Walkden (Forthcoming) adds a ‘shift’ phase, lasting from 1042 until the point at which Old Norse ceased to be a living language in Britain. The date of this is unknown, and would have varied from place to place: in the northern isles of Orkney and Shetland, for instance, Old Norse continued to be spoken and transmitted until the eighteenth century (see Kinn and Walkden, this volume). However, for most of England we can assume that Old Norse ceased to be acquired by children some time during the Middle English period. Lexical and phonological evidence has been taken to suggest a change in the dominant mechanics of the contact situation during the eleventh century, from one favouring recipient-language agentivity in transfer, that is, borrowing, to one favouring source-language agentivity, that is, imposition (Townend 2002: 201–210; see also Adamczyk, this volume). If so, then, following the considerations outlined in Sect. 1.1, we should  For a very similar quadripartition of phases of Old Norse contact in a different context, see Timofeeva (2016: 87). 15  The scale of settlement has been a matter of some debate among historians, with Sawyer (1971) particularly sceptical. The current consensus is that the settlement was indeed substantial: see Hadley (1997), and Kershaw and Røyrvik (2016). 14

254 

G. Walkden et al.

expect to see morphosyntactic contact influence from the eleventh century onwards. The key evidence as to whether this hypothesis is borne out or not must come from the cluster of texts usually described as Late Northumbrian Old English (see Fernández Cuesta and Senra Silva 2008): the tenth-century glosses to the Lindisfarne (London, British Library, MS Cotton Nero D.iv) and Rushworth Gospels (Oxford, Bodleian Library, MS D. 2. 19) and Durham, Cathedral Library, MS A.IV.19 (the so-called Durham Ritual or Collectar). This evidence is, however, difficult to interpret, and arguments have been made both for morphosyntactic contact influence in these texts (e.g. Kroch et al. 2000; Millar 2000; Adamczyk, this volume) and against it (e.g. Cole 2018; Walkden, Forthcoming). As regards the sociolinguistic typology of Norse contact, Trudgill (2011: 53) concludes categorically that ‘Contact between Old Norse and Old English was not of the sociolinguistic type that makes for simplification’. His main piece of evidence for this conclusion is the putative transfer of the third-person plural pronouns from Norse into English, since (he argues) pronominal transfer only occurs under conditions of long-­ term co-territorial contact and proficient bilingualism. However, the case of the pronouns has been called into question (see discussion in Sect. 3.2), and so has Trudgill’s broader analysis of the contact situation. Warner (2017) broadly accepts the framework of sociolinguistic typology but argues in detail that contact with Old Norse did in fact lead to simplification. His conclusion is that koineisation (Siegel 1985)—a process that leads to mixing of mutually intelligible varieties in the context of increased interaction and integration among speakers of those varieties— is the primary process characterising structural changes in English induced by contact with Old Norse. Crucially, koineisation can lead to simplification, but via levelling and accommodation rather than as the consequence of adult second-language acquisition. This is good, since given the mutual intelligibility of Old Norse and Old English as well as the historical circumstances in which the contact arose, it is unlikely that adult second-language acquisition of Old English by speakers of Old Norse was ever a significant phenomenon, except perhaps during the shift phase to a limited extent. As will be seen below, if one major result of contact with Old Norse was a koineisation process, this fits well with recent research on

9  Overview of Contact-Induced Morphosyntactic Changes 

255

individual phenomena arguing that the outcome of the contact situation in English was not transfer in the literal sense but rather reinforcement of competing West Germanic variants that were formally and functionally more similar to what was found in Old Norse (Versloot, this volume; Cole and Pons-Sanz, this volume).

3.2 Transfer of Functional Lexical Items The poster children for Scandinavian influence on medieval English have long been the third-person plural pronouns they / their / them. Ever since Kluge (1899) and Björkman (1900), these pronouns have featured in textbook treatments of contact influence, and have also played an important role in assessments of the scale and level of contact influence. For Thomason and Kaufman, for instance, the transfer of pronouns suggests ‘an intense contact situation’ (1988: 281), either level 3 or level 4 on their scale. For Trudgill (2011), this putative borrowing is at the heart of his argument that Old Norse contact on the whole involved complexification, not simplification. The basic argument for transfer is that the th-­ forms cannot be the descendants of Old English hīe / hira / him simply because of the initial consonant; by contrast, Old Norse has the forms þeir / þeira / þeim, providing a plausible source. Added to this is the fact that the spread of the th-forms seems to proceed from the north and east, providing a circumstantial argument for contact influence. The nature of the transfer has been debated: Buccini (1992) makes the case that imposition, not borrowing, was the key process. Morse-Gagné (2003) presents a detailed empirical study. Until recently, there was a near-consensus that these forms were transferred from Old Norse. Cole (2018) challenges this conventional wisdom, however, reviving the view that the th-forms instead derive directly from Old English demonstratives. In support of this, she notes that the grammaticalisation pathway from distal demonstratives to third-person pronouns is well trodden, and that a purely anaphoric function for the ‘demonstrative’ forms is already well attested in Old English, especially in Old Northumbrian. The main objection to deriving the th-forms from demonstratives has been that it cannot account for the [eː] and

256 

G. Walkden et al.

diphthongal vocalism in the Middle English forms; Cole shows, however, that there is far more variation here than traditionally assumed, and that these vocalisms are not problematic when that is taken into account. Cole’s argument is not intended to rule out contact influence; rather, she aims to show that an explanation in terms of transfer is not necessary. She does allow (2018: 187) for the possibility of ‘interlingual identification’ (Weinreich 1968 [1953]), in which the functional and formal similarity of the emerging Old English anaphoric th-forms with the Old Norse third-person plural pronouns led to convergent development and mutual reinforcement. This sits well with a scenario of koineisation as proposed by Warner (2017), with imposition also potentially playing a role during the Middle English period. A fine-grained examination of the situation suggests both West and North Germanic input in different language users at different times, that is, polygenesis; see Pons-Sanz and Cole (this volume), who reach this conclusion based on an examination of pronoun use in seven manuscripts of the early Middle English La estorie del evangelie from different regions.

3.3 Inflectional Morphology and Agreement Contact with Old Norse has long been implicated in the massive reduction in inflectional forms that English underwent during the medieval period (see Bradley 1904: 32). Poussa (1982) goes so far as to propose that this reduction is a symptom of creolisation, with Old Norse contact as the trigger. Potentially relevant developments include the spread of plural -s and genitive -s from a tiny corner of the Old English inflectional system to become system-wide defaults, the rise of third-person singular verbal -s, the loss of most of the case system (especially on nouns) and the complete evaporation of the system of grammatical gender. In works such as Allen (1997), Trudgill (2011), Warner (2017) and Adamczyk (this volume), the question of morphological influence has been the subject of renewed attention from a modern contact-linguistic perspective. As with the function words discussed above, the core issue is that a scenario of direct transfer is difficult to motivate, as precise formal isomorphisms that have real diagnostic value are not found (compare, for instance, the striking overlaps seen in the discussion of the twofold

9  Overview of Contact-Induced Morphosyntactic Changes 

257

paradigm of the verb ‘to be’ discussed in Sect. 2.2.1). Furthermore, many of these developments—especially the general loss of inflectional endings—are of a kind that is seen in language after language. Early Old English, for instance, already exhibits a complete loss of person distinctions in plural verb forms; this is a development common to all North Sea Germanic languages (see Ringe and Taylor 2014: 158–160), and hence cannot be attributed to influence from Scandinavian settlement in the British Isles. Facts like these are behind Thomason and Kaufman’s (1988: 303) succinct statement that ‘Norse did not stimulate simplification in English’. Nevertheless, the geographical distribution and the nature of the contact situation make for strong circumstantial arguments that some of these changes were at least contact-accelerated. As regards nominal inflection, Allen (1997) investigates the simplification and loss of case-marking morphology and categories in detail. In Old Northumbrian, in particular, sound changes such as the merger of unstressed vowels and the loss of final -n are already underway, and would necessarily have reduced the number of formal distinctions available in the system, opening the door also to further analogical restructurings.16 Still, ‘it is probably no coincidence that case-marking reduction proceeded fastest in the area with the most contact with a closely related language’ (Allen 1997: 73), and she goes on to outline a scenario of dialect levelling and koineisation (though she does not use the term). A particularly striking reduction is the loss of genitive inflection in English dialects (e.g. my father boots ‘my father’s boots’), the geographical distribution of which maps well onto the areas of Scandinavian influence in England (Klemola 1997), and which can also be found in substantial numbers in the Old English gloss to the Lindisfarne Gospels (Rodríguez Ledesma 2016) and in the Northern Middle English Cursor Mundi (Allen 1998). Adamczyk (this volume) carries out a detailed study of nominal inflection that directly compares Old Northumbrian, West Saxon and Old Norse, and concludes that at this stage we see no direct transfer, though structural parallels can be observed—one of these is the ‘superstable’ plural exponent -Vr, which Adamczyk suggests is parallel to the emerging  For further discussion of simplification of nominal morphology in Old Northumbrian texts, see Millar (2016), and Fernández Cuesta and Rodríguez Ledesma (2020). 16

258 

G. Walkden et al.

default plural -(e)s in medieval English.17 In northern Middle English, we also see plurals marked by i-mutation, for example hend ‘hands’ and breþer ‘brothers’, which Adamczyk suggests may be favoured by the fact that Norse exhibits a similar morphologisation of i-mutation; some of these plural forms, such as hend, could in principle even be instances of direct lexical transfer. Such a development is hard to characterise as simplification, as she notes—though it is consistent with a scenario of imposition. As regards verbal inflection, a major candidate for Norse influence is the emergence of third-person singular verbal -s, which replaces the -þ inherited from Old English. Samuels (1985: 276) points out that in Old Norse the second- and third-person singular were identical, and suggests that this would have provided a structural template for the extension of the Old English second-person singular -s to the third-person singular.18 Kroch et  al. (2000: §5.1) pursue an alternative hypothesis of contact influence, in which the crucial factor is misperception of the voiceless fricative allophone [θ] in unstressed verbal endings as the similar but less marked [s] by native speakers of Old Norse. Both types of hypothesis fail to explain, however, why the -s and -þ endings in the Lindisfarne glosses show grammatical conditioning by subject type and adjacency, in the manner of the Northern Subject Rule (Cole 2014; see also Miller 2002 for arguments against contact influence). Moreover, second-language acquisition by native speakers of Old Norse is arguably not the right way to think about the contact situation that gave rise to the Lindisfarne glosses, given the time period, geographical location and mutual ­intelligibility of the varieties involved (Walkden Forthcoming). As in the nominal domain, however, the case can be made that the earlier and faster simplification of verbal morphology in the north and east during the Middle English period is a consequence of koineisation; by contrast, the West Midlands Middle English texts of the Katherine Group are relatively  See also Hotta (2009) and Warner (2017: 328–332, 345–348).  Samuels claims that the Old Norse ending, later -r, would still have been pronounced -z at the time of Scandinavian settlement in the British Isles, but this is not the consensus view: see Cole (2014: 32–33) for discussion. 17 18

9  Overview of Contact-Induced Morphosyntactic Changes 

259

conservative in their verbal morphology. This trend towards simplification consists of the loss not only of inflectional endings but also of the derivational -i(j)- formative of class 2 weak verbs (Warner 2017) and of the Old English prefix ġe-, which Old Norse had already lost entirely (along with other verbal prefixes) before the textually attested period.

3.4 Syntax: Constituent Order Emonds and Faarlund (2014) catalogue a number of syntactic properties which, they argue, have their origins in Old Norse. From these they conclude that modern English is descended from (anglicised) Old Norse, and is thus genealogically a North Germanic rather than West Germanic language. While there are reasons to doubt their conclusion (see Bech and Walkden 2016; Crisma and Pintzuk 2019), the syntactic similarities they demonstrate between Present-Day English and the Scandinavian languages are numerous and striking, and that contact with Old Norse is implicated in the historical development of some of these phenomena is hard to dispute. This subsection zooms in on a constituent order feature, verb-second (V2). V2, in which only one constituent can precede the finite verb, is widespread among Germanic languages. Yet, as is well known, Old English was not a strict V2 language. Instead, it exhibited an information-­ structurally driven alternation: when the subject is given information, and especially when it is a pronoun, as in (5), verb-third order can be found, without subject-verb inversion. Walkden (Forthcoming) dubs this ‘information-structural V2’ as opposed to the more familiar strict V2 characteristic of, for example, modern German. Kroch and Taylor (1997) show that there are strict V2 texts in the history of English, however. One such is the Northern Prose Rule of St Benet, in which subject-verb inversion is obligatory, as in (6). (5)    æfter his gebede he ahof  þæt cild   up    after  his prayer  he  lifted the  child up    ‘After his prayer he lifted the child   up’   (Ælfric, Catholic Homilies)

260 

G. Walkden et al.

(6)  Lauerd, of me haue  I noht,   bot þu  sende it me    lord   of me have  I  nothing but you  send  it me    ‘Lord, by myself I have nothing unless you send it to me’     (Northern Prose Rule of St Benet).

Middle English texts in particular show variation between these two systems. Texts with more strict V2 tend to be those from the north and east, including the Ormulum, where Old Norse influence is well established (Trips 2002). Old Norse exhibits strict V2 from its earliest textual records, and so Kroch and Taylor posit that Old Norse influence led to strict V2 (see also Miller 2012).19 Walkden (Forthcoming) questions some aspects of this scenario: in particular, strict V2 can be found, albeit to a lesser extent, in Old Northumbrian texts, which otherwise do not reflect a contact situation of the kind likely to lead to direct syntactic transfer. Moreover, strict V2 can be found in some contexts in the earliest Germanic texts (Eyþórsson 1995). Transfer of strict V2 constituent order cannot be the whole story, then. It is very plausible, however, that koineisation led to a substantial increase in use of strict V2, including beyond the contexts in which it was typical in Old English.

4 French French was brought to England as the language of the conquering Normans in 1066, and at least some speakers retained active knowledge of insular French, that is, Anglo-French,20 until around the end of the  Kroch and Taylor’s causal story is more nuanced than this, involving a mediating role for the loss of verbal agreement, as discussed in the previous subsection. See Walkden (Forthcoming) for detailed discussion. 20  Some scholars (e.g. Ingham 2012) prefer the term ‘Anglo-Norman’, since the variety of French spoken in England was closely related to the Norman dialect. Others (e.g. Blake 1992 in the Oxford History of the English Language) associate ‘Anglo-Norman’ with the earliest phase of French in England, and ‘Anglo-French’ with the language in the thirteenth and fourteenth centuries, ‘which was essentially an administrative language which had to be acquired as a foreign language by the English’ (1992: 5). We do not share Blake’s view that the situation in the thirteenth and fourteenth centuries was fundamentally different to the earlier phase, as we discuss below, and for clarity adopt the term ‘Anglo-French’ to refer to the variety of French spoken in England throughout its history. 19

9  Overview of Contact-Induced Morphosyntactic Changes 

261

fourteenth century (see Sect. 4.1). Contact thus directly precedes the emergence of Middle English in the mid-twelfth century and the wide-­ ranging morphological and syntactic changes that this entailed. Nevertheless, while lexical borrowing from French, or Latin via French, into Middle English is significant (see Dance et al., this volume), whether or not French had any significant influence on English morphosyntax is subject to lively debate. In Sect. 4.2, we discuss possible effects on word order, showing that while some contact influence from Old French can be demonstrated, it does not trigger fundamental syntactic change in Middle English. However, in Sect. 4.3 we review recent research on change in argument structure and suggest that contact with French has more substantial effects on Middle English syntax than previously supposed.

4.1 Contact between English and French Following the Norman conquest of 1066, Anglo-French was the native spoken and, to some extent, written language of the new feudal elite and those who came over from Normandy with them. However, as this situation did not last more than a few generations, there has been considerable debate as to what kind of contact situation took its place during the crucial period in the thirteenth and fourteenth centuries when textual records of Middle English emerge. To take two diametrically opposed positions, while the Anglo-French scholar Legge writes that ‘by this time [1170] most people, down to the very poorest, were bilingual’ (1963: 4), Thomason and Kaufman reject any possibility of major morphosyntactic contact influence, asserting that ‘[t]here is no reason to suppose that any large proportion of native English learned French between 1066 and 1250’ (1988: 308; see also discussion in Trotter 2012: 1787). However, subsequent research, in particular that of Rothwell (1976, 1978, 1998, 2001), Trotter (2006, 2012) and Ingham (2009, 2012), has led to the emergence of something of a consensus around an intermediate position, which acknowledges that bilingualism in English (L1) and French (L2) was the norm for educated, literate speakers well into the fourteenth century (Trotter 2012: 1785).

262 

G. Walkden et al.

Ingham (2012), focusing primarily on the linguistic development of Anglo-French, shows that a significant number of fluent English-French bilinguals remained until the latter half of the fourteenth century. He argues that the transmission of Anglo-French after the Conquest was punctuated by two ruptures. The first rupture occurs in the first half of the thirteenth century and is marked by increased phonological influence from Middle English (2012: 53–71), a development likely related to the loss of English royal domains in France in 1204 and a corresponding lack of input from monolingual French speakers to children in English aristocratic households (2012: 161). However, Ingham convincingly argues that, even after this first rupture, Anglo-French continued to be acquired in early childhood as a spoken second language, permitting native-like transmission of morphology and syntax. Although there is little evidence about the context in which such transmission took place—perhaps in elementary schola cantus ‘song school’ run by the Church (2012: 33)—it is clear that French remained the primary language of instruction in grammar schools (see also Reed, this volume). Moreover, Anglo-French does not show undue morphological or syntactic influence from English at this time, continuing to develop in much the same way as L1 varieties of French spoken in France, for example with regard to the decline of verb-second after adjuncts from the second half of the thirteenth century (2012: 116) or in the increased frequency of postposed adjectives (2012: 135). The second rupture, marking the end of spoken bilingualism in England, came with the Black Death in the mid-fourteenth century, which took a heavy toll on the clergy who were the primary agents of the transmission of Anglo-French (2012: 35). The contact situation is very different to the Celtic and Scandinavian cases described above. First, the primary mechanism of transfer is borrowing in van Coetsem’s (1988) terms, and consequently lexis is affected to a far greater extent than morphosyntax, as many scholars have observed (e.g. Miller 2012; Fischer 2013). Second, French was a prestige language spoken initially by the ruling class and later only by literate, educated bilinguals. Third, although not standardised, Old French was a written language and was extensively used in literature and administration in

9  Overview of Contact-Induced Morphosyntactic Changes 

263

both England and France. Nevertheless, it is questionable whether it is justified to call this ‘fairly similar’ to contact between English and Latin (Fischer 2013: 23), as the fact that Latin itself was taught through the medium of French (Ingham 2012: 34–35) indicates far more widespread spoken competence in the latter than the former.

4.2 Word Order If borrowing is the core mechanism of transfer, contact influence on an area of core syntax largely independent of the lexicon such as word order is perhaps unlikely. This is borne out by recent reviews of the literature: Miller (2012: 185) concludes that ‘French influence on English syntax is very limited’, while Fischer (2013: 40) opines that ‘the arguments are usually not fully persuasive’ where studies attempt to demonstrate syntactic influence. In this section, we briefly consider possible French influence on two developments in Middle English word order: verb-subject inversion and the use of postposed adjectives. Both Old English and Old French show verb-subject inversion in declaratives as part of their verb-second grammar, and in both languages this is lost by the end of the medieval period, although the development in French occurs slightly later than in Middle English. Haeberli (2010) examines a number of possible areas in which French may have influenced specific aspects of the moribund Middle English verb-second system. Due to ‘information-structural V2’ (see Sect. 3.4), verb-subject pronoun inversion was uncommon in many varieties of Old English, yet it is found in some Middle English texts, as in (7): (7) And many mervayles shall he do      and  many miracles    shall  he do     ‘and he will do many miracles’ (fifteenth century, Middle English, Malory, 47.79; Haeberli 2010: 148).

Contrary to Old English, pronominal subjects were regularly inverted in Old French (Vance 1997), for example:

264 

G. Walkden et al.

(8) Voirement sont ce  des  aventures del  saint Graal     truly     are     this some   adventures   of.the   holy   Grail     ‘truly these are adventures of the Holy Grail’ (thirteenth century, continental Old French, La Queste del Saint Graal 52; Ingham 2012: 102).

Haeberli (2010) studies the frequency of verb-subject pronoun inversion in V2 contexts in the Penn-Helsinki Parsed Corpus of Middle English (PPCME2) (Kroch and Taylor 2000) and shows that it is more frequent in all periods of Middle English than in Old English. However, there are a number of caveats. First, there is no text in which the frequency of verb-­ subject pronoun is ever as systematic as it is in thirteenth-century Old French (Vance 1997: 350) or Anglo-French (Ingham 2012: 109–114), where it occurs in 90–100% of contexts. Second, there are vast differences between the Middle English texts: while Chaucer’s works show a relatively high rate of verb-subject pronoun inversion (50%), other texts from the same period, such as the Middle English translation of the Old Testament by John Wycliffe and his followers, show no inversion at all. While Haeberli (2010: 156–159) shows that texts translated from French or written by authors such as Chaucer, who were known to be proficient in French, tend to have higher rates of verb-subject pronoun inversion, a clear link to French cannot be established for all such texts. Finally, as we discussed in Sect. 3.4, contact with Old Norse is at least equally plausible as a source for strict V2, in particular given that it is more frequently found in northern and eastern dialects of English. Overall, Haeberli cautiously concludes that ‘French influence seems as likely a hypothesis as others that have been proposed’ (2010: 161) to account for the attested changes. A second aspect of Middle English word order for which French influence has been argued to play a role is the use of postnominal adjectives. Modern English shows noun-adjective (N-A) order in a number of idiomatic noun phrases (NPs) where both elements were borrowed from French, such as heir apparent or proof positive (Trips 2014: 74). In Middle English, while adjectives predominantly occur before the noun (A-N), N-A order is also attested, and more widely than in modern English. However, opinions are divided as to whether this reflects French influence on Middle English syntax or not, in particular since, in contrast to

9  Overview of Contact-Induced Morphosyntactic Changes 

265

modern French, N-A order was clearly less frequent than A-N order in Old French (cf. Buridant 2000; Trips 2014: 78). On the side of the ‘French-minimisers’, Fischer claims that cases of N-A order involving French lexemes in Middle English are, like modern English heir apparent, ‘fixed [and] used by English authors […] as a unit’ (2006: 271), that is, French influence is limited to word order in lexical borrowings and has no effect on Middle English syntax. While other cases of N-A order in Middle English do exist, these are considered an internal development, since Old English also (rarely) allowed the postposition of strongly rhematic adjectives, even in definite NPs (9). This strongly rhematic reading continues to characterise Middle English N-A orders, for example where two adjectives are contrasted (10): (9)  Þone ilcan ceaddan iungne      the  same Chad    young.str     ‘the same Chad, when young’ (The Life of St Chad 1.184; Fischer 2006: 258) (10) to þe    lyf   bodilyche and to  þe    lyf   gostliche      to the life bodily   and to    the life    spiritual     ‘to the bodily and the spiritual life’ (The Mirror of St Edmund; Fischer 2006: 272).

Trips (2014), on the other hand, highlights several aspects of the Middle English data which are difficult to account for if contact influence is excluded. First, she argues that N-A order with borrowed adjectives does not just occur in ‘fixed phrases’ but also in NPs with a non-French head noun. This is even true of the rare cases where French plural concord of the adjective is also copied into English, for example lordes Arabiens ‘Arabian lords’ from Chaucer’s Treatise on the Astrolabe (Trips 2014: 84). Second, she shows that Old French also favours N-A order with rhematic adjectives and is thus a possible source for the Middle English pattern. Third, like Haeberli (2010), she shows that texts which can be linked to French—such as Chaucer’s works and the Ayenbite of Inwyt, a translation of the Old French La somme le roi—show a higher

266 

G. Walkden et al.

percentage of N-A order. Moreover, French influence in these texts is also confirmed by the occasional presence of plural concord. She therefore concludes that ‘grammatical replication cannot be excluded’ as a source of Middle English N-A order (2014: 91). Overall, while contact with French is a plausible source for changes in word order, it is important to note that both verb-subject pronoun inversion and N-A order are transitory developments in Middle English which did not lead to language change. In the case of N-A order in particular, it is uncertain to what extent it was ever a phenomenon of the spoken language as opposed to a stylistic feature used in writing.

4.3 Argument Structure The development of argument structure in Middle English is an exciting and dynamic current research area (Trips and Stein 2019; Ingham 2020; Trips 2020; García García and Ingham, this volume), and being situated at the interface between syntax and the lexicon, it is also an area where extensive lexical borrowing can plausibly have triggered further syntactic change. Trips and Stein identify two core research questions, which can in our view be taken as the starting point in evaluating all studies of this type: i) to what extent did the English system retain and integrate the argument structure of verbs copied from French? ii) did the argument structure of these verbs influence the argument structure of native verbs, and if so to what extent […]? (Trips and Stein 2019: 234).

We consider these questions in turn looking at two case studies: the use of to + NP to mark a formerly dative argument (Trips and Stein 2019) and the spread of Experiencer subjects in verbs of psychological state (Trips 2020; Trips and Rainsford 2022; García García and Ingham, this volume). The evidence that verbs borrowed from French retain their argument structure in Middle English, at least initially, is strong. Trips and Stein (2019) develop Allen’s (1995: 330) observations on the borrowing of Old

9  Overview of Contact-Induced Morphosyntactic Changes 

267

French plaire ‘please’ into Middle English, focusing in particular on the use of to + NP to mark the Experiencer (11), a structure unattested in Old English but obligatory in Old French with non-pronominal arguments (12): (11) For  God  wasted  þe  bones of  hem  þat  plesen    tomen for  God    rejected  the bones  of them  that please.3pl to men ‘because God rejected the requests of those who please men’ (Earliest Complete English Prose Psalter; Trips and Stein 2019: 251) (12) […] pour plere   mauvesement   au    monde      for    please.inf wickedly  to.the world ‘[…] to wickedly please the world’ (thirteenth century, continental Old French, La Somme le Roi; Trips and Stein 2019: 250).

A corpus study of the PPCME2 confirms that the first occurrences of plesen (M2, 1250–1350) are found with the French pattern of to + NP, as in (9). The native pattern with a non-prepositional NP Experiencer, on the other hand, is not attested until period M3 (1350–1420) (13): (13) men axed  hym how that men sholde plese the peple men asked him  how that men should please the people ‘people asked him how men should please the people’ (Chaucer, The Parsonʼs Tale; Trips and Stein 2019: 251).

A further example in which verbs borrowed from French clearly retain their argument structure in English is the case of labile change-of-state and change-of-location verbs (Ingham 2020).21 Lability is common in continental French and Anglo-French verbs of this type (Ingham 2020: 457), and this is reflected in borrowings into Middle English, where

 Labile verbs are defined as those whose direct object argument in a transitive construction may be realised as the subject in an intransitive construction without any change in the verb form, for example PDE you broke it / it broke (Ingham 2020: 447). 21

268 

G. Walkden et al.

twenty-five of the thirty verbs of change-of-state and change-of-location of French origin are labile (Ingham 2020: 459). Trips and Stein’s second research question, however, is both more central and more challenging to evaluate. One possible conduit for contact influence affecting the argument structure of native verbs is direct translation. Trips and Stein (2019: 254–256) show that in the Ayenbite of Inwyt, OFr. plaire is translated with the native verbs ME liken and quemen, or the earlier French borrowing ME paiien. With ME liken and quemen, but not ME paiien, the Old French à + NP construction (14a) can be translated into English by a to + NP construction (14b): (14) a.    par quoi    l’ame     plet  a  Dieu  by   which  the=soul  pleases   to   God  (La Somme le Roi; Trips and Stein 2019: 255) b.  […] hueruore  þe    zaule  to  god   like   wherefore      the soul    to   God like  ‘because of which the soul pleases God’  (Ayenbite of Inwyt; Trips and Stein 2019: 255).

While contact influence seems clear in this case, it does not ultimately change the argument structure of ME liken and plesen, which do not retain to + NP Experiencers in modern English. A second possibility is therefore to show that a successful change in the argument structure of a native verb is initially or predominantly attested in French-based Middle English texts. A number of such cases are found in the literature. Trips and Stein (2019: 259–260), for example, show that the emergence of to + NP Goal arguments with the native verb give are significantly more frequent in French-based texts in the M2 and the M3 periods of the PPCME2 (1250–1420). This is taken as evidence that French ‘served as a model language’ for the introduction of this construction, while at the same time its occurrence in other Middle English texts proves that it is not simply a translation effect (2019: 261). A further possibility is to compare the frequency of a particular argument structure configuration in sets of French-based versus native ME verbs. Ingham (2020) and

9  Overview of Contact-Induced Morphosyntactic Changes 

269

García García and Ingham (this volume) adopt this approach with different classes of labile verbs. In the case of change-of-state or change-of-­ location verbs, a very limited number of labile verbs in Old English (16%) contrast with a high proportion of labile verbs borrowed from French in Middle English (83%). Crucially, lability is also shown to have spread among Middle English verbs of native origin (67%), which Ingham (2020: 462) interprets as an expansion of lability in English under the influence of French. Yet caution is required: destroy-verbs and psych verbs show a similar, if temporary, increase in lability in Middle English, but here the same approach does not show higher rates of lability among French-based verbs. García García and Ingham (this volume) therefore reject a contact explanation of change in these verb classes.

5 Conclusion The three language contact situations surveyed in this chapter have all left their mark on the morphology and syntax of medieval English, but have done so in different ways. Widespread intermingling between speakers of British Celtic and Old English led to acculturation, bilingualism and eventually language shift, with the effects best characterised as instances of imposition. In the case of Norse, the close relationship between the two Germanic languages paved the way for koineisation, favouring those structures that were similar or identical in Old Norse and medieval English; at a later stage, imposition through shift probably also played a role. Finally, contact with French, while by no means a purely elite phenomenon as argued in earlier literature, was of a type conducive to borrowing rather than imposition, and therefore the morphological and syntactic changes that can be observed are less fundamental and are in part mediated by lexical transfer. Our overview highlights both the ubiquity of contact in shaping the grammar of medieval English and the diversity of outcomes and processes involved.

270 

G. Walkden et al.

References Ahlqvist, Anders. 2010. Early English and Celtic. Australian Celtic Journal 9: 43–73. Allen, Cynthia L. 1995. Case Marking and Reanalysis: Grammatical Relations from Old to Early Modern English. Oxford: Clarendon Press. ———. 1997. Middle English Case Loss and the ‘Creolization’ Hypothesis. English Language and Linguistics 1: 63–89. ———. 1998. Genitives and the Creolization Question. English Language and Linguistics 2: 129–135. van der Auwera, Johan, and Inge Genee. 2002. English do: On the Convergence of Languages and Linguists. English Language and Linguistics 6: 283–307. Bech, Kristin, and George Walkden. 2016. English is (Still) a West Germanic Language. Nordic Journal of Linguistics 39: 65–100. Björkman, Erik. 1900. Scandinavian Loan-Words in Middle English, Part 1. Halle: Max Niemeyer. Blake, Norman F. 1992. Introduction. In The Cambridge History of the English Language,  II: 1066–1476, ed. Norman F.  Blake, 1–22. Cambridge: Cambridge University Press. Bowern, Claire. 2008. Syntactic Change and Syntactic Borrowing in Generative Grammar. In Principles of Syntactic Reconstruction, ed. Gisella Ferraresi and Maria Goldbach, 187–216. Amsterdam: John Benjamins. Bradley, Henry. 1904. The Making of English. London: Macmillan. Buccini, Anthony F. 1992. Southern Middle English Hise and the Question of Pronominal Transfer in Language Contact. In Recent Developments in Germanic Linguistics, ed. Rosina Lippi-Green, 11–32. Amsterdam: John Benjamins. Buridant, Claude. 2000. Grammaire nouvelle de l’ancien français. Paris: Sedes. Capelli, Cristian, et al. 2003. A Y Chromosome Consensus of the British Isles. Current Biology 13: 979–984. Carver, Martin. 2019. Formative Britain: An Archaeology of Britain, Fifth to Eleventh Century AD. Cambridge: Cambridge University Press. van Coetsem, Frans. 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. ———. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter. Cole, Marcelle. 2014. Old Northumbrian Verbal Morphosyntax and the (Northern) Subject Rule, NOWELE Supplement Series, 25. Amsterdam: John Benjamins.

9  Overview of Contact-Induced Morphosyntactic Changes 

271

———. 2018. A Native Origin for Present-Day English They, Their, Them. Diachronica 35: 165–209. Crisma, Paola, and Susan Pintzuk. 2019. The Noun Phrase and the ‘Viking Hypothesis’. English Language and Linguistics 31: 219–246. Dance, Richard. 2012. English in Contact: Norse. In English Historical Linguistics: An International Handbook, II: 1724–1737, ed. Alexander Bergs and Laurel J. Brinton, 202–219. Berlin: De Gruyter Mouton. A Dictionary of the Older Scottish Tongue. 1931–2002. Ed. William Craigie, A. J. Aitken, James Stevenson, and Marace Dareau. 12 vols. Oxford: Oxford University Press. Emonds, Joseph E., and Jan Terje Faarlund. 2014. English: The Language of the Vikings. Olomouc: Palacký University. Erschler, David. 2009. Possession Marking in Ossetic: Arguing for Caucasian Influence. Linguistic Typology 13: 417–450. Evans, D.  Simon. 1976. A Grammar of Middle Welsh. Dublin: The Dublin Institute for Advanced Studies. Eyþórsson, Þórhallur. 1995. Verbal Syntax in the Early Germanic Languages. Unpublished doctoral thesis, Cornell University. Farrar, Kimberley, and Mari C. Jones. 2002. Introduction. In Language Change: The Interplay of Internal, External, and Extra-linguistic Factors, ed. Mari C. Jones and Edith Esch, 1–16. Berlin: Mouton de Gruyter. Fernández Cuesta, Julia M., and Inmaculada Senra Silva. 2008. Does Old Northumbrian Exist? Northern Varieties of Old English. In Multidisciplinary Studies in Language and Literature: English, American and Canadian, ed. María F.  García-Bermejo, Pilar Sánchez Giner, Consuelo Montes García, Elvira Pérez Iglesias Granado, and Juan Andrés Jurado Torresquesana, 49–58. Salamanca: Ediciones Universidad de Salamanca. Fernández Cuesta, Julia M., and Nieves Rodríguez Ledesma. 2020. Reduced Forms in the Nominal Morphology of the Lindisfarne Gospel Gloss: A Case of Accusative / Dative Syncretism? Folia Linguistica Historica 54: 37–65. Filppula, Markku, Juhani Klemola, and Heli Paulasto. 2008. English and Celtic in Contact. London and New York: Routledge. Fischer, Olga. 2006. On the Position of Adjectives in Middle English. English Language and Linguistics 10: 253–288. ———. 2013. The Role of Contact in English Syntactic Change in the Old and Middle English Periods. In English as a Contact Language, ed. Marianne Hundt and Daniel Schreier, 18–40. Cambridge: Cambridge University Press.

272 

G. Walkden et al.

Fischer, Olga, Hendrik De Smet, and Wim van der Wurff. 2017. A Brief History of English Syntax. Cambridge: Cambridge University Press. Hadley, Dawn. 1997. ‘And They Proceeded to Plough and to Support Themselves’: The Scandinavian Settlement of England. Anglo-Norman England 19: 69–96. Haeberli, Eric. 2010. Investigating Anglo-Norman Influence on Late Middle English Syntax. In The Anglo-Norman Language and Its Contexts, ed. Richard Ingham, 143–163. Woodbridge: York Medieval Press. Härke, Heinrich. 2003. Population Replacement or Acculturation? An Archaeological Perspective on Population and Migration in Post-Roman Britain. In The Celtic Englishes, ed. Hildegard L.C. Tristram, III, 13–28. Heidelberg: Universitätsverlag C. Winter. Heine, Bernd. 2009. Identifying Instances of Contact-Induced Grammatical Replication. In Topics in Descriptive and African Linguistics: Essays in Honor of Distinguished Professor Paul Newman, ed. Samuel Gyasi Obeng, 29–56. Munich: Lincom Europa. Higham, Nicholas. 1992. Rome, Britain and the Anglo-Saxons. London: Seaby. Holthausen, Ferdinand. 1913. Negation statt Vergleichungspartikel beim Komparativ. Indogermanische Forschungen 32: 339–340. Hotta, Ryuichi. 2009. The Development of the Nominal Plural Forms in Early Middle English. Tokyo: Hituzi Syobo Publishing. Ingham, Richard. 2009. Mixing Languages on the Manor. Medium Ævum 78: 80–97. ———. 2012. The Transmission of Anglo-Norman: Language History and Language Acquisition. Amsterdam: John Benjamins. ———. 2020. How Contact with French Drove Patient-Lability in English. Transactions of the Philological Society 118: 447–467. Jackson, Kenneth H. 1953. Language and History in Early Britain. Edinburgh: Edinburgh University Press. Joly, André. 1967. Negation and the Comparative Particle in English. Quebec: Les Presses de l’Université Laval. Keller, Wolfgang. 1925. Keltisches im englischen Verbum. In Anglica: Untersuchungen zur englischen Philologie,  I: Sprache und Kulturgeschichte, 55–66. Leipzig: Mayer & Müller. Keller, Jonas. 2020. The Leipzig-Jakarta List as a Means to Test Old English / Old Norse Mutual Intelligibility. NOWELE 73: 252–275. Kershaw, Jane, and Ellen C.  Røyrvik. 2016. The ‘People of the British Isles’ Project and Viking Settlement in England. Antiquity 90: 1670–1680.

9  Overview of Contact-Induced Morphosyntactic Changes 

273

Klemola, Juhani. 1997. Dialect Evidence for the Loss of Genitive Inflection in English. English Language and Linguistics 1: 349–353. Kluge, Friedrich. 1899. Geschichte der englischen Sprache. 2 vols. Berlin: Felber. Kroch, Anthony, and Ann Taylor. 1997. Verb Movement in Old and Middle English: Dialect Variation and Language Contact. In Parameters of Morphosyntactic Change, ed. Ans van Kemenade and Nigel Vincent, 297–325. Cambridge: Cambridge University Press. ———. 2000. The Penn-Helsinki Parsed Corpus of Middle English 2. Philadelphia: University of Pennsylvania, Department of Linguistics. Kroch, Anthony, Ann Taylor, and Don Ringe. 2000. The Middle English Verb-­ Second Constraint: A Case Study in Language Contact and Language Change. In Textual Parameters in Older Languages, ed. Susan C.  Herring, Pieter van Reenen, and Lene Schøsler, 353–391. Amsterdam: John Benjamins. Laing, Lloyd, and Jennifer Laing. 1990. Celtic Britain and Ireland, AD 200–800: The Myth of the Dark Ages. Dublin: Irish Academic Press. Laker, Stephen. 2008a. Changing Views about Anglo-Saxons and Britons. In Six Papers from the 28th Symposium on Medieval Studies Held at the Vrije Universiteit Amsterdam on 15 December 2006, ed. Henk Aertsen and Bart Veldhoen, 1–38. Leiden: Leiden University, Department of English. ———. 2008b. The English Negative Comparative Particle. Transactions of the Philological Society 106: 1–28. Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Legge, Mary D. 1963. Anglo-Norman Literature and Its Background. Oxford: Clarendon Press. Leslie, Stephen, et  al. 2015. The Fine-Scale Genetic Structure of the British Population. Nature 519: 309–314. Lucas, Christopher. 2012. Contact-Induced Grammatical Change: Towards an Explicit Account. Diachronica 29: 275–300. ———. 2014. Contact-Induced Language Change. In The Routledge Handbook of Historical Linguistics, ed. Claire Bowern and Bethwyn Evans, 519–536. London: Routledge. Lutz, Angelika. 2009. Celtic influence on Old English and West Germanic. English Language and Linguistics 13: 227–249. McIntosh, Angus. 1994. Codes and Cultures. In Speaking in Our Tongues: Proceedings of a Colloquium on Medieval Dialectology and Related Disciplines, ed. Margaret Laing and Keith Williamson, 135–137. Cambridge: D. S. Brewer.

274 

G. Walkden et al.

Millar, Robert McColl. 2000. System Collapse, System Rebirth: The Demonstrative Pronouns of English 900–1350 and the Birth of the Definite Article. Berlin: Peter Lang. ———. 2016. At the Forefront of Linguistic Change: The Noun Phrase Morphology of the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 153–167. Berlin: Mouton de Gruyter. Miller, D.  Gary. 2002. The Origin and Diffusion of English 3SG -S. Studia Anglica Posnaniensia 38: 353–362. ———. 2012. External Influences on English: From Its Beginnings to the Renaissance. Oxford: Oxford University Press. Morse-Gagné, Elise. 2003. Viking Pronouns in England: Charting the Course of THEY, THEIR, and THEM. Unpublished doctoral thesis, University of Pennsylvania. Myres, John. 1986. The English Settlements. Oxford: Oxford University Press. [OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). https://www.oed.com. Accessed 24 May 2023. Pons-Sanz, Sara. 2013. The Lexical Effects of Anglo-Scandinavian Linguistic Contact on Old English, Studies in the Early Middle Ages, 1. Turnhout: Brepols. Poussa, Patricia. 1982. The Evolution of Early Standard English: The Creolization Hypothesis. Studia Anglica Posnaniensia 14: 69–58. Ringe, Don, and Ann Taylor. 2014. A Linguistic History of English, II: The Development of Old English. Oxford: Oxford University Press. Rodríguez Ledesma, Nieves. 2016. Dauides sunu vs. filii david: The Genitive in the Gloss to the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 213–238. Berlin: Mouton de Gruyter. Rothwell, William. 1976. The Role of French in Thirteenth-Century England. Bulletin of the John Rylands University Library of Manchester 58: 445–466. ———. 1978. À quelle époque a-t-on cessé de parler français en Angleterre? In Mélanges de philologie romane offerts à Charles Camproux, II, 1075–1089. Montpellier: C.E.O. ———. 1998. Arrivals and Departures: The Adoption of French Terminology into Middle English. English Studies 79: 144–165.

9  Overview of Contact-Induced Morphosyntactic Changes 

275

———. 2001. The Teaching and Learning of French in Later Medieval England. Zeitschrift für französische Sprache und Literatur 111: 1–18. Samuels, M.L. 1985. The Great Scandinavian Belt. In Papers from the Fourth International Conference on English Historical Linguistics, Amsterdam, April 10–13, ed. Roger Eaton, Olga Fischer, Willem F. Koopman, and Frederike van der Leek, 269–281. Amsterdam: John Benjamins. Sawyer, Peter H. 1971. The Age of the Vikings. 2nd edn. London: Arnold. Schrijver, Peter. 2002. The Rise and Fall of British Latin: Evidence from English and Brittonic. In The Celtic Roots of English, ed. Markku Filppula, Juhani Klemola, and Heli Pitkänen, 87–110. Joensuu: University of Joensuu, Faculty of Humanities. ———. 2007. What Britons Spoke Around 400 AD. In Britons in Anglo-Saxon England, ed. Nick Higham, 164–171. Woodbridge: Boydell and Brewer. ———. 2014. Language Contact and the Origins of the Germanic Languages. London: Routledge. Schumacher, Stefan. 2007. Die Deutschen und die Nachbarstämme: Lexikalische und strukturelle Sprachkontaktphänomene entlang der keltisch-­germanischen Übergangszone. In Johann Kaspar Zeuss im kultur- und sprachwissenschaftlichen Kontext (19. bis 21. Jahrhundert), ed. Hans Hablitzel and David Stifter, 167–207. Vienna: Praesens Verlag. Siegel, Jeff. 1985. Koines and koineization. Language in Society 14: 357–378. Sims-Williams, Patrick. 1983. Gildas and the Anglo-Saxons. Cambridge Medieval Celtic Studies 6: 1–30. Small, George W. 1924. The Comparison of Inequality: The Semantics of the Comparative Particle in English. Baltimore: Johns Hopkins Press. Thomason, Sarah G., and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Los Angeles, CA: University of California Press. Timofeeva, Olga. 2010a. Anglo-Latin Bilingualism Before 1066: Prospects and Limitations. In Interfaces Between Language and Culture in Medieval England: A Festschrift for Matti Kilpiö, ed. Alaric Hall, Olga Timofeeva, Ágnes Kiricsi, and Bethany Fox, 1–36. Leiden: Brill. ———. 2010b. Non-Finite Constructions in Old English with Special Reference to Syntactic Borrowing from Latin. Unpublished doctoral thesis, University of Helsinki. ———. 2016. The Viking Outgroup in Early Medieval English Chronicles. Journal of Historical Sociolinguistics 2: 83–121.

276 

G. Walkden et al.

Tolkien, J.  R. R. 1963. English and Welsh. In Angles and Britons, 1–41. O’Donnell Lectures. Cardiff: University of Wales Press. Townend, Matthew. 2002. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English. Turnhout: Brepols. Trips, Carola. 2002. From OV to VO in Early Middle English. Amsterdam: John Benjamins. ———. 2014. The Position Proper of the Adjective in Middle English: A Result of Language Contact. In Adjectives in Germanic and Romance, ed. Petra Sleeman, Freek van de Velde, and Harry Perridon, 73–93. Amsterdam: John Benjamins. ———. 2020. Impersonal and Reflexive Uses of Middle English Psych Verbs Under Contact Influence with Old French. Linguistics Vanguard 6: 20190016. https://doi.org/10.1515/lingvan-­2019-­0016. Trips, Carola, and Achim Stein. 2019. Contact-Induced Changes in the Argument Structure of Middle English Verbs on the Model of Old French. Journal of Language Contact 12: 232–267. Trips, Carola, and Thomas Rainsford. 2022. Tolerating subject-experiencers? Yang’s Tolerance Principle Applied to Psych Verbs under Contact in Middle English. Journal of Historical Syntax 6 (12): 1–43. Tristram, Hildegard L. C. 2002. Attrition of Inflections in English and Welsh. In The Celtic Roots of English, ed. Markku Filppula, Juhani Klemola, and Heli Pitkänen, 111–149. Joensuu: University of Joensuu, Faculty of Humanities. Trotter, David. 2006. Language Contact, Multilingualism, and the Evidence Problem. In The Beginnings of Standardization: Language and Culture in Fourteenth-Century England, ed. Ursula Schaefer, 73–90. Frankfurt am Main: Peter Lang. ———. 2012. English in Contact: Middle English Creolization. In English Historical Linguistics: An International Handbook, ed. Alexander Bergs and Laurel J. Brinton, II, 1781–1793. Berlin: De Gruyter Mouton. Trudgill, Peter. 2010. Investigations in Sociohistorical Linguistics: Stories of Colonisation and Contact. Cambridge: Cambridge University Press. ———. 2011. Sociolinguistic Typology: Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. Vance, Barbara S. 1997. Syntactic Change in Medieval French: Verb-Second and Null Subjects. Dordrecht: Kluwer. Walkden, George. Forthcoming. Scandinavians and Verb-Second in Northumbrian Old English. In The Legacy of Medieval Scandinavian

9  Overview of Contact-Induced Morphosyntactic Changes 

277

Encounters with England and the Insular World, ed. Richard Dance, Sara M. Pons-Sanz, and Brittany Schorn. Turnhout: Brepols. Warner, Anthony. 2017. English-Norse Contact, Simplification, and Sociolinguistic Typology. Neuphilologische Mitteilungen 118: 317–404. Weinreich, Uriel. 1968 [1953]. Languages in Contact: Findings and Problems. The Hague: Mouton. White, Donald A. 1971. Changing Views of the Adventus Saxonum in Nineteenth and Twentieth Century English Scholarship. Journal of the History of Ideas 32: 585–594. Winford, Donald. 2003. An Introduction to Contact Linguistics. Oxford: Wiley Blackwell. ———. 2005. Contact-Induced Changes: Classification and Processes. Diachronica 22: 373–427.

10 Traces of Language Contact in Nominal Morphology of Late Northumbrian and Northern Middle English Elżbieta Adamczyk

1 Introduction In contrast to lexicon and more recently morphosyntax, morphology has hitherto received relatively little scholarly attention in studies exploring Norse influence on the structure of early English. One of the factors contributing to such a state of affairs is the fact that contact phenomena in the area of morphology are relatively less frequent and thus have largely been disregarded in scholarly research (Gardani 2020: 96). Another factor is that the effects of Scandinavian influence that extend beyond the lexicon are fairly difficult to capture (Dance 2012: 1736). The major structural feature of early English commonly ascribed to Anglo-­ Scandinavian contact of the eighth and ninth centuries is the linguistic innovativeness of the northern dialectal area (Northumbrian and Mercian) when compared to other dialects (e.g. Toon 1992: 429–433;

E. Adamczyk (*) University of Wuppertal, Wuppertal, Germany e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_10

279

280 

E. Adamczyk

Milroy 1992: 11, 181–183; cf. Allen 1995: 178–179). This refers not only to the lexical level, where the North shows a unique, Norse-­ dominated profile, with the degree of Norsification depending on time and region, but also to changes occurring at the structural level. The most prominent morphological features typically interpreted as emerging from contact with Old Norse involve simplification of the nominal categories such as grammatical gender, and generalisation of the plural marking in -(e)s (< OE -as) and the genitive ending -(e)s. The extension of the ending -(e)s is commonly believed to have been completed in the North and Midlands already in the twelfth and thirteenth centuries, respectively, in marked contrast to the situation in the South(-West) (Brunner and Johnston 1970: 48–49; Fulk 2017: 58–59; Millar 2002, 2016; Newman 2008: 116). Among the phonological features of tenth-century Northumbrian ascribed to Norse influence is the loss of the final ending -n after an unaccented syllable, visible in nominal and adjectival inflection, and infinitives (Brunner 1962: 108; Hogg 1992: 298–299; see also Walkden et al., this volume). The dissemination of the above-mentioned endings is an essential component of a more general process of analogical reorganisation of nominal inflection, which led to the reduction and eventual decline of declensional class diversity. This development will be the focus of the present study and the testing ground for the evaluation of the potential impact of Old Norse on early English nominal morphology. The process entailed a gradual analogical extension of productive inflectional markers to the paradigms of nouns affiliated with unproductive classes. The aim of the investigation is to evaluate whether the patterns of analogical restructuring identified in the northern material, i.e. in Late Northumbrian sources, diverge substantially from those found in the material of southern provenance from approximately the same period (early eleventh-­ century West Saxon), and if they do, whether this divergence can be ascribed to external factors, i.e. linguistic contact with Scandinavians. With a view to making the picture more complete, but also recognising the fact that morphological developments, which tend to have a slower pace than changes in the lexicon, may not yet be fully reflected in the Late Northumbrian texts, a further layer of analysis will involve an examination of the material from Cursor Mundi (c.1300), representing

10  Traces of Language Contact in Nominal Morphology 

281

Northern Middle English, but attested also in copies from the Midlands. Such a methodological approach will enable a comparison of the findings and thus allow a more precise assessment of the scale of language contact influence on early English morphology. This potential influence cannot only be expected to have involved a direct transfer of morphological material from Old Norse. Therefore, the study attempts to identify parallel developments in the two languages, potentially entailing structural (pattern) transfer, which may have enhanced the dynamics of analogical developments in the nominal system of Late Northumbrian and subsequently Northern Middle English. The findings of the analysis will additionally offer further insight into the nature of the linguistic contact between Old Norse and Old English. The chapter is organised as follows: Section 2 discusses theoretical aspects of contact-induced developments in morphology, zooming in on the Anglo-Scandinavian context. Section 3 offers a brief overview of the analogical restructuring of the Old English declensional system, focusing on the changes affecting the minor unproductive paradigms and their Norse parallels. Section 4 introduces the methodology of the empirical part of the study and presents the results of the investigation of the Late Northumbrian (Sect. 4.2) and Cursor Mundi material (Sect. 4.3). Section 5 summarises the findings and discusses them in the context of language contact theory.

2 Language Contact and Morphology According to the traditional view, morphology, and particularly inflectional morphology, is considered to be one of the more stable domains of language, more resistant to change due to language contact than other areas of grammar (e.g. Winford 2005: 387; Matras 2015: 48). This implies that transfer of morphological features must reflect a strong pressure of one language on another, which in turn presupposes a high intensity of contact in compliance with the borrowing scale postulated by Thomason and Kaufman (1988), where the scope of (structural) borrowing is a function of intensity of (cultural) contacts (cf. Matras 2015). Contact-induced morphological change entails typically direct structural

282 

E. Adamczyk

transfer, which, in contrast to lexical borrowing, is strongly constrained by a range of factors.1 Accordingly, morphological change is believed to occur essentially in contexts where the interacting languages are typologically similar (van Coetsem 1988, 2000; cf. Winford 2005: 387). Only then can a substitution of an element (morpheme) in the recipient language by one from the source language occur. This complies with the observations of Matras (2015: 75), according to whom, if diffusion of borrowed morphemes occurs (to inherited lexemes), it involves ‘structural modification of inherited forms based on morphology’, which is facilitated (and in fact only then possible) when the borrowed form and the inherited affix are structurally similar. Additionally, social and cultural factors have been viewed as predictors of contact-induced change. Consequently, contact-induced change in morphology must be considered a complex phenomenon where all these factors interact (Gardani 2020: 113). Importantly, morphological elements marking nominal plural category have a relatively high borrowability rating, which can be ascribed to the fact that plural is ‘a prototypical category of inherent morphology, thus closer to derivation (and the lexicon) than contextual inflection, and as such is characterised by a higher semantic load’ (Gardani 2020: 109; see also Gardani 2020: 116–117; Myers-Scotton 2002: 16–18). Finally, morphological (inflectional) borrowing can be enhanced by internal structural factors, involving, among others, reinforcement, i.e. replacement of endingless forms and phonetically weaker forms by stronger ones (Gardani 2020). The hypothesis that some morphological features or structural patterns may have been transferred from Old Norse into early English necessarily involves certain assumptions about the nature of the contact between the two communities. Given that borrowing beyond the lexicon is more likely when the level of bilingual proficiency is higher, traces of morphological influence would imply that contacts between communities must have been intense and involved a high degree of bilingualism between speakers. Such a constellation is more likely to occur in long-lasting  The degrees of borrowability differ for inflectional and derivational morphology, the former being less prone to contact-induced change (e.g. Thomason and Kaufman 1988: 74–75); for reasons of this resistance, see e.g. Matras (2015: 48 and 66). 1

10  Traces of Language Contact in Nominal Morphology 

283

contact-situations, facilitating structural borrowing. Another aspect of language contact to be considered in the context of morphological change is related to the effects of contact on the overall typological profile of a language. The effect of contact, involving short- or long-term bilingualism, as a factor leading to morphological simplification and complexification, respectively, has been widely recognised (e.g. van Coetsem 1988; Trudgill 2011). Linguistic complexity has been defined (among others) in terms of second-language acquisition and associated with the difficulty of language acquisition for adult learners (Trudgill 2011: 41; cf. relative complexity discussed by Miestamo et al. 2008). The loss of morphological categories and inflectional markers would be characteristic of morphological simplification. The mechanism underlying such simplification involves imperfect learning, which occurs in adult second-language acquisition, and is based on imposition, as defined in van Coetsem (1988: 26). In its extreme form, it may promote ‘catastrophic modification’ of the recipient language (van Coetsem 1988: 20; cf. Townend 2002 for the Anglo-Scandinavian context). Linguistic elements which typically get transferred in the process of language imposition belong to the more stable linguistic domains such as phonology and grammar. The interpretation of the manifold changes that English inflectional morphology underwent, encompassing both phonological reductions of vowels in inflectional syllables and ensuing analogical levelling in the transition period between Old and Middle English, can be framed in this model as a consequence of Anglo-Scandinavian contact. Such an interpretation has been questioned, e.g. by Trudgill (2014), according to whom simplification occurs typically in contexts when the interaction between two languages is short-term, which was not the case with the Anglo-­ Scandinavian contacts. Additionally, the acquisition, given that the contact was prolonged, must have taken place rather as (proficient) child language acquisition, which would typically lead to complexification rather than simplification of linguistic structures (see also Fox et al., this volume; and Walkden et al., this volume).2  Earlier contact with the Celtic substrate can arguably be a factor contributing to the incipient simplification, only later enhanced by the contact with Old Norse (White 2002; Tristram 2004; Miller 2012: 147; Warner 2017: 352–371; see also Walkden et al., this volume). 2

284 

E. Adamczyk

Given the structural proximity of Old English and Old Norse on many levels (phonological, morphosyntactic and lexical), the relevance of the current models of contact-induced change (relying largely on the evidence of languages more distant typologically and genetically) for the specific Anglo-Scandinavian contact-situation has been called into question (Dance 2012: 1727). Alternatively, the contact situation may have been one of ‘extreme “dialect contact”’, with new forms functioning as ‘variants of native ones’ in the recipient language (Dance 2012: 1727; cf. Townend 2002: 205–210). Additionally, in light of the relatively long time-span of interaction, the intensity of these contacts must have fluctuated over time (e.g. Townend 2002; Pons-Sanz 2013; Emonds and Faarlund 2014; Bech and Walkden 2016). An argument in favour of intense contact in tenth-century Northumbria comes from the lexical domain: not only do lexical loans appear very early in the Northumbrian material, but they also attest to an extensive use of non-technical Norse-­ derived vocabulary (Pons-Sanz 2013, 2016: 307). This implies close interaction between the two communities, which may have involved language shift, with the loans being the result of imposition rather than borrowing. It is such shift-based interference, entailing source language agentivity (van Coetsem 1988), that has been viewed as a potential underlying mechanism responsible for transfer of elements from more stable domains like morphology (Townend 2002: 201–207; see also Walkden et al., this volume). In view of the findings of the research exploring possible impact of language contact on morphology, and the observations about the nature of the Anglo-Scandinavian contacts, three different forms of morphological influence can be hypothesised for the Anglo-Scandinavian context, reflecting several levels of language contact influence: 1. Redistribution of inherited morphological affixes with a parallel development in the contact language (Old Norse), in compliance with their lexically or phonologically conditioned distribution in the contact language. 2. Borrowing of morphological affixes from the contact language, involving a ‘direct’ adoption of Old Norse morphemic material (cf. Dance 2012: 1734).

10  Traces of Language Contact in Nominal Morphology 

285

3. Extensive simplification of grammatical categories (case and gender) and allomorphic variation, and ensuing collapse of the inherited morphological system. The decline of grammatical gender, the reduction in case agreement and the dissemination of -(e)s as nearly the sole plural marker in Northern Middle English dialects may be taken as an expression of the third scenario, involving an ‘indirect’ result of simplifications and accommodations, associated with imperfect language acquisition (adult language learning) (cf. Dance 2012: 1734–1735). The other two scenarios involve long-term language contact effects, which can be expected to be found in the linguistic material from the area north of the Danelaw, approximately a century after the Danish invasion and settlement, and are associated with proficient child language acquisition and widespread bilingualism. The present study evaluates these scenarios in the two-stage procedure mentioned earlier. The comparison of the patterns of changes found in the early attestations with those identified for the period when the Scandinavian impact was more discernible in the textual material will allow the capture of the dynamics of changes in nominal morphology and evaluation of their potentially contact-induced origin. Any claims about the impact of contact with Scandinavians made in this study involve making inferences from the potential differences between the northern and southern varieties of Old and early Middle English. The investigation will thus aim to assess whether the limited dialectological evidence coincides with the geographical distribution of Scandinavian settlement arguing in favour of a contact-induced scenario for the restructuring of nominal inflection in early English.

3 Late Old English Nominal System on the Move and Its Norse Parallels The Late Old English nominal paradigms were subject to various analogical developments which led to loss of their transparency and ultimately large-scale restructuring of the system. Instigated by phonological

286 

E. Adamczyk

changes, especially reduction of the inventory of unstressed vowels in Late Old English (eleventh and twelfth centuries), some syncretisms emerged within and across paradigms. The former resulted in (near-)elimination of the inherited morphophonemic alternation in the paradigms, i.e. the i-mutated vowels and consonantal alternations, which served as case-number exponents, and substantial uniformisation of the paradigms. The latter triggered analogical extension of inflectional markers from the productive classes (masculine and neuter a-stems, feminine ō-stems and marginally n-stems) to the minor, unproductive paradigms (i-, u-, s-, r-, þ-, nd-stems and root nouns). Some interparadigmatic realignments can also be observed in the productive classes, e.g. the plural masculine a-stem maker -as is occasionally extended to n-stems and to neuter a-stems. These analogical reshufflings abound in Late Old English, and there is evidence for their origin in the North, where phonological changes were also more advanced (e.g. the reduction of final unstressed -n, the merger of final unstressed /e/ and /a/ with a compromise spelling , Hogg 1992: 242; Ross 1937). Given this overall progressive shape of Late Northumbrian English with respect to phonology and morphology, it can be expected that the dynamics of analogical transition of minor stems to productive classes were also different in the North and the South. The essential question in the context of potential contact with Scandinavian is to what extent parallel developments can be found in Old Norse. The analogical reorganisation of inflections is attested in the available Old Norse material too; however, the patterns are not parallel with those found in Late Old English. Old Norse paradigms display considerable formal variation across classes. The factor that complicates and delays analogical levelling is the widespread presence of mutated forms, both in the singular and plural. While the singular is often characterised by traces of u-mutation, the plural abundantly attests to i-mutation, appearing alongside the widespread plural ending -r, in most paradigms, including all minor stems except the i-stems, where the i-mutated vowel was levelled throughout the entire paradigm (e.g. u-stem hǫnd ~ hendr ‘hand’, root noun maðr—menn ‘man’, bók—bœkr ‘book’, r-stem faðir— feðr ‘father’, nd-stem gefandi—gefendr ‘giver’). The shape of the paradigms is complicated by further assimilatory processes, such as u- and a-­breaking (e.g. in the ō-stems, grǫf ‘ditch’, graf-, pl. grafar; u-stems vǫllr ‘ground’, pl.

10  Traces of Language Contact in Nominal Morphology 

287

vellir, gen. vallar; Gordon and Taylor 1981: 283–289). At the same time, there seems to be considerable homogeneity when it comes to the suffixal marking of the nom.pl., with the prevalent ending -(V)r (< Proto-Norse *-z; with varying vowel qualities). A major analogical interparadigmatic shift affected feminine i-stems and u-stems, which transferred to feminine ō-stems very early (Heusler 1967: 67). The most productive paradigm is the masculine a-stem, characterised by gen.sg. -s and nom.pl. -ar, as in armr ‘arm’ (Gordon and Taylor 1981: 283–284; cf. Heusler 1967: 61). Note that the paradigm shows only limited syncretism, and the acc.sg. and pl. do not equal the nom.sg. and pl., as it is the case in Old English. sg.

pl.

Nom. armr Gen. arms Dat. armi Acc. arm

armar arma ǫ rmum arma

The plural marker -ar also appears in the feminine ō-stems, which, after the merger with the (feminine) i-stems, became a productive template for feminine nouns, with a plural in -ir. Another productive class were the masculine and feminine n-stems, with two distinct cases in the singular (masc. nom. bogi ‘bow’—gen. / dat. / acc. boga, fem. saga ‘story’—sǫgu) and the marker -ar / -ur in the nom.pl., extended from the productive a-stems (nom. bogar, nom. / acc. sǫgur) (Heusler 1967: 70). Although no direct one-to-one formal correspondence exists in the inflectional patterns of Old Norse and Old English, one can observe a structural similarity in the plural paradigm, in that one inflectional exponent came to prevail across various classes: ON -Vr, OE -(a)s. Even if system-internal factors can explain many of the attested patterns and changes (as triggered by drift, e.g. Allen 1995), the fact that traces of inflectional reorganisation occur earlier in the North renders it plausible that language-contact may have accelerated the restructuring. With regard to other morphological developments, it has been assumed that Old Norse functioned as an agent of changes rather than a supplier of direct material for borrowing of morphological features (e.g. Miller 2012: 145; cf. Poussa 2006: 323–324). Accordingly, in the restructuring process, the potential impact is not likely to involve direct transfer of Old

288 

E. Adamczyk

Norse inflectional affixes, but rather to have an indirect effect, resulting, most likely, from imperfect learning (interference). At the same time, some structural innovations can be expected to have been mediated by lexical borrowing (Winford 2005: 386).

4 Patterns of Restructuring of Nominal Inflection in Late Northumbrian and Northern Middle English 4.1 Methodology of the Study The analysis of the data entailed two stages: First, the tenth-century Northumbrian material was examined and compared to the West Saxon textual sample deriving from approximately the same period. The former comprised the Lindisfarne Gospels gloss (Li; London, British Library, MS Cotton Nero D IV; c.970), the Rushworth Gospels gloss (Ru2; Oxford, Bodleian Library, MS Auct. D. 2. 19; late tenth century) and the gloss to Durham, Cathedral Library, MS A.IV.19 (Durham Ritual or Collectar; c.970); the latter was confined to the corresponding southern text of the West Saxon Gospels (WSCp; Cambridge, Corpus Christi College, MS 140; c.1025).3 The selection was made with a view to comparing relatively similar material in terms of content and dating. The Northumbrian texts, and especially the variation found in the Lindisfarne gloss, are expected to offer evidence for change in progress (see e.g. Fernández Cuesta and Pons-Sanz 2016). Although Late Northumbrian does not represent a homogeneous entity (Hogg 2006: 405–406; Fernández Cuesta 2016: 139), a fine-grained scrutiny of differences between individual texts goes beyond the purview of this study. The guiding question addressed in this part of the analysis is whether and to what extent the patterns and scope of analogical restructuring in nominal paradigms  The analysis was conducted on the text editions used in the Dictionary of Old English Corpus (DOEC). Also the title abbreviations of the texts come from this source. For criticism on relying on Skeat’s nineteenth-century edition of the Lindisfarne gloss, used in the corpus, see Fernández Cuesta (2016). 3

10  Traces of Language Contact in Nominal Morphology 

289

found in the Northumbrian and West Saxon sources diverge or agree with each other. The second part of the investigation entails an examination of patterns and dynamics of changes in a sample from the Northern Middle English poem Cursor Mundi, composed around the turn of the thirteenth century (Morris 1961–1966). As the text is attested in several manuscripts, including the northern copies and one more southern copy, it lends itself well to an investigation entailing a regional dimension, which can reveal potential dialectal disparities in the inflectional patterns.4 While the analysis of the Late Northumbrian material is concentrated on the changes in the unproductive classes, in which the analogical transformation was most advanced, the investigation of Cursor Mundi focuses on the residue inflections, zooming in on several more prominent patterns.

4.2 Morphological Archaisms and Innovations in Late Northumbrian The present section is devoted to an overview and a comparison of the tendencies found in the Old Northumbrian and West Saxon nominal systems with respect to the level of inflectional archaism (i.e. inherited inflection) and innovation (i.e. analogical inflection). The starting point for the analysis is the shape of the nominal paradigms as expected to have regularly developed from Proto-West Germanic (Adamczyk 2018: 113–116). The overview is organised in two parts reflecting two layers of analysis, qualitative and quantitative, which are aimed at identifying potential structural traces of contact-induced inflectional changes, and capturing their dynamics, as these could reflect the potential impact of the contact situation. The overview of tendencies for individual  Five of nine extant manuscripts are included in the edition of Morris (1874–1893), four of which represent Northern or North Midlands dialect. MS C (London, British Library, MS Cotton Vespasian A III) is the closest to the original (Horrall 1978); the other manuscripts in the edition include: Fairfax (F; Oxford, Bodleian Library, MS Fairfax 14), Göttingen (G; Göttingen University Library, MS theol. 107), Trinity College (T; Cambridge, Trinity College, MS R.3.8) (representing a southern Midlands version, copied around 1400), and fragments of the Edinburgh manuscript (E; Edinburgh, Royal College of Physicians). The focus of this analysis lay in MS C as compared to MS T (the latter representing a more southern variety), while the forms identified in other manuscripts were used additionally when they departed from those found in MSS C and T. 4

290 

E. Adamczyk

declensions adopts a taxonomy based on etymological criteria, even if its relevance is limited due to a range of analogical developments that had occurred already by the tenth century. The distinction with respect to syllable structure of nouns is made only occasionally, when relevant. For many subclasses there are no relevant forms attested which could provide information about the restructuring patterns. It must be noted that the i-stems, which were an unproductive class in Old English, are not included in the analysis, since the class underwent considerable restructuring already at the prehistoric stage.

u-Stems The u-stems are scantily attested in the investigated sample, with only a few nouns displaying vestiges of the etymological inflection. The transition of masculine u-stems to a-stem declension is almost completed in the heavy-syllable nouns, and archaisms are found primarily (though not exclusively) in the light-syllable stems. The acc.pl. endings in -o (hondo, duro, wintro), found in Li alongside historical forms in -a, are interpreted as a continuation of etymological -a (> -o). The original u-stem inflection is also present in Ru2, with forms of nom. / acc.pl. honda, dura, winter / -u(o, a). A sporadic innovation is found in the acc.pl. endingless form winter, which may have been triggered by the numeral, by analogy to endingless (collective) neuter a-stem gear ‘year’ (e.g. winter siofune, LkGl(Ru) 2.36). In contrast, the few singular forms mostly resemble the a- and ō-stem inflection. This can be seen in the  gen.sg.  eardes ‘earth’ (MkGl(Li) 13.27), dat.sg.  sune ‘son’ (e.g. JnGl(Li) 4.5, 5.26), earde (JnGl(Li) 4.44), honde ‘hand’ (e.g. MkGl(Li)14.58), cweorne ‘hand-mill’ (coernæ) (MtGl(Li) 24.41), and, very occasionally,  acc.sg. dure door (MtGl(Li) 7.13).5 Sporadic analogical inflections appear also in the paradigm of sunu (not included in Table  10.1 on account of its high frequency), which otherwise exhibits an archaic inflectional profile. The genitive ending -es is found in only one form, sunes (MtArgGl(Li) 3). The archaism of the genitive can be ascribed to ‘indeclinability’ of the singular  Archaic inflection of the u-stem dēað ‘death’, which otherwise transferred to the a-stems at the prehistoric stage, is found twice in JnGl(Ru): dat.sg. dēoða, 11.13 and 21.14. 5

50% (N = 22) 11% (N = 75) 96% (N = 25) 76% (N = 283) 91% (N = 11) 69% (N = 42) 65% 56%

18% (N = 17) 39% (N = 27) 100% (N = 6) 1% (N = 105) 100% (N = 1) 94% (N = 63) 37% 38%

sg.

30% (N = 10) 12% (N = 49) 35% (N = 31) 49% (N = 109) 97% (N = 32) 100% (N = 11) 42%

sg.

15% (N = 20) 4% (N = 24) 100% (N = 13) 91% (N = 22) 0% (N = 16) 67% (N = 3) 40%

pl.

a

The dental stems are excluded from the overview due to their scarce attestation in the selected material. The reorganisation of this class is very advanced, with analogical inflection reaching 73% in Late Old English (Adamczyk 2018: 234) b The plural does not include the thirty-four feminine tokens which show a formal overlap with the ō-stem ending -a (for a discussion see Adamczyk 2018: 169). Likewise, the counting excludes the attestations of relatively frequent sunu. In my analysis of the seventy-four relevant forms, four (6.3%) show analogical inflection (cf. Ross 1937). The feminine paradigm is dominated by relatively frequent hand

total

u-stemsb root nouns s-stems r-stems nd-stems (monosyl.) nd-stems (disyl.) total sg. / pl.

West Saxon pl.

Late Northumbrian

Table 10.1  The distribution of analogical inflections in minor classes in Late Northumbrian and West Saxon Gospelsa

10  Traces of Language Contact in Nominal Morphology 

291

292 

E. Adamczyk

paradigm of sunu, resulting from the merger of back vowels (Ross 1937: 79–82). The singly attested gen.pl. sunana (MtHeadGl(Li) 68) indicates a rare n-stem influence, possibly triggered by the analogical nom.sg. suna (alongside sunu, suno) (cf. Brunner 1965: 229). The presence of analogical inflections in sunu in Old Northumbrian is, however, meaningful given that in the entire DOEC, including the Late material, the instances which show innovation in the paradigm of sunu constitute only 6.5% of all forms (Adamczyk 2010: 379). The text of the West Saxon Gospels shows more innovation in this class, with plural forms such as dure (Jn(WSCp) 20.19) and wintre (Mk(WSCp) 5.42; instead of expected -a-ending), whose interpretation is ambiguous, as they can well be ascribed to phonological factors rather than analogy (or both). The influence of the n-stems appears only in the  dat.sg.  duran (Mk(WSCp) 1.32), found alongside a few archaic forms (wintra) (see Table 10.1). In Old Norse, this class is characterised by the presence of i-mutated vowels in the plural as in: hǫnd—hendr, dyr (plurale tantum). No direct structural parallels can be found in the Northumbrian material (cf. Sect. 4.3).

Root Nouns The root noun paradigm is the most conservative of all the unproductive declensions, with i-mutated plural forms retained until today. Analogical inflections in Northumbrian appear in heavy-syllable feminine nouns with the masculine marker -as (burgas, 50%). The masculine paradigm is dominated by the lemma fōt ‘foot’, which is entirely archaic in the plural (foet, Li, Ru2), and the only analogical form is the  dat.sg.  toðe ‘tooth’ (MtGl(Li) 5.38). In the plural the i-mutation is consistently retained except the acc.pl. fota.6 More insight into the shape of the dat.sg. can be gained from the feminine nouns, where the analogical -e appears occasionally (4.3%, e.g. næhte, DurRitGl1 36.15), alongside more frequent  The single attestation of foeta in LkGl(Li) 7.45 (L non cessauit osculari pedes meos, rendered as ne blann cossetunges ł foeta mine) has been interpreted as a contamination of the i-mutated nom. / acc. pl. and the regular gen.pl. in -a (Griepentrog 1995: 159), but it may well represent a genuine genitive plural form. 6

10  Traces of Language Contact in Nominal Morphology 

293

i-mutated (boec, byrig) and endingless forms (næht). A closer scrutiny of the data indicates that, for the feminine nouns, the relatively high innovation level is due to the innovative shape of burg; other lemmas, bōc ‘book’, gāt ‘goat’, mūs ‘mouse’, wlōh ‘fringe’ and (neuter) bū ‘dwelling’, show the inherited i-mutated plural (boec, gæt, mys, wloeh, by). Analogical masc.gen. -es appears only in neaht ‘night’ (næhtes, DurRitGl1 36.18; LkGl(Li) 2.8) (cf. gen.sg. boec, JnGl(Ru) 21.25). The paradigms of the West Saxon Gospels show a dichotomy with archaic plural (4% of innovation) and innovative singular (39%). The latter appears with the -e-­ending in the dat.sg. and a single gen.sg. nihtes (Mt(WSCp) 28.13), while in the former a weak form fotan (Jn(WSCp) 11.44) is singly recorded, but no analogical -as-forms are found. As the figures for the singular are very low, they may not reflect the actual state of affairs adequately. As regards the Old Norse patterns in this class, the noun burg has a feminine cognate in OIc borg, affiliated with the productive ō- / i-stem declension (pl. borgir), with a regular plural marker -ir and lack of i-­ mutation. I-mutated plurals are attested for OIc bók, fót and ló, which are also affiliated with root nouns in Old Norse, and where the i-mutated vowel is accompanied by -r in the plural: OIc bœkr, fœtr, lœr.

r-Stems The distribution of analogical inflections in this class is quite divergent for the singular and plural. The Northumbrian sources show a considerable share of analogical forms in the singular (76%), reflected in the extension of gen.sg. -es (also in the feminine paradigm, 54%: e.g. moderes LkGl(Li) 1.15, modres and MtHeadGl(Li) 54) and dat.sg. -e. In the plural, the preferred endings -u / -o (fadero, JnGl(Li) 6.31, broþero LkHeadGl(Li) 29, dohtero LkGl(Li) 23.28) appear alongside analogical forms in -as (fadoras, LkGl(Li) 6.23) and the historical endingless plurals (broþer, dohter). Significantly, Northumbrian evinces ablaut-alternation in the singular, which points to considerable phonological archaism of this paradigm (Hogg and Fulk 2011: 56–57). Subtle differences can be detected between individual texts: JnGl(Li) 6.31:nom.sg. / nom.pl. fæder—fadero, the rest of Li:  nom.sg. / nom.pl. fader / fæder—fadoras

294 

E. Adamczyk

(LkGl(Li) 6.23, 6.26); Ru2: nom.sg. / nom.pl. fæder—fædras / -es (LkGl(Ru) 2.49, JnGl(Ru) 6.49). In West Saxon the paradigms display opposite tendencies in the singular and plural, with the former being entirely conservative, retaining endingless genitive and endingless i-mutated dat.sg. forms, the latter very innovative, with the nom. / acc.pl. in -as (fadaras, Lk(WSCp) 6.23) and in -u / -o (broþru, Lk(WSCp) 14.25). A characteristic feature of Old Norse r-stems is the presence of i-mutated plurals: faðir—feðr, bróðir—brœðr. Li shows occasional occurrences of mutated feder, but no further parallels can be found. The shape of ‘sister’ has been traditionally considered to be an instance of phonological substitution through contact with Old Norse, cf. OE sweoster, swuster and OIc systir (Gersum, s.v. half-suster; OED, s.v. sister). The paradigm is, however, scantily attested and little can be concluded about analogical inflection therein. The presence of endingless genitive forms, which abound in early Middle English, has been considered a typical feature of the r-stems in Li (e.g. Rodríguez Ledesma 2016). These endingless forms, however, can arguably be interpreted as vestiges of the etymological pattern, where the genitive singular had no inflectional marker. Their appearance in the earlier material, and their frequent occurrence in West Saxon lend credence to such an interpretation. Additionally, the high token frequency of nouns of relationship may have acted towards conservation of historical inflection. A potential scenario of the  development at the later stage could have involved a reinforcement of the historical inflectional pattern and a merger with the new tendency occasioned by a rapid reduction in inflectional endings.

s-Stems The Old Northumbrian s-stems emerge as very innovative in the singular, but fairly archaic in the plural, where the consonantal formative element -r- is often preserved (65%: e.g. lombor(o), dogor, calfero). The innovative singular in Li is solely due to the hybrid dat.sg. / acc.sg. doege ‘day’, with archaic i-mutation, accompanied by an analogical ending -e (MkGl(Li) 2.20). Further innovative forms are found in the genitive

10  Traces of Language Contact in Nominal Morphology 

295

singular (cælfes, celfes, LiProlMt 14, 21; lombes, DuRitGl1 47.21). The extension of the masculine plural marker -as is evinced in the acc.pl. ehras ‘ear of grain’ (LkGl(Li) 6.1), and in the paradigm of cild, commonly affiliated with the s-stems. Here the nom. / acc.pl. cildas (MtGl(Li) 19.13), cildes (MtHeadGl(Li) 4) are found alongside neuter endingless plural cild (MkHeadGl(Li) 31) and acc.pl. cildo, which also lacks the r-formative (LkGl(Li) 18.15). A single weak gen.pl. cildena appears in DurRitGl1 104.5 along with cilda. While the retention of the formative -r- in this class is a common pattern in Northumbrian, the West Saxon inflection is much more innovative, and the original formative is consistently eliminated from the paradigm. It is also a scantily attested class, with the lemma cild dominating the attestations. West Saxon evinces endingless neuter plurals: cild (Mt (WSCp) 2.16), lamb (Lk(WSCp) 10.3), genitive singulars in -es (cildes, Mt(WSCp) 2.21) and dative singulars in -e (eare (Mk(WSCp) 4.28, cilde (Mt(WSCp) 2.8). In Old Norse the s-stems do not form a coherent class and traces of the historical pattern are not retained; they show endingless forms in the plural instead, which indicates that the analogical reorganisation of this class occurred much earlier (cf. OIc nom. / acc. sg. / pl. dœgr,  dat. sg. dœgri; nom. / acc.sg. lamb, nom. / acc.pl. lǫmb). Plurals in -Vr are in fact common for nouns affiliated with various declensions, but the Northumbrian -Vr-plurals have no direct corresponding forms in Old Norse. It seems feasible, however, that the preservation of the -r marker in the plural, which is much better evinced in Northumbrian than in West Saxon, may have been enhanced by the availability of the Old Norse plural forms in -r, which seem to have had the nature of a superstable plural marker there.

nd-Stems The nd-stems from two subclasses depending on the syllable structure: monosyllabic stems with i-mutation in the dative singular and nominative / accusative plural, and disyllabic stems without i-mutation. Analogical nominative / accusative plural appears mostly with plural

296 

E. Adamczyk

ending -as, and singly as frionde (nom.pl., LkHeadGl (Li) 46) and friondo (acc.pl., LkGl(Li) 2.44). The genitive and dative plural forms were considered neutral as their inflections are ambiguous, but it is worth noting that a single adjectival form byendra ‘dweller’ is recorded in DurRitGl1 (121.22). The plural paradigm is very innovative, with 97% of analogical forms in monosyllabic stems and no archaisms retained in the disyllabic stems. The genitive singular shows no variation, with -es-ending throughout the paradigm, while in the dative singular a competition between a zero maker and -e-ending is found. Also the lack of i-mutated forms in Northumbrian points to the innovative inflectional profile of this class. The West-Saxon material turns out to be much more conservative in the monosyllabic stems; no analogical forms are found in the nominative / accusative plural where the i-mutated variant is consistently retained (except one non-mutated form) and the singly attested genitive singular form shows the -es ending (feondes, Lk(WSCp) 10.18). The disyllabic subtype, in contrast, is substantially affected by analogy, with the genitive and dative singular dominated by a-stem inflections, and the plural evincing some more variation. None of the patterns found in the analysed material corresponds to the development of these stems in Old Norse, where the singular was restructured on the template of the n-stems, while the plural followed a consonantal pattern with i-mutated vowel in the nominative and accusative, and the dominant -r plural marker (e.g. gefandi—gefendr).

Summary of Major Tendencies Table 10.1 presents the quantitative distribution of analogical inflections in individual classes in tenth-century Northumbrian, juxtaposed against the inflections from early eleventh-century West Saxon material. The figures for individual case-number categories are aggregated, with the contrast between the singular (genitive and dative) and plural (nominative and accusative). Clearly, the reorganisation of the nominal inflection in minor classes had a greater scope in Late Northumbrian. This is particularly visible in the singular paradigm of u-stems and r-stems, and in the plural of

10  Traces of Language Contact in Nominal Morphology 

297

nd-­stems. The notable archaism of root nouns is attested for both examined varieties. At the same time, considerable discrepancy can be observed in the restructuring of the singular and plural in Late Northumbrian, with the plural being in general more conservative. This refers not only to the root nouns, but also s-stems and u-stems. The pattern in the s-stems is particularly interesting in the context of potential Old Norse influence, as it stands in sharp contrast to the innovative profile of the West Saxon material. The retention of the -r-marker in the plural can be interpreted as an effect of the availability of the Norse -(V)r plural marker, whereby the native historical inflection is reinforced by structural similarity to the corresponding Norse pattern. What can be then concluded about the potential traces of Norse influence? Patterns of restructuring of minor classes do not evince any explicit effects of direct borrowing of morphological material from Old Norse; however, several structural parallels could be identified. In the root nouns the inflectional detachment of burg from the root noun paradigm in contrast to simultaneous retention of historical inflection in other lemmas affiliated with this class corresponds to the Norse situation, where borg(ir) no longer belongs to the root noun declension. Further, the plural exponent -Vr in Old Norse, just like the productive plural marker -es in English, seems to have had the nature of a ‘superstable marker’ (Dammel and Nübling 2006) in that most of the nouns from various historical declensions tended to adopt it. In Late Northumbrian the dissemination of the -as-ending is still only incipient (e.g. masc. r-stems, nd-stems). It is primarily the classes marked by less salient inflectional exponents, such as zero or vocalic endings, that become susceptible to superstable markers.7 The factor triggering analogical extension of inflections thus may have been the functionality of the system, working towards maintaining a grammatical contrast (in number). Given the dating of the loss of -(e)n endings in the North, the incipient changes leading to the dissemination of the -(e)s plural may have occurred prior to the Old Norse input and may have been thus system-internal (cf. Warner 2017: 333–335). At the  This refers also to the class of n-stems, which were not investigated here. The loss of final -n (in Late Northumbrian) rendered their inflections vocalic and thus potentially more prone to replacement by the salient inflection -as (Hogg and Fulk 2011: 125–126). 7

298 

E. Adamczyk

same time, the spread of the superstable plural marker -es could have been enhanced by a parallel development, the generalisation of -Vr ending, in the corresponding Norse system. A similar interpretation was postulated by Hotta (2009: 57, 195), who shows that the incidence of s-plurals ranged from 30 to 36% for Northumbrian and Mercian, respectively, compared to 11 and 13% of n-plurals in South-West Midlands. Contact with Old Norse, Hotta argues, contributed to the extension of s-plurals in English on account of the parallel formation of the Norse plural with the marker -R (later -r), which had a similar distribution to the English -s plural.8 It must be observed that the analogical influence of the p ­ roductive n-stems is very limited in the examined sources. This is true also of the West Saxon material, where, given the later developments in the south, traces of incipient analogical pressure from this class could be expected.

4.3 Morphological Archaisms and Innovations in Cursor Mundi (Manuscripts C, T) One of the inflectional features characterising the Middle English period is the advancing spread of the analogical nominative / accusative plural marker, with historical n-stem ending -en (< -an) in the South, and the a-stem -(e)s (< -as); the latter was extended to nouns of other classes first in the North and North Midlands, and in the South in the thirteenth (Southwest) and fourteenth (Southeast) centuries (Wełna 2012: 417). The restructuring of the nominal inflection at this stage is quite advanced, with only the genitive and the plural being inflectionally marked (e.g. Brunner and Johnston 1970: 3). Given the limited scope of this study, the ending -e, occasionally found in the dative singular and potentially plural is not taken into account in the investigation due to its problematic phonological status (Fulk 2017: 48).

 Phonetic similarity between OE -s and ON -r has also been postulated (Hotta 2009; Miller 2012), adducing the origin of ON -r < *z and its merger with /r/ from *r. As the completion of the merger is dated by the end of the early Viking Age (Nielsen 2000: 258), this direct correspondence is debatable, in contrast to the postulated functional correspondence, which seems a more likely explanation (also Warner 2017). 8

10  Traces of Language Contact in Nominal Morphology 

299

The examination of restructuring patterns in this part of the study was confined to a few samples extracted from Cursor Mundi, with special emphasis on the Cotton manuscript (C) (c.1340), showing a strong northern linguistic profile, and the Trinity College (T) manuscript, which has a more southern character (Morris 1961–1966; see fn. 4). The manuscripts Fairfax (F) and Göttingen (G), both also with a northern flavour, were examined as well, but did not testify to any substantially divergent inflectional patterns. The following sections from Book 3 (Morris 1961–1966) were analysed exhaustively for relevant genitive and plural forms (of all nouns, irrespective of historical affiliation): ll. 19301–19748 (448 lines) and ll. 19973–20254 (282 lines). Additional search queries in Books 1, 2 and 3 were performed for lemmas originally affiliated with minor stems, including: hō̆ nd(e), sone, dōr(e), bōk, night, fā̆ der, brōther, doughter, mōder, suster, chīld / bā̆ rn, lō̆ mb, frēnd, fēnd. Of these, hō̆ nd(e) was covered exhaustively for the attestations in Book 3. The Middle English Dictionary (MED) was used as a starting point to identify inflectional forms of nouns that could be relevant for the present analysis (many historical minor stems had disappeared from the lexicon by then).

Default Plural Marking Prior to discussing the inflection of the selected lemmas, a few general observations are in order. Both in the northern manuscript C and the southern manuscript T, the s-plural is the dominant, default plural marker, with three allomorphs: -s, -es and -is, whose distribution is largely defined by synchronic phonological conditions. This allomorphic alternation has no relation to the historical declensions, nor is it in any way related to Old Norse inflectional morphology and will therefore not be considered. The weak -en-ending was only found in two historical -n-stems: eien (C), eȝen (T) (ll.19437, 19691) and eren (F, T; l. 19452, not found in C, but in F attested as eres), and in these forms no clear North-South contrast can be identified. Both varieties (as represented by C and T) share the endingless forms in a couple of historical neuter heavy-syllable a-stems, such as thing / þing ‘things’ (l. 20159) and yeir / ȝeer ‘years’ (l. 20026), which is a feature found also in early twentieth-century northern English dialects.

300 

E. Adamczyk

Genitive Marking A contrastive feature of the northern varieties is lack of the -s-ending in the genitive of many lemmas, irrespective of the original class affiliation (cf. l. 20177C: Er tu mi sun messeger vs. l. 20177T: Art þou my sones messangere). The lack of relation of this pattern to any specific historical declension implies an innovation. As Old Norse shows regular -s-­genitives, and modern Scandinavian languages such as Danish and Norwegian have a phrasal genitive -s, the absence of the ending can therefore not in any direct way be related to Scandinavian influence.

Inflection of Historical Minor Stems in Cursor Mundi u-Stems The analysed sample contained the historical u-stem lemmas dō̆ r(e), hō̆ nd(e), and sone. The northern form of sone is sun, in the South attested as son (OIc sonr). The endingless gen.sg. sun is found in C, while the southern T regularly shows sones. In the plural, the marker -(e)s is found regularly: suns / sones, dors / dores. The attestations of hō̆ nd(e) display a different inflectional pattern, with a clear dialectal contrast. The southern forms (T) are: honde (sg.) and honde(s) (pl.), where the lack of ending -s in the plural testifies to relative inflectional archaism. C shows hand in the singular and i-mutated hend in the plural, with incidental instances of handes. Such i-mutated plural forms are characteristic of this lemma in Old Norse: hǫnd—hendr, which could point to Norse influence and an extension of this structural pattern. The fact that this lemma appeared often in the collocation ‘hands and feet’ (cf. OIc fótr—fœtr) may also have played a role, triggering analogical extension of the mutated vowel to hō̆ nd(e) (see also Lass et al. 2013–: s.v. $hand/n). These language external and internal factors may have worked synergetically, supporting the mutated plural variant. Alternatively, the form can also be interpreted as an instance of direct lexical borrowing (OED, s.v. hand). The fact that no parallel extension of the mutated pattern occurred in the other u-stems, dō̆ r(e) and sone (cf. OIc dyr, synir), seems indeed to speak in favour of this interpretation.

10  Traces of Language Contact in Nominal Morphology 

301

r-Stems Nouns of relationship fā̆ der and mōder are relatively infrequently recorded in forms relevant for the present study. The plural of mōder is found with the -s-marker in both manuscripts (l. 8596C moders, l. 8596T modris, cf. endingless moder in F). Some variation between the -s marker in C and zero marking in T is identifiable, although the single endingless attestation in T, which could point to a retained historical inflection, is problematic, as it does not unambiguously refer to a plural meaning (l. 6814C moders, l. 6814T modir; cf. G moderis, vs. F moder). An endingless genitive singular form is found for the lemma fā̆ der in the southern manuscript (l. 8538C his faders bij vs. l. 8538T his fadir by). The plural of brōther with analogical i-mutation, breþer (broiþer), appears in both C and T, with no sharp geographical contrast identifiable. Also for doughter, a form which may point to i-mutated plural, with an additional -es-­ marker, is attested: deghteres (l. 9623C; cf. l. 9623T douȝtres), but it seems to be the only form found along various instances with an unmutated vowel, also in C (doughtirs, dohutris, doghtres). Similar i-mutated forms are also recorded in other Middle English texts (MED, s.v. doughter). Weak forms in -en were not found in the analysed material, but the attestations of suster in MED evince two weak plural forms (l. 9532Ld sustyrne and l. 9684Ld sustirn) in Oxford, Bodleian Library, MS Laud Misc. 416, which was not part of the analysis, attested alongside -es in l. 9532C sisteres. The Norse r-stems are characterised by i-mutation in the plural, which is absent from the Old English forms, e.g. OIc brœðr, OE brōðer, brōðru. If this feature were to be evaluated as an instance of contact-induced influence, it seems that this irregular allomorph was most successful in the lemma brōther, and to some extent in doughter, which are relatively frequent in the plural (cf. single attestation of nom.pl. doehter */dø:xtər/ in MtGl(Li) 21.5, out of four forms). Significantly, mutated vowels were present in the Old English forms of these lemmas in the dative singular, which may have been an additional internal source for the dissemination of i-mutation in the plural, although this direction of analogy is cross-­ linguistically rather rare (but cf. German Stätte ‘place, location’, derived from i-mutated dat.sg. form of Stadt ‘town, city’ in the locative function; OHG nom.sg. stat— dat.sg. stetti).

302 

E. Adamczyk

Root Nouns Historical root nouns in the sample that preserve irregular plural are fōt, man and tōth: fete / feet, men / men, tethe / teeþ. Interesting here are instances of endingless pl. night after a numeral: thre night / þre nyȝt (cf. collective þousonde ȝeer, l. 20026T). Plurals in -(e)s were not found in the examined material, except in MS F (l. 14941F sex niȝtes). The morphology of bōk is difficult to interpret. Both C and T attest to bo(o)k and bokes, but in various contexts where T shows bokes, one finds an endingless bok(e) in C, as in l. 20775: C sum book sais, vs. T somme bokes seis. In l. 4178, for instance, a plural reading seems to be the most logical interpretation: C als in boke we rede vs. T as we in bokes fonde. However, in certain plural contexts with numerals (ll. 8460 and 21205) also C shows bokes. An i-mutated plural form of bōk was not found, while Old Norse, just as Old English, shows i-mutation in the plural of this lemma.

s-Stems The sample contained genitive singular forms of lō̆ mb: l.21677 C lambs, T lambis, and a plural lambes (10380C). The lemma chīld / chīlde shows a historical plural childer / children, and there are no differences between the northern and southern texts in that in both the -r formative surfaces. In two instances, the northern texts have forms of barn, bern vs. T childe (ll. 22341 and 22811), but barn also appears incidentally in T (20050). The parallel with Old Norse (cf. barn—bǫrn) is lexical rather than inflectional, and no plural forms of barn were identified in the sample.

nd-Stems Monosyllabic nd-stems, frēnd, fēnd, are regularly endingless and mutated in the plural in Old English. Endingless plurals of frēnd appear in 20245C freind (cf. freinde; MED, s.v. frēnd) alongside more frequent freindes (20183C), and they are not found in the parallel text of T, where frendes

10  Traces of Language Contact in Nominal Morphology 

303

(20183T) is consistently attested. Old Norse has a mutated plural frændr, but it does not seem to enhance the retention of the historical inflection.

Summary of the Patterns in Cursor Mundi The analysed material allows us to conclude that, with respect to the productivity of the -(e)s-marker, there are no substantial differences in plural morphology between C and T.  Further overlapping patterns entail the restriction of the -en-plural to only ēre and eie, the preservation of several endingless plurals, the -r-plural in chīld, as well as a handful of mutated plural forms, still found in modern English. The major differences were identified in the inflection of four lemmas: • The plural of hō̆ nd(e) has an i-mutated form in the North and tends to be endingless in the South, alongside -(e)s-plurals in both varieties; • The plural of frēnd can appear as an archaic endingless form in the North, alongside an analogical, more frequent form in -es; • The plural of doughter appears once with a mutated root vowel in C, but not in the other manuscripts, whereas a mutated plural of brōther is common in all texts, including T; • The plural of bōk can be endingless in C, but not in T. Moreover, the dialects differ with respect to the distribution of gen.sg. -(e) s, which tends to be less common in C. The nominal morphology of the northern manuscript does not emerge as radically different from that of the southern manuscript (South Midlands). In fact, the former tends to preserve (or develop) a number of irregular forms not found in the language of the southern version. No traces of a massive transfer or influence of Norse plurality patterns are identifiable, but positive evidence of contact with Norse is present in several innovative plural-allomorphs, which may have been enhanced by contact on (incidental) lexical basis. The attestation of hend, which has no source in Old English, may have been triggered by Norse hendr, possibly as a lexically based feature rather than a structural phenomenon, and supported by the inherited fote—fete contrast. Similar incidental i-mutated

304 

E. Adamczyk

forms in brether and degtheres can be viewed as inflectional pattern extension or transfer. Here the i-mutated forms existed already in the dative singular, and their appearance in the plural may have involved a morphological enhancement by the Old Norse inflectional pattern. In all three cases, the phonologically regular forms in Middle English would have been endingless (OE handa, brōðru, dohtor; Li hondo, broðro, dohter(o)) due to progressive final vowel apocope in Middle English. The resulting loss of inflection obliterated the formal marking of singular-plural contrast and rendered these forms vulnerable to analogical pressure or potential external influence. Taking the evidence from other declensional classes into account, it can be concluded that this morphological number ambiguity was mitigated by employment of two strategies, namely (a) the application of the regular -(e)s plural marker (as in suns, handes, lambes) or / and (b) transfer or enhancement of mutated plural pattern (as in brether, hend and a hybrid formation deghteres).

5 Conclusion The present chapter investigated potential traces of Old Norse influence in the nominal morphology of Late Northumbrian and early Northern Middle English. The aim of the study was to evaluate whether the patterns and dynamics of analogical changes in the nominal inflection found in both periods in the northern material were different from those in the South, which could potentially be indicative of contact-induced change. The focus of the investigation lay on the inflectional shape and restructuring of inflection of nouns traditionally affiliated with unproductive declensions. The results of the study indicate that several inflectional patterns or tendencies of early English, whose distribution coincides with the area of Scandinavian settlement, may be ascribed to contact with Old Norse, and its effects are manifested on several levels. One of them is the more abstract structural level, where the redistribution of inflections shows a parallel model in English and Norse. This entails the parallel in the spread of the superstable plural marker -es, mirroring the abundance of the Norse ending -Vr, only incipient in the Late Northumbrian material. A further structural parallel is potentially identifiable in the retention

10  Traces of Language Contact in Nominal Morphology 

305

of the r-plural marking in the historical s-stems, attested for Late Northumbrian (and not for West Saxon), where the pattern can be interpreted as reinforced by the presence of the Old Norse inflectional marker -Vr. A similar enhancing effect can be adduced for the plural marking by i-mutation extended to a number of lemmas in Northern Middle English (hende, brether, deghter(es)). The latter parallel seems to be strongly lexically conditioned and in the case of hende likely to be an instance of direct lexical borrowing. In fact, ‘direct’ adoption of Old Norse morphemic material (e.g. an inflectional ending) could not be identified in the analysis. The retention of endingless r-stem plurals may have been supported by a corresponding endingless Old Norse inflection, additionally marked by i-mutation, which is later evinced in Cursor Mundi. The third level of interaction refers to the issue of contact-induced changes in the dynamics and scope of inflectional restructuring. Diffusion and more rapid dynamics of inflectional changes in the North is only partly confirmed in the analysed data. The patterns identified in the Late Northumbrian material indicate that the restructuring of the plural paradigm was more advanced in some classes but not in others, compared to the West Saxon data. The lack of substantial differences may result from the delayed visibility of the Anglo-Scandinavian influence in textual material, and the nature of the examined material, i.e. Bible translations, which in general tend to represent a formal genre, possibly characterised by more conservatism than other genres, even if individual texts show already some innovation. The patterns of restructuring found in the nominal morphology of Cursor Mundi are understandably much more advanced; however, quite remarkably, the discrepancies between the northern and southern varieties are less prominent. In fact, the plural morphology of the northern manuscript(s) of Cursor Mundi is characterised by more variation than the southern (South Midland) text. Accordingly, the higher incidence of the s-plurals in the north, which is traditionally interpreted as contact-induced, is not directly corroborated in the analysed Cursor Mundi sample, but it is the other mentioned inflectional features that point to external influence. As regards the implications of these findings for the nature of Anglo-­ Scandinavian interactions, the identified traces of external influence on nominal inflection in the northern texts imply that the contact between

306 

E. Adamczyk

the speakers of Old Norse and Old English must have been intense. The described developments and identified patterns appear to comply with the scenario of koineization, whereby similar features and patterns between the two languages are emphasised and enhanced, while differences tend to be circumvented (cf. Millar 2016; Warner 2017; also Fox et al., this volume; and Walkden et al., this volume). Significantly, the structural parallelisms lead to instances of activation of the unproductive patterns of i-mutation and -r-plurals, and bring about a relative increase in complexity of plurality patterns rather than simplification, which tends to be triggered typically in the context of long-term bilingualism (cf. Miestamo et al. 2008; Trudgill 2011). This is in particular visible in the material from Cursor Mundi. Koineization as a simplification-­inducing mechanism, involving considerable levelling of variation, is better reflected in the Late Northumbrian material, although the scale of simplification is still limited there. In this sense, the differences between the southern and northern varieties at this early stage could broadly fit into the scenario of imposition effects, with some, though limited, external impact involving imperfect language acquisition. At the same time, the analysed material seems to indicate that—when sufficiently zoomed in— the process of morphological simplification was not straightforwardly unidirectional, but rather nuanced, with local and temporary instances of apparent complexification and resistance to analogy.

References Adamczyk, Elżbieta. 2010. Morphological Reanalysis and the Old English u-Declension. Anglia: Zeitschrift für englische Philologie 128: 365–390. ———. 2018. Reshaping of Nominal Inflection in Early Northern West Germanic Languages: A Comparative Corpus Study, NOWELE Supplement Series, 31. Amsterdam: John Benjamins. Allen, Cynthia. 1995. Case Marking and Reanalysis. Oxford: Oxford University Press. Bech, Kristin, and George Walkden. 2016. English is (Still) a West Germanic Language. Nordic Journal of Linguistics 39: 65–100. Brunner, Karl. 1962. Die englische Sprache. 2 vols. Tübingen: Max Niemeyer.

10  Traces of Language Contact in Nominal Morphology 

307

———. 1965. Altenglische Grammatik. Tübingen: Max Niemeyer. Brunner, Karl, and Grahame Johnston. 1970. An Outline of Middle English Grammar. Oxford: Blackwell. van Coetsem, Frans. 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. ———. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: C. Winter. Dammel, Antje, and Damaris Nübling. 2006. The Superstable Marker as an Indicator of Categorial Weakness? Folia Linguistica 40: 97–113. Dance, Richard. 2012. English in Contact: Norse. In English Historical Linguistics: An International Handbook, ed. Alexander Bergs and Laurel J. Brinton, II, 1724–1737. Berlin: De Gruyter Mouton. [DOEC =] The Dictionary of Old English Corpus in Electronic Form. 2009. Ed. by Antonette diPaolo Healey, Joan Holland, David McDougall, Ian McDougall, and Xin Xiang. Toronto: Toronto University Press, on CD-ROM. Emonds, Joseph Embley, and Jan Terje Faarlund. 2014. English: The Language of the Vikings. Olomouc: Modern Language Monographs. Fernández Cuesta, Julia. 2016. Revisiting the Manuscript of the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 257–285. Berlin: De Gruyter. Fernández Cuesta, Julia, and Sara M.  Pons-Sanz, eds. 2016. The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, Anglia Book Series, 51. Berlin: De Gruyter. Fulk, Robert D. 2017. An Introduction to Middle English, Grammar and Texts, reprinted with corrections. Peterborough, ON: Broadview Press. Gardani, Francesco. 2020. Morphology and Contact-induced Language Change. In The Oxford Handbook of Language Contact, ed. Anthony Grant, 96–122. Oxford: Oxford University Press. [Gersum =] Dance, Richard, Sara M. Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge, Cardiff and Sheffield. https://www.gersum.org/. Accessed 28 May 2023. Gordon, Eric V., and Arnold R. Taylor. 1981. An Introduction to Old Norse, 2nd edn. Oxford: Clarendon Press. Griepentrog, Wolfgang. 1995. Die Wurzelnomina des Germanischen und ihre Vorgeschichte. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft.

308 

E. Adamczyk

Heusler, Andreas. 1967. Altisländisches Elementarbuch,  3rd edn. Heidelberg: C. Winter. Hogg, Richard. 1992. A Grammar of Old English, I: Phonology. Oxford: Wiley-Blackwell. ———. 2006. Old English Dialectology. In The Handbook of the History of English, ed. Ans van Kemenade and Bettelou Los, 395–416. Oxford: Blackwell. Hogg, Richard M., and Robert D. Fulk. 2011. A Grammar of Old English, II: Morphology. Oxford: Wiley-Blackwell. Horrall, Sarah M., ed. 1978. The Southern Version of Cursor Mundi. Ottawa: University of Ottawa Press. Hotta, Ryuichi. 2009. The Development of the Nominal Plural Forms in Early Middle English. Tokyo: Hituzi Syobo. Lass, Roger, Margaret Laing, Rhona Alcorn, and Keith Williamson. 2013–. A Corpus of Narrative Etymologies from Proto-Old English to Early Middle English and accompanying Corpus of Changes, version 1.1. Edinburgh: The University of Edinburgh. http://www.lel.ed.ac.uk/ihd/CoNE/CoNE.html. Accessed 28 May 2023. Matras, Yaron. 2015 [2014]. Why Is the Borrowing of Inflectional Morphology Dispreferred? In Borrowed Morphology, ed. Francesco Gardani, Peter M. Arkadiev, and Nino Amiridze, 47–80. Berlin: De Gruyter. [MED =] Middle English Dictionary. 1952–2001. Ed. by Robert E.  Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner. Ann Arbor: University of Michigan Press. https://quod.lib.umich.edu/m/med. Accessed 28 May 2023. Miestamo, Matti, Kaius Sinnemäki, and Fred Karlsson, eds. 2008. Language Complexity: Typology, Contact, Change, Studies in Language Companion Series, 94. Amsterdam: John Benjamins. Millar, Robert McColl. 2002. After Jones: Some Thoughts on the Final Collapse of the Grammatical Gender System in English. In Studies in English Historical Linguistics and Philology: A Festschrift for Akio Oizumi, ed. Jacek Fisiak, 293–306. Bern: Peter Lang. ———. 2016. At the Forefront of Linguistic Change: The Noun Phrase Morphology of the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 153–168. Berlin: De Gruyter. Miller, D. Gary. 2012. External Influences on English: From Its Beginnings to the Renaissance. Oxford: Oxford University Press.

10  Traces of Language Contact in Nominal Morphology 

309

Milroy, James. 1992. Middle English Dialectology. In The Cambridge History of the English Language, II:  1066–1476, ed. Norman Blake, 156–206. Cambridge: Cambridge University Press. Morris, Richard, ed. 1961–1966. Cursor Mundi: A Northumbrian Poem of the XIVth Century, Early English Text Society, o.s. 57, 59, 62, 66, 68, 99, 101. London: Oxford University Press. Myers-Scotton, Carol. 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Newman, John G. 2008. The Spread of the s-Plural Formative in Old and Middle English Nouns. Warsaw: Institute of English Studies, University of Warsaw. Nielsen, Hans Frede. 2000. The Early Runic Language of Scandinavia: Studies in Germanic Dialect Geography. Heidelberg: C. Winter. [OED =] Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 28 May 2023. Pons-Sanz, Sara M. 2013. The Lexical Effects of Anglo-Scandinavian Linguistic Contact on Old English, Studies in the Early Middle Ages, 1. Turnhout: Brepols. ———. 2016. A Study of Aldred’s Multiple Glosses to the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 301–328. Berlin: De Gruyter. Poussa, Patricia. 2006. The Relatives Who and What in Northern East Anglia. In Types of Variation: Diachronic, Dialectical and Typological Interfaces, ed. Terttu Nevalainen, Juhani Klemola, and Mikko Laitinen, 321–350. Amsterdam: John Benjamins. Rodríguez Ledesma, Nieves. 2016. Dauides sunu vs. filii david: The Genitive in the Gloss to the Lindisfarne Gospels. In The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, ed. Julia Fernández Cuesta and Sara M. Pons-Sanz, 213–238. Berlin: Mouton de Gruyter. Ross, Alan S.C. 1937. Studies in the Accidence of the Lindisfarne Gospels. Leeds: School of English Language. Thomason, Sarah Grey, and Terrence Kaufman. 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press.

310 

E. Adamczyk

Toon, Thomas E. 1992. Old English Dialects. In The Cambridge History of the English Language, I: The Beginnings to 1066, ed. Richard M. Hogg, 409–451. Cambridge: Cambridge University 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. Tristram, Hildegard. 2004. Diglossia in Anglo-Saxon England or What Was Spoken Old English Like? Studia Anglica Posnaniensia 40: 87–110. Trudgill, Peter. 2011. Sociolinguistic Typology: Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. ———. 2014. Contact-Related Processes of Change in the History of the English Language. In The Cambridge Handbook of English Historical Linguistics, ed. Merja Kytö and Päivi Pahta, 318–334. Cambridge: Cambridge University Press. Warner, Anthony. 2017. English–Norse Contact, Simplification, and Sociolinguistic Typology. Neuphilologische Mitteilungen 118: 317–404. Wełna, Jerzy. 2012. Middle English: Morphology. In English Historical Linguistics: An International Handbook, ed. Alexander Bergs and Laurel J. Brinton, I, 415–434. Berlin: De Gruyter Mouton. White, David L. 2002. Explaining the Innovations of Middle English: What, Where, and Why. In The Celtic Roots of English, ed. Markku Filppula, Johani Klemola, and Heli Pitkänen, 153–174. Joensuu: Joensuu University Press. Winford, Donald. 2005. Contact-Induced Changes: Classification and Processes. Diachronica 22: 373–427.

11 Origin and Spread of the Personal Pronoun They: La Estorie del Evangelie, a Case Study Marcelle Cole and Sara M. Pons-Sanz

1 Introduction There are few features of the English language that have traditionally epitomised the influence of Old Norse on English more than the 3pl. personal pronouns they, their, them. The modern-day forms derive from the þei(-), þai(-) and þe(-) type 3pl. pronouns that appeared in Middle English and gradually replaced the reflexes of the Old English pronouns hīe, hira, him. The traditional view holds that the Middle English þ-type pronouns derive from the Old Norse 3pl. pronouns þeir, þeira, þeim. The transfer of such grammatical material is considered indicative of the intensity of Anglo-Scandinavian contact and the typological proximity of

M. Cole University of Utrecht, Utrecht, Netherlands e-mail: [email protected] S. M. Pons-Sanz (*) Cardiff University, Cardiff, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_11

311

312 

M. Cole and S. M. Pons-Sanz

the two languages  (Thomason and Kaufman 1988: 274–304). Recent studies, however, indicate that there is scope for a re-evaluation of the Norse influence on the development of English (e.g. Pons-Sanz 2013; Dance et  al. 2019; see also Versloot’s paper in this volume). This also applies to the origin of they, their, them. Developing some tentative suggestions in the literature (e.g. Ogura 2001), Cole (2018) is the first to systematically challenge the pronouns’ Norse origin. She argues that the Northern Middle English þ-type pronouns derived from the Old English demonstratives þā (þāra, þām) via a process of reanalysis from demonstrative to personal pronoun. Internal changes to the functional distribution of the Old Northumbrian demonstratives thus created a system that lined up with the distribution of the Norse cognate forms and was reinforced, but not triggered, by language contact (see also Cole Forthcoming). By shifting the focus away from the categorical þ-type paradigm of Northern Middle English discussed in Cole (2018) to investigate  the mixed h-type and þ-type paradigms of the West and South-West Midlands, this chapter offers a new account of the dynamics of the pronominal replacement process and their implications for the etymology of the þ-type pronouns. We depart from scholarly tradition in two respects. In terms of etymology, we do not presuppose  two common  assumptions:  (i) the þ-type pronouns were the product of language contact alone and (ii) the appearance of þ-type pronouns beyond the heavily Scandinavianised areas was the result of dialect contact alone. Not taking these assumptions for granted and adopting the innovative, multifaceted methodology outlined below, enables us to challenge the seemingly strong foundations on which the common understanding of the pronouns’ etymology has rested for many years. Our findings indicate that the þ-type pronouns were the result of polygenesis, i.e. that reflexes of Old English demonstrative pronouns co-existed with and were reinforced by their Norse-derived cognates. To investigate the replacement process itself, we examine the linguistic, stylistic and scribal pathways that led to it. Several studies analyse the internal and external factors that motivated the substitution process based on pronoun usage either in a particular text / dialect (e.g. Johannesson 1995; Ritt 2001; Bergs 2005) or in a collection of texts, thus providing diatopic and diachronic scope but not in-depth analysis of

11  Origin and Spread of the Personal Pronoun They 

313

textual choices (e.g. Smith 2001; Morse-Gagné 2003; Stenroos 2005). The present study takes a different  approach, focusing instead on the pronominal usage recorded in the seven manuscripts of the Middle English poem known as La estorie del evangelie (hereafter Estorie). The examination of 3pl. pronouns in various dialectal rewritings of a single text provides us with directly comparable data that allow for the exploration of the pronouns’ distribution and diffusion in relation to other dialects. This focus also opens up several layers of analysis, including the potential influence of linguistic and stylistic factors, and that of scribal behaviour. We study how scribes from different dialectal areas responded to the (near-)categorical þ-type system of the early exemplars of Estorie, and  how scribes from the same dialectal area adapted (and possibly adopted) the different pronominal systems of their exemplars at a local level. At the intrascribal level, our method allows for a comparison of pronominal patterns in Estorie to those in other texts penned by the same scribe, a task facilitated by the electronic version of the Linguistic Atlas of Late Mediaeval English (eLALME; Benskin et al. 2013–). In addition to providing a diatopic and scribal perspective, we also contextualise pronoun usage in Estorie diachronically—particularly in the South-West and West Midlands versions—by comparing pronoun use in relevant versions to earlier usage in the same dialectal areas on the basis of the Linguistic Atlas of Early Middle English (LAEME; Laing 2013–). The paper is structured as follows. Section 2 introduces the text and its manuscripts. Section 3 uses our novel method to thoroughly scrutinise the 3pl. pronoun forms found in the various versions of Estorie and takes a nuanced approach to considering how scribal responses to the exemplar interact with linguistic and stylistic conditioning. The paper concludes with a summary of findings in Sect. 4.

2  La estorie del evangelie and Its Manuscripts Estorie is a metrical account of Christ’s life. In spite of its Anglo-French title, which derives from the brief introduction to its fragment in the Vernon manuscript, it is an early Middle English text. Its exact date and place of composition remain unknown, although its origin is generally

314 

M. Cole and S. M. Pons-Sanz

placed in late thirteenth-century North-West Norfolk  /  North-East Cambridgeshire / South Lincolnshire because of the date of its earliest witness and the linguistic features shared by the various manuscripts (Millward 1998: 56–64). Perhaps because of its fragmentary character, and in spite of what it has to offer for our understanding of Middle English dialectology (e.g. Pons-Sanz 2021), this text has  received hardly  any scholarly attention from historical linguists, who tend to focus on other compositions from its general dialectal area, such as the Ormulum, Genesis and Exodus or Havelok the Dane. Estorie is preserved in seven manuscripts, none of which is a copy of another (Millward 1998: 45):1 D = Dulwich College, MS XXII (519 lines—beginning to l. 528—in fols 81v–85v; c.1300, South Lincolnshire; LAEME dulwicht.tag): the fragment of Estorie is the only English text written by the relevant scribe. This version of Estorie is the closest to the original in dialect and date but it seems to incorporate a number of additions (e.g. ll. 69–144, 151–178 and 269–282), as these lines, which are unique to D, differ in content and metrical structure from those around them (see Millward 1998: 25). P  =  Oxford, Bodleian Library, MS Rawlinson C 655 (262 lines—ll. 853–1057, 1388–1389, 1727–1844—in various folios; c.1350, Somerset; eLALME LP 5220): the whole manuscript, written in one hand, contains a version of the Northern Passion that integrates lines from Estorie. This version of Estorie is not directly connected to any of the other extant texts. The language of the manuscript shows various dialectal layers, with northern forms co-existing alongside Midland and southern forms (Foster 1914: 24–26). Both Estorie and the Northern Passion show dialectal mixture. This suggests that the Somerset scribe relied (directly or ultimately) on an exemplar copied further north than South Lincolnshire  /  North-West Norfolk which had already incorporated lines from Estorie into the Northern Passion (Millward 1998: 73). While copying his exemplar, the scribe substi Millward’s (1998) edition is the primary source for our discussion of the textual coverage of the fragments and stemmatic relations. Line numbers refer to her reconstructed text. 1

11  Origin and Spread of the Personal Pronoun They 

315

tuted some of his southern forms, which resulted in a Mischsprache (e.g. we find occasional examples of the southern voicing of /f/ together with typically northern unpalatalised instances of the modal auxiliary shal). V  =  Oxford, Bodleian Library, MS Eng. Poet a.1, also known as the Vernon Manuscript (396 lines—beginning to l. 574—in fol. 105r–v; c.1390, North Worcestershire; eLALME LP 7630, based on a sample of texts by Scribe B not including Estorie): this manuscript is the work of two scribes. Our scribe, known as Scribe B, copied most of the approx. 350 texts in this impressive collection of religious texts, while Scribe A was responsible for rubrics, foliation and some materials in the first quire (eLALME LP 7670; see further Horobin 2013). The texts copied by Scribe B originated from different dialectal areas: e.g. ll. 9899–10818 from Robert Mannyng’s Handlyng Synne (South Lincolnshire), with some adaptation to a western audience (Perry 2013: 86–87, fn. 46); texts by Richard Rolle (South Yorkshire); and a copy of the A-text of Langland’s Piers Plowman, a text with Worcestershire connections but probably copied from an eastern exemplar (Samuels 1985). The texts also have chronological depth: e.g. besides Estorie, the manuscript includes a copy of Ancrene Riwle, which was probably originally penned in Herefordshire around two centuries before. Despite this diversity, the fact that the scribe’s language can be associated with late fourteenth-century North Worcestershire suggests that he (and / or the authors of his exemplars; cf. Doyle 1990: 9–10) undertook both dialectal translation and modernisation to make the texts comprehensible and relevant to contemporary local audiences. However, his texts also show significant phonological and morphological variation, probably as a result of constrained scribal response to the exemplars (Görlach 1974: 103; Horobin and Smith 2011; Smith 2013, 2020: 85–88).2 Notably, though, Smith (2013: 63) points out  Benskin and Laing (1981: 75) define constrained usage as ‘the accommodation of a scribe’s own repertoire to that of the exemplar, which accommodation does not extend to the reproduction of exotic forms’ and explain that ‘the relative frequencies obtaining between variants in the scribe’s spontaneous usage may be much altered by adherence to the usage of an exemplar’. Although they focus on orthographic and phonological matters, this explanation can also account for morphosyntactic and lexical issues. 2

316 

M. Cole and S. M. Pons-Sanz

that this variation does not go beyond what one would expect in South-West Midland usage. This observation also applies to the scribe’s pronominal usage (see Sect. 3.1). S  =  London, University of London Library, MS Sterling V. 17, also known as the Clopton Manuscript (1764 lines—beginning to l. 1974—in fols 97v–111v; c.1400, North Worcestershire; eLALME LP 7650): this manuscript was part of a larger manuscript written in a single hand which also comprised Washington, D C, Folger Library, MS V.b.236 and Princeton University Library, R. H. Taylor MS 10. The Sterling part also includes a copy of the C-text of Piers Plowman and the Assumption of Our Lady, the Folger part includes Mannyng’s Handlyng Synne and Meditations on the Supper of Our Lord and the Taylor part has a copy of Mandeville’s Travels. The texts exhibit different dialects: while Handlyng Synne and Meditations were copied in a non-uniform East Midlands dialect (Perry 2007: 142), the other texts reflect a range of West Midlands dialects; eLALME places Estorie in North-West Worcestershire but Piers Plowman in South-West Worcestershire (eLALME LP 7780), the area to which Perry (2007: 141) attributes the language of Mandeville’s Travels. He suggests that the West Midlands dialect in Assumption could also be associated with Worcestershire, although this association is less certain (2007: 141–142). This leads him to conclude that it is likely that ‘the scribe copied these texts literatim rather than imposing dialectal conformity on the exemplars used’ (2007: 142) and that we might actually be dealing with a scribe from London (2007: 138, 140, 142). The Clopton and Vernon manuscripts share a number of texts; while their copies of various other texts (e.g. Handlyng Synne; see Sullens 1983: xxii) are not directly related, those of Estorie can be said to go back to a shared exemplar (see further Turville-Petre 1990; and Sect. 3). B = Oxford, Bodleian Library, MS Additional C 38 (1703 lines—beginning to l. 2435—in fols. 71v–82r; 1410–1420, Worcestershire): Estorie is here preceded and followed by parts of the South English Legendary. Even though its exemplar is likely to share the same source as the exemplar of V and S, this version of Estorie differs significantly from those in V and S because it has been shortened and its language has been normalised and modernised, as is the case with the other texts in

11  Origin and Spread of the Personal Pronoun They 

317

the manuscript (Görlach 1974: 106; Millward 1997; and Pons-­ Sanz 2021). R  =  London, British Library, MS Royal 17 C XVII (501 lines—ll. 1905–2441—in fols 152v–155v; early fifteenth century, North Lincolnshire; eLALME LP 45, based on a sample of texts not including Estorie): this large late-medieval collection of devotional texts including Estorie was penned by a single hand and exhibits overall dialectal consistency across the various texts (Millward 1998: 73–75). Thus, even though eLALME’s linguistic profile is not based on the extract from Estorie, the forms for the 3pl. pronouns that it reports (mainly yai for they, yam, for them and yer for their, with for ) are aligned with those in our text. L  =  London, British Library, MS Lansdowne 388 (515 lines—ll. 1879–2404—in fols 373r–380v; fifteenth century, Central East Midlands; Pickering 1972: 46): besides the lines from Estorie, fols 368r–380v, also by the same hand, include a collection of doctrinal, moralistic and medicinal prose texts, and an account of how to interpret dreams known as Daniel’s Dreams. Most of the didactic material remains unedited and the short treatise on dreams (Bühler 1962: 266–267) only includes two examples of 3pl. pronouns, viz. two h-type objective forms (ll. 6 and 20). Accordingly, it is difficult to place the pronominal use in our text into a wider context of scribal trends. L and R share a number of readings for Estorie not found in the other versions, but they are less closely aligned than V and S, or even S and B.

3 Third-Person Pronoun Usage in the Estorie Texts 3.1 General Distribution of the Data in Their Manuscript Context Table 11.1 summarises the pronoun variants for they, their, them in the seven versions of Estorie; all instances of the relevant pronouns were

þ-type

þey (50) þei (1)

þei (2)

þei (148)

þey (3) þay (1) þei (1) þe (1)

þei (15) þe (2)

þai (40) þei (1)

þey (44) they (6) þei (4)

MS

S (N. Wor., c.1400)

V (N. Wor., c.1390)

B (Wor., 1410–1420)

P (Somerset, c.1350)

D (S. Lincs., c.1300)

R (N. Lincs., early 15th cent.)

L (Central East Midlands, 15th cent.)

they

_

he (1)

_

hii (10)

he (2)

heo (18) he (1)

heo (47) he (40) e (1) a (1) ȝeo (1)

h-type

þer (1) þere (1) there (1)

þer (8) þare (2)

þer(e) (6) þeire (5)

þar (2)

_

_

_

þ-type

their

Table 11.1  Þ-type and h-type pronouns in the versions of Estorie

here (6) her (5)

here (1)

_

here (4) hire (1) hore (1)

her (38)

heore (9) heor (2) her (1)

here (42) her (6)

h-type

þem (8)

þam (33)

þeim (6) þeym (1) þem (1)

_

_

_

_

þ-type

them

hem (36) hemselue (1)

hymselfe (1)

hem (2)

hem (9) hom (1)

hem (89)

hem (10)

hem (84)

h-type

318  M. Cole and S. M. Pons-Sanz

11  Origin and Spread of the Personal Pronoun They 

319

identified and studied.3 Throughout, we use þ-type to denote pronouns in both and , and h-type to denote pronouns in (including rare , and ). The distribution of forms for they, their, them across the seven manuscripts is, in the main, as expected on the basis of what we know about the textual history and origin of the manuscripts (see Sect. 2). The data are consistent with the general picture in the literature on the dialectal development of 3pl. forms: the replacement of h-type pronouns by the þ-type occurs first in the nominative, followed, more often than not, by the possessive and lastly the object forms (e.g. Morse-Gagné 2003: 134–138; Bergs 2005: 103). The more northerly D and R, despite their significant age gap, share a near-categorical þ-type system across the whole paradigm. Until LAEME’s compilation, D was often attributed to North-West Norfolk because McIntosh (1987) identified a number of features commonly found in North Norfolk texts in D’s copy of Estorie. However, its pronominal system is closer to that of the Ormulum (Johannesson 1995), an early Middle English text commonly attributed to South Lincolnshire (Laing 1978: 20–23; Parkes 1983), than to that of near-contemporary Norfolk manuscripts. In fact, Norfolk significantly lagged behind in the use of þ-type forms even by the end of the Middle English period (Morse-Gagné 2003: 149; Bergs 2005: Chapter 2; Cole Forthcoming). The Central East Midlands manuscript L has categorical þ-type subject forms but a mixed oblique paradigm, with h-type forms outnumbering þ-type forms (cf. Morse-Gagné 2003: 148–149). In the Somerset text P, h-type forms predominate, but þ-type forms occur in the subject and even the possessive. The pronouns in Estorie are in keeping with the set of forms that Foster (1914: 25) cites for the whole manuscript: they thai, þey, þei, þe, hii, hy; their here, hire, hore; them þem, hem (cf. eLALME LP 5220, which, for the Northern Passion text,  Some versions of Estorie have 3pl. pronouns where other versions have 3sg. pronouns in corresponding lines. As a result, the ambiguous form he occasionally posed problems of interpretation. Based on verbal inflexion and / or contextual clues, he at l. 410V, ll. 1828B, 2310B, l. 2178R was tagged as plural; he at ll. 1579B, 1581B and l. 397SVB was considered 3sg.; and he at l. 2251R, which involved a referent that could be singular or plural (ilke a man ‘each man’), was excluded. We follow LAEME in considering here at l. 418D a fem. sg. possessive (cf. Millward 1998: 68, who considers it a 3pl. possessive). 3

320 

M. Cole and S. M. Pons-Sanz

records they hii, hy, þey; their hire, hore, here, þar; them hem, hom, þem). As a whole, the language of P’s Estorie can be said to represent a Mischsprache (see Sect. 2), but the extent to which its pronominal forms should be analysed as an example of dialect mixing is less obvious. The distribution of h-type and þ-type subject forms is consistent with usage found in V and S, and early South-West and West Midlands Middle English (see below), and is suggestive of variation within the confines of the scribe’s own (passive) repertoire. However, the occurrence of possessive þar appears to be indicative of dialect mixing.4 The Worcestershire texts V, S and B have categorical h-type oblique forms but subject forms that range from the (near-)categorical þ-type in B to near-categorical h-type in V and a mix of variants in S. The exemplar ultimately shared by V, S and B is likely to have been an early Middle English Worcestershire rewriting of a northerly copy; it probably had h-type pronouns predominately for they, but also low-frequency þ-type forms, and only h-type pronouns for their and them. While the predominance of subject þ-type forms in B can be easily explained as one of the outcomes of its consistent modernising, V and S reflect the mixed paradigm of the exemplar more closely.5 The distribution that we suggest for the exemplar is supported by the pronominal forms shared by V and S in the section where they overlap up to l. 574. Subject h-type forms are almost categorical, except at l. 45, where both S and V have a þ-type form. The line is virtually identical across V, S and B, which also suggests that the þ-type pronoun was in the exemplar. eLALME’s profile for V’s Scribe B, viz. LP 7630, records a mixed paradigm for the subject, with a predominance of þ-type forms (together with some examples of heo and a), and h-type forms for the object and the possessive. Close examination shows some variation across the sampled texts: e.g. þ-type forms are categorical in the legends of St Bernard and St Paula, while the dialogue Ypotis includes thirteen þ-type forms and three h-type forms (Horstmann 1878: 3–8 and 41–61, and 1881: 341–348).  There is only one instance of a þ-type form for their recorded in the South-West and West Midlands in early Middle English: þaire in LAEME text tr323ct.tag, Central Herefordshire, 1250–1275 (see Cole Forthcoming). 5  In l. 1605B modernisation has been taken one step too far: B seems to have replaced an original h-form (cf. he in S) with þei even though the pronoun was 3sg., referring to Christ. 4

11  Origin and Spread of the Personal Pronoun They 

321

However, not all the texts that he copied have this general profile: in his version of Ancrene Riwle (Zettersten and Diensberg 2000) and the texts from the South English Legendary h-type forms are near-categorical in the subject (see e.g. Morris 1871: 18–61; cf. Görlach 1974: 102–103). This suggests that this scribe was happy to retain h-type forms fairly consistently (but see below). The pronominal distribution posited for an early Middle English Worcestershire copy is in line with that found in West and South-West Midlands texts from the period 1150–1325. A quantitative study of 3pl. subject pronoun systems in this area based on LAEME (Cole Forthcoming) suggests that þ-type pronouns already occurred as a low-frequency variant for they. Even though þ-type forms are not recorded in Worcestershire in early Middle English, the North-East Gloucestershire text LAEME iacobt.tag and the North-West Gloucestershire text LAEME digby86mapt.tag have five cases of þ-type forms each (vs. 63 and 135 instances of h-forms, respectively). Similarly, þa occurs in LAEME tr323ct (Herefordshire); þai in LAEME layamonBOt.tag (North-West Wiltshire); and þei in LAEME laud108at.tag (Oxfordshire). The þ-type subject pronouns in early Middle English West and South-­ West Midlands texts can arguably be traced back to þæge and þage, late West Saxon variant forms of the Old English demonstrative þā. These variants are found in various late West Saxon texts fulfilling the same independent and dependent functions as OE þā, including that of plural determiner, antecedent to a relative clause and personal pronoun. Thus, þæge and þage have been argued to be the source of þai(e), a very low-­ frequency form found for those (pronoun), the and they in the text of Laȝamon’s Brut recorded in the thirteenth-century manuscript London, British Library, MS Cotton Otho C XIII (LAEME layamonBOt.tag). Þai is also recorded for the in the other manuscript of Laȝamon’s Brut, viz. London, British Library, MS Cotton Caligula A IX, from nearcontemporary Worcestershire (Morse-Gagné 2003: 231). The phonological and morphosyntactic evidence thus supports a native origin for early Middle English (South-)West þai(e) (cf. Förster 1941, 1942; Dance 2003: 456–457; Morse-Gagné 2003: 217–219, 231; Pons-Sanz 2013: 397–398; Cole 2018: 189–191). The West Midlands þey-here-hem system in late Middle English can therefore ultimately be traced back to

322 

M. Cole and S. M. Pons-Sanz

usage in late West Saxon and early Middle English, although the use of þ-forms for the subject and the rest of the paradigm is also likely to have been boosted by þ-forms from the North and East following interdialectal contact. Establishing the existence of low-frequency þ-type forms for they, but not for their and them in early Middle English in the South-West Midlands allows us to explain why an early Middle English Worcestershire scribe working on the ultimate exemplar of V, S and B would have categorically replaced oblique þ-type forms with h-type forms, and would have replaced most—but not all—of the þ-type subject forms with h-type forms. Thus, we propose that the scribal practice in the exemplar, and by extension in S and V, represents a single coherent system and an example of ‘constrained’ scribal response to an exemplar. The same can be argued for the use of þ-type subject forms in P (see Sect. 2). In their overlapping section, S and V only differ at l. 51, where S has he and V þei. It is impossible to identify the source of the disparity. However, it is likely that V’s pronoun represents a later substitution, on the one hand, because this scribe appears to have been more willing to alter and adapt his texts in terms of both content and linguistic features than the S scribe (cf. Doyle 1990: 9–10; Turville-Petre 1990; and Smith 2013, 2020: 85–88); and, on the other, because the subject þ-type forms, and not the h-type forms, are likely to have been dominant in the active repertoires of the V and S scribes. This is suggested by the preference for þ-type subject pronouns in late medieval Worcestershire and London (Morse-Gagné 2003: 147–148). The general distribution of dialectal usage is more helpful in establishing scribal preferences in this case than a comparison with the other texts that they copied. With regard to V, the dominance of þ-type subject forms in eLALME LP 7630 is generally replicated in texts that were not sampled by eLALME’s compilers: e.g. Handlyng Synne (Sullens 1983: 247–269 and 365–370), Richard Rolle’s Ego Dormio (Ogilvie-Thomson 1988: 26–33 and 135–149) and the A-text of Piers Plowman (Skeat 1867: Text A). However, as noted above, the scribe also retains h-type subject forms in several texts. As long as the variants form part of his active or passive linguistic repertoires, he is happy to reproduce them.

11  Origin and Spread of the Personal Pronoun They 

323

As far as S is concerned, eLALME LP 7780, the profile for the C-text of Piers Plowman, has a categorical þey-here-hem system. The scribe is said to have copied his texts literatim, although the extent to which this is true for his pronominal usage is unclear on the basis of the pronominal distribution in Estorie. The number of þ-type subject forms in S increases significantly as the text progresses: in ll. 1–987, h-type forms occur forty-six times and þ-type forms just six times, while in the second half (ll. 988–1974) the different pronoun types have a near-identical quantitative distribution, appearing forty-four and forty-five times, respectively. We take the increasing number of þ-type forms as an indication that the text was penned by someone for whom subject þ-type forms had become more common than for the author of the early Middle English Worcester exemplar; as the later scribe progressed in the copying process, more þ-forms slipped in (cf. Benskin and Laing 1981: 66). Given that we no longer have V for comparison, it is impossible to know whether the substitutions should be associated with the author of the exemplar shared by S and V, or the S scribe; the latter would force us to reconsider to what extent he can be identified as a literatim scribe. Linguistic and stylistic conditioning also appears to play a role in explaining the increased use of þ-type subject pronouns, as we will argue in the following section.

3.2 Accounting for Intratextual Variation Since certain versions of Estoire offer interesting variation in some forms of the 3pl. personal pronoun paradigm, this section explores whether, in addition to the existence of different layers of scribal intervention with concomitant dialectal variation, linguistic and stylistic factors can also help account for such variation. Based on the findings of previous literature, we carried out an analysis of several linguistic factors that could influence the choice of pronoun variant; our five foci are presented in the subsections below. Given the low counts and / or near categorical distribution of variants in certain environments (see Table 11.1), we restricted our analysis of transient pronominal systems to they variants in S and P, and them variants in L. There are only three þ-type their forms in L and two in P, and only two h-type them forms in D: too few to include in a quantitative variationist analysis, but we refer to these forms in our

324 

M. Cole and S. M. Pons-Sanz

discussion where relevant. The analysis of S is based on the second half (ll. 988–1974), where h-type and þ-type pronouns occur in almost equal measure.

Syntactic Context To understand the syntactic behaviour of h- and þ-type subject pronouns during the Middle English period, it is helpful to consider some of the factors that have been put forward to explain the use of Old English demonstrative pronouns as personal pronouns (Cole 2018). As noted in Sect. 3.1, þ-type pronouns are likely to be the result of polygenesis, with reflexes of Old English demonstratives in personal pronoun use in the West and South-West Midlands being reinforced by þ-type pronouns spreading from the North and East Midlands via dialect contact, themselves of multiple origins (see Cole 2018, Forthcoming). Thus, the use of þ-type pronouns probably reflects some continuation of the patterns characteristic of the native demonstrative pronouns, as well as the association of the native and Norse-derived paradigms because of their formal similarity. The use of demonstrative pronouns in Old English broadly conforms to accessibility hierarchies for pronominal material (Ariel 2001; Los and Kemenade 2018), whereby pronoun use is determined by the saliency of the antecedent in terms of retrieval and accessibility of meaning. Anaphoric expressions like personal pronouns have reduced referentiality, so they typically pick up salient antecedents like the topic in a previous stretch of discourse. Newly introduced referents are less salient, and thus less accessible; they favour demonstrative anaphoric expressions that have a stronger reference and phonetic form than personal pronouns. Stressed personal pronouns fall somewhere between (Ariel 2001: 31; see the discussion of phonological factors below). Clause initial demonstratives in Old English therefore generally refer to less accessible referents. Typically, they involve the second mention of a newly introduced protagonist and develop the narrative further (Los and Kemenade 2018: 138). This could explain why both S and V use the þ-type pronoun at l. 45 (see Sect. 3.1). The clause-initial pronoun þey picks up the newly introduced referent

11  Origin and Spread of the Personal Pronoun They 

325

wise men and develops the narrative’s description of their customs (see further the discussion on reference below): Wise men in olde dawis Bokis made of gode lawis How me sholde hem with riȝt leden & wyse to ben in alle here deden But þey noþynge ne wiste of þe. (ll. 41–45S) ‘Wise men in old days / Books made of good law / How men should themselves with righteousness lead / and be wise in all their deeds / But they nothing knew of you’.

Clause initial position is closely linked to line initial and thematic positions. Not surprisingly in this respect, in a study of the distribution of þ-type and h-type pronouns in the mid-fourteenth-century London / Middlesex variety of Hand I in the Auchinleck MS, Smith (2001: 232–233) finds that line-initial position favours þ-type pronouns, whereas h-type pronouns are rare there, and appear most frequently after subordinating conjunctions and topicalised adverbial phrases. He argues that this reflects the vestiges of a system where þ-type pronouns were originally used in thematic position but ‘where h-types were an option in non-thematic positions’ (Smith 2001: 233). In keeping with these arguments, we tagged they variants in S and P depending on whether they occurred in main or subordinate clauses. We also coded they variants in S and P for syntactic context depending on whether the subject pronoun appeared: (i) in clause-initial position;6 (ii) following a topicalised element (object or adverbial), e.g. To helle he wende withouten strif (l. 190S: ‘to hell they went without strife’); (iii) in non-­ clause initial position either involving inversion, e.g. ne hedden heo (l. 194S: ‘they had not’), or following a subordinating conjunction or relative pronoun, e.g. Þe child þat heo abyde longe (l. 330S: ‘the child that they had long waited for’); (iv) as an antecedent to a restrictive relative clause,  We follow the literature in considering personal pronouns after the conjunctions but / and clause initial (cf. Los and van Kemenade 2018). We also coded subject pronouns modified by al(le), e.g. alle heo comen (l. 410S: ‘they all came’), as clause initial because alle is part of the subject. 6

326 

M. Cole and S. M. Pons-Sanz

Table 11.2  MS S, ll. 988–1974 (they): Clause type Main Subordinate Total

h-type

þ-type

Total

29 15 44

29 16 45

58 31 89

Table 11.3  MS S, ll. 988–1974 (they): Syntactic context Cl.int. Top.obj. / adv. Non-cl.int. Ant.rel. Total

h-type

þ-type

Total

5 17 20 2 44

5 18 19 3 45

10 35 39 5 89

e.g. he þat byforen & aftur ȝede (l. 1068S: ‘they that went before and after’), as this was a typical environment for the use of a demonstrative pronoun in Old English. The results indicate that the different pronoun variants in S are similarly distributed across main and subordinate clauses (Table 11.2). The figures for syntactic context in S do not point in any particular direction either. In the second part of S, where the rates of h-type and þ-type subject forms are almost identical, the distribution of pronoun variants across the different syntactic contexts is remarkably similar (Table 11.3). In P, five out of six þ-type subject tokens appear in main clauses and four of these tokens occur in clause initial position or as an antecedent to a relative clause, typical demonstrative pronoun contexts in Old English, although the dataset is too small to be conclusive in any way. H-type subject pronouns occur just once in clause initial position, a tendency in line with Smith’s (2001) findings (Tables 11.4 and 11.5). The only þ-type forms for their in P also occur in initial position. Regarding variants for them and their syntactic function, Bergs (2005) finds that hem is favoured as a direct-object pronoun in the fifteenth-­ century Norfolk Paston Letters, and most individual speakers demonstrate a ‘slight in some cases, but still noticeable’ tendency to use them more as a prepositional object and hem as a simple object (Bergs 2005: 115–116). A similar trend is found in late Old Northumbrian, whereby

11  Origin and Spread of the Personal Pronoun They 

327

Table 11.4  MS P (they): Clause type Main Subordinate Total

h-type

þ-type

Total

8 2 10

5 1 6

13 3 16

h-type

þ-type

Total

– 1 3 6 10

2 2 – 2 6

2 3 3 8 16

Table 11.5  MS P (they): Syntactic context Ant.rel. Cl.int. Non-cl.int. Top.obj. / adv. Total

Table 11.6  MS L (them): Syntactic role

do io po

Total

h-type

þ-type

Total

8 11 18 37

– 3 5 8

8 14 23 45

prepositions favour þ-type (demonstrative) pronoun objects over h-type pronouns (Cole 2018: 183–184). However, no such effect is found in the Ormulum; the LAEME sample of the text has eleven prepositional object contexts, where hemm occurs six times, hemmsellfenn four times and þeȝȝm once. In L, the only text with a reasonable degree of object pronoun variation, variants for them were coded for the following syntactic functions: (i) direct object (do); (ii) indirect object (io); (iii) prepositional object (po). The results are striking in that, similarly to Bergs’s findings, þ-type forms occur as indirect objects and prepositional objects, but not in direct object function (Table 11.6). Only h-type pronouns function as direct objects. Contrastively, the instances of hem in D are restricted to the only prepositional contexts in the text (ll. 68D, 240D), but this also suggests a syntactic constraint that differentiates prepositional objects from other object types using different surface morphology.

328 

M. Cole and S. M. Pons-Sanz

Bergs (2005: 116) suggests that the tendency for prepositional objects to involve them rather than hem in the Paston Letters could be the result of psycholinguistic principles but takes the suggestion no further. The use of periphrastic structures instead of monomorphemic forms has been argued to be motivated by a speaker’s desire to talk in a way that gets them noticed (Haspelmath 1999: 1057). Periphrastic structures, such as prepositional phrases, are more expressive and transparent than opaque or reduced constructions (Lehmann 2015: 4). In a Middle English pronoun system involving competing variants, a periphrastic prepositional structure might initially have favoured the phonetically more salient pronoun object. In a system where h-type forms had become rare, however, as in the Lincolnshire dialect of D, the contrast between po and do could be maintained and reinforced by using h-type forms in po contexts.

Referent In considering whether the anaphoric pronominal form is determined by the nature of the referent, a starting point is Bergs’s (2005) study of animacy and gender as possible influences in the Paston Letters. We were not able to investigate animacy because the 3pl. pronouns in Estorie refer categorically to animate referents, but subject pronouns in S and P were coded for gender of the referent as follows: male for male-only groups with explicit male referents (e.g. men, kings, princes, knights, shepherds, etc.); female for female-only groups with explicit female referents (e.g. Mary Magdalene [and] sweet Mary, women, etc.); mixed for groups explicitly comprising both men and women (e.g. Zachariah and Elizabeth); neuter for neuter groups (e.g. referents such as folk, friends, Jews, the meek); and impersonal for contexts in which 3pl. pronouns are used impersonally with an implied but not stated antecedent: e.g. Þo leyden heo on Ihesu honde (l. 1186S: ‘then they laid hands on Jesus’), where þey implies the soldiers but the referent is not explicit in the narrative. There is also a small indefinite group of the type þey þat put loue bytwene… (l. 1008P: ‘they that put love between…’), where the 3pl. pronoun is the antecedent to a restrictive relative clause.

11  Origin and Spread of the Personal Pronoun They 

329

Table 11.7  MS S, ll. 988–1974 (they): Referent h-type

þ-type

Total

12 2 _

5 4 1 5 27 3 45

17 6 1 16 44 5 89

h-type

þ-type

Total

indefinite

10 _ _

Total

10

2 2 2 6

12 2 2 16

male female mixed neuter impersonal indefinite

Total

11 17 2 44

Table 11.8  MS P (they): Referent

male neuter

The results for S indicate that h-type pronouns clearly prefer male referents (70.5%, N = 12/17) and neuter referents (68.7%, N = 11/16; Table 11.7). Similarly, in P, h-type anaphoric pronouns are only used to refer to male referents (Table  11.8). However, þ-type forms are the favoured impersonal pronoun in S (61%, N  =  27/44; Table  11.7). Notably, all bar one of the impersonal contexts in S occur during the second half, which might help explain the far higher rate of þ-type pronouns in this later section. Ariel (2001: 31–33) notes that, in addition to the inherent salience of an antecedent, the accessibility of an antecedent is also determined by clear cohesion and unity between the antecedent and its anaphor. The looser the unity between the antecedent and its anaphor, the lower the accessibility of the relevant mental representation. Thus, the preference in S for impersonal þ-type pronouns is fully in keeping with the accessibility hierarchy discussed above. Shared origin or formal similarity with the demonstrative pronoun can also account for the fact that þ-type pronouns occur in five out of the seven indefinite contexts across S and P. Tokens for them in L were also tagged for gender of referent; h-types predominate in all contexts, but this is expected given the relatively low occurrence of þ-type pronouns (Table 11.9).

330 

M. Cole and S. M. Pons-Sanz

Table 11.9  MS L (them): Referent

male female neuter impersonal

Total

h-type

þ-type

Total

18 2 8 9 37

5 1 _

23 3 8 11 45

2 8

Phonological Factors Segmental phonotactic factors have been shown to condition pronominal variation. Johannesson (1995) demonstrates that, in the Ormulum, only þeȝȝre can be used after words ending in vowels or /j/, and þeȝȝre is also favoured if the final segment of the preceding word is /l/. The phonological conditioning of þeȝȝre and heore maximises the sonority contrast across word boundaries and avoids vowel hiatus. Hiatus avoidance also determines the distribution of þeȝȝm and hemm: only þeȝȝm can occur after words with a non-elidable final vowel (e.g. þatt he þeȝȝm), whereas hemm appears after word-final consonants. The findings highlight the importance of perceptual salience in determining the appearance of þ-type pronouns instead of h-type pronouns and are consistent with arguments that have stressed the role of phonological prominence in explaining the spread of the þ-type (see below). Having said that, Bergs (2005: 124) finds that usage in the Paston Letters shows no strong link between pronoun variation and preceding segment. He suggests that, as the replacement process advanced, the phonological motivations for adopting þ-type forms (clear perception, creation of prominent onsets) may have become less important than other factors, such as paradigmatic (analogical) pressure (Bergs 2005: 125–126). To test for the possible effect of segmental phonotactics in our data, they variants in S and P and them variants in L were tagged for preceding and following phonological context.7 Preceding environments in S  Coding for preceding segment involved the challenge of deciding whether final (elidable) -e was pronounced or not. As a general guideline, we assumed that, as in syllabic verse, final -e is elided when a vowel, /h/, or /w/ follows. For the purposes of coding, we assumed that initial and final were pronounced. Millward (1998: 75–76) notes that endings in S occur at a higher rate in non-rhyme position than in rhyme, which suggests that final was not being used simply to preserve rhymes. 7

331

11  Origin and Spread of the Personal Pronoun They  Table 11.10  MS S, ll. 988–1974 (they): Phonological environment Preceding segment /b/ /d/ /ð/ /f/ /g/ /h/ /j/, /ʤ/ /k/ /l/ /m/ /n/ /p/ Pause /r/ /s/ /ʃ/ /t/ Vowel /w/ Total

h-type _ 8 1 _ _ _ _ _

þ-type _ 4 _

Following segment Total _

1 _

12 1 1 _

1 _

1 _

2 3 5 _

1 5 5 6 1 1 3 3 _

1 8 7 15 1 3 6 8 _

8 3 _

8 6 _

16 9 _

44

45

89

3 2 9 _

h-type 4 1 3 _ _ 4 4 1 1 1 6 _ _ _ 3 2 3 5 6 44

þ-type

Total

4 2 1 2 2 5 1 3 3 5 2 _ _ _

8 3 4 2 2 9 5 4 4 6 8 _ _ _

4 5 _

7 7 3 6 11 89

1 5 45

involving /l/ or a vowel tend to favour þ-type subject pronouns disproportionally; this effect is particularly notable in the latter section of S at N  =  5/8 and N  =  6/9, respectively. Following environments involving vowels, however, favour h-type subject pronouns (Table  11.10). In L, þ-type pronouns for them only occur after /d, t, ð/ and vowels, and they favour following segments in /ð/ or sonorous segments like vowels, /j/ or /l/ (Table 11.12), leading to strings like with þem þey (l. 1918L), to þem þe (l. 1951L), which appear to reflect a preference for dental consonance. Four out of six þ-type they pronouns in P are also preceded by segments in /d, t/ and two out of the six are followed by /ð/ (Table 11.11). However, as the pronoun appears in sequences like and þey and þey þat (twice each), there could be an interaction here with the tendency for þ-type pronouns in P to favour clause initial and antecedent to relative clause contexts (cf.

332 

M. Cole and S. M. Pons-Sanz

Table 11.11  MS P (they): Phonological environment Preceding segment /b/ /d/ /ð/ /f/ /j/ /l/ /m/ /n/ Pause /r/ /s/ /t/ Vowel Total

h-type _ 2 _ _ _ _

þ-type _

Following segment Total _ 5 _ _ _ _

1 _

3 _ _ _ _ _ _

1 1 _

2 _ _

3 1 _

2 3 10

1 _

3 3 16

6

1 _

h-type _ 2 1 _ 1 1 1 2 _ _ 1 _ 1 10

þ-type 1 _ 2 1 _ 1 1 _ _ _ _ _ _ 6

Total 1 2 3 1 1 2 2 2 _ _ 1 _ 1 16

Table 11.12  MS L (them): Phonological environment Preceding segment h-type /b/ /θ/, /ð/ /f/ /h/ /j/ /k/ /l/ /m/ /n/ Pause /r/ /s/ /t/ Vowel /w/ Total

5 1 4 _ _ _ 3 _ 3 3 3 3 2 10 _ 37

þ-type 1 2 _ _ _ _ _ _ _ _ _ _ 2 3 _ 8

Following segment Total 6 3 4 _ _ _ 3 _ 3 3 3 3 4 13 _ 45

h-type 4 8 _ 4 _ 1 1 1 1 _ _ 6 4 3 4 37

þ-type _ 3 _ _ 1 1 1 _ _ _ _ 1 _ 1 _ 8

Total 4 11 _ 4 1 2 2 1 1 _ _ 7 4 4 4 45

the discussion on syntactic context and reference  above). S, however, appears to avoid dental consonance; h-type pronouns are preferred after preceding environments involving /d, ð/ (N = 9/13) and following environments involving /d, t, ð/ (N  =  7/10). Similar stylistic preferences

11  Origin and Spread of the Personal Pronoun They 

333

might also explain the disproportionately high rate of þ-type pronouns following segments in /m/ (N = 5/7). In all these contexts, þey is preceded by him (or hem), which suggests that þey is used to avoid the sequence him heo. Following environments involving /m/ also favour þ-type pronouns, but four out of the five cases involve the verb mouen, which suggests lexical conditioning (Table  11.10; cf. the discussion on verbal morphology and semantics below). There is also good reason to believe that the development of pronoun systems interacts with suprasegmental features such as stress, as pronouns typically have a stressed form and an unstressed form. In this respect, the emphatic function of demonstrative pronouns fits the suggestion that þ-type pronouns were first used in stressed positions. After all, they had a stronger onset and offered greater phonological prominence (Samuels 1972: 71; Ritt 2001; Bergs 2005: 94–95 and 122). Recovering the stress and intonation patterns of historical written data, however, is extremely problematic. Estorie was written as four-stress accentual verse rather than syllabic verse, and each line has an indeterminate number of unstressed syllables (Millward 1998: 80–81). Later substitutions might also have disrupted the original stress patterns. All of this makes the classification of stress in such data too impressionistic to be entirely reliable; we therefore decided not to code for stress. Rhyme has also been shown to have an effect on the distribution of þ-type and h-type forms (Smith 2001: 229–230), but 3pl. pronouns do not occur in rhyming position in Estorie.

Verbal Inflection and Semantics The demise of h-type pronouns in favour of þ-type pronouns has often been argued to involve an example of therapeutic change that ‘fixed’ ambiguities in the language. Middle English pronoun paradigms often had identical 3sg. and 3pl. h-type personal pronouns. S’s Estorie is a case in point; he occurs for he, she and they, and heo for she and they. This lack of a third-person singular-plural pronoun distinction, together with the loss of distinctive singular-plural verbal morphology, created a situation in which þ-type pronouns had the potential advantage of disambiguating a pronoun system with opaque number distinctions (Samuels 1972;

334 

M. Cole and S. M. Pons-Sanz

Thomason and Kaufman 1988: 324; Howe 1996; Lass 1999: 160–166). However, the association of þ-type forms with disambiguation is problematic: the Ormulum, the earliest Middle English text with consistent use of þ-type pronouns, employs þeȝȝ categorically and yet retains distinctive singular-plural verbal morphology. Smith (2001: 231) also finds no link between the distribution of h- and þ-type forms and the loss of verbal inflection; when h-type pronouns follow a verb in his data, final -n is required in order to avoid elision (e.g. wenten he) but not as a disambiguating strategy (Smith 2001: 231). We tested a possible link between pronoun type and number marking on the verb by tagging pronouns for the number morphology of the verb that they co-occurred with.8 As in other studies, no correlation was found between the loss of distinctive verbal number morphology and the use of þ-type subject pronouns. If replacement were therapeutic, higher rates of þ-type subject pronouns would occur with verbs unmarked for number, yet the distribution of þ-type and h-type pronouns with numberless verbs is almost identical in S (Table 11.13). Interestingly, the only example of a distinctive plural verbal form in P occurs with a þ-type pronoun (Table 11.14). Table 11.13  MS S, ll. 988–1974 (they): Number morphology + pl − pl Total

h-type

þ-type

Total

20 24 44

20 25 45

40 49 89

Table 11.14  MS P (they): Number morphology + pl − pl Total

h-type _ 10 10

þ-type

Total

1 5 6

1 15 16

 The loss of -en neutralises the number distinction in weak preterites. However, ablaut contrasts in the stem continued to distinguish number in some strong verbs regardless of the loss of final -en: e.g. bēde < ME bēden < OE bǣdon is distinctly plural, as opposed to bad (< OE bæd), which occurs in both singular and plural contexts; cf. also plural (by)gunne versus singular bigan / bigon, and plural abyde < ME abiden. 8

335

11  Origin and Spread of the Personal Pronoun They  Table 11.15  MS S, ll. 988–1974 (they): Lexical conditioning Both

h-type

þ-type

Stative / intransitive: stōnden, understōnden (1); biginnen, gōn (2); witen (3)

Stative / intransitive: faren, ginnen, hēren (1); sēn (2); comen, willen (3)

Other: bēren, brēken, casten, clēpen, mēten, nimen, setten, stingen, swinken (1); leien, tāken (2)

Stative / aux. / intrans. h-type bēn (aux.) 1 bēn (lex.) 5 Other: hāven (aux.) 1 bīnden, drauen, graunten, hāven (lex.) 2 liften, māken, shāken, mouen 2 sprēden (1); lēden, shulen 2 seien (2) wenden 2 Other: bidden 2 bringen 3 dōn 1 fīnden 1

þ-type 1 4 3 1 5 4 1 1 1 1 1

Table 11.16  MS P (they): Lexical conditioning h-type

þ-type

Stative and intransitive: gōn, shulen, understōnden (1); hāven (lex.) (2)

Stative, auxiliary and intransitive: fallen, lēven, mouen (1); bēn (lex.) (2)

Other: lēten, māken, tellen (1); dōn (2)

Other: putten (1)

Unlike morphology, verbal semantics does, however, appear to be an influencing factor, but once again the low numbers make the findings rather impressionistic. Þ-type subject pronouns in S seem to favour stative, auxiliary and intransitive verbs disproportionately: þ-type variants occur with these verb types at a rate of 67% (N = 30/45) compared with h-type pronouns at 50% (N = 22/44) (Table 11.15). Similarly, five out of six þ-type subject pronouns in P co-occur with these types of verbs (Table 11.16). These observations are in line with early Middle English pronominal systems in Oxfordshire, Ely, West Suffolk and Gloucestershire with low-frequency þ-type subjects; þ-type tokens occur more frequently with stative, auxiliary and intransitive verbs (see Cole Forthcoming).

336 

M. Cole and S. M. Pons-Sanz

Stylistic Factors A qualitative analysis indicates that other stylistic factors may also determine pronominal choice. Scribes appear to use h-type and þ-type variants to create a stylistic effect when referring to the same referent twice within the line and / or across the sentence, as illustrated in (1–4). At times, the scribe avails himself of the different pronoun variants to avoid repetition, as in (2), in much the same way that he uses other synonyms (cf. wende and come): (1) Þo he wiste þey weren nye (l. 658S: ‘Then they knew they were near’) (2) Al heo aȝeyns hym wende / Aȝeyn him þey come al bydeene… (ll.  1061–1062S: ‘And they to him went / To him they came all, offering…’ ) (3) And þey left here bot anon / Þar nettus walweþ þat ssolde hem fede (ll. 965–966P: ‘And they left their boat immediately / They rolled their nets that should feed them’) (4) þat Ihesus byfore to hem seide / To þem þe whiles þat he was alyue (ll. 1950–1951L: ‘That Jesus said to them before / To them the time that he was alive’). Scribes also seem to have avoided the stylistically awkward sequence þey þey ‘though they’ (cf. Samuels 1972: 71–72; Smith 2001: 228).

4 Conclusion This paper has taken a new approach to the history of 3pl. pronominal forms by triangulating data associated with various aspects of the distribution of these forms  in the seven manuscripts that record the early Middle English text La estorie del evangelie. We have combined  (i) the in-depth analysis of the seven versions of Estorie with (ii) scribal behaviour across various texts, and (iii) a diatopic and diachronic overview of pronominal usage in Middle English made possible by bringing eLALME

11  Origin and Spread of the Personal Pronoun They 

337

and LAEME data together. This has produced a new understanding of the etymology of 3pl. þ-type pronominal forms and their textual use in our corpus and  beyond. The existence of the near contemporary Norse-­ derived þeȝȝ- forms in the Ormulum and the native þai(e) in (South-)West Midlands texts, and the patterning of þ-type forms in our texts argue in favour of polygenesis, as opposed to the traditional etymological explanation based on language contact at the expense of native derivation. Even though the number of tokens is relatively low and therefore we can only talk about trends rather than statistically significant results, it is notable that þ-type forms in our corpus tend to dominate in contexts where Old English demonstrative pronouns were commonly used as personal pronouns. Þ-type variants do not show the same clear preference for clause initial position found in Smith’s (2001) fourteenth-century London data, except in the case of P, but there is a strong tendency for þ-type pronouns to function as anaphors for referents that lack saliency and as antecedents of relative pronouns. The phonological salience of the þ-type forms also seems to have contributed to their use in particular phonological environments and in marked syntactic structures, such as prepositional phrases, as opposed to monomorphemic constructions. Moreover, our analysis has shown that, while the avoidance of pronominal or verbal ambiguity cannot easily explain the use of þ-type forms, lexical conditioning and stylistic factors might play a role in pronominal choice. The impact of these factors can also be seen beyond our corpus, but this is an area which still requires further analysis (cf. Sect. 3.2). There is, however, some variation within these trends. For instance, the syntactic behaviour of þ-type subject forms in S differs from that in P, possibly as a result of the interaction between syntax and stylistics: e.g. the two texts appear to exhibit different preferences when it comes to dental consonance. The tendency for S to avoid dental consonance might explain why there aren’t more clause-initial strings in S of the type but þey… / and þey… Other differences might be attributable to the make-up of the pronominal profile in a particular dialectal area; this is exemplified by the choice of h-type as opposed to þ-type forms as a prepositional object in D, which differs from most other manuscripts in our corpus in having near-categorical þ-type forms across the whole paradigm. The

338 

M. Cole and S. M. Pons-Sanz

shared trends and the differences between the various versions demonstrate that both general overviews (cf. Morse-Gagné 2003) and the nuanced analysis of intratextual variation are fundamental to understanding how and why þ-type pronominal forms replaced h-type forms. This work also has implications beyond the confines of personal pronouns. On the one hand, the analysis of intrascribal variation in Sect. 3.1 points to the need for a significant reconsideration of scribal behaviour. The scribe responsible for writing Estorie in S is said to have copied his texts literatim, but the distribution of þ-type subject forms across this version of the text raises the intriguing possibility that he might have had more input into his texts that he is given credit for. Indeed, his behaviour does not seem to have been very different from that of V’s Scribe B, who is generally presented as a moderniser who imposed some dialectal regularity in the texts that he copied. Furthermore, the multifaceted approach used here could  be applied to other texts for which there are multiple manuscripts, and doing so could shed light on other linguistic features, both grammatical and lexical. 

References Ariel, Mira. 2001. Accessibility Theory: An Overview. In Text Representations: Linguistic and Psycholinguistic Aspects, ed. Ted Sanders, Joost Schilperoord, and Wilbert Spooren, 29–89. Amsterdam: John Benjamins. Benskin, Michael, and Margaret Laing. 1981. Translations and Mischsprachen in Middle English Manuscripts. In So Meny People Longages and Tonges: Philological Essays in Scots and Mediaeval English Presented to Angus McIntosh, ed. Michael Benskin and Michael Samuels, 55–106. Edinburgh: Middle English Dialect Project. Benskin, Michael, Margaret Laing, Vasilios Karaiskos, and Keith Williamson. 2013–. An Electronic Version of A Linguistic Atlas of Late Mediaeval English. Edinburgh: University of Edinburgh. http://www.lel.ed.ac.uk/ihd/elalme/ elalme.html. Accessed 27 May 2023.

11  Origin and Spread of the Personal Pronoun They 

339

Bergs, Alexander. 2005. Social Networks and Historical Sociolinguistics: Studies in Morphosyntactic Variation in the Paston Letters (1421–1503), Topics in English Linguistics, 51. Berlin: Mouton de Gruyter. Bühler, Curt Ferdinand. 1962. Two Middle English Texts of the ‘Somnia Danielis’. Anglia 80: 264–273. Cole, Marcelle. 2018. A Native Origin for Present-Day English They, Their, Them. Diachronica 35: 165–209. ———. Forthcoming. Personal Pronoun They: A Native Origin Account. Houndmills: Palgrave Macmillan. Dance, Richard. 2003. Words Derived from Old Norse in Early Middle English: Studies in the Vocabulary of the South-West Midlands Texts, Medieval and Renaissance Texts and Studies, 246. Tempe, AZ: Arizona Center for Medieval and Renaissance Studies. Dance, Richard, Sara M. Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge: Cardiff and Sheffield. https://gersum.org. Accessed 27 May 2023. Doyle, A.I. 1990. The Shaping of the Vernon and Simeon Manuscript. In Studies in the Vernon Manuscript, ed. Derek Pearsall, 1–13. Oxford: D. S. Brewer. Förster, Max. 1941. Die spätae. deiktische Pronominalform þæge and ne. they. Beiblatt zur Anglia Beiblatt 52: 274–280. ———. 1942. Nochmals ae. þæge. Beiblatt zur Anglia 53: 86–87. Foster, Frances Allen. 1914. A Study of the Middle-English Poem Known as the Northern Passion and Its Relation to the Cycle Plays: A Dissertation. London: Clay & Sons. Görlach, Manfred. 1974. The Textual Tradition of the South English Legendary, Leeds Texts and Monographs, n.s. 6. Leeds: University of Leeds, School of English. Haspelmath, Martin. 1999. Why is Grammaticalization Irreversible? Linguistics 37: 1043–1068. Horobin, Simon. 2013. The Scribes of the Vernon Manuscript. In The Making of the Vernon Manuscript: The Production and Contents of Oxford, Bodleian Library, MS Eng. Poet. a. 1, ed. Wendy Scase, 27–47. Turnhout: Brepols. Horobin, Simon, and Jeremy Smith. 2011. Language. In A Facsimile Edition of the Vernon Manuscript: Oxford, Bodleian Library MS Eng. Poet. A. 1, Bodleian Digital Texts, ed. Wendy Scase, III. Oxford: Bodleian Library.

340 

M. Cole and S. M. Pons-Sanz

Horstmann, C., ed. 1878. Sammlung altenglischer Legenden. Heilbronn: Henninger. Howe, Stephen. 1996. The Personal Pronouns in the Germanic Languages: A Study of Personal Pronoun Morphology and Change in the Germanic Languages from the First Records to the Present Day. Berlin: Walter de Gruyter. ———, ed. 1881. Altenglischer Legenden: Neue Folge. Heilbronn: Henninger. Johannesson, Nils-Lennart. 1995. Old English versus Old Norse Vocabulary in the Ormulum: The Choice of Third Person Plural Personal Pronouns. In Studies in Anglistics, ed. Gunnel Melchers and Beatrice Warren, 171–180. Stockholm: Almqvist & Wiksell International. Laing, Margaret. 1978. Studies in the Dialect Material of Medieval Lincolnshire. Unpublished doctoral thesis, University of Edinburgh. ———. 2013–. A Linguistic Atlas of Early Middle English, 1150–1325, version 3.2. Edinburgh: University of Edinburgh. http://www.lel.ed.ac.uk/ihd/ laeme2/laeme2_framesZ.html. Accessed 27 May 2023. Lass, Roger. 1999. Phonology and Morphology. In The Cambridge History of the English Language, III: 1476–1776, ed. Roger Lass, 56–186. Cambridge: Cambridge University Press. Lehmann, Christian. 2015. Thoughts on Grammaticalization,  3rd edn. Berlin: Language Science Press. Los, Bettelou, and Ans van Kemenade. 2018. Syntax and the Morphology of Deixis: The Loss of Demonstratives and Paratactic Clause Linking. In Atypical Demonstratives: Syntax, Semantics and Pragmatics, ed. Marco Coniglio, Andrew Murphy, Eva Schlachter, and Tonjes Veenstra, 129–157. Berlin: de Gruyter. McIntosh, Angus. 1987. The Middle English Estoire del Evangelie: The Dialect of the Original Version. Neuphilologische Mitteilungen 88: 186–191. Millward, Celia. 1997. The Medieval Scribe as Editor: The Case of La Estorie del Evangelie. Manuscripta 41: 155–170. ———, ed. 1998. La Estorie del Evangelie: A Parallel-Text Edition, Middle English Texts, 30. Heidelberg: C. Winter. Morris, Richard, ed. 1871. Legends of the Holy Rood, Symbols of the Passion, and Cross-Poems, Early English Text Society, o.s. 46. London: Trübner & Co. Morse-Gagné, Elise. 2003. Viking Pronouns in England: Charting the Course of THEY, THEIR, and THEM. Unpublished doctoral thesis, University of Pennsylvania.

11  Origin and Spread of the Personal Pronoun They 

341

Ogilvie-Thomson, S.J., ed. 1988. Richard Rolle: Prose and Verse, Edited from MS Longleat 29 and Related Manuscripts, Early English Text Society, o.s. 293. Oxford: Oxford University Press. Ogura, Michiko. 2001. Late West Saxon Forms of the Demonstrative Pronouns as Native Prototypes of They. Notes and Queries 48: 5–6. Parkes, Malcolm B. 1983. On the Presumed Date and Possible Origin of the Manuscript of the Ormulum. In Five Hundred Years of Words and Sounds: A Festschrift for Eric Dobson, ed. Eric G. Stanley and Douglas Gray, 115–127. Cambridge: D. S. Brewer. Perry, Ryan. 2007. The Clopton Manuscript and the Beauchamp Affinity: Patronage and Reception Issues in a West Midlands Reading Community. In Essays in Manuscript Geography: Vernacular Manuscripts of the English West Midlands from the Conquest to the Sixteenth Century, ed. Wendy Scase, 131–159. Turnhout: Brepols. ———. 2013. Editorial Politics in the Vernon Manuscript. In The Making of the Vernon Manuscript: The Production and Contents of Oxford, Bodleian Library, MS Eng. Poet. a. 1, ed. Wendy Scase, 71–95. Turnhout: Brepols. Pickering, O.S. 1972. A newe lessoun off Crystys ressurrectoun: A Second Manuscript, and Its Use of La Estorie del Evangelie. English Philological Studies 13: 43–48. Pons-Sanz, Sara. 2013. The Lexical Effects of Anglo-Scandinavian Linguistic Contact on Old English, Studies in the Early Middle Ages, 1. Turnhout: Brepols. ———. 2021. Norse-Derived Vocabulary in La Estorie del Evangelie. Folia Linguistica Historica 42: 461–491. Ritt, Nikolaus. 2001. The Spread of Scandinavian Third Person Plural Pronouns in English: Optimisation, Adaptation and Evolutionary Stability. In Language Contact in the History of English, ed. Dieter Kastovsky and Arthur Mettinger, 279–304. Frankfurt am Main: Peter Lang. Samuels, Michael. 1972. Linguistic Evolution, with Special Reference to English. Cambridge: Cambridge University Press. Samuels, M.L. 1985. Langland’s Dialect. Medium Aevum 54: 232–247. Skeat, Walter William, ed. 1867. William Langland: The Vision of Piers Plowman, Early English Text Society, o.s. 28. London: Trübner & Co. Smith, Jeremy. 2001. Pronominal Systems in the Auchinleck Manuscript. In Lexis and Texts in Early English: Studies Presented to Professor Jane A. Roberts, ed. Christian J. Kay and Louise M. Sylvester, 225–235. Amsterdam: Rodopi. ———. 2013. Mapping the Language of the Vernon Manuscript. In The Making of the Vernon Manuscript: The Production and Contents of Oxford,

342 

M. Cole and S. M. Pons-Sanz

Bodleian Library, MS Eng. Poet. a. 1, ed. Wendy Scase, 49–70. Turnhout: Brepols. ———. 2020. Transforming Early English: The Reinvention of Early English and Older Scots. Cambridge: Cambridge University Press. Stenroos, Merja. 2005. The Spread of They, Their and Them in English: The Late Middle English Evidence. In Naked Wordes in Englissh, ed. Marcin Krygier and Liliana Sikorska, 66–96. Frankfurt am Main: Peter Lang. Sullens, Idelle, ed. 1983. Robert Mannyng: Handlyng Synne, Medieval and Renaissance Texts and Studies, 14. Binghamton, NY: Center for Medieval and Early Renaissance Studies. Thomason, Sarah Grey, and Terence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Turville-Petre, Thorlac. 1990. The Relationship of the Vernon and Clopton Manuscripts. In Studies in the Vernon Manuscript, ed. Derek Pearsall, 29–44. Cambridge: D. S. Brewer. Zettersten, Arne, and Bernhard Diensberg, eds. 2000. The English Text of the Ancrene Riwle: The ‘Vernon’ Text, Edited from Oxford, Bodleian Library MS Eng. Poet. A. 1, Early English Text Society, o.s. 310. Oxford: Oxford University Press.

12 Language Contact Effects on Verb Semantic Classes: Lability in Early English and Old French Luisa García García and Richard Ingham

1 Introduction A very large body of research bears out how significant contact with other languages has been as a driver of changes in English over the centuries. Yet linguistic change does not always succeed. For one reason or another, an innovation that is initially taken up by a speech community, or some members of it, may fail to generalise and oust its competitor forms. Research into so-called failed changes has been gaining some traction in recent years. Postma (2010) discussed two cases: the temporary replacement of hem ‘him’ in a Middle Dutch border dialect by the innovative

L. García García University of Seville, Seville, Spain e-mail: [email protected] R. Ingham University of Westminster, London, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_12

343

344 

L. García García and R. Ingham

lexical reflexive sick, borrowed from Low German,1 and the temporary extension of do-support to non-emphatic declarative clauses in Early Modern English. He noted that the maximum point of the failed innovations was connected in time with a strong rise of the successful innovations that accompanied them: these were another reflexive form in the Dutch dialect case, and the rise of do-support in other clause types. Bacovcin (2017) studied the short-lived V - to NP - NP construction with dative verbs in Middle English, which he accounted for as a temporary by-product of successive changes in the Middle English Case system. At the level of lexis, Sylvester et al. (2022) have researched lexical variation in Late Middle English, showing that numerous innovations, especially loan items, did not succeed in displacing established lexemes. In this study we consider changes to the argument structure of English verbs in two semantic domains where contact influence may well have made some impact, but not sufficiently to alter their character long-term. Importantly, the innovative syntax which did not survive with these verbs emerged at a time when, as found by Postma (2010), the change in question was becoming successful in comparable linguistic domains. The phenomenon in question was the ability of a verb to appear in morphologically unmarked form in both causal-transitive and noncausal-intransitive frames, as discussed more fully in the next section.

2 Lability According to Haspelmath’s (1993) typology, languages possess a variety of causativisation and anticausativisation strategies involving verbs of change of state or location,2 including P(atient)-lability, where the Patient or Undergoer argument can appear either as subject or direct object and the verb can adopt a noncausal or causal sense without alteration of the verb form, e.g.:

 Sick was ousted by another reflexive form, sich, which successfully replaced hem.  ‘Causativisation’ is the process by which a noncausal verb is rendered causal, as in ‘laugh’ > ‘make laugh’. ‘Anticausativisation’ turns a causal verb into non-causal, as in ‘scare (so.)’ > ‘get scared’. 1 2

12  Language Contact Effects on Verb Semantic Classes 

345

(1) a. The branch broke (noncausal-intransitive). b.  The man broke the branch (causal-transitive). (2) a. The ship sank (noncausal-intransitive). b.  The pirates sank the ship (causal-transitive).

It is found most notably in the semantic sub-classes of change-of-state verbs and change-of-location verbs, as illustrated above in (1)–(2), respectively, where the positionally alternating argument denotes the entity undergoing the relevant change. Lability in Modern English has been well-studied descriptively and theoretically (e.g. Visser 1963; Pinker 1989; Levin 1993; Rappaport and Levin 1995; McMillion 2006). The semantic criteria model of Pinker (1989)—(see also Rappaport & Levin 1995)—catalogued verbs on the basis of semantic constraints on their ability to alternate. Thus, verbs such as break involve a change of state which could occur independently of the action of an agent. Verbs such has cut, however, do not alternate because their semantics necessarily involve the action of an agent. In Modern English lability is virtually absent from the semantic class of verbs with psychological meaning (henceforth psych verb) (Levin 1993, van Gelderen 2011, Alexiadou 2016). The semantic role of Experiencer (roughly, the entity undergoing the psychological process) can appear either as subject or direct object of a psych-verb clause, e.g.: (5)  John feared a lightning strike. (6)  A lightning strike frightened John.

However, the verbs in question do not allow both constructions: (7)  *John frightened a lightning strike. (8)  *A lightning strike feared John.

Lability is also absent from the class of destroy-verbs (Levin 1993), even though their meaning involves a change of state in the Patient argument which otherwise permits lability (cf. (1)):

346 

L. García García and R. Ingham

  (9)  The storm wrecked the ship outside the harbour. (10)  *The ship wrecked outside the harbour.

Following the criteria-based approach, Levin (1993) observed that destroy-class verbs differ semantically from change-of-state verbs in that the latter encode some specific information about the resulting state of the entity that is e.g. broken or sunk (cf. (1)–(2) above), whereas destroy-­ class verbs encode simply the fact of the entity’s destruction. In the Middle English period, the position seems to have been interestingly different from Present-Day English. It was noted observationally by one of the present authors that some verbs with psychological meaning, as well as certain verbs with meanings similar to ‘destroy’, also showed lability. These observations prompted the questions pursued in this research, namely, (i) how widespread lability was among Middle English verbs belonging to these semantic classes, (ii) whether their Old English forerunners were also labile and (iii) whether contact with French might account for any changes that occurred, as with change-of-state / position verbs.

3 Lability in Earlier English: Previous Research The literature on lability in Old English is somewhat limited but adequate for a basic understanding of the situation. The first consistent contribution to the study of the argument structure of Old English verbs (known to us) is Hermodsson (1952), which contrasted the valency of a large number of verbs in Old High German, Old Frisian, Old English and Old Saxon. Hermodsson (1952: 210) pointed out the radical spread of labile verbs in English, when compared to other Germanic languages. He concluded from that comparison that lability is an inherited trait that originates in the Ingvaeonic language group, comprising Old Frisian, Old English, Old Saxon (and Middle Dutch), and that is clearly visible in the Old English period. He acknowledged, though, like Visser (1963: 99), the abundance of intransitive verbs in Old English, compared to

12  Language Contact Effects on Verb Semantic Classes 

347

Present-Day English. McMillion (2006) claimed that Old English is not especially labile.3 In recent research into valency in the history of English, Ottosson (2013) and van Gelderen (2011), following Hermodsson, also trace back Present-Day English use of labile coding to Old English, while noting at the same time the much higher number of labile verbs in Present-Day English. More specifically, van Gelderen (2011: 122) links the rise of lability in Old English to the loss of the causative formation. This link has been studied by García García (2020), who points out that, since the causative formation is by far the main valency-changing mechanism in Proto-Germanic, the decline of its transitivising function in favour of labile coding in Old English demonstrates the impetus of labilisation already in this early period (2020: 170). This has been confirmed by a recent corpus study that shows that the number of labile verbs increases significantly both in token and type in late Old English with respect to early Old English (García García and Ruiz Narbona 2021). Note that most of the Old English causatives that become labile denote a change of state or change of position: e.g. byrnan ‘to burn (intr.); to burn (sth.)’, gecelan ‘to be cool; to make cool’. The number of labile verbs massively increases over time. Ingham (2020) has shown that a great expansion of lability took place in Middle English: less than 20% of native verbs surviving into Middle English in the semantic areas of change of state and change of position were found to have been labile in Old English, while in Middle English two-thirds of them were attested as labile, with an even greater proportion of verbs in these semantic areas borrowed from French. He argued that intense French influence in the later medieval period was responsible for the large-scale adoption of lability by Middle English. Medieval French, unlike the modern language, made extensive use of lability in those same areas, so the case for seeing French contact influence at work in the large-­ scale extension of lability at the very time when the influence of French  He came to this conclusion by applying Haspelmath’s (1993) questionnaire of thirty-one inchoative / causative verb pairs to Old English. However, the reliability of such questionnaires for extinct languages has been rightly disputed, because of the absence of native speakers who can confirm the most usual gloss (see van Gelderen 2011: 115). 3

348 

L. García García and R. Ingham

on Middle English generally was at its height (thirteenth and fourteenth centuries) appears solid. Arising from that research, we wish to consider how far contact with French might have affected the argument structure of verbs in other semantic areas where lability could have, but did not, become dominant. The two lexical domains to which the present study is devoted are psych verbs and destroy-verbs (Levin 1993). Psych verbs in earlier English have received considerable attention in recent years from a number of authors, notably Allen (1995), Möhlig-Falke (2012), van Gelderen (2014) and Miura (2014), particularly focusing on developments that led to rearranged linking of semantic roles to grammatical functions. Such studies have often highlighted the process by which initially object-Experiencer verbs such as (modern spelling) like, e.g. (11), shifted to the subject-­Experiencer type, e.g. (12): (11) Đam    wife    þa     word wel      Det.-Dat. woman-dat.  Det.-Nom.Pl.  words well licodon  (Beo 491) please-pret.-3Pl.   ‘Those words pleased the woman well’. (12)  The woman really liked those words (PDE).

Studies have also examined the impersonal construction with psych verbs in Old English (see (24) below) and its disappearance in Middle English and beyond. Certain authors (Guidi 2011; van Gelderen 2014; Alexiadou 2016) have commented on lability in psych verbs. Alexiadou (2016) discussed the absence of what she called the ‘causative alternation’, i.e. lability, in this semantic class; she mentioned that, whereas both psych verbs and change-of-state verbs could take part in morphological alternations of the causative type in Old English, only psych verbs ‘were for some reason incompatible with lability’ (Alexiadou 2016: 7). This alleged disparity is one of the points of departure of the present article. Thus, in Old English, change-of-state, psych- and also destroy-verbs participated in the causative morphological alternation (e.g. abylgan ‘to anger’ vs. belgan ‘to swell with anger’, wyrdan ‘to injure, hinder’ vs. forweorþan ‘to perish, go off’). The morphological alternation was already declining in Old English in favour of labile coding. However, all three

12  Language Contact Effects on Verb Semantic Classes 

349

classes were not affected by the spread of lability in the same way. The different development of verb semantic classes in pre-Modern English regarding lability is the focus of our study. Argument structure in destroy-verbs has received little or no attention. Our work appears to be the first to consider lability or otherwise in the semantic domain from a diachronic perspective, and also to relate linguistic change in these domains to contact influence from French, more specifically from the variety of Old French used in England during the centuries following the Norman Conquest, and conventionally known as Anglo-French (or Anglo-Norman; Trotter 2003; Rothwell 1993). Although in some respects, especially phonology, it deviated from the continental language, in terms of syntax it remained broadly similar to it (Wilshere 1993; Ingham 2012).

4 Analysis of Old English and Middle English destroy-verbs In this section we describe and compare the argument structure of Old English, Middle English and Anglo-French destroy-verbs. First, we present a detailed account of the data collection process of the Old English and Middle English material, and then discuss the results. The data collection procedure used for Old English material was first to filter all the verbs in Bosworth and Toller’s (1898) Anglo-Saxon Dictionary (ASD) which contained in their definition ‘destroy’ or any of the other members of the ‘destroy’ class identified by Levin (1993: 239). We used the advanced search tool in the electronic version of the dictionary, which turned up 122 verbs with the word ‘destroy’ in their definition, thirtynine verbs with ‘waste’, eleven with ‘ruin’, nine with ‘demolish’, eight with ‘ravage’, and zero with ‘wreck’. We then determined the argument frames in which the verbs were attested, consulting the citations in the ASD and, when available, in the Dictionary of Old English (DOE), completed up to letter I so far. Only four of the verbs thus selected were attested as labile, that is, were used as both causal-transitive, with the Patient in object function, and noncausal-intransitive, with the Patient as subject. These are brecan ‘to break, demolish (sth.)’ / ‘to force a way out of confinement’,

350 

L. García García and R. Ingham

slītan ‘to slit, tear, rend, destroy, waste (sth.)’ / ‘to tear, get torn’, losian ‘to lose, destroy’ / ‘to be lost, perish’ and forfaran ‘to perish, be destroyed’ / ‘to kill, destroy’. For the first three verbs, the meaning category ‘destroy’ is clearly secondary and results from the translation of specific contextual usages of the respective primary meanings, namely ‘break’, ‘tear’ and ‘lose’. OE forfaran ‘to perish, be destroyed’ / ‘to kill, destroy’ is the only labile verb in the destroy-class in Old English retrieved by our searches. Examples of its noncausal-intransitive and causal-­transitive usages follow:     Noncausal-intransitive: (13) a. & mycel sciphere hider com suðan of Lidwicum, & hergodon swyðe be Sæfern, ac hi þær mæst ealle syþþan forforan (ChronD 910.10)4 ‘And a big fleet of war came hither from the south of Brittany, and they pillaged much along the Severn, but afterwards almost all were destroyed there’.    Causal-transitive: b. …þæt man þa sawla ne forfare, þe Crist mid his agenum life gebohte (LawVAtr 2) ‘…so that man does not destroy the soul, which Christ bought with his own life’.

Unsurprisingly, our results closely coincide with previous findings, such as Hermodsson’s (1952: 196ff.) and Visser’s (1963: 97); neither study includes any destroy-verbs in its list of labile verbs in Old English. Additional proof for the (near-)absence of lability in destroy-verbs in Old English is supplied by the fact that causative verb pairs with this and similar meanings (‘injure’, ‘harm’, ‘kill’) are kept intact, unlike causative pairs denoting change of state or change of position, which frequently become labile, as noted above. Some examples of causative pairs around the meaning ‘destroy’ are ā-hry¯ran ‘to cause to fall, destroy’ vs. hrēosan ‘to fall, be destroyed’, wyrdan ‘to injure’ vs. for-weorðan ‘to perish, vanish, go off’ (intrans.) or cwellan ‘to kill’ vs. cwelan ‘to die’ (see complete list in García García 2020: 175–6). This suggests that the incompatibility of  Old English works are cited after the Dictionary of Old English Corpus.

4

12  Language Contact Effects on Verb Semantic Classes 

351

lability with destroy-verbs is probably (nearly) categorial in Old English, and not merely the result of attestation gaps. The method used to find destroy-verbs in Middle English was first to probe the Middle English Dictionary (MED) using the ‘definition and notes’ search tool so as to identify all the verbs in whose definition the words ‘destroy’, ‘waste’, ‘ruin’, ‘wreck’, ‘ravage’ or ‘demolish’ occurred.5 Some 263 hits were obtained in this way, the majority for ‘destroy’. For each verb, the primary sense was then examined to identify whether labile syntax was attested in the citations, that is, whether the argument bearing the Patient role was attested in subject and also in object function of the same lexeme without morphological alteration. This procedure required some refinement in order to arrive at appropriate target verbs: the MED adopts a very wide range of definiens terms, seemingly in order to give an idiomatic translation of particular uses of lexemes in the citations provided, e.g. ‘destroy’ is proposed as a translation of voiden and misleden respectively, in the following citations:6 (14)  … in caas hir baner… were voided or rased (Vegetius 89/4) ‘In case their banner were to be destroyed or cut down’. (15) Icham adred, Þurth his folye he is misled (Arthur & Merlin (Auch.) 4988) ‘I am afraid that through his folly he is destroyed’.

The search also provided numerous hits where the ‘destroy’ meaning arose only thanks to a phrasal use, e.g. s.v. putten in the combination put down, as in: (16) They wolde putt downe mannes reason in fulfillyng of their worldly desires (Chartier, Treat. Hope 59/20–21) ‘They would overturn human reason while fulfilling their worldly desires’.  Rarer words listed by Levin (1993: 239) such as decimate and obliterate were left aside as it was expected that one of the five terms selected would suffice to generate the corresponding Middle English lexis. 6  All Old English, Middle English and Anglo-French citations are referenced in the short forms given in the respective dictionaries.

5

352 

L. García García and R. Ingham

The great majority of hits did not correspond to items with ‘destroy’, ‘ruin’, and so on as the primary meaning (taken as being mentioned in the first sense listed) and were therefore culled. Thirty-six verbs remained, some of them labile, e.g.: (17) a. Right as wormes shende a tree, Right so a wyf destroyeth hir housbonde (Chaucer CT.WB 376) ‘Just as grubs destroy a tree, so does a wife destroy her husband’. b. Sonne myght there none discende Lest the tender grasses shende (RRose 1400) ‘No sun could reach down there lest the tender grass would be destroyed’. (18) a. Ye han nat anientissed or destroyed hem (Chaucer CT.Mel 2438) ‘You have not eliminated or destroyed them’. b. With þe lest winde [...] Þe note anentis [...] & all to noʒt worthis (Wars Alex.3302) ‘With the least wind the fabric of a spider’s web is destroyed and is nothing’.

Table 12.1 below lists all the verbs belonging to the destroy-class and their valency in Middle English. NB: No lexeme meaning ‘wreck’ was Table 12.1  Labile and non-labile Middle English verbs in the destroy-class in MED Labile

Non-labile

anienten, forfaren, astruien, awerden, corrupten, diliʒen, destroien, perishen, shenden, fordon, forspillen, forwirchen, outenden, squachen, spillen, tinen: ‘destroy’ stroien, teren, tocasten, todreosen, toshenden, tospillen, towasten: ‘destroy’ westen: ‘waste’

adwinen, aswinden awesten, forclingen, iwesten towesten, wansen ‘waste’ amerren, defeten: ‘ruin’ forherghien: ’ravage’ abaten, toslen: ‘demolish’

Labile: 7

Non-labile: 29

12  Language Contact Effects on Verb Semantic Classes 

353

obtained that was not already identified by the use of the probe ‘destroy’, leaving five meaning categories: ‘destroy’, ‘waste’, ‘ruin’, ‘ravage’ and ‘demolish’. As can be seen, only seven out of the altogether thirty-six ME destroy-­ class verbs (19%) were attested in MED as labile; two of them, namely anienten and perish, are of Anglo-French origin. Four native destroy-verbs (shenden, spillen, anienten and perishen) became labile, the first two of which had been non-labile in Old English, e.g. intransitive uses of shenden in (17b) above and in (19): (19)  Whi nult þu hider wenden? þu lezst us her scenden (Lay.Brut (Clg A.9) 9015) ‘Why will you not come here? You cause us to be destroyed here’.

Turning to destroy-class verbs in Anglo-French, target items were obtained by reverse look-up in the Anglo-Norman Dictionary (AND) using the same procedure as with Middle and Old English stated above. Table 12.2 shows that the majority of A-Fr. destroy-verbs (five out of nine) were labile. Examples of lability in Anglo-French were: (20) a. Si tost come il vous veit hardy et bald en la grace Dieu, sa force anentera et il tost s’en fuiera (Ancren 1186.21) ‘As soon as he sees you brave and bold in God’s grace, his force will vanish and he will quickly run away’.

Table 12.2 Anglo-French destroy-class verbs Labile

Non-labile

agravanter, anienter, degaster ‘to destroy’ / ‘to waste away’

anientisser, annuller, destrure ‘to destroy’

gaster ‘to ravage, destroy’ / ‘fall into disrepair’

corrumpre ‘to waste away’

perir ‘to cause the loss or destruction of’ / ‘perish’ Labile: 5

Non-labile: 4

354 

L. García García and R. Ingham

b.  […] pur ceo k’il avait ennenty par arsouns Viles et eglises (Langtoft, Wright ii, 362) ‘[…] because he had destroyed by fire towns and churches’. (21) a. Le cors degaster ne purra Pur ceo ke jamés ne murra (Lum Lais ANTS 11849) ‘The body will not be destroyed because it will never die’. b. Beneuré sunt cil ki cest fu eschaufe e enlumine, nent cil k’il embrase e deguaste (Joshua II 14) ‘Blessed are those who this fire warms and lights up, not those that it burns and destroys’.

The absence or near-absence of lability in Old English destroy-verbs and its greater type-frequency in Anglo-French strongly suggest the determining influence of the latter on the occurrence of lability in Middle English destroy-verbs. Given the data from Old and Middle English, the most likely hypothesis seems to us to be that lability in ME destroy-verbs was a short-term tendency brought to life by Anglo-French models, which did not last after the main period of influence, as there were no Old English exemplars on which to sustain it. In contrast, contact influence has had a large-scale lasting effect on English in verbs of change of state and change of position. Here, Old English already had some models to act as bridgeheads for the developing syntactic trend, but where it lacked them entirely, as with destroy-verbs, contact influence was not lasting. The temporary spread of lability into the ME destroy-verb class was plausibly influenced by the presence of labile verbs belonging to this semantic class in Old French.

5 Psych Verbs in Old and Middle English It will be recalled that lability is virtually non-existent among Present-­ Day English psych verbs, but that observational evidence of the extent of its presence in earlier stages of the language suggested that it merited

12  Language Contact Effects on Verb Semantic Classes 

355

more thorough investigation. In this section, appropriate analytic categories for this enquiry are first described, then results of similarly designed analyses to those pursued in the previous section of the paper are presented. There were in Old English three main kinds of linking between the Stimulus and Experiencer roles and grammatical function: (i) 

Stimulus as subject, Experiencer as (in)direct object, e.g. lician ‘to please’, as in (11), repeated here as (22):

(22)  Đam    wife      þa      word wel licodon (Beo 639) det.-dat. woman-dat. det.-nom.pl. words well please-pret.-3PL. ‘Those words pleased the woman well’. (ii) Experiencer as subject, Stimulus as direct object, e.g. wundrian (‘to admire’), e.g.: (23) Ic wundrige þa     stihtunge  þære    godcundan (GDPref 3 (C) 28.233.28) I admire-1sg. det.-acc. dispensation det.-gen. divine mildheortnesse mercy. ‘I admire this dispensation of divine mercy’. (iii)  I mpersonal null subject, Experiencer as indirect object, Stimulus as genitive-case object, e.g. (24)  Him ofhreow ðæs mannes (ÆCHom I, 13 281.12) He-dat.  feel pity-pret.-3sg. det.-gen. man-gen. ‘He was sorry for the man’.

Type (iii) verbs did not permit either Experiencer or Stimulus argument to be linked to subject. A verb appearing both in construction (ii) and in one of (i) or (iii) was labile, since alternative linkings of the Experiencer role were available, either to subject or to (in)direct object. The method used to obtain Old English psych verbs was first to identify the forerunners of Middle English psych verbs of native origin (see below and Appendix 3), which make up a large part of the Old English

356 

L. García García and R. Ingham

Table 12.3  Psych-verb types as attested in ASD and DOE Subject Exp. only Total 24 % 34

Object Exp. only

Labile, but ‘Object’ Exp. only in Labile impers. clause Total

33 46

11 16

3 4

71 100

psych verbs in the list in Appendix 1. These were collated and if necessary augmented by the items in Guidi’s (2011) study on Old English psych verbs, as well as by the psych verbs listed by García García (2020). The meaning and valency of the verbs from A to I was ascertained in the DOE. For the remainder of the alphabet, the ASD was our main source of information. A total of seventy-five psych verbs were found in Old English (Appendix 1). Their classification according to types (i)–(iii) above, and their combination, is presented in Table 12.3. The first column (‘subject Experiencer only’) corresponds to type (ii), the second includes types (i) and (iii), the third and fourth display the combination of types (i) and (ii), and (ii) and (iii) respectively. The table does not take into account Old English psych verbs which are attested only once (gæstan ‘to torment’ and wrāþian ‘to be angry’) or whose claim to lability rests on a single gloss example (gramian ‘to anger (so.)’ / ‘burn with anger’, gryllan ‘to provoke’ / ‘to gnash the teeth, rage’). Labile psych verbs in Old English made up 20% of the total. They included those with the meanings of ‘anger’, ‘trouble’, ‘doubt’, ‘shame’, ‘gladness’, ‘pleasure’ and ‘rejoicing’. In the examples that follow, gebelgan (25a and 25b) and gladian (26a and 26b) illustrate plain lability, where the Experiencer can be linked to subject or object (types (ii) and (i) above); (27) combines types (ii) and (iii) above, that is, the Experiencer can be linked to subject or object, but in the latter case only with an impersonal subject. gebelgan ‘to be, get angry’ / ‘to make (so.) angry’ (25) a. & gebealh heo swiðe eorlice wið hire sunu (Ch 1462 17) ‘and she got very angry with her son’

12  Language Contact Effects on Verb Semantic Classes 

357

b. […] þæt he […] ecean dryhtne bitre gebulge (Beo 2329) ‘[…] that he […] the eternal lord bitterly made angry’ gladian ‘to be, get glad’ / ‘to make glad’ (26)   a.  ahebbað þonne eowre heafda. þæt is gladiað on eowrum mode (ÆCHom I, 40 526.77) ‘raise then your heads. That is, rejoice (be glad) in your heart’ b.  mid þan we magon ure mod gastlice glædien (B) 35.176)

(ÆCHom II, 4

‘with which we can gladden our heart spiritually’ tweogan ‘to doubt’ / ‘to inspire doubt into a person (accusative)’with impersonal construction: (27) a. Hine wile tweogan, hwæðer heo him soð secge (HomU 21 35) ‘him will inspire doubt, whether she told him the truth’. b. Ic nat […] ymbe hwæt ðu giet tweost (Bo 5.12.26) ‘I don’t know […] about what else you doubt’.

All three construction types in (i), (ii) and (iii) continue into Middle English, though the genitive case is now rarely found expressing the Experiencer role. Lability continues to be found, e.g.: (28)  He wile wreðe wið þe (Lamb. Hom. I.33,8) ‘He will become angry with you’. (29)  We […] wreþe hyne wyþ sennes (Shoreham 124/270) ‘We anger him with sins’.

A new development is the appearance of verbs taking an Experiencer subject together with a co-referential reflexive pronoun, of the type ‘He

358 

L. García García and R. Ingham

wraths him’. Verbs of this kind were counted separately if the subject linking of the Experiencer was restricted to the reflexive construction. Other verbs also appearing optionally with a reflexive pronoun were noted but will not be further considered for the purpose of this paper. To establish the argument structure possibilities of Middle English psych verbs, the procedure was as in previous sections to use the Modern English psych-verb lists in Levin (1993: §31.1.–31.3) as search terms, in this case in the MED’s reverse look-up facility (search by definition). Whenever a verb in those lists occurred in the definition (of one of the senses) of a Middle English verb lexeme, the latter was included in the data analysis. Once again, verbs with single citations were not counted, as in such cases no alternation would be observable. Altogether 123 Middle English psych-verb lexemes were identified in this way, of which forty-nine came from French (36%). Numerous French loans restocked the object Experiencer category, as noted by van Gelderen (2011), e.g. surprise, please, delight, annoy, disturb, trouble, astonish, grieve, dismay and discomfort. French also contributed loans to the subject Experiencer category, such as adore, rejoice, despise, regret, repent and marvel. Forty-four verbs (35%) were inherited from Old English, the remainder being innovations or borrowings from other languages, not only French. The breakdown of these 123 verb lexemes into the analytic categories presented above was as in Table 12.4 below. Native verb types (Old Table 12.4 Psych-verb types as attested in MED entries, 1150–1450 Labile, but ‘Object’ Exp. Object Exp. / Subject Verb valency Subject Object only in Subject refl. Exp. + refl. type: Exp. only Exp. only Labile impers. clause pron. pron. Total Native French loan Total %

19 15 34 27.6

20 17 38 30.9

29 12 41 33.3

4 2 6 4.9

0 2 2 1.6

2 1 3 2.5

74 49 123

12  Language Contact Effects on Verb Semantic Classes 

359

English-origin, Scandinavian-origin and coinages in Middle English) are shown separately from French-origin loans.7 It can be seen that labile verb types made up a third of the psych verbs established by our search procedure. The situation in Old English, in which lability was limited among psych verbs, thus gave way to one in which the psych-verb lexicon accommodated lability a good deal more readily. However, it was still considerably less frequent than among verb types of change of state / position, where 72% are attested in the MED as labile (Ingham 2020). We next pursue the hypothesis of French influence as a reason for this development, by examining lability in psych verbs used in the French of England.

6 Anglo-French Psych Verbs In Old French lability was found by Heidinger (2010) to have been common in verbs of change-of-state / position (see also Hatcher 1942), but his study did not cover psych verbs. Research into the syntactic behaviour of these verbs in insular French (Anglo-French) was accordingly undertaken for the purpose of the present enquiry. The reverse look-up facility of the AND was used to mine data on the basis of Levin (1993) in the same way as described above for Middle English. In this way a total of seventy-six psych-verb lexemes was obtained. Depending on their argument structure patterning, citations were categorised into those taking only an object Experiencer,8 those taking only an object Experiencer with an impersonal subject, those taking only an Experiencer subject, and those that were labile, e.g. respectively (30), (31), (32) and (33a)–(33b) below: (30)  E, s’um le desturbast, ne seüst parfiner (Becket 2365) ‘And had someone disturbed him, he would not have managed to finish’.  Old Norse loans are classified here as ‘native’ on the basis that by the thirteenth century they had been in the language for up to half a millennium since the first Viking incursions. In addition, they are not always easy to distinguish etymologically from Old English lexemes (Durkin 2014). 8  Passive forms, e.g. il fu esbai (‘he was surprised’) were ignored, as they were analysable in a stative sense as consisting of estre ‘to be’ + adjectival past participle, and so did not necessarily attest to a transitive use. 7

360 

L. García García and R. Ingham

(31)  De vostre maladie me peise (Fabliaux 21.276) ‘I am troubled by your illness’. (32)  “Sire Reis”, fait li il, “bien devum doluser” (Becket 155) ‘“(My) Lord King”, he said to him, “We must feel great sorrow”’. (33)  a. De la voiz del tuen tuneire espowenterunt (Oxf Ps. 1.103.8) ‘By the voice of your thunder (they) will (be) frighten(ed)’. b. É Deu chalt pas sur les Phistiens tuná é forment les espoentá (Quatre Livres, p. 15) ‘And God quickly thundered over the Philistines and greatly frightened them’.

The results of the analysis illustrated in (30)–(34) above are shown in Table 12.5. As can be seen, verbs in the labile group were in a minority, though at 22% not a negligible one. Effectively, Anglo-French psych verbs were divided roughly between those taking a subject Experiencer only, and those taking an object Experiencer, some of which allowed lability, while some did not. The last attestation dates of each construction involved in lability were also recorded, and these showed an interesting temporal progression. Verbs with meanings involving ‘sadness’, ‘fear’, ‘enjoyment’, ‘delight’, ‘trouble’, ‘astonishment’, ‘dismay’ and ‘shame’ were attested in the twelfth century as labile, appearing not only in Experiencer object examples but also attesting to the Experiencer subject non-­ reflexive construction:

Table 12.5  Psych-verb categories as attested in AND entries (twelfth to fourteenth century) Subject Exp. only

Object Exp. only

Labile

Total

38 (50.7%)

20 (26.7%)

17 (22.6%)

75

12  Language Contact Effects on Verb Semantic Classes 

361

(34)  Esjoirat Jacob e eslecerat Israel (Camb Ps 18.XIII.10) ‘Jacob will rejoice and Israel will rejoice’. (35)  Ne nul homme ne coroucera ové celuy qui la portera (Lapid 281) ‘And no man will get angry with the one who shall carry it’.

Examples of this construction were found in Old English, as noted above, with verbs of often directly comparable meanings. However, with nine of the seventeen Anglo-French verb types collected, the non-­reflexive subject Experiencer construction no longer occurred in insular French after the twelfth century. By the thirteenth century, the majority of verb types previously labile now used the reflexive subject Experiencer construction, e.g.: (36)  Quant homme se esjoit de suen mal (Mirour Egl 24.17) ‘When a man rejoices in his evil’. (37) Car li jugges se couroussent volunters quant om apele de eus (Lett de Rois 51) ‘For the judges easily grow angry when one appeals against them’.

In the fourteenth century, hardly any were still attested as labile with the non-reflexive subject Experiencer construction as in (33a)–(33b) above. Thus, Anglo-French was tending to lose lability in this verb domain by the time the period of intense French contact influence on English is encountered, roughly 1220–1400. French influence on English favouring lability would hence not be expected on the scale that took place with change-of-state / location verbs, where a great number of Anglo-French verbs continued to be labile throughout this period (Ingham 2020). In short, insular French would not have provided a strong model for psych-verb lability in Middle English. There was already some use of lability in Old English, which was not massively increased by contact with insular French, unlike with verbs of change of state and change of position. The much weaker incidence of lability in the psych-verb domain is in line with the hypothesis that argument structure in Middle English was most open to contact influence from French during the period of

362 

L. García García and R. Ingham

greatest lexical borrowing from that language. Lability among Anglo-­ French verbs of change of state and change of position remained very common throughout that time, whereas as we have seen among psych verbs it declined quite sharply. Indeed, French influence may even, by the thirteenth century, have operated as a brake on the use of lability with Middle English psych verbs, favouring instead the subject Experiencer + reflexive construction.

7 Conclusions Our analysis of the two verb semantic domains investigated here has enabled some nuanced conclusions regarding the role of external and internal factors on language change, at least as regards verb argument structure. It was found that in the destroy-verb class a certain number of items became temporarily labile in Middle English, whereas Old English had almost no exemplars among destroy-verbs. The finding that a good proportion of Anglo-French destroy-verbs were found to be labile appears not to have been decisive: any contact influence, attested by the Middle English citations attesting to lability, was clearly not lasting. Thus, although there may have been some external pressure from French towards lability, the language lacked an internal basis for lability in that semantic area. In the psych-verb semantic domain, Middle English made considerably more use of lability. Here the same two factors were found to be important. Old English already had some exemplars, which we would see as having constituted a kind of ‘bridgehead’, from which lability could expand its scope, in a similar way to what transpired with change-of-­state / position verbs, and here a substantial number of verbs became labile in Middle English. However, the model that French could have constituted in this domain occurred too early for contact influence to have played a decisive role; by the time bilingualism (and thus contact influence) became established among the educated members of the English speech community, lability among French psych verbs was already very recessive. Whereas in Anglo-French nearly a quarter of psych verbs were found to be labile, the non-reflexive subject Experiencer construction gave way to the subject reflexive Experiencer construction after the twelfth century. Nevertheless,

12  Language Contact Effects on Verb Semantic Classes 

363

we would judge from our findings that the combined effect of contact influence from Anglo-French and the existence of lability in a fair number of Old English psych verbs provided the language with the momentum to develop lability somewhat further in this semantic domain than with destroy-class verbs. In both semantic domains, the temporary rise of lability appears to have been a failed change, never going to completion as it has largely done in the change-of-state / position semantic domains. As noted, the combined influence of Old English and Anglo-French was too weak to allow lability to put down strong roots among destroy-verbs, where even in Middle English the innovation was clearly failing. With psych verbs, we have, however, noted informally from OED entries that lability continues to be attested quite often in the Early Modern period until apparently growing obsolescent in the seventeenth century. It has not been possible in this study to take detailed account of developments in the post-medieval period, which would undoubtedly merit further attention. One suggestive point that has emerged from this inquiry is that psych-­verb lability in the source language (French) was being challenged strongly, and crucially at the period of maximal contact influence on English, by a competitor, the reflexive construction, as discussed in Sect. 6. This factor of competition may have had the effect of putting a brake on the spread of lability right across the English psych-verb domain, similar to Postma’s (2010) finding that a competitor form led to a failed change in the context of languageinternal development in Dutch. Some limitations must be acknowledged. Dictionaries as a data source do not normally provide evidence of the relative textual frequencies of the lexemes they feature, nor of the constructions in which they appear. Useful information on the development of verb alternations could be gained by searching a Middle English text corpus, which in particular cases might show that one or other alternant never achieved more than very minimal representation, and quickly died out. It could be that in some cases the lability noted in the dictionaries consulted for this research could be hapax attestations. There again, since available Middle English texts, especially prior to the fourteenth century, are not greatly numerous and are strongly skewed towards particular text types, it could not be claimed that the frequencies to be derived from them are representative of language use at the time the English language was undergoing contact influence from French.

364 

L. García García and R. Ingham

A further limitation is that account has not been taken of possible dialectal or idiolectal variation. Conceivably, the appearance of lability taken from our sources could be a compilation of a speaker variety in which a verb was used only in one construction by one type of speaker, and only in the other construction by another type of speaker. The resources at our disposal hardly allow us to resolve this issue. Nevertheless, some broader conclusions can be drawn. The semantic criteria model of Pinker (1989) and Levin (1993) may be valid in synchronic terms, but evidently can hardly have applied in the same terms to earlier stages of the language. Some verbs which now conform to the semantic constraints on a given alternation evidently did not in the past, while other verbs have done so for a very long time, including the Old English verbs of change of state researched by García García (2020). The constraint now ruling out lability with destroy-class verbs may have existed in Old English, but then appears to have been relaxed temporarily in Middle English, under contact influence as we argue, only to be reinstated later. In both cases, the semantic criteria valid for Present-Day English may be seen over time as emergent, appearing via a process whereby over time more and more verbs were used according to those criteria, while those that did not fit the criteria ceased being used in ways that would have violated them. This is naturally speculative and attendant on further research into questions of argument structure diachrony.

Appendix 1: Old English Psych Verbs9 OE psych verbs

Valency OE psych verbs

Valency

byldan ‘to embold’

O

geirsian ‘to be angry’; ‘to make L angry’

belgan ‘to swell with anger’

S

lāþian ‘to be hateful or loathed’

S

abelgan ‘to cause so. to swell up in anger’

O

mislǣdan ‘to mislead’

O (continued)

 Key: L = labile; O = Experiencer object only; S = Experiencer subject only.

9

365

12  Language Contact Effects on Verb Semantic Classes  (continued) OE psych verbs

Valency OE psych verbs

Valency

gebelgan ‘to become angry’; ‘to anger, enrage’

L

lician ‘to please’; ‘to take pleasure in’ (= be pleased by)

O

bliþian ‘to rejoice’; ‘to gladden’

L

gelician ‘to please’

O

ābylgan ‘to anger, offend’

O

mislician ‘to displease’

O

carian ‘to sorrow’ (int.); ‘to take care’ (tr.)

S

oflician ‘to displease, be displeasing (to so.)’

O

cwacian ‘to shake’

S

lustfullian ‘to rejoice, be glad’

S

cweman: imp. ‘to please, satisfy’

O

gelustfullian ‘to be delighted’; ‘to delight’

L

gecweman ‘to please, satisfy’

O

lufian ‘to love’

S

miscweman ‘to displease’

O

lystan ‘to cause pleasure or desire’ (pers., impers.); ‘to desire’

L

dreccan ‘to afflict oppress’

O

gelystan ‘to please, cause a desire’ (impersonal)

O

drefan ‘to disturb, agitate, oppress’

O

gemǣdan ‘to make insane’

O

(a)drǣdan ‘to dread, fear’ (adrǣdan w. dat. reflex)

S

(be)mǣnan ‘to lament, mourn, complain’

S

dwellan ‘to lead into error; err (intr.)’

L

niþerian ‘to bring low, humiliate’

O

dyrfan ‘to be diligent’; ‘to torment’

O

sceamian ‘to feel or cause shame’ (impersonal)

L

eargian ‘to be or become timid’; ‘to fear’ (tr.)

S

forsceamian ‘to be or make greatly ashamed’ (impersonal)

L

faegnian ‘to rejoice’; ‘to welcome’

S

swencan ‘to cause a person to labour, harass, afflict’

O

(a)faeran ‘to terrify’

O

geswencan ‘to disturb, vex, afflict’

O

forhtian ‘to fear’

S

swincan ‘to toil, labour, work with effort’

S

aforhtian ‘to be, become afraid, fear, dread’

S

scendan ‘to put to shame’

O

frefran ‘to console, comfort’

O

sorgian ‘to care, grieve’

S (continued)

(continued) OE psych verbs

Valency OE psych verbs

Valency

gæstan ‘to torment’ (only 1x)

O

teonian ‘to vex, irritate’

O

gladian ‘to be glad, make glad’

L

tirgan ‘to vex, provoke’

O

gramian ‘to anger (so.)’; ‘to L? burn with anger’ (gloss)

tregian ‘to vex, annoy’

O

gremman ‘to enrage, provoke’

tweogan ‘to feel, cause doubt’ (imp.)

L

O

grimman ‘to rage, vent fury S (intr.)’

getweogan ‘to doubt, hesitate’ S

agrīsan ‘to be afraid, dread’ S

for-þencan ‘to despair’ (always refl.)

S

gryllan ‘to provoke’; ‘to gnash the teeth, rage’ (gloss)

L?

of-þyncan ‘to cause regret or displeasure’ (imp.)

O

hatian ‘to hate’

S

ā-þrēotan ‘to be wearisome, tedious, distasteful’

S

ā-þry¯tan ‘to weary, tire out (so.)’

O

hefigian ‘to become heavy’; L ‘to make heavy, burden’ gehefigian ‘to become heavy’; ‘to make heavy, burden’

L

wlatian ‘to cause a person (ac.) loathing’ (imp.)

O

heortan ‘to hearten, encourage’

O

wōrian lit. ‘to wander’; fig. ‘to hesitate, err’

S

O hreowan ‘to distress’, ‘to affect (so.) with compassion for so. else’,

wrāþian ‘to be angry’ (only 1x)

S

gehreowan ‘to distress’, ‘to cause to regret’

O

gewræþan ‘to be wroth, savage’; ‘to attack, molest’ (BT & BT Addenda)

L

hreowsian ‘to grieve, lament, regret’

S

wundrian ‘to wonder at’; ‘to make wonderful’

S

behreowsian ‘to regret, lament, repent, feel pity’,

S

weallan ‘to bubble forth’ (cp. meaning wyllan ‘to boil (sth.); torment, agitate (so.)’)

O

irsian ‘to be angry’, ‘to make angry’

L

Total

74

Labile Experiencer object Experiencer subject Not conclusive

14 34 24 2

12  Language Contact Effects on Verb Semantic Classes 

367

 ppendix 2: Encoding of Experiencer Role, AngloA French Verbs, as Attested in AND Subject Experiencer

Object Experiencer

Labile

Total

abhomer ‘to abhor’

acurager ‘to encourage’

contrister ‘to sadden, become sad’

Subj. Experiencer 38

admirer ‘to admire’

agreger ‘to aggravate’

corucer ‘to anger, become angry’

Obj. Experiencer 20

aduluser ‘to deplore’ apeiser ‘to calm’ deduire ‘to amuse, be amused’

Labile 17

affiancer ‘to promise’

asuager ‘to soothe’

deliter ‘to delight, be delighted’

Total 75

affier ‘to promise’

commovoir ‘to upset’

rejoir ‘to gladden, rejoice’

agarder ‘to observe’

conforter ‘to comfort’

travailler ‘to torment suffer’

amer ‘to love’

cuivrer ‘to vex’

vergonder ‘to shame, be ashamed’

aurer ‘to honour’

desturber ‘to disturb’

enpourir ‘to frighten, take fright’

aviler ‘to dishonour’

embaudir ‘to embolden’

esmaier ‘to dismay, be dismayed’

chericer ‘to cherish’

encumbrer ‘to harass’

espourir ‘to frighten, take fright’

cherir ‘to cherish’

enoier ‘to trouble’

esbahir ‘to amaze, be amazed’

confier ‘to trust’

enorter ‘to exhort’

effreer ‘to frighten’

creire ‘to believe’

entarier ‘to irritate’

esjoir ‘to rejoice’

criembre ‘to fear’

esmovoir ‘to stir esleecer ‘to rejoice’ up’

dedeignier ‘to disdain’

estoner ‘to astonish’

heitier ‘to gladden’ (continued)

368 

L. García García and R. Ingham

(continued) Subject Experiencer

Object Experiencer

se desafier ‘to distrust’

grever ‘to afflict’

irer ‘to anger’

desesperer ‘to despair’

hunir ‘to humiliate’

espoenter ‘to frighten’

despire ‘to despise’

peser ‘to calm’

doleir ‘to suffer’

trubler ‘to trouble’

doluser ‘to deplore’

turmenter ‘to torment’

duter ‘to fear’ endoler ‘to afflict’ envier ‘to envy’ esbanoier ‘to amuse oneself’ esmervoiller ‘to amaze’ esmirer ‘to admire’ hair ‘to hate’ honurer ‘to honour’ joir ‘to enjoy’ leescer ‘to delight’ pleindre ‘to complain’ preiser ‘to appreciate’ reduter ‘to fear’ regreter ‘to regret’ repentir ‘to repent’ reverencer ‘to worship’ reviler ‘to revile’ s’asseurer ‘to feel confident’

Labile

Total

369

12  Language Contact Effects on Verb Semantic Classes 

 ppendix 3: Native (OE-Origin and Newly A Coined) Psych Verbs Attested in MED, with Old English Source Verb where Relevant10 Labile in OE

Labile in ME

Labile ME accenden ‘to stir, become stirred’ agasten ‘to frighten, become frightened’ angren ‘to anger, become angry’ arghen ‘to intimidate, be timid’ agrisen ‘to terrify, be terrified’ blithen ‘to be / make happy’ bolden ‘to be / make bold’ dreden ‘to frighten, fear’ fainen ‘to be / make happy’ feren ‘to frighten, fear’ forþinken ‘to be / make sorry’ gamen ‘to be / make merry’ gladen ‘to be / make glad’ grillen ‘to offend, suffer’ hevien ‘to trouble, be troubled’ iwreþen ‘to anger, become angry’

Labile in OE

Labile in ME

X

S

–

S

X

S

X

S

X

S

–

✓

X

✓

–

✓

–

✓

X

✓

X

✓

Subject Experiencer ME belwen ‘to become angry’ abhorren ‘to abhor’ arwurðen ‘to respect’ bimenen ‘to deplore’ caren ‘to grieve’ haten ‘to hate’

X

S

–

✓

lipnen ‘to trust’ –

S

X

✓

–

S

X

✓

–

S

X

✓

–

S

X

✓

–

S

X

✓

–

S

✓

✓

–

S

X

✓

X

S

✓

✓

liten ‘to delight’ mistresten ‘to mistrust’ mistrouen ‘to disbelieve’ mornen ‘to grieve’ overhouen ‘to despise’ overtrouen ‘to mistrust’ quaken ‘to be upset’ reuen ‘to rue’

X

S

X

✓

X

S

sorwen ‘to feel sorrow’

(continued) 10

 Key: ✓ = labile, X = non-labile, S = Experiencer subject only, O = Experiencer object only.

370 

L. García García and R. Ingham

(continued) Labile in OE

Labile in ME

irken ‘to make / become weary’

–

✓

liken ‘to like, please’

X

✓

listen ‘to cause / feel desire’ loþen ‘to loathe’ madden ‘to make / be mad’ ofþinken ‘to grieve, regret’ shamen ‘to cause / feel shame’ souen ‘to cause / feel sorrow’ swenchen ‘to afflict, suffer’ tenen ‘to anger, become angry’ wroþen ‘to anger, become angry’ misliken ‘to displease / dislike’ gramen ‘to make / be angry’ herten ‘to encourage’ trusten ‘to trust’

✓

✓

X –

✓ ✓

X

✓

✓

✓

–

✓

X

✓

X

✓

–

✓

X

✓

X

✓

–

S + refl.

–

S + refl.

–

S, O + impers. S S, O + impers. S S, O + impers. S S, O + impers. S

biloven ‘to be inclined’ lusten ‘to please’

–

uggen ‘to be fearful’ – wondren ‘to amaze’

X

Labile in OE

Labile in ME

traisten ‘to trust’

–

S

wlaten ‘to feel disgust’ worshipen ‘to worship’

X

S

–

S

Object Experiencer ME auen ‘to terrify’ –

O

cheren ‘to console’ derfen ‘to harass’ drecchen ‘to oppress’ dreven ‘to trouble’ frevren ‘to comfort’ gasten ‘to frighten’ ihevien ‘to afflict / suffer’ iswenchen ‘to harass’ mirþen ‘to amuse’ misleden ‘to mislead netheren ‘to humiliate’ quemen ‘to please’ shenden ‘to shame’ terren ‘to vex’ treien ‘to annoy’

–

O

X

O

X

O

X

O

X

O

X

O

✓

O

X

O

–

O

–

O

–

O

X

O

X

O

– X

O O (continued)

371

12  Language Contact Effects on Verb Semantic Classes  (continued) Labile in OE

Labile in ME vexen ‘to vex’ woren ‘to confuse’ wrixlen ‘to confuse’

Labile in OE

Labile in ME

X –

O O

–

O

Total Old English Labile Non-labile Verb unattested Total

Middle English 5 38 31 74

Subj. Experiencer only Subj. Experiencer + reflexive Obj. Experiencer only Subj. Experiencer/O + impers S Labile Total

19 2 20 4 29 74

 ppendix 4: Encoding of Experiencer Role, A French-Origin Psych Verbs Attested in MED up to 1500 Labile affraien ‘to harass / be afraid’

✓

enjoien ‘to rejoice’

S + refl.

agasten ‘to make / become afraid’

✓

envien ‘to envy’

S

deliten ‘to give / take pleasure’

✓

humblisshen ‘to feel humble’

S

dismaien ‘to dismay, be dismayed’

✓

maien ‘to be frightened’

S

greven ‘to harass / be angry’

✓

prisen ‘to esteem’

S

masen ‘to confuse / be confused’

✓

recomforten ‘to comfort’

S

merveillen ‘to surprise / be surprised’

✓

sotten ‘to be stupid’

S

rejoicen ‘to give / take pleasure’

✓

remorden ‘to cause / feel remorse’

✓

Object Experiencer (continued)

372 

L. García García and R. Ingham

(continued) savouren ‘to give pleasure / relish’

✓

acoien ‘to soothe’

O

stonen ‘to astonish / be astonished’

✓

acombren ‘to burden’

O

troublen ‘to make / become agitated’

✓

agreven ‘to trouble’

O

amesen ‘to soothe’

O

comforten ‘to comfort’ O Subject Experiencer

destourben ‘to disturb’ O

douten ‘to fear’

S

disesen ‘to inconvenience’

O

endeinen ‘to take offence’

S

egren ‘to excite’

O

adouren ‘to worship’

S

mispaien ‘to irritate’

O

affien ‘to trust’

S

misplesen ‘to displease’ O

aggrugen ‘to resent’

S

noien ‘to afflict’

O + impers. S

cherishen ‘to appreciate’

S

paien ‘to satisfy’

O + impers. S

despisen ‘to despise’

S

perturben ‘to disturb’

O

discomforten ‘to discomfort’

O, S + refl.

plesen ‘to please’

O

disdeinen ‘to disdain’

S

recomforten ‘to comfort’

O, S + refl.

dolen ‘to suffer’

S

solasen ‘to comfort’

O

stourben ‘to disturb’

O

tarien ‘to vex’

O

Total Labile Subj. Experiencer only Obj. Experiencer only Obj. Experiencer + impers. Subj. Obj. Experiencer/Subj. reflexive Subject Experiencer + reflexive Total

12 15 17 2 2 1 49

12  Language Contact Effects on Verb Semantic Classes 

373

References Alexiadou, Artemis. 2016. English Psych Verbs and the Causative Alternation: A Case Study in the History of English. Questions and Answers in Linguistics 3 (2): 1–13. Allen, Cynthia. 1995. Case Marking and Reanalysis: Grammatical Relations from Old to Early Modern English. Oxford: Clarendon Press. [AND =] Anglo-Norman Dictionary. 1977–1992. Ed. by Louise W. Stone, T. B. W. Reid and William Rothwell. Revised edition, A-E. 2005. Ed. by William Rothwell, Stewart Gregory and David Trotter. AND2 Online Edition. 2021. Ed. by David Trotter (–2015) and Geert de Wilde (2015–). Aberystwyth: Aberystwyth University. https://anglo-­norman.net. Accessed 31 May 2023. Bacovcin, Hezekiah Akiva. 2017. Parameterising Germanic Ditransitive Variation: A Historical-Comparative Study. Unpublished doctoral thesis, University of Pennsylvania. Bosworth, J., and T. Toller (eds). 1898. An Anglo-Saxon Dictionary. Oxford: Clarendon Press. Online edition available at https://bosworthtoller.com/. Accessed 31 May 2023. Dictionary of Old English Corpus in Electronic Form. 2009. Ed. by Antonette diPaolo Healey, Joan Holland, David McDougall, Ian McDougall, and Xin Xiang. Toronto: Toronto University Press, on CD-ROM. [DOE =] Dictionary of Old English: A to I online. 2018. Ed. by Angus Cameron, Ashley Crandell Amos, and Antonette DiPaolo Healey. Toronto: Dictionary of Old English. https://www.doe.utoronto.ca/pages/index.html. Accessed 31 May 2023. Durkin, Philip. 2014. Borrowed Words. Oxford: Oxford University Press. García García, Luisa. 2020. The Basic Valency Orientation of Old English and the Causative ja-Formation: A Synchronic and Diachronic Approach. English Language and Linguistics 24: 153–177. García García, Luisa, and Esaúl Ruiz Narbona. 2021. Lability in Old English Verbs: Chronological and Textual Distribution. Anglia 139: 283–326. van Gelderen, Elly. 2011. Valency Changes in the History of English. Journal of Historical Linguistics 1: 106–143. ———. 2014. Changes in Psych-Verbs: A Reanalysis of Little v. Catalan Journal of Linguistics 13: 99–122. Guidi, Luiz Guilherme. 2011. Old English Psych Verbs and Quirky Experiencers. York Papers in Linguistics, 2nd ser. 11: 29–48.

374 

L. García García and R. Ingham

Haspelmath, Martin. 1993. More on the Typology of Inchoative/Causative Verb Alternations. In Causatives and Transitivity, ed. Bernard Comrie and Maria Polinsky, 87–120. Amsterdam: John Benjamins. Hatcher, Anna Granville. 1942. Reflexive Verbs: Latin, Old French, Modern French. Baltimore: Johns Hopkins Press. Heidinger, Stefan. 2010. French Anticausatives: A Diachronic Perspective, Linguistische Arbeiten, 537. Berlin: de Gruyter. Hermodsson, Lars. 1952. Reflexive und intransitive Verba im älteren Westgermanischen. Uppsala: Almqvist & Wiksells. Ingham, Richard. 2012. The Transmission of Anglo-Norman: Language History and Language Acquisition, The Language Faculty and Beyond, 9. Amsterdam: John Benjamins. ———. 2020. How Contact with French Drove Patient-Lability in English. Transactions of the Philological Society 118: 447–467. Levin, Beth. 1993. English Verb Classes and Alternations: A Preliminary Investigation. Chicago: University of Chicago Press. McMillion, Alan. 2006. Labile Verbs in English. Unpublished doctoral thesis, University of Stockholm. [MED =] Middle English Dictionary. 1952–2001. Ed. by Robert E. Lewis. Online edn in Middle English Compendium. 2000–2018. Ed. by Frances McSparran and Paul Schaffner. Ann Arbor: University of Michigan Press. https://quod.lib.umich.edu/m/med/. Accessed 31 May 2023. Miura, Ayumi. 2014. Middle English Verbs of Emotion and Impersonal Constructions: Verb Meaning and Syntax in Diachrony. Oxford: Oxford University Press. Möhlig-Falke, Ruth. 2012. The Early English Impersonal Construction: An Analysis of Verbal and Constructional Meaning. Berlin: de Gruyter. Ottóson, Kjartan. 2013. The Anticausative and Related Categories in the Old Germanic Languages. In Diachronic and Typological Perspectives on Verbs, ed. F. Josephson and I. Söhrman, 329–381. Amsterdam: John Benjamins. Pinker, Steven. 1989. Learnability and Cognition: The Acquisition of Argument Structure. Cambridge, MA: MIT Press. Postma, Gertjan. 2010. The Impact of Failed Changes. In Continuity and Change in Grammar, ed. Anne Breitbarth, Christopher Lucas, Sheila Watts, and David Willis, 269–302. Amsterdam: John Benjamins. Rappaport, Malka, and Beth Levin. 1995. Unaccusativity. Cambridge: Cambridge University Press. Rothwell, William. 1993. The ‘Faus franceis d’Angleterre’. In Anglo-Norman Anniversary Essays, ed. Ian Short, 309–326. London: Birkbeck College.

12  Language Contact Effects on Verb Semantic Classes 

375

Stroobant, J. 2014. Factors Triggering P-Lability from Old English to Present-­ Day English. Unpublished master’s thesis, University of Ghent. Sylvester, Louise, Megan Tiddeman, and Richard Ingham. 2022. Lexical Borrowing in the Middle English Period: A Multi-Domain Analysis of Semantic Outcomes. English Language and Linguistics 26: 237-261. Trotter, David. 2003. L’Anglo-Normand: Variété insulaire ou variété isolée? Médiévales 45: 43–54. Visser, Frederik. 1963. An Historical Syntax of the English Language, I. Leiden: Brill. Wilshere, A. 1993. A Plea for Syntax. In Anglo-Norman Anniversary Essays, ed. Ian Short, 395–404. London: Birkbeck College.

13 Exploring Norn: A Historical Heritage Language of the British Isles Kari Kinn and George Walkden

1 Introduction In this chapter, we investigate the syntax of Norn, a descendant of Old Norse spoken in the Northern Isles of Scotland during the medieval and early modern periods. In doing so, we focus on its status as a contact language—not, as several others have, from the perspective of its

K. Kinn (*) University of Bergen, Bergen, Norway e-mail: [email protected] G. Walkden University of Konstanz, Konstanz, Germany e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_13

377

378 

K. Kinn and G. Walkden

influence on Scots, but rather from the perspective of potential contact influence from (Older) Scots on Norn itself.1 Our novel theoretical contribution is that we argue that Norn is a historically attested heritage language, and hence that it can be studied using some of the methods and questions of modern heritage language research. Empirically, we contribute a corpus study of possessives in Norn texts, and compare this to a baseline corpus of texts from contemporary Norwegian. Our results show some stability, but also change. Focusing on a syntactic phenomenon in Norn is also a novelty; previous research on Norn has focused overwhelmingly on phonological, morphological and lexical phenomena.2 The structure of the chapter is as follows. Section 2 provides background on Norn and its sociolinguistic and textual history, and Section 3 outlines our assumptions and rationale for treating Norn as a heritage language. The core of the paper is Sect. 4, which presents our case study of possessive constructions. Section 5 discusses and contextualises our findings, and Sect. 6 concludes.

2 Norn: Background and History Norn is a North Germanic variety, descended from Old Norse (ON), which was spoken in the Orkney and Shetland Islands off the northern coast of mainland Scotland. It now qualifies as a dormant language: There is evidence that it was spoken until at least the middle of the eighteenth

 Versions of this work were presented at the Workshop on Medieval English in a Multilingual Context, Cardiff, January 2020, at the Forum for Germanic Language Studies (FGLS), Bristol, January 2020, and at the RUEG conference ‘Dynamics of Language Contact’, Berlin (online), February 2021. We are grateful to audiences at all of these events and to various other colleagues, in particular Tanja Kupisch for her encouragement of this project, Arjen Versloot for drawing our attention to relevant literature, Sarah Einhaus for editorial work, the anonymous reviewer of this chapter for their comments, and the network organisers for bringing us together in Seville and Cardiff. None of these people should be assumed to agree with us. 2  The only exceptions that we are aware of are unpublished works by Eekman (2012, 2015). These focus primarily on clausal constituent order, though Eekman (2015: 10–11) does mention in passing that there is variation in possessive constructions. His conclusion is that Norn predominantly retains Scandinavian syntax, even at a late stage. 1

13  Exploring Norn 

379

century, but it currently has no native speakers, and has not done for a long time. Historical discussion of Norn can be found in Barnes (1998), on which the following summary is largely based. North Germanic varieties were first introduced to the Northern Isles of what is now Scotland in around 800 CE by settlers, predominantly from western Norway. Orkney and Shetland were ruled by Scandinavian earls until the thirteenth century; it was not until the fourteenth century that a Scotsman was first appointed Earl of Orkney (and Shetland). Formally, the Northern Isles did not become part of Scotland until 1472, and the first official documents in Scots date to the fifteenth century. Before this time, Old Norse / Norn and Latin were used in official contexts. The history of Norn after Orkney and Shetland became part of Scotland is murkier, and there is some debate around when exactly the variety stopped being spoken. Rendboe (1984) argues that it was still spoken as late as the nineteenth century (see also discussion in Wiggen 2002). This position is not widely accepted today; Knooihuizen (2008), synthesising and building on previous research, suggests that Norn stopped being learned at home some time between the late seventeenth century and 1774. Peripheral though the Northern Isles might seem from a present-day British perspective, their position meant that in medieval times they were well connected by water to a variety of locations around the North Sea and beyond. With these connections, of course, came a wide variety of language contact scenarios. The Neolithic village at Skara Brae on the Orkney mainland shows that the isles have been inhabited for at least 5000 years; at the time that speakers of North Germanic varieties settled there, the previous inhabitants would have been speakers of Pictish and / or Gaelic, though we have no written records of those varieties from that time and place.3 Between 1400 and 1700, speakers of Norn also enjoyed substantial contact with (Low) German and Dutch-speaking sailors (Melchers 1987).  The current scholarly consensus is that Pictish was a Celtic (hence, Indo-European) language more closely related to Brittonic than to Gaelic; see Rhys (2015) for discussion. On Celtic influence on Norn, see Lindqvist (2015). 3

380 

K. Kinn and G. Walkden

The contact scenario that is most key for our purposes, however, is contact with (Older) Scots. Norn and Scots would have been in contact from the fourteenth century at the latest, with varying degrees of intensity. Of course, Norn ended up being supplanted entirely by Scots as the home language of the speech communities of the Orkney and Shetland Islands; North Germanic influence on the Scots spoken and written on the Northern Isles is well documented and studied (see van Leyden 2004; Melchers 2008a and 2008b, for instance). This language shift must have taken several centuries to run its course. It is during this period of shift that we have textual evidence for Norn. The transmitted texts date to two periods specifically. On the one hand, during the late medieval period—from roughly 1300–1500—Norn was a language used for administrative and official purposes, and several charters of Orkney or Shetland provenance in the Diplomatarium Norvegicum (hereafter DN) can be dated precisely to this period:4 The earliest is from 1299, and the latest is from 1509. Goudie (1904) presents a further charter from 1516–1545, not included in DN; Barnes (1998: 12) additionally mentions a document  from 1586 (Orkney and Shetland Records (OSR)  74) and an unpublished copy of a 1431 document (c. 1550) An issue relating to these documents is that the scribes are likely to have received training in Norway (Barnes 1998: 11). This implies that developments in Norwegian scribal practices may be reflected also in documents from Orkney and Shetland; such developments include influence from contemporary Danish, especially in the later charters (see further discussion in Sects. 4 and 5). The remaining textual material is of a very different nature. George Low, a clergyman, visited Orkney and Shetland in 1774; the report of his travels was published in 1879. Low collected an amount of fragmentary material, including a ballad of thirty-five stanzas, now known as Hildina. A sample is given in (1), following the edited version in Hægstad (1900).5  DN is a twenty-two-volume collection of charters written in Old and Middle Norwegian; see Blaxter (2017b: Ch. 2, 2021). 5  The translation to English is from the website https://nornlanguage.x10.mx/ (accessed 25 May 2023). The author of this website is anonymous; the translation follows Kershaw (1921: 217). 4

13  Exploring Norn 

381

(1) Da vara Iarlin o Orkneyar       ‘It was the Earl from Orkney, for frinda sı ̆n spirde ro,        And counsel of his kin sought he, whirdì an skildè meun          Whether he should the maiden or vannaro eidnar fuo –          Free from her misery. Or glasburyon burtaga.        “If thou free the maid from her gleaming ‘Tega du meun our glasburyon,     hall, kere friendè min,             O kinsman dear of mine, yamna meun eso vrildan stiendi,    Ever while the world shall last gede min vara to din’.        Thy glory still shall shine’”.

Low’s material is also highly informative as regards the date at which Norn ceased to be acquired, as the following two quotes from Low (1879 [1774]: 105) illustrate: The Norse Language is much worn out here, yet there are some who know a few words of it; it was the language of the last age, but will be entirely lost by the next … . None of them can write their ancient language, and but very few speak it: the best phrases are all gone, and nothing remains but a few names and two or three remnants of songs which one old man can repeat, and that but indistinctly.

Low is hardly an unbiased observer; he also describes Norn as ‘entirely confined to the lower class of people, who cannot be supposed to have a thorough knowledge either of one Language or the other’ (Low 1879 [1774]:107). The nature of the material itself, however, yields more information. William Henry, Low’s speaker, was a farmer in Guttorm on the outlying island of Foula in Shetland. He was able to provide a translation to Hildina, but the translation mapped only partially to the ballad itself (Hægstad 1900: 31–32). Knooihuizen (2005, 2008) suggests that Henry had no active proficiency in Norn, but was rather a ‘rememberer’ in the terms of Dorian (1982), with competence limited to a (potentially extensive) list of set phrases and texts. The list of texts that we draw upon for the purposes of our own study is given in Table 13.1.6 Obviously, there are substantial limitations on the  The document OSR 74 (1586) and the unpublished copy of a 1431 document ( c. 1550, Barnes 1998: 12) remain to be investigated. 6

382 

K. Kinn and G. Walkden

Table 13.1  Investigated texts from Orkney and Shetland Type

Text

Date

Provenance

Words

Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Ballad Total

DN I.89 DN I.109 DN II.168 DN II.170 DN I.340 DN III.284 DN III.310 DN I.404 DN II.691 DN II.797 DN II.859 DN VI.651 Goudie (1904) Hildina

1299 1307 1329 1329 1354 1355 1360 1369 (b)1426 1452 1465 1509 1516–1545 (b)1774

Shetland Shetland Orkney Orkney Shetland Shetland Shetland Orkney Orkney Shetland Shetland Shetland Shetland Orkney

366 528 259 286 115 253 317 623 2,441 75 135 197 393 677 6,665

conclusions that can be drawn from any syntactic study based on such material, which amounts to well under 10,000 words. The texts are not evenly distributed across time and space, and the crucial late witness— Hildina—is of a different text type. Moreover, both charters and ballads pose challenges for syntactic research, albeit for different reasons. As a legal text type, charters are known for their use of formulaic language; for use of this genre for syntactic and other linguistic studies, see, for example, Farasyn et al. (2018) and Blaxter and Kinn (2018). Meanwhile, in poetic texts, syntactic alternations may be affected by metrical requirements. Any study of syntactic variation in this kind of text must be sensitive to these factors. For all these reasons, our conclusions can only be tentative ones.

3 Norn as a Heritage Language Our starting point for considering Norn as a heritage language is the widely adopted definition of Rothman (2009: 156), repeated here: ‘[A] language qualifies as a heritage language if it is a language spoken at home or otherwise readily available to young children, and crucially this language is not a dominant language of the larger (national) society.’ Some

13  Exploring Norn 

383

examples of heritage languages spoken worldwide today include Heritage English spoken in France, Israel, Japan and Thailand (Polinsky 2018: Chapter 2), Pennsylvania Dutch (e.g. Brown and Putnam 2015), American Heritage Russian (e.g. Dubinina and Polinsky 2013; Polinsky 2018), and American Heritage Norwegian (e.g.  Haugen 1953; Johannessen 2018). The past fifteen years in particular have witnessed a flurry of work on heritage languages from a variety of different standpoints, with the result that our understanding of the mechanisms that shape heritage languages has advanced considerably. Up to now, however, heritage language research has focused largely on languages of the present day, or at least of the last century or so. The papers in Brown (2019) are a welcome exception to this, bringing historical sociolinguistics and heritage language research together, with a focus on ego-documents. These papers, however, focus on more recent and indisputably heritage-language situations of the mid-to-late nineteenth century. Our contribution aims to demonstrate the applicability of the concepts and methods of heritage language research much further back in time. Against this backdrop, the characterisation of Norn as a historical heritage language is relatively straightforward. At least in Orkney, language shift from Norn to (Older) Scots must have begun in the fourteenth century (Barnes 1984) and was complete around the end of the seventeenth. During this period, Norn would have been a language acquired in the home context, regardless of its wider societal status. Crucially, the dominant language of the larger (national) society would have been (Older) Scots from at least 1472 onwards, when the Northern Isles formally came under Scottish rule. The fact that Norn was no longer used in administrative contexts after 1516 is a strong indicator that Norn itself had begun to occupy a societally subordinate position by this point. Another definition stems from Polinsky’s recent survey of the field: ‘[a] heritage language speaker (for short, heritage speaker) is a simultaneous or sequential (successive) bilingual whose weaker language corresponds to the minority language of their society and whose strong language is the dominant language of that society’ (2018: 9). This definition is narrower in that it requires the heritage language to be weaker in terms of psycholinguistic proficiency. Here, too, there is no doubt that there were such

384 

K. Kinn and G. Walkden

speakers of Norn during the process of language shift, insofar as Orkney and Shetland Scots must have been the dominant variety in wider society long before Norn ceased to be spoken in the home. To summarise, we see that Norn conforms to both Rothman’s and Polinsky’s definitions in all respects for at least 200 years, and hence we can conclude that Norn was a heritage language of Orkney and Shetland in the late medieval and early modern periods.7 Rothman (2009: 156) adds that ‘an individual qualifies as a heritage speaker if and only if he or she has some command of the heritage language acquired naturalistically ... although it is equally expected that such competence will differ from that of native monolinguals of comparable age’. This raises the question of heritage I-languages and whether they differ systematically from those of speakers who grew up in other acquisition scenarios. In this connection it has been observed that variation and diachronic change in heritage languages can often be related to the multilingual context in which they are acquired and used, and this influence may take a number of forms. The first and most obvious type of influence is simply direct transfer of features or properties from the societally dominant / majority language, or cross-linguistic influence. This type of influence is extremely well attested in heritage languages and beyond: see, for example, Polinsky (2018: 18–21), Aalberse et al. (2019: 155–157 and references therein) and Lloyd-Smith et al. (2021). In historical linguistics, this type of influence in contact situations is normally known as ‘transfer’ (van Coetsem 1988, 2000; Winford 2003, 2005; Lucas 2012, 2014). However, transfer does not exhaust the space of logical or attested possibilities in heritage-­ language contact situations. Two types of indirect influence are also found. One of these has been referred to as cross-linguistic  The only potential caveat relates to the use of the word national. As reviewers of early versions of this work have observed, the pre-modern state was a different kind of animal from the nation state of today (see Hobsbawm 1992 for a historical perspective). In particular, the one-nation-one-­ language ideology that determines the social status of present-day heritage languages first found full expression in the works of Johann Gottfried Herder in the eighteenth century, rapidly gaining traction after that (see Rutten 2019). Clearly, fifteenth- and sixteenth-century Scotland as state cannot be equated with today’s United States, for instance. Whether these sociopolitical differences find any reflection in morphosyntactic variation and change is an interesting question for future research. 7

13  Exploring Norn 

385

overcorrection; this entails that speakers seem to overstress differences between the heritage language and the majority language, extending patterns that are already present in their heritage language to new contexts (Kupisch 2014; Anderssen et al. 2018). The other possibility is spontaneous innovation, changes not modelled on any existing pattern either in the majority or heritage language, but which nevertheless seem to be catalysed by the contact situation; this latter type of change in heritage languages is documented by Benmamoun et  al. (2013) and Polinsky (2018). Spontaneous innovations can arise during acquisition (divergent attainment, Polinsky 2018: 24); heritage speakers may also, in the course of their lifetimes, lose linguistic skills that were once acquired as a consequence of reduced input and use (attrition, Polinsky 2018: 22). Given that Norn plausibly meets the definition of a heritage language, the general research question we attempt to answer in the remainder of this chapter is: Do the Norn textual records show any (direct or indirect) morphosyntactic influence of the sort found in present-day heritage languages? For this purpose, we focus on possessive constructions.

4 Possessive Constructions As our case study of the morphosyntactic properties of Norn, we have chosen possessive constructions. This is an area in which the ancestor language, Old Norse, and the contact language, Older Scots, differ; thus, it is particularly interesting to investigate the outcomes in Norn. Moreover, it is an area that has been previously studied in a partially similar, present-day heritage language context, namely American heritage Norwegian, which provides an interesting comparative backdrop (we return to this in Sect. 5). We present an overview of possessive constructions in Old Norse and Older Scots in Sect. 4.1. We then present our corpus of texts from Norway in Sect. 4.2; these texts serve as a basis for comparison, or baseline, for the results from Norn presented in Sect. 4.3. Our dataset and supplementary materials can be found online at https://doi.org/10.5281/ zenodo.7544507.

386 

K. Kinn and G. Walkden

4.1 Possessive Constructions in Old Norse and Older Scots Possessive constructions are diverse, both in terms of structure and meaning. In our study, we focus on noun-phrase-internal possessive constructions, more precisely nominals consisting of a head noun and an adnominal possessive determiner or genitive pronoun / NP of the semantic type ‘possessive’ or ‘argumental’ (Faarlund 2004: 61).8 The possessive genitive is used for possession in the narrow sense of ownership, but also other abstract relationships, such as kinship. Argumental genitives correspond to the subject or an object of a deverbal noun, for example, ferð Óláfs ‘Olaf ’s journey’ (Faarlund 2004: 62). Genitives of other semantic types, such as partitives, are excluded (e.g. tunna bjórs ‘a barrel of beer’), as are possessors that are embedded in a prepositional phrase. (This pattern started to appear in Norwegian documents in the late medieval period, concomitant with a weakening of case morphology, e.g. far åt Olav, literally ‘father to Olav’, meaning ‘Olav’s father’; see Indrebø 2001: 243). In Old Norse, possessors could either precede or follow the head noun; the latter option, that is, postnominal possessors, was the most common (Falk and Torp 1900: 311; Faarlund 2004: 59–60). Examples of the two patterns in Old Norse are given in (2) (from Óláfs saga ins helga; Jónsson 1965): (2) a. Hann lagði sín skip milli boða nǫkkura     he put his.refl. ships between rocks some    ‘he put his ships between some rocks’ (ON – prenominal possessor)    b. at  Sveinn konungr […] varð bráðdauðr um nótt í    that  Sveinn king    became sudden.death at night in   rekkju sinni bed his.refl.        ‘that king Sveinn suddenly died one night in his bed’ (ON – postnominal possessor).  The difference between possessive determiners and adnominal genitive pronouns is that possessive determiners agree with the head noun in case. There are possessive determiners in the first and second person and the third-person reflexive (Faarlund 2004: 59). 8

13  Exploring Norn 

387

Occasionally, the definite marker -inn is found on the head noun; in other words, definiteness marking and possessors may co-occur in Old Norse. When the noun is marked with -inn, the possessor is usually postnominal, as shown in (3) (from Heimskringla, cited in Faarlund 2004: 60): (3) um   tilfǫng-in  búsins     about supplies-def. household.def.gen.    ‘about the household’s supplies’ (ON – definite marker + possessor).

Old Norse has separate reflexive forms of third-person possessive determiners. The reflexive forms are generally used when the possessor is bound by the subject of the clause, as illustrated in (4) (from Finnboga saga hins ramma, cited in Faarlund 2004:280). (4) reið hanni heim með flokk   sinni    rode   he    home    with    people  his.refl.    ‘He rode home with his people’ (ON – reflexive possessor).

In (4), the reflexive form sinn is used to refer back to the subject hann ‘he’; the non-reflexive form would be hans. In Older Scots, possessive constructions are in some ways different from those in Old Norse. Possessors are normally prenominal, as shown in (5) (from Moessner 1997: 118–122): (5)  a. my querrel        my quarrel ‘my quarrel’ (Older Scots, The Historie of ane Nobil and Wailʒeand Squyer, William Meldrum 1273)    b. the  pure  howlatis  appele    the  poor  owl’s     appeal    ‘the poor owl’s appeal’ (Older Scots, The Buke of the Howlat 850).

388 

K. Kinn and G. Walkden

Occasionally, a genitive NP can be split between the prenominal and postnominal position, but postnominal possessors of the Old Norse type are not generally found. Moreover, possessors do not normally co-occur with definiteness marking (although Moessner 1997: 119 notes that they ‘very rarely’ may combine with a demonstrative). Furthermore, there is no distinction between reflexive and non-reflexive possessive pronouns / determiners. The differences between possessive constructions in Old Norse and Older Scots give rise to certain expectations. For example, if cross-­ linguistic influence from Older Scots has taken place in Norn, we may expect the proportion of prenominal possessors to rise; we may also expect possessors and definiteness markers to become mutually exclusive and the distinction between reflexive and non-reflexive possessives to be vulnerable. However, in order to get a clear and accurate picture of whether Norn underwent change, it is not sufficient to just compare the Norn data to general descriptions of Old Norse and Older Scots. It is necessary to establish a more detailed baseline for comparison. We address this in the immediately following Sect. 4.2.

4.2 A ‘Homeland’ Baseline for Comparison A widely discussed question in the field of heritage-language studies is which variety, or varieties, should be used as the reference point, or baseline, when a heritage language is described (e.g. Montrul 2016: 168–175; Polinsky 2018: 13–16; Aalberse et  al. 2019: 111–118). The answer to some extent depends on the research questions. To identify any changes or special features of Norn in our study, we have attempted to establish a baseline which takes the language of the settlers as its starting point (the ‘first generation of immigrants’, Polinsky 2018: 12), and which is otherwise as similar as possible to the available Norn data in all respects apart from the heritage-language status (cf. Aalberse et al. 2019: 111). The available Norn texts range over a period of more than two centuries,9 and they represent particular genres (charters and a ballad). For comparison, we created a baseline corpus of texts written in Norway,  Or even longer if we take the date of transcription of the Hildina ballad at face value.

9

13  Exploring Norn 

389

also mainly taken from the Diplomatarium Norvegicum (DN). The texts selected for this corpus match the investigated Norn texts (Table 13.1) as closely as possible in terms of dates and genre; the charters can be characterised as (late) Old Norse or Middle Norwegian. The corpus mainly consists of charters from Western Norway, where most of the settlers came from. Two documents from Eastern Norway were included (DN II.681 and DN II.683), as they strongly resemble one of the longest Norn charters (DN II.691) in terms of contents (these documents contain complaints about abuse of power by officials). Using the tool Annotated DN Online (Blaxter 2017a, 2017b), we matched the social status of the first signatory of the Norwegian charters to the Norn charters. As a Norwegian ‘equivalent’ of the Hildina ballad, we included the ballad of Falkvor Lommanson (M. B. Landstad’s transcription from 1846, available in the online archive of ballads hosted by the University of Oslo).10 Landstad’s transcription of Falkvor Lommansson is more recent than Low’s transcription of Hildina (1774); however, the ballads are of a similar type (historical ballads), and in both cases we can assume that the date of origin is older than the date of transcription. The texts in the homeland baseline corpus are listed in Table 13.2. We now turn to possessive constructions in Norn, in comparison with this corpus.

4.3 Possessive Constructions in Norn As mentioned, our study focuses on possessive constructions consisting of a head noun and an adnominal possessive determiner or genitive pronoun / NP of the semantic type ‘possessive’ or ‘argumental’ (see the introduction to Sect. 4). We have excerpted possessive constructions manually from the texts; most of the texts are not normalised, and they can at times be difficult to analyse. For the purposes of the study, patronyms and matronyms (e.g. in the name Falkvor Lommansson ‘Falkvor, Lomman’s son’) were excluded; these are very frequent in the charters, for example, in lists of people who act as witnesses or signatories, and they do not seem to  The archive is available at https://www.dokpro.uio.no/ballader/lister/arkiv_gml.html (accessed 25 May 2023).

10

390 

K. Kinn and G. Walkden

Table 13.2  Investigated texts from Norway Type

Text

Date

Provenance

Words

Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Charter Ballad Total

DN II.54 DN I.117 DN II.165 DN I.338 DN IV.374 DN I.370 DN XVIII.27 DN XVIII.25 DN II.683 DN II.681 DN XV.55 DN I.709 DN III.680 DN I.706 DN VII.446 DN IV.959 DN IV.1051 DN XII.283 Falkvor Lommansson

1299 1308 1328 1353 1355 1360 1370 1368 1425 1424 1426 1425 1425 1425 1452 1465 1510 1516 (b)1846

Stavanger Bergen Bergen Giske (Sunnmøre) Stavanger Stedje (Sogn og Fjordane) Tønjum (Sogn og Fjordane) Ljøvik (Møre og Romsdal) Rakkestad Rakkestad Suldal Kinsarvik (Hordaland) Volda (Møre og Romsdal) Giske (Sunnmøre) Valle (Vest-Agder) Kvalbein (Rogaland) Barskaar (Telemark) Bø (Rogaland) Telemark

379 340 609 132 252 262 115 170 475 371 146 122 131 217 222 222 138 142 504 4,949

exhibit any variation with respect to word order (the possessor is always prenominal). Nominals that can be easily read as compounds were also excluded (e.g. vitnis burðr ‘testimony’, literally ‘witness’ bearing’). In determining which nominals should be classified as compounds, we used the Old Norse dictionary of Heggstad et al. (2004) as a guide; if in doubt, we excluded nominals listed as compounds in this dictionary. Some unclear examples were also excluded.

Prenominal and Postnominal Possessors As a first observation, we note that both prenominal and postnominal possessors are attested in Norn. Some examples are given in (6):11

 While the translation of  the sample from Hildina in example (1) (Sect. 2) followed Kershaw (1921), the translations in what follows are sometimes more literal. 11

13  Exploring Norn 

391

(6)  a. Nu  Iarlin    an genger i   vadlin     fram ukadnar   sìna    mien     now jarl.def. he goes  to field.def. forth and calls his.refl men ‘Now the Earl steps forward onto the duelling ground and addresses his men’ (Hildina – prenominal possessor)   b. for  frinda sı̆n       spirde ro,      for  relative his.refl. asked   advice   ‘He asked for advice from his kinsman’ (Hildina – postnominal possessor)    c. som   hennarh ffader Koÿss siig     that  her     father    chose  himself      ‘which her father chose for himself ’ (Goudie 1904: 81, 1516–45 – prenominal possessor)    d. ek adr    nefndr    Markus  ok Helga kona min      I  before mentioned Markus and  Helga wife     my    ‘I, aforementioned Markus, and Helga, my wife’ (DN III.310, 1360 – postnominal possessor).

The distribution of prenominal vs. postnominal possessors in the investigated Norn texts and the texts from Norway is shown in Fig. 13.1.12 In this figure, each circle represents a text; the size of the circles reflects the number of possessive constructions that each text contains. Note that the y axis, which indicates the proportion of prenominal possessors, starts at 0.25. The patterns that emerge from Fig. 13.1 are not strikingly clear; however, the Norn texts have somewhat more and earlier prenominal possessors (until c.1500). We observe a rise in the proportions of prenominal possessors in the texts from Norway too; this could, at least in part, be related to Danish influence, as Danish, which predominantly had prenominal possessors (Falk and Torp 1900: 311; Skautrup 1944: 276), gradually took over as the written standard in Norway in the fifteenth and sixteenth century as a consequence of the political union in which  For a recent study on the development of prenominal vs. postnominal possessors in a different selection of Norwegian charters, mainly from eastern Norway, see Alieva (2017). 12

392 

K. Kinn and G. Walkden

1.0

Proportion of possessors prenominal

0.8

Total 20 40 60 0.6

Origin Norn Norway

0.4

1300

1400

1500

1600

1700

1800

Date

Fig. 13.1  Possessors in texts from the Isles and from Norway

Norway eventually lost its sovereignty (see Mørck 2018: 346–347).13 Barnes (1998:16) notes that a ‘gradual Danicising’ can also be observed in Norn fifteenth- and sixteenth-century documents, as a consequence of contact with Norway. However, at face value, the fact that the increase of  Danish influence seems particularly relevant for possessors that are pronouns/determiners. In present-day Norwegian, these possessors exhibit variation, but the prenominal position is clearly more common in the Danish-influenced written standard Bokmål than in the spoken language (Lødrup 2012). In the spoken language, the proportion of prenominal possessors is only around 25% (Westergaard and Anderssen 2015: 39). Possessors that are full NPs, on the other hand, are invariably prenominal in all spoken and written varieties of present-day Norwegian; it is not clear if this shift had any connection with Danish influence (in general, Danish influence primarily affected the written language). 13

13  Exploring Norn 

393

prenominal possessors appears to be earlier and stronger in Norn could be interpreted as a result of cross-linguistic influence from Scots, which only has prenominal possessors (see further discussion in Sect. 5). The clearest outlier in Fig. 13.1 is the Hildina ballad, which seems to have a very low proportion of prenominal possessors for its date (1774). However, recall that this date reflects the time of transcription rather than the time of origin. Also, the distribution in the Hildina ballad might be influenced by metrics.

Interaction between Possessors and Definiteness Marking Recall that Old Norse occasionally combines the definite marker -inn with postnominal possessors. This pattern is also attested in Norn, as shown in (7): (7)  feyr-in   sien   father-def. her.refl.    ‘her father’ (Hildina).

The Hildina ballad additionally exhibits a different pattern, whereby a definite marker is combined with a prenominal possessor: (8)  sin  shall lathì  min hera-n  i  bardagana fwo   as self let-you my lord-def.  in battle.def get    ‘as you yourself let my lord receive in battle’ (Hildina).

This pattern is not commonly known in Old Norse, and it is not found in the Norwegian charters in the baseline corpus. However, Aasen (1864: 291) notes similar examples in ballads from the area of Telemark; this includes the Falkvor Lommansson ballad, as shown in (9): (9)  hans Folk-et   det falt ifraa   his  people-def. it  fell from    ‘his people retreated’ (Falkvor Lommansson).

394 

K. Kinn and G. Walkden

Thus, the combination of prenominal possessor + definite marker is not exclusively found in Norn. In the investigated sources, it seems to be a genre-specific feature associated with ballads; it is not immediately clear if it ever had support in the everyday spoken language. As the co-occurrence of a possessor and a definite marker is very infrequent, we do not have enough data to make any quantitative generalisations about this phenomenon. However, based on the evidence that we have, it seems that Norn is similar to the homeland baseline. The pattern of prenominal possessor + definite marker found in Hildina and in the homeland ballad Falkvor Lommansson highlights the importance of establishing a baseline that is as similar as possible to the heritage language sample in terms of genre and other relevant variables.

Reflexive Possessives The Norn texts contain several examples of reflexive possessive determiners (sinn) that are used in the same way as in the baseline corpus, that is, when the possessor is bound by the subject of the clause. Some examples are given in (10): (10)  a. thet Jon af Baddy... thok ater sin      hæst    af Mikel Magy     that Jon of Baddy... took back his.refl. horse of Mikel Magy     ‘that Jon of Baddy took his (own) horse back from Mikel Magy’ (DN II.691, b. 1426)    b.  ...ath sira Gregorius Iuarsson haffde nockrar thœr sakir     that sir   Gregorius   Iuarsson   had    any    those  things giort    i  sina       daga done  in his.refl. days     ‘that sir Gregorius Ivarsson had done any of those things in his days’ (DN VI.651, 1509).

However, the non-reflexive form hans ‘his’ is also attested in similar contexts, as shown in (11):

13  Exploring Norn 

395

(11) a.  Jtem  forde      Thomos kærde    oppa  hans      moreover aforementioned Th.   complained on  his eghna wæghnæ... own  behalf     ‘Moreover, the aforementioned Thomas complained on his own behalf...’ (DN II.691, b. 1426)     b. An cast ans huge ei fong ednar      he threw his head in lap  her      ‘He threw his head into her lap’ (Hildina, 1774).

This use of non-reflexive forms is not found in the homeland baseline corpus, and it could be interpreted as cross-linguistic influence from Scots. An alternative explanation could be influence from medieval Danish, which also exhibited non-reflexive possessive forms bound by a subject (Falk and Torp 1900: 134–135; Barnes 1998: 14). However, Danish influence seems less plausible for the Hildina ballad than for charters: Danish influence is mainly known to have affected the written language (at least in the time period and in the sociolinguistic contexts that are relevant to our study; see, e.g., Nesse and Torp 2018 for further discussion), while Hildina is part of an oral tradition.

Morphological Marking of Possessors In Old Norse, adnominal possessive pronouns and NPs had genitive case marking. In Norn, there are occasional examples of unmarked possessors: (12) a. herra biskup jnsigle     lord   bishop    seal      ‘the lord bishop’s seal’ (DN I.404, 1369)     b. Alexander tomesszonn skolgetinn   dotthr      A.   T.                    lawfully.born  daughter      ‘A. T.’s lawfully born daughter’ (Goudie 1904, 1516–45).

396 

K. Kinn and G. Walkden

This morphological reduction could be interpreted as ongoing restructuring or attrition. However, it is not unique to Norn; unmarked possessors have also been observed in texts from Norway in the late medieval period (Mørck 2013: 654).

5 Discussion Overall, we observe a high degree of stability in the morphosyntax of possessive constructions in Norn compared to Old Norse and the homeland Norwegian baseline corpus, even many centuries after the Scandinavian settlements. The Norn texts do, however, exhibit some special features and developments which can potentially be related to Norn’s status as a heritage language. In this section, we discuss these features in further detail. We found that the Norn texts exhibit somewhat more and earlier prenominal possessors than the Norwegian texts (Fig. 13.1). This development is compatible with cross-linguistic influence from Older Scots, which only has prenominal possessors. However, it is worth keeping in mind that prenominal possessors are not a fundamentally new pattern that emerges in Norn; they are already present in Old Norse and in the earliest text in the baseline corpus (alongside with postnominal possessors). Thus, rather than treating the increase in prenominal possessors as cross-­linguistic influence in the sense of ‘copying’ of a pattern, it might be more accurate to analyse it as an incipient change which was latent in the baseline (Polinsky 2018: 33), and which was triggered or accelerated by reduced input and convergence with Older Scots. This is, perhaps, particularly relevant for possessors that are full NPs: These eventually became exclusively prenominal in homeland Norwegian too, while possessors that are pronouns or determiners still vary in terms of position (see fn. 13). An interesting comparative backdrop to the development of possessive constructions in Norn is constituted by the present-day variety American Heritage Norwegian, spoken in the United States (Anderssen et al. 2018). The contact language of American Heritage Norwegian is American English, and the comparison is relevant because possessive constructions in homeland Norwegian (the ‘ancestor’ of American Heritage Norwegian)

13  Exploring Norn 

397

and American English differ in a way that resembles Old Norse and Older Scots: Homeland Norwegian allows both prenominal and postnominal possessive pronouns / determiners, while American English only has the prenominal option.14 In American Heritage Norwegian, crosslinguistic influence is actually not a very common pattern in possessive constructions; it is only a minority of speakers that ‘overuse’ prenominal possessors compared to the homeland baseline, while the remaining speakers use postnominal possessors to an even greater extent than homeland speakers (exhibiting cross-linguistic overcorrection; Anderssen et al. 2018: 755). Overuse of prenominal possessors is a feature which is associated with low proficiency (Anderssen et al. 2018: 758). The implications of this for our study are not entirely clear, but it poses the question for future research about whether other apparent signs of low proficiency can be found in Norn charters. Another particular feature of Norn is the use of the non-reflexive pronoun hans instead of the reflexive sinn, which is not found in the baseline corpus of texts from Norway. Cross-linguistic influence from Older Scots seems to be a plausible explanation. As mentioned, medieval Danish has been suggested as an alternative source of influence (Barnes 1998: 14); however, this idea raises some questions. First, as mentioned, it is not equally plausible for all attested cases, as Danish influence primarily affected the written language, while the Hildina ballad, which exhibits this feature, is part of an oral tradition. Second, if medieval Danish was the source of influence, one might expect to see more of this feature in documents from Norway. However, as previously pointed out, it is not attested in the baseline corpus; moreover, it is not mentioned in Indrebø’s (2001: 193) overview of Danish pronominal forms borrowed into Norwegian charters, or in other previous literature, to the best of our knowledge. Finally, we noted some Norn instances of possessors without genitive case marking. Morphological reduction is common in heritage languages, particularly in nominal morphology (see Polinsky 2018: 197ff). Divergent morphology could follow from individual irregularities in performance  A difference between Old Norse and modern Norwegian is that in the latter the head noun is required to appear with a definite suffix when the possessor is postnominal. 14

398 

K. Kinn and G. Walkden

(attrition), or ongoing restructuring of the system. In the heritage language context, divergent morphology is typically analysed as a consequence of reduced input, and this may very well be relevant in Norn. However, it is difficult to draw any firm conclusions, as a similar reduction in case morphology is also observed in documents from Norway around the same time, and as the scribes producing the Norn documents probably were familiar with Norwegian scribal traditions (Barnes 1998: 16).

6 Conclusion In this chapter, we have made the case that, from a sociohistorical perspective, Norn plausibly meets the definition of a heritage language (at least during the period 1472–1700). As a morphosyntactic case study we looked at possessives, for which in other heritage languages (specifically present-day American heritage Norwegian) it has been shown that divergence from the baseline may arise. Our findings are mixed. Quantitatively, the Norn texts exhibit somewhat more use of prenominal possessors at an early date, but the difference is not huge, and variation is present in both the baseline corpus and the Norn corpus from the earliest texts onwards. Thus, we may be witnessing cross-linguistic influence in the sense of contact with Older Scots accelerating an incipient change, but the inference is by no means secure. We also find other features—use of non-reflexive pronouns where a reflexive would be expected in Old Norse, and use of unmarked possessors—which may be the consequence of the heritage language contact situation but could also be amenable to other explanations. It should come as no surprise that these developments cannot be exclusively and securely attributed to the heritage language status of Norn. Indeed, one of the major findings of heritage language research over the past decades is that outcomes in heritage languages vary: In some circumstances we find stability, in other circumstances various types of change. Our case study is an exploratory one, aiming to generate new hypotheses and open up a new way of looking at historical heritage languages.

13  Exploring Norn 

399

Certainly, other areas of the syntax of Norn may benefit from being viewed through this lens. One such is verb-second, a very robust feature of Old Norse (Faarlund 2004: 191). Eekman (2015) points out that there are a handful of exceptions to verb-second in Hildina, for example, (13), in which the subject follows a temporal adverbial, and is in turn followed by the finite verb laghdè in third position. (13)  sien  on     laghdè  gloug   i  otsta     jatha  port    then    she laid    ember in outmost street entrance     ‘then she lit a fire in the main entrance’ (Hildina, 1774).

Such examples are a minority, but also occur in American Heritage Norwegian (Eide and Hjelde 2015; Alexiadou and Lohndal 2018; Westergaard and Lohndal 2019; Westergaard et al. 2021), as well as in other contact varieties where transfer from English is implausible as an explanation (Walkden 2017; Meelen et al. 2020). In particular, we would like to encourage future researchers to consider what other insights could be gleaned from revisiting historically attested varieties in the light of modern heritage language research.

References Aalberse, Suzanne, Ad Backus, and Pieter Muysken. 2019. Heritage Languages: A Language Contact Approach. Amsterdam: John Benjamins. Aasen, Ivar. 1864. Norsk Grammatik. Christiania: Malling. Alexiadou, Artemis, and Terje Lohndal. 2018. V3 in Germanic: A Comparison of Urban Vernaculars and Heritage Languages. In Non-Canonical Verb Positioning in Main Clauses, ed. Mailin Antomo and Sonja Müller, 245–264. Hamburg: Helmut Buske Verlag. Alieva, Dinara. 2017. Adnominale genitivskonstruksjoner i mellomnorske diplomer. Maal og Minne 109: 79–92. Anderssen, Merete, Bjørn Lundquist, and Marit Westergaard. 2018. Crosslinguistic Similarities and Differences in Bilingual Acquisition and Attrition. Bilingualism: Language and Cognition 21: 748–764.

400 

K. Kinn and G. Walkden

Barnes, Michael P. 1984. Orkney and Shetland Norn. In Language in the British Isles, ed. Peter Trudgill, 352–366. Cambridge: Cambridge University Press. ———. 1998. The Norn Language of Orkney and Shetland. Lerwick: The Shetland Times Lt. Benmamoun, Elabbas, Silvina Montrul, and Maria Polinsky. 2013. Heritage Languages and Their Speakers: Opportunities and Challenges for Linguistics. Theoretical Linguistics 39: 129–181. Blaxter, Tam, ed. 2017a. Annotated DN Online. http://www.icge.co.uk/dn_ online/. Accessed 25 May 2023. ———. 2017b. Speech in Space and Time: Contact, Change and Diffusion in Medieval Norway. Unpublished doctoral thesis, University of Cambridge. Blaxter, Tamsin. 2021. Diachronic Dialectology: New Methods and Case Studies in Medieval Norwegian. Publications of the Philological Society, 53. Chichester: Wiley. Blaxter, Tam, and Kari Kinn. 2018. On Ek and Jak in Middle Norwegian: Mixed Methods in Historical Sociolinguistics. Transactions of the Philological Society 116: 383–409. Brown, Joshua R., ed. 2019. Historical Heritage Language Ego-Documents: From Home, from Away, and from Below, special issue of Journal of Historical Sociolinguistics. Amsterdam: John Benjamins. Brown, Joshua R., and Mike Putnam. 2015. Functional Convergence and Extension in Contact: Syntactic and Semantic Attributes of the Progressive Aspect in Pennsylvania Dutch. In Germanic Heritage Languages in North America: Acquisition, Attrition and Change, ed. Janne B.  Johannessen and Joseph C. Salmons, 135–160. Amsterdam: John Benjamins. van Coetsem, Frans. 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. ———. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter. Dorian, Nancy C. 1982. Linguistic Models and Language Death Evidence. In Exceptional Language and Linguistics, ed. Lorraine K. Obler and Lise Menn, 31–48. London: Academic Press. Dubinina, Irina, and Maria Polinsky. 2013. Russian in the USA.  In Slavic Languages in Migration, ed. Michael Moser and Maria Polinsky, 130–158. Vienna: University of Vienna Press. Eekman, Stephan. 2012. De Syntaxis van het Norn in Vergelijking met die van het Scandinavisch en het Engels. Unpublished manuscript, University of Amsterdam.

13  Exploring Norn 

401

———. 2015. A Descriptive and Comparative Overview of Norn Syntax, the Ancient Language of Orkney and Shetland. Unpublished manuscript, University of Amsterdam. Eide, Kristin M., and Arnstein Hjelde. 2015. Verb Second and Finiteness Morphology in Norwegian Heritage Language of the American Midwest. In Moribund Germanic Heritage Languages in North America: Theoretical Perspectives and Empirical Findings, ed. B. Richard Page and Michael Putnam, 64–101. Leiden: Brill. Faarlund, Jan Terje. 2004. The Syntax of Old Norse. Oxford: Oxford University Press. Falk, Hjalmar, and Alf Torp. 1900. Dansk-Norskens Syntax i Historisk Fremstilling. Oslo: H. Aschehough. Farasyn, Melissa, George Walkden, Sheila Watts, and Anne Breitbarth. 2018. The Interplay between Genre Variation and Syntax in a Historical Low German Corpus. In Diachronic Corpora, Genre, and Language Change, ed. Richard Jason Whitt, 181–200. Amsterdam: John Benjamins. Goudie, Gilbert. 1904. The Celtic and Scandinavian Antiquities of Shetland. Edinburgh: William Blackwood and Sons. Hægstad, Marius. 1900. Hildinakvadet: Med Utgreiding um det Norske Maal paa Shetland i Eldre Tid. Oslo: Jacob Dybwad. Haugen, Einar. 1953. The Norwegian Language in America. 2 vols. Pennsylvania: University of Pennsylvania Press. Heggstad, Leiv, Finn Hødnebø, and Erik Simensen. 2004. Norrøn ordbok, 4th edn. Oslo: Samlaget. Hobsbawm, Eric J. 1992. Nations and Nationalism since 1780: Programme, Myth, Reality, 2nd edn. Cambridge: Cambridge University Press. Indrebø, Gustav. 2001. Norsk målsoga, 2nd edn. Bergen: Norsk bokreidingslag. Johannessen, Janne Bondi. 2018. Factors of Variation, Maintenance and Change in Scandinavian Heritage Languages. International Journal of Bilingualism 22: 447–465. Jónsson, Finnur, ed. 1965. Óláfs saga ins helga. Oslo: Universitetsforlaget. Kershaw, Nora. 1921. Stories and Ballads of the Far Past, Translated from the Norse (Icelandic and Faroese) with Introductions and Notes. Cambridge: Cambridge University Press. Knooihuizen, Remco. 2005. The Norn-to-Scots Language Shift: Another Look at Socio-Historical Evidence. Northern Studies 39: 105–117.

402 

K. Kinn and G. Walkden

———. 2008. Fishing for Words: The Taboo Language of Shetland Fishermen and the Dating of Shetland Language Death. Transactions of the Philological Society 106: 100–113. Kupisch, Tanja. 2014. Adjective Placement in Simultaneous Bilinguals (German–Italian) and the Concept of Cross-Linguistic Overcorrection. Bilingualism: Language and Cognition 17: 222–233. van Leyden, Klaske. 2004. Prosodic Characteristics of Orkney and Shetland Dialects: An Experimental Approach. Amsterdam: LOT. Lindqvist, Christer. 2015. Norn im keltischen Kontext. Amsterdam: John Benjamins. Lloyd-Smith, Anika, Henrik Gyllstad, Tanja Kupisch, and Stefano Quaglia. 2021. Heritage Language Proficiency Does Not Predict Syntactic CLI into L3 English. International Journal of Bilingual Education and Bilingualism 24: 435–451. Low, George. 1879 [1774]. A Tour through the Islands of Orkney and Schetland: Containing Hints Relative to Their Ancient, Modern, and Natural History, Collected in 1774. Kirkwall: William Peace and Son. Lucas, Christopher. 2012. Contact-Induced Grammatical Change: Towards an Explicit Account. Diachronica 29: 275–300. ———. 2014. Contact-Induced Language Change. In The Routledge Handbook of Historical Linguistics, ed. Claire Bowern and Bethwyn Evans, 519–536. London: Routledge. Lødrup, Helge. 2012. Forholdet mellom prenominale og postnominale possessive uttrykk. In Grammatikk, bruk og norm, ed. Hans-Olav Enger, Jan Terje Faarlund, and Kjell Ivar Vannebo, 189–203. Oslo: Novus. Meelen, Marieke, Khalid Mourigh, and Lisa Lai-Shen Cheng. 2020. V3  in Urban Youth Varieties of Dutch. In Syntactic Architecture and Its Consequences, II: Between Syntax and Morphology, ed. András Bárány, Theresa Biberauer, Jamie Douglas, and Sten Vikner, 327–355. Berlin: Language Science Press. Melchers, Gunnel. 1987. On the Low German and Dutch Element in Shetland Dialect. In Sprachkontakt in der Hanse: Aspekte des Sprachausgleichs im Ostseeund Nordseeraum, ed. P. Sture Ureland, 276–324. Tübingen: Max Niemeyer. ———. 2008a. English Spoken in Orkney and Shetland: Morphology, Syntax, and Lexicon. In Varieties of English, I: The British Isles, ed. Bernd Kortmann and Clive Upton, 285–298. Berlin: Mouton de Gruyter. ———. 2008b. English Spoken in Orkney and Shetland: Phonology. In Varieties of English, I: The British Isles, ed. Bernd Kortmann and Clive Upton, 35–47. Berlin: Mouton de Gruyter.

13  Exploring Norn 

403

Moessner, Lilo. 1997. The Syntax of Older Scots. In The Edinburgh History of the Scots Language, ed. Charles Jones, 47–111. Edinburgh: Edinburgh University Press. Montrul, Silvina. 2016. The Acquisition of Heritage Languages. Cambridge: Cambridge University Press. Mørck, Endre. 2013. Mellomnorsk språk. In Handbok i Norrøn Filologi, ed. Odd Einar Haugen, 640–690, 2nd edn. Bergen: Fagbokforlaget. ———. 2018. Seinmellomalderen (1350–1536). In Norsk språkhistorie: Tidslinjer, ed. Agnete Nesse, 293–356. Oslo: Novus forlag. Nesse, Agnete, and Arne Torp. 2018. Dansketiden (1536–1814). In Norsk språkhistorie: Tidslinjer, ed. Agnete Nesse, 357–424. Oslo: Novus forlag. Polinsky, Maria. 2018. Heritage Languages and Their Speakers. Cambridge: Cambridge University Press. Rendboe, Laurits. 1984. How ‘Worn-Out’ or ‘Corrupted’ Was Shetland Norn in Its Final Stages? NOWELE 3: 53–88. Rhys, Guto. 2015. Approaching the Pictish Language: Historiography, Early Evidence and the Question of Pritenic. Unpublished doctoral thesis, University of Glasgow. Rothman, Jason. 2009. Understanding the Nature and Outcomes of Early Bilingualism: Romance Languages as Heritage Languages. International Journal of Bilingualism 13: 155–163. Rutten, Gijsbert. 2019. Language Planning as Nation Building: Ideology, Policy and Implementation in the Netherlands, 1750–1850. Amsterdam: John Benjamins. Skautrup, Peter. 1944. Det Danske Sprogs Historie, I. Copenhagen: Gyldendal. Walkden, George. 2017. Language Contact and V3 in Germanic Varieties New and Old. Journal of Comparative Germanic Linguistics 20: 49–81. Westergaard, Marit, and Merete Anderssen. 2015. Word Order Variation in Norwegian Possessive Constructions: Bilingual Acquisition and Attrition. In Germanic Heritage Languages in North America, ed. Janne Bondi Johannessen and Joseph C. Salmons, 21–45. Amsterdam: John Benjamins. Westergaard, Marit, and Terje Lohndal. 2019. Verb Second Word Order in Norwegian Heritage Language: Syntax and Pragmatics. In Variable Properties in Language: Their Nature and Acquisition, ed. David W.  Lightfoot and Jonathan Havenhill, 91–102. Washington, DC: Georgetown University Press. Westergaard, Marit, Terje Lohndal, and Björn Lundquist. 2021. Variable V2 in Norwegian Heritage Language: An Effect of Cross-Linguistic Influence? In Linguistic Approaches to Bilingualism. https://doi.org/10.1075/lab.20076.wes

404 

K. Kinn and G. Walkden

Wiggen, Geirr. 2002. Norns død, især skolens rolle: Kommentarer til en disputt om nedgangen for det nordiske språket på Orknøyene og Shetland. Oslo: Det Norske Videnskaps-Akademi/Novus. Winford, Donald. 2003. An Introduction to Contact Linguistics. Oxford: Wiley Blackwell. ———. 2005. Contact-Induced Changes: Classification and Processes. Diachronica 22: 373–427.

Part IV Textual Manifestations of Medieval Multilingualism

14 Textual and Codicological Manifestations of Multilingual Culture in Medieval England Ad Putter, Joanna Kopaczyk, and Venetia Bridges

1 Introduction This chapter provides an overview of developments and desiderata in scholarship on multilingualism with specific reference to medieval multilingualism as reflected in the literature and literary manuscripts of the period. What changes of approach and emphasis can be seen in recent work on this topic, and what avenues for future work look promising? The multilingual miscellany London, British Library, MS Harley 2253, Ad. Putter University of Bristol, Bristol, UK e-mail: [email protected] J. Kopaczyk (*) University of Glasgow, Glasgow, UK e-mail: [email protected] V. Bridges Durham University, Durham, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_14

407

408 

Ad. Putter et al.

has often been at the centre of discussions of these questions, and the body of critical literature on this manuscript and its contents provides an excellent means of exemplifying changing trends and future opportunities, and so we use this manuscript and its texts to illustrate new departures and new possibilities.

2 Editing Multilingual Texts It should be said at the start that readers of medieval literature today are in a much better place to appreciate its multilingual texts and contexts than earlier generations of readers were, because multilingualism was not high on the scholarly agenda until the 1990s. Editions set the tone, and recent editions, such as Fein’s (2014–2015) and Turville-Petre’s (2015), discussed below, have been much more sensitive to the company that texts keep in the codex in which they are found. The policy that was characteristic of earlier editorial practice, that of extracting Middle English poems from their multilingual codicological environment and presenting them in monolingual isolation, has fortunately been revised. It is instructive in this context to compare old and new editorial practices in the case of two multilingual miscellanies, both compiled in the second quarter of the fourteenth century: MS Harley 2253, from the West Midlands, the manuscript in focus in this chapter, and London, British Library, MS Harley 913 (the ‘Kildare manuscript’), from Leinster in Ireland. Both manuscripts contain texts in Latin, French and English (and texts that mix languages, which is another area in need of further exploration), but one would not know it from the earliest editions. Heuser (1904), under the title Die Kildare Gedichte, did not in fact edit all Gedichte (poems) from MS Harley 913, but only the Middle English ones. The recent edition by Turville-Petre (2015), however, makes one realise that the ‘Kildare poems’ include texts in French and Latin, as well as shorter Latin prose items. Editions of these French-language poems, including ‘The Walling of New Ross’, beginning Talent me prent de rimanceir (‘I wish to compose a song in French’), and two alliterating French poems, together with summaries of the Latin items, are helpfully included in Turville-Petre’s edition. The multilingualism of MS Harley 2253 is again only now easily accessible to readers of modern editions. Böddeker (1878) was content to

14  Textual and Codicological Manifestations of Multilingual Culture 

409

print the English poems in splendid isolation from their multilingual context, and Brook (1978) followed suit. Again, the title of Brook’s influential edition, The Harley Lyrics, might make the unwary reader think that the manuscript contains only English lyrics but, thanks to the edition of the complete manuscript by Fein (2014–2015), it is now possible to appreciate that the Middle English poems are outnumbered by texts (including lyrics) in French and Latin.

3 Revisiting ‘Trilingual England’ Because of these advances and because of the burgeoning interest in medieval multilingualism more broadly (see, for instance, Trotter 2000; Schendl and Wright 2011; Pahta et al. 2018; Wright 2021), the polyglot nature of ‘English’ literary culture is now well appreciated. Still, much more work is needed to understand the relative positions of and interactions between the languages concerned. An important contribution to this topic was Rothwell (1994), which emphasised the interrelations between Anglo-French, Latin and English in what Rothwell called ‘the trilingual England of Geoffrey Chaucer’, but there are issues which call for further examination. One issue is that the distinction that Rothwell drew between French and English, both living vernaculars in the fourteenth century, and Latin, which he labelled a ‘dead construct’, is too crude. It needs to be emphasised that Latin, too, was at least in some circles an everyday spoken language (Sharpe 1996; Putter 2016). The thirteenth-century Franciscan friar Roger Bacon said he had three mother tongues, Latin, French and English (Lusignan 1986: 67; Coseria and Meisterfeld 2003: 257), and he and others who conversed in Latin on a daily basis would surely not have regarded it as dead. A second important issue is the co-presence of other languages besides English, French and Latin. As shown in the essays by Callander and Fulton (on the place of Welsh in the multilingual repertoire) and Putter (on cross-linguistic English-Dutch influence in William Caxton’s language), the multilingualism of medieval England was always more than trilingualism, and this was true for Chaucer’s England too. Anyone living in London in Chaucer’s day would have heard other languages, such as

410 

Ad. Putter et al.

those spoken by the sizeable communities of Flemish and Italian migrant workers and merchants (Tiddeman 2017, 2018, 2020; Ormrod et al. 2019), while Celtic languages were spoken in parts of England: Cornish in western Cornwall (Padel 2017), and Welsh not just in Wales but in the counties of England that bordered on Wales, especially Hereford and Shropshire (Smith 2000). The so-called trilingual miscellany that is MS Harley 2253 reflects this more complicated linguistic reality. Compiled on the Welsh border of Herefordshire and Shropshire, the manuscript mixes Latin, French and English. The English also contains a few borrowings from Welsh, such as miles (‘animals’), from Welsh mil, and wolc (‘hawk’), from Welsh gwalch (Brook 1933);1 and from Middle Dutch, such as polle (‘head’; Critten 2022), a word surviving in this sense in the word ‘poll tax’ and first found in English in Harley 2253 (‘Of rybauds Y ryme’, l. 3). Moreover, one of the macaronic poems in the manuscript, labelled ‘The Flemish Insurrection’ in the edition by Robbins (1959) and ‘Lustneth, lordinges, bothe yonge ant olde’ in the edition by Fein (2014–2015), contains a pun that hinges on the difference between the Dutch and French word conyng. The poem tells the story of the Battle of the Golden Spurs at Courtrai (1302), where a group of Flemish artisans defeated the flower of French chivalry. The leader of the Flemish rebels was the weaver Pieter de Coninck. In Dutch Conick means ‘the king’, and the poet clearly plays on that sense when he describes how he became the rebels’ ringleader: The webbes ant the fullaris assembleden hem alle Ant makeden huere consail in huere commune halle, Token Peter Conyng huere ‘kyng’ to calle       Ant beo huere cheveuteyn (ll. 17–20). ‘The weavers and the fullers all assembled And held their council in their common hall, Chose Peter Coning their ‘king’ to be called       And be their leader’ (Text and translation from Fein 2014–2015).

 Bergner’s (1978) suggestion that miles is an error for males is unnecessary and implausible.

1

14  Textual and Codicological Manifestations of Multilingual Culture 

411

In medieval French, however, connin was the word for rabbit, and that double entendre is clearly meant when the Earl of Boulogne, who has been sent to quash the rebellious townspeople, seeks to rouse his cavalry: Nus ne lerrum en vie chanoun ne moyne … We shule flo the Conyng and make roste his loyne (ll. 65–69). ‘We are not going to leave any canon or monk alive. We will flay the Rabbit and roast its loin’ (our translation).

The audience of MS Harley 2253 would have found it easier to get the pun because the French word was taken up in English and because English king (< OE cyning) once had forms closer to Dutch koning, but even so the wordplay depends on a metalinguistic awareness that Dutch and French speakers would have understood conyng very differently (Levelt and Putter 2021: 44–45). Another issue that merits further investigation concerns the synchronic and diachronic fluctuations in the relations between the different languages in different texts and codices of the period. Turville-Petre (1996: 182) described the relationship between the ‘three languages’ of MS Harley 2253 as a harmonious one, but not surprisingly (since the manuscript itself is a compilation of heterogenous sources) one language may dominate in one particular item, as English does in ‘The Flemish Insurrection’, while in others it is as marginal as it is in the manuscript as a whole, where French and Latin texts dominate. The question of a dominant language is in itself a complex one that has much been debated in scholarship on language mixing, medieval (e.g. Schendl and Wright 2011; Ingham et al. 2021) and modern (e.g. Myers-Scotton 2002; Auer and Muhamedova 2005). If one attempted to calculate the ratio of Latin vs. French vs. English in MS Harley 2253, what would be the basis for comparison? Would it be enough to count the individual lexical items (in which case one has to face the perennial difficulty of distinguishing between established borrowings and more ad hoc code-switches) or should we be concerned with larger narrative structures and discourse units? Or would it be more illuminating to carry out assessments on each plane of language complexity, even down to the spelling practices potentially influenced by another language in the scribe’s repertoire?

412 

Ad. Putter et al.

Finally, there is scope for connecting the multilingual character of Middle English texts with current research on the visual aspects of text construction and presentation on the page (Machan 2011; Sebba 2012; Jefferson 2013; Carroll et al. 2013; Peikola et al. 2017; Rogos-Hebda 2020; Kopaczyk 2018, 2022). As observed by Lerer (2008: 238), ‘[the critic’s responsibility] should be to see how the meaning of […] literary texts lies not just in the words transmitted on the page but in the page itself. The physical appearance of the manuscript—its script, its layout, its selection of texts, its size and so on—may be shaped by social, economic and aesthetic values’. We now have a much better understanding of the role of visual pragmatics in the creation of the text’s overall meaning in its material, cultural and social context. In multilingual manuscripts there are multiple ways in which the languages involved may or may not be visually indicated, which calls for a systematic approach to observing and analysing visual cues.

4 ‘One Culture in Three Voices’? Linguistic and Cultural Ramifications of Code­Switching in MS Harley 2253 and Beyond 4.1 Multilingualism and Monocultural Englishness The idea that MS Harley 2253’s ‘three languages’ are harmonious is based on several important assumptions about language relationships. One important and continuing assumption inherent in the common phrase ‘trilingual England’ is that there is a balance between the three languages, certainly in terms of each pertaining to a particular sphere of the population and / or literary function. Turville-Petre’s description (given here in full) demonstrates this clearly: ‘three languages existed in harmony, not just side by side but in symbiotic relationship, interpenetrating and drawing strength from one another; not three cultures but one culture in three voices’ (Turville-Petre 1996: 182). This formulation, whilst useful in its evocative description of linguistic interaction, also in effect smooths over differences between the ‘three voices’, which are assimilated into ‘one

14  Textual and Codicological Manifestations of Multilingual Culture 

413

culture’.2 The effect is notable, since although Turville-Petre’s (1996: 200) analysis does highlight the three languages as ‘different in function and character’, the bulk of his commentary focuses on the English material, which appears in slightly under half of the items. The emphasis upon English gives it an importance that may be justified in terms of MS Harley 2253, but which is not necessarily applicable more widely.3 This emphasis, which remains emblematic of ongoing scholarly approaches, frequently means in practice that English’s presence is more prominent in scholarly discourse than in the historical written record, being given an equal (if not greater) share of literary culture as Latin and French in numerical terms. In addition, the elision of cultural difference between languages (‘one culture in three voices’) allows a monocultural ‘Englishness’ to be constructed multilingually (‘trilingual England’), which may similarly misrepresent medieval cultural and linguistic complexities. These two factors, combined with the high profile of the English language in medieval studies as a discipline, mean that English in effect represents this cultural ‘Englishness’, despite the latter’s acknowledged multilingual nature. This narrowing of focus is clearly visible in the studies of multi- and trilingual manuscripts: collections like MS Harley 2253 and its slightly earlier contemporary, Oxford, Bodleian Library, MS Digby 86, are still studied mostly for their English works rather than the Anglo-French and Latin that they contain.4 In practice, then, multilingualism in medieval England is conceptualised in terms of linguistic and cultural Englishness, which may well inhibit our understanding of its impact in both of these spheres. Crucially, this approach also overlooks the fact that language  Critten (2020: 229) uses a similar metaphor to critique Turville-Petre’s phrase, describing it as ‘flatten[ing] out the matter of the book’s multilingualism’. 3  Turville-Petre’s (1996: 186–193) analysis of London, British Library, MS Add. 46919 in the same chapter makes this distinctiveness clear. 4  See results of the British Library database search for research on Digby 86: https://explore.bl.uk/ primo_library/libweb/action/search.do?fn=search&ct=search&initialSearch=true& mode=Basic& tab=local_tab&indx=1&dum=true&srt=rank&vid=BLVU1&frbg=&tb=t&vl%28freeText0%29= digby+86&scp.scps=scope%3A%28BLCONTENT%29&vl%28297891280UI0%29=any&v l%28297891280UI0%29=title&vl%28297891280UI0%29=any [accessed 9 December 2022]. The essays collected in Fein (2019) account for much of the total critical work on Latin and French material in the manuscript. 2

414 

Ad. Putter et al.

relationships may alter over time, as Sect. 4.3 illustrates. The linguistic interactions visible in MS Harley 2253 are not representative of the postConquest Middle Ages in England as a whole, nor indeed of any single era, given the book’s multi-temporal nature as an anthology bringing together works from a variety of periods. Our critical approach to trilingualism needs to take account of these conceptual and historical facts.

4.2 Trilingualism and Monoculture in Action A Manuscript Miscellany as a Multi-layered Multilingual Text The multi-temporal nature of collections of medieval texts alluded to above is best appreciated when experiencing the manuscripts as material objects. The physical differences between parts of the manuscript—in the quality of vellum or paper, in scribal hand, in the layout, in the ink colours and other visual aspects—testify to the disparate circumstances in which they were produced. The fact that such distinct parts had come together as part of a Sammelhandschrift adds another layer of interpretation with the focus shifted to the manuscript’s readership and ownership. Consider the two pages from MS Harley 2253 in Fig. 14.1.5 The open book view illustrates the transition between quires, and in this case also between Booklets 3 and 4, two of the seven booklets bound into a single codex. They both contain entertaining verse in Anglo-­French, introducing a new scribal hand—Scribe B, also known as the Ludlow scribe—and new genres to the compilation of so far mostly religious texts by Scribe A. Fein (2014–2015) in her introduction to the second volume of her edition draws attention to the interaction between the two scribes evident in Scribe B filling in blank titles in Scribe A’s part of the compilation. What is interesting is that Scribe B used red ink for his additions and assigned titles both in Latin (Vitas patrum, while the text is in AngloFrench) and in Anglo-French (e.g. Epistle a claudie lemperour). Fein concludes that he must have owned the previous booklets and known their  All images from Harley MS 2253 are taken from the British Library Digitised Manuscripts website (http://www.bl.uk/manuscripts/FullDisplay.aspx?ref=Harley_MS_2253), and are reproduced with permission from the British Library. 5

14  Textual and Codicological Manifestations of Multilingual Culture 

415

Fig. 14.1  London, British Library, MS Harley 2253, fols 52v–53r

contents, while the use of a Latin title shows a formulaic reference to the Latin hagiographic tradition, which Scribe B evokes in his choice of the linguistic resource. Moving on to the break between Booklets 3 and 4, the empty space in fol. 52v was subsequently filled in by Scribe C. Here we observe not only the change of hand and topic—the inserted text consists of eight paint recipes—but also the change of language on the level of genre (cp. the Structural Model of Multilingual Practice for a more elaborated view on the change of linguistic resource at different structural levels; Kopaczyk 2018, 2022). The language of the recipes is Middle English, but on closer inspection one can glean evidence of Scribe C’s multilingual repertoire as well (Fig. 14.2). Consider our transcription of the fragment in Fig. 14.2, with abbreviations expanded in italics and line breaks as in original:

416 

Ad. Putter et al.

Fig. 14.2  London, British Library, MS Harley 2253, fol. 52v (close-up on the bottom half of Column A)

[…] vorte hit beo i molten . & seþþe cast þerof in to þin asure, & sture ham togedere. & ʒef þer beþ bobeles þeron . tac a lutel erewax . & pute þer in . & þenne writ Et ne grynnt þu nout þin asure neuirmore . Et ȝef hit nis noht fin tac itempret gleyr . & cast þerto & let hit stonden & resten . vorte alþe asure beo i vallen a doun to grounde . Et bote þu seo hit fin cast out þe gleyr softelech . & cast þer to more gleyr . & wash hit eft sonus i þe selue maner . Et whan hit is wel i puret & þe gleyr i hald out clene þenne cast þer to þi gummetwater, & writ . as ic seyde er

14  Textual and Codicological Manifestations of Multilingual Culture 

417

‘[…] until it melts. And then cast some of this into your azure, and stir them together. And if there are bubbles in it, take a little earwax, and put it in, and then write. And you need never again grind your azure. And if it is not pure, take tempered egg white, and cast it in and let it stand and rest, until all the azure has fallen down to the bottom. And until you see it pure, remove the glair gently, and cast into it more glair, and wash it quickly in the same manner. And when it is well purified and the glair comes out clean, then add to it your gum-water, and write, as I explained earlier’ (Fein 2014–2015).

The red ink helps to identify the beginning of a clause, although it is not used consistently for this purpose. These paratactically coordinated clauses all start with the equivalent of the English conjunction and, but the coordination is either expressed with the Latin et (e.g. Et ne grynnt þu nout þin asure neuirmore. Et ȝef hit nis noht fin tac itempret gleyr; ‘And you need never again grind your azure. And if it is not pure, take tempered egg white’), or with an abbreviation (& cast þerto & let hit stonden & resten; ‘and cast it in and let it stand and rest’) whose linguistic status is ambiguous. Here we witness the switch of languages within the clause on the level of lexicon, which is arguably enhanced with a visual cue, the red ink. Multilingualism in Harley MS 2253 can go even deeper into the linguistic structure. It is often taken for granted in the study of multilingual practices that native vocabulary, especially proper names, would readily undergo grammatical adaptation to the main language, Latin. Consider this excerpt from Column A in fol. 53r (Fig. 14.3), penned by Scribe B after he left off mid-page on folio 52v, leaving space for the Middle English recipes, and moved on to the next booklet. Consider our transcription of the fragment in Fig. 14.3, with abbreviations expanded in italics and line breaks as in original: Incipit vita sancti Ethelberti Gloriosus ac summo regi acceptus rex ethelbertus regali prosapia oriundus a redwaldo rege in Es stanglia regnante . cuius meminit

418 

Ad. Putter et al.

Fig. 14.3  London, British Library, MS Harley 2253, fol. 53r (close-up of the top half of Column A)

sanctus Beda . in anglorum historia ori entalium anglorum regno undecimo loco prefuit Genitor illius rex mag nificus Ethelredus extitit . geni trix illius regina Leofruna alto san guine progenita . Non hos regie dig nitatis summa potencia . ut crebro solet filios regni fecerat obliuisci .

14  Textual and Codicological Manifestations of Multilingual Culture 

419

quam ineffabilia sint gaudia celesti bus . hec adquirere . hec possidere tota mente affectabant ‘Glorious King Ethelbert, pleasing to the highest king, arisen of royal stock from King Redwald who ruled in East Anglia, whom saintly Bede recalls in The History of the English, was the eleventh to hold dominion over the East Anglians. His father was the magnificent King Ethelred, his mother Queen Leofruna born of noble blood. The highest power of royal office did not cause, as often happens, these children of royal power to forget how indescribable are the joys of those in heaven. They aspired with their whole heart to acquire and possess these’ (Fein 2014–2015).

Scribe B is now writing in Latin. He had already provided a taster of his Latin ability in the red-ink title inserted into the previous scribe’s work. Interspersed throughout the Latin life of King Ethelbert are Old English personal names and topographic terms, adapted to Latin nominal morphology (Ethelberti, Redwaldo, Ethelredus) and Latin spelling conventions (Esstanglia). Instead of treating these instances as borrowings into Latin, one can think of them as linguistic ingredients in the multilingual repertoire of the scribe, whose practice then involves switching the linguistic resource on the lexical level according to need and context, and applying grammatical and orthographic rules of the matrix language to construct a coherent message (see also Keller and Seiler, this volume).

‘Dum ludis floribus’ Such layers of multilingual complexity are famously present in the poetic parts of MS Harley 2253. The following analysis considers examples of both ‘macaronic’ and monolingual poetry in this manuscript and related collections in order to explore the linguistic and cultural nature(s) of multilingualism beyond the idea of ‘Englishness’ interrogated in previous sections. It should be emphasised that the individual examples are not used to make any wider claims about language interaction in England, but rather to highlight the sorts of findings that may be useful in analysing other individual contexts. Demonstrating the kinds of linguistic interaction that occur is important for developing not just our linguistic insight into medieval multilingualism in this particular context, but also

420 

Ad. Putter et al.

for understanding how the models that we use to do so may be influenced by wider cultural assumptions about language relationships during the Middle Ages. The best-known example of macaronic poetry in MS Harley 2253 is ‘Dum ludis floribus’, a short poem composed in Latin, Anglo-French and English and therefore logically described as trilingual. An unusual trilingual work, since most such poems are composed in two languages, it has been the focus of attention because its last two lines are in English (Critten 2020: 224), despite the fact that all four of its preceding stanzas and the first half of the final one are a mixture of Anglo-French and Latin. The nature of this mixture has undergone a shift in critical perception over the last half-century, moving from disorder (Brook 1978: 84) to ‘virtuosic improvisation’ (Critten 2020: 223), but the interaction of the languages has not often been studied at a close syntactic level. Below is the full text of the poem, laid out in stanzaic form (following traditional editorial practice for easier understanding; cp. Fig. 14.4 below): Dum ludis floribus velud lacivia, Le dieu d’amour moi tient en tiel angustia, Merour me tient de duel e de miseria, Si je ne la ay quam amo super omnia.

‘While you play in flowers as if in wantonness The God of Love binds me in such anguish, Holding for me a mirror of sorrow and misery, Since I don’t have her whom I love above all.

Eius amor tantum me facit fervere Qe je ne soi quid possum inde facere; Pur ly covent hoc seculum relinquere, Si je ne pus l’amour de li perquirere.

Love of her makes me burn so fervently That I don’t know what I can do about it; For her I must give up this world, If I can’t be worthy of her love.

Ele est si bele e gente dame egregia, Cum ele fust imperatoris filia; De beal semblant e pulcra continencia, Ele est la flur in omni regis curia.

She’s a lady so superbly beautiful and refined, As if she were an emperor’s daughter; Of lovely appearance and beautiful demeanour, She’s the flower in every king’s court.

Quant je la vey, je su in tali gloria, Come est la lune celi inter sidera! Dieu la moi doint sua misericordia, Beyser e fere que secuntur alia.

When I see her, I’m in such ecstasy, Like the moon among the stars of heaven! May God grant her to me by his mercy, To kiss and do the other things that follow.

Scripsit hec carmina in tabulis. Mon ostel est enmi la vile de Paris. May Y sugge namore, so we[i]a me is; Yef Hi deye for love of hire, duel hit ys!

I’ve written these songs on a tablet. My lodging’s amid the city of Paris. I may say no more, I am so distraught; Should I die for love of her, sad it is!’

(Ed. and trans. Fein 2014–2015) a The manuscript and previous editors read wel, but emendation to wei is indicated by sense and context. We have therefore adapted Fein’s text and translation.

14  Textual and Codicological Manifestations of Multilingual Culture 

421

As Critten (2020: 223) has observed, the rhyming words are always in Latin in the first four stanzas, but the French and Latin in the rest of the lines vary in their proportions. There is more to be gleaned from the poem than this point about the predominance of the languages, however, since the ways in which they interact are also varied. At first glance, the two languages seem to be grammatically interdependent, as a line will start in French and switch into Latin to finish, as in Si je ne la ay quam amo super omnia (l. 4). The switch here can be interpreted as a change of linguistic resource within the conditional clause. In this example the Latin word quam refers back to the French pronoun la, the desired woman, connecting the two languages grammatically. However, there is still grammatical separation between the two languages here, since the Latin is a relative clause embedded in a conditional subclause: it could be removed without altering the overall sense of the verse. In support of this point about separation, a closer look at the poem reveals that this grammatical segregation is the norm. Some of the Latin phrases are adverbial tags, as in in ll. 3.4 and 4.3: Ele est la flur in omni regis curia (‘She is the flower in every king’s court’) and Dieu la moi doint sua misericordia (‘May God give her to me in his mercy’). Other Latin phrases enrich the description: e.g. ‘De beal semblant e pulcra continencia’ (‘Of lovely appearance and beautiful demeanour’; l. 2.3). Here the Latin is syntactically parallel to the French in the first half of the line and could again be removed without altering the line’s grammatical structure or general sense. Changes in languages thus coincide with minor and major grammatical divisions. In many macaronic lyrics this distinction is even more pronounced: different languages often occupy separate lines, as in Harley lyric no. 28, ‘Mayden moder milde, / oiez cel oreysoun’, which moves between English and Anglo-French from one verse to the next. In ‘Dum ludis floribus’, on the other hand, the presence of both languages in a single line gives rise to greater complexity, involving both grammatical integration and separation. This observation nuances both Brook’s and Critten’s points about the interaction of French and Latin. The two languages do not often mix in grammatical terms and, whilst their switching certainly displays compositional / scribal improvisation of a skilled nature that draws on the multilingual resources available, this poem, too, ends up by embracing the

422 

Ad. Putter et al.

alternating line approach seen in other lyrics. The last stanza of ‘Dum ludis floribus’ departs from the previous ones by confining the different languages to separate verses but also by inverting the hierarchy of languages that marks the earlier stanzas and indeed the macaronic lyric in general. In almost all macaronic verse from medieval England, the vernacular comes first and Latin last. This procedure ensured that Latin had the last word, as it does in the first four stanzas of ‘Dum ludis floribus’.6 The last stanza, however, achieves closure by putting the expected sequence of languages in reverse. Latin comes first, then French, and finally we have English in the poet’s emotive closing couplet. In these final two lines, the English is grammatically set apart at the level of discourse (rather than clause) from the French that precedes them. The ‘postscript’ effect is reminiscent of multilingual letters of the period, in which writers not uncommonly switched to English to add a valediction or an urgent personal note (Schendl 2002; Putter 2015). Schendl’s observation about code-switches to English at the end of letters written in Anglo-French applies to ‘Dum ludis floribus’ as well: ‘[i]t seems justified to ascribe […] the function of personalization to English’ (Schendl 2002: 257). ‘Trilingualism’ in this lyric, then, appears to confirm the idea of overt language separation and its attendant points about balance and equality critiqued above.7 However, it is possible to argue that, in many ways, this is a poem whose assumptions and structure are essentially Anglo-French, potentially aligning it with Turville-Petre’s other point about ‘one culture in three voices’. Considering the precise moments where code-switching occurs (usually mid-line) reveals that frequently the words involved, whether French or Latin, are cognates, sharing the same linguistic root. For example, the line Ele est la flur in omni regis curia (l. 3.4) has flur as the last French word, which derives from L flos, floris. Likewise, the line Come est la lune celi inter sidera! (‘Like the moon among the stars of the sky’; l. 4.2) has at the intersection two words that share lexical roots: Fr.  As noted by Wehrle (1933), a rare exception is the macaronic song ‘Exultemus et letemus’ (cited in Harvey 1978), which has Latin first, followed by Anglo-French in the tail-lines. 7  It is notable that recent discussions of ‘Dum ludis floribus’ and other similar lyrics all focus on this idea of separation; see e.g. Putter (2009) and Butterfield (2013), both cited in Critten (2020: 225 and 240–241, notes 16–17). 6

14  Textual and Codicological Manifestations of Multilingual Culture 

423

lune derives from L luna (‘moon’) and L celi (the genitive form of celus ‘sky’) is also the etymon of Fr. ciel. Thus the movement between French and Latin is frequently enabled by the shared lexis of both languages. This is hardly a surprise given French’s Latinate origins, but the crucial fact here is that, at such moments of transition, the distinction between the two languages is minimal; significantly, the multilingualism of the poem is enabled by lexical cognates and underlying grammatical structure. Even the English, despite its roots in a different language family (Germanic), sometimes participates in this shared lexicon. Thus the word duel (l. 5.4) is a repetition of the same word found at the beginning (l. 1.3), where it appears in a French context. The consciousness of shared lexical heritage is thus vital throughout the poem for enabling its linguistic multiplicity, even in its brief English lines. A new appreciation of Scribe B’s multilingual practice is enabled by considering the physical context of his text. In contrast to the edited text presented above for easier comprehension, Figure 14.4 demonstrates clearly that the poem was written in continuous lines. There is no visible break or any special visual cue when the scribe switched from French to Latin, from Latin to French, and indeed, to Middle English at the end of the poem.

Fig. 14.4  London, British Library, MS Harley 2253, fol. 76r (bottom part)

424 

Ad. Putter et al.

Despite the critical analysis of Turville-Petre’s characterisation of trilingualism above, it seems that ‘Dum ludis floribus’ upholds the points made there about three languages with distinct functions that nonetheless collude to create a single culture, since the poem’s Latin, French and English lines are separated grammatically but combine lexically to produce a unified trilingual work. The lack of visual cues separating the linguistic resources in the poem strengthens this point. That single culture should not be mistaken for a narrowly insular one, however. As discussed above, the trilingual culture envisaged is often, despite its multilingual formation, seen by critics in terms of a single proto-national identity, so that (for example) MS Harley 2253’s linguistic and cultural variety is in effect—if not in intention—reduced to a sense of ‘Englishness’ represented mostly by the English-language poems (Turville-Petre 1996: 206–215). However, this proto-national Englishness is not the culture constructed by the witty and complex yet ordered multilingual exchanges in ‘Dum ludis floribus’. Instead, just as the poem is rooted linguistically in Latinity, so too its multilingualism points to a culture predominantly located in the schools and the newly established universities (note the reference to ‘Paris’ in the last stanza), an intellectual culture that is promulgated by both Latin and vernacular works by the time of MS Harley 2253’s copying around 1340 but that is based upon the former language. This ‘learned’, fundamentally Latinate context is demonstrated beyond language itself by the poem’s literary style and form. The rhyming and accentual Latin defines the work, aligning it with similar if metrically more complex works in the well-known Carmina Burana collection (nos. 110 and 111), and some of the code-switches into Latin, such as its depictions of the pains of love and the beauty of the lover likewise owe much to international Latin traditions, even if here they appear in Anglo-French. Despite the English lines, which have caught critics’ attention and have enabled them to develop the concept of ‘trilingual England’ in nationalistic terms, the poem’s point of reference is at heart an international one, showing that a ‘single culture in three voices’ does not automatically result in a local, vernacular-centric ‘Englishness’ but may in fact relate much more to a multifaceted intellectual culture shared across

14  Textual and Codicological Manifestations of Multilingual Culture 

425

Europe (including England) via French and Latin.8 It is a culture that is rooted in the educational structures and developments of the thirteenth century, a timely reminder that, despite MS Harley 2253’s later date, at least some of its textual traditions hark back to a different era.

‘ABC a femmes’ ‘Dum ludis floribus’ reveals that the relationships between multilingual repertoires and cultural identities are more complex than ideas of separate linguistic functions and a unifying nationalism allow. Language alone is not necessarily a straightforward indication of political or social affiliation. This point opens up interesting possibilities for monolingual poems, which naturally enough are usually interpreted in ‘local’ terms both linguistically and culturally. Translation into a vernacular with a restricted geographical remit is habitually, and often accurately, read as relating primarily to a local context, but this does not inevitably mean that other wider and multilingual contexts are excluded. A good example of a monolingual text that nevertheless gestures towards such contexts is the alphabetically acrostic poem ‘ABC a femmes’. This poem appears in MS Harley 2253 in Anglo-French (item 8), but it has a close English-language relation in the ‘ABC of Women’ found in the Auchinleck manuscript (Edinburgh, National Library of Scotland, MS Advocates 19.2.1). This manuscript is best known as the first large anthology of mostly English-­ language verse, compiled in the 1330s (for the full online edition with facsimile images see Burnley and Wiggins 2003). Consider the following two corresponding stanzas from ‘ABC a femmes’ (with our translation below it) and the ‘ABC of Women’. Stanza 5: Harley 2253          Auchinleck Eux ont le corps de bel entayle,      Eiȝen gray & browes bent En tous poyntz tres bien assis.         Þat bere þis birddes briȝt on ble; Um ne vaudreit une mayle        In eueri lond, þer þai be lent Si femme ne fust, ce m’est avys.      Is ful of mirþe & jolifte. Donque dussum nous, sauntz fayle,     It is a sond, þat God haþ sent,  Critten (2020: 239–240) makes the same argument in slightly different terms, focusing on French.

8

426 

Ad. Putter et al.

De tiele chose tenir grant pris,     In erþe to gladi man wiþ gle. Quar il n’y a rien que a femme vayle   Were wimen out of lond ywent, Desouz la joie de parays,          Al our blis were brouȝt on kne    En terre.              Wel lawe: Yl n’y a nulle terrene            Hou schuld men ani corn repe,     Que purra a tous plere.            Þer no sede is souwe. ‘They have beautifully formed bodies and are so nice in all respects that a man would not be worth a halfpenny if women did not exist, so I think. Therefore, we ought to treasure such a thing, without doubt, for here on earth, below the joy of paradise, nothing is worth as much as a woman. No other earthly thing can please everyone’.

It is evident that these poems are closely related. Both portray women in formulas that are reminiscent of descriptions of beauty in other Anglo-­ French and English poems of similar date. For example, Eux ont le corps de bel entayle; in Middle English: Eiȝen gray & browes bent) can be paralleled in Anglo-French narrative and Middle English lyric, respectively. The two works are unsurprisingly rooted in their own linguistically distinct if often overlapping literary traditions. Yet it is notable that, despite this relationship to different linguistic and stylistic spheres, the form of both poems is very similar in rhyme scheme and metrics. The same stanza form is found in other English poems, such as Sir Tristrem and Laurence Minot’s Siege of Tournay, but neither in continental French nor in Anglo-­ French verse (Pickering 2005; Putter 2018). Whether this means the Middle English version was the original one, as Dove (2000: 331–332) has argued, is debatable (see Pickering 2005); but, in any case, the formal identity shows that the translatio process preserved the metre of the original version. This approach to translatio, which carries across features that may be characteristic of a particular linguistic tradition into the new work, is itself inherently multilingual, even if both original and translated versions are monolingual, since both the single-language works are embedded in, and enabled by, a linguistically plural textual culture. Although the translation of the ‘ABC’ might well have been driven by local vernacular needs, its inheritance derives not only from its own language’s stylistic traditions but also from a different literary and linguistic

14  Textual and Codicological Manifestations of Multilingual Culture 

427

context. In this sense, such poetics are a sign of continuing multilingual and cultural influences even within an apparently monolingual work. Like ‘Dum ludis floribus’, but in a different way, the ‘ABC’ poem demonstrates that language interaction cannot only be categorised in terms of separate functions, and that multilingualism inevitably results in cultural pluralism rather than literary nationalism.

4.3 The Diachrony of Multilingualism in Medieval England: The Question of French The diachronic variability with regard to the languages that mix in texts and codices is another matter that deserves further attention. The ‘trilingual’ culture evoked by Rothwell and Turville-Petre provides a good fit for the primary audience of MS Harley 2253 and for the world that Chaucer encountered when he was a page at the household of Countess Elizabeth of Ulster (Pearsall 1992: 34–40). Chaucer’s mother tongue was English, but the language of Countess Elizabeth’s court was probably still French, and the language of record was Latin (interlarded with French and English). The first mention of the ‘Father of English poetry’ is thus found in the Countess’ Latin household accounts. But the influence of French in England’s multilingual matrix diminished rapidly. It used to be thought that French had become an artificial language that had to be taught to English people within a generation or two after the Norman Conquest (Kibbee 1991), but there is good evidence that this decline of Anglo-French into a ‘foreign’ language that English people could no longer speak with native-like competence occurred around the middle of the fourteenth century. An important and neglected source of evidence is to be found in language change. Dead languages do not evolve, and so changes in a language are signs that it is alive. This principle has been the starting point for more recent research carried out by Ingham (2012). His findings show that, until the middle of the fourteenth century, Anglo-­ French underwent some of the same structural syntactical changes as did continental French. This warrants the conclusion that Anglo-French remained an organic part of a French language continuum and was subject to some of the same language changes as was continental French.

428 

Ad. Putter et al.

Another revealing indicator of this is accuracy with regard to grammatical gender in Anglo-French. In English grammatical gender died out in the thirteenth century (Mustanoja 1960: 43–44). French, of course, still has it, and often there is no general rule whether a noun is masculine or feminine. For non-native speakers the only solution is to memorise lexical items and their genders on a case-by-case basis (le lait, la mort, etc.). One would expect non-native speakers to make many mistakes in genderassignment but, interestingly, until the mid-fourteenth century such mistakes are rare in Anglo-French texts. Thus, when we look at pronoun selection (which should dictate mon, ton, son for a masculine noun, but ma, ta, sa for a feminine one), Anglo-French writers are largely accurate until around the middle of the fourteenth century, after which they lose their grip on grammatical gender. Ingham suggests that the Black Death wiped out a generation of older people able to pass on their knowledge of French to the next generation. However, whatever the causes, there is good evidence to corroborate Ingham’s findings that French was spiralling into decline around 1350. While Ranulf Higden in his Latin Polychronicon (c.1357) records the fact that the children of noblemen learn French from the cradle, John Trevisa in his English translation of the work (c.1385) thought it necessary to add the rider that it is no longer so: gentilmen haueth now moche i-left for to teche here children Frensche (Babington and Lumby 1865–1886: i, 161). He also notes the decline of French and the rise of English in the schoolroom. It had been customary—Higden and others such as Robert Holcot (Richter 1979: 36–37) observe—to teach French to English school pupils and to use French as a medium for teaching Latin (the main business of elementary education), but Trevisa notes that English has now replaced French. The result was that students of the ‘new school’ were no longer fluent in French. William Langland complained in Piers Plowman (B.15.372) about a generation of newe clerkes who cannot construe but [except] in Latyn or Englissh (Schmidt 1995), and that complaint can be understood in this historical context. The gradual demise of French as a living vernacular in England, a language acquired with native-like fluency in family settings and by young schoolchildren, also explains the rise of educational texts designed to teach French. Reed (in this volume) examines the manuscripts of the

14  Textual and Codicological Manifestations of Multilingual Culture 

429

earliest Anglo-French conversation manuals, Manières de langage, the earliest dating from the late fourteenth century, and explores what these manuscripts can tell us about the settings in which French had to be acquired now that it was no longer acquired from the cradle. Multilingual manuscripts and macaronic poetry also register the crisis of French as a living vernacular in the mid-fourteenth century. The precariousness of that ‘one culture in three voices’ (Turville-Petre) becomes evident if we compare multilingual manuscripts like MS Harley 2253 with later ones. A Sammelhandschrift of various genres, MS Harley 2253 is the swansong of a particular kind of multilingual codex, one that changes linguistic resource on the genre level, mixing English-language stories for literary entertainment, such as the romance of King Horn, with stories in French, such as fabliaux and the witty tale of the encounter and dialogue between the Jongleur of Ely and the King of England (Le jongleur d’Ely e le roi d’Angleterre). Before MS Harley 2253, we find the same mixture of French and English tales in London, British Library, MS Cotton Vitellius D III, now damaged and fragmentary, but once a large compilation combining narratives in English (e.g. Floris and Blanchefleur) and French (Amis et Amile). Despite the fact that the vast majority of miscellanies (about 70%) containing Middle English are from the fifteenth century, we do not find comparable French-English miscellanies from this period (Griffith and Putter 2014). Language-mixing in lyric poetry reveals the same trend. The three lyrics in MS Harley 2253 that Fein (2014–2015) labels as macaronic, ‘Dum ludis floribus’, ‘Mayden, moder milde’ and ‘Dieu, roy de magesté’ (also known as ‘Against the King’s Taxes’), are linguistically diverse, the first being trilingual, the second English and French, and the third French and Latin. As in the macaronic prose items in this manuscript, ‘Prayer for Protection’ (Latin and French) and ‘Prayer on the Five Joys of Our Lady’ (French and Latin), the one constant here is the co-presence of French. Other macaronic poems composed before 1400 show themselves to be similarly hospitable to French (Harvey 1978). By the fifteenth century, however, macaronic poems composed in England are almost exclusively in English and Latin. Noting that macaronic poems with French become rare in the later medieval period, Wehrle (1933: 62) could find only two from the fifteenth century. These statistics can certainly be improved

430 

Ad. Putter et al.

upon,9 but since so many more extant macaronic lyrics date from the fifteenth century, Wehrle’s conclusion that French disappeared from the multilingual repertoire in the fifteenth century remains valid. Medieval writers registered this fact too. When Dan Michel of Northgate imagined mixed language in his translation of Laurent d’Orléans’s La somme le roi, he conceived of it as taking the form that was normal when he was writing, c.1340: Hoe that bit god with-oute deuocion of herte, he spekth to god patroyllart, ase the ilke thet spekth half englis and half urenss (Morris 1965: 211). ‘Whoever prays to God without heart-felt devotion speaks a mixed language to God, like someone who speaks half-English and half-French’ (our translation).

But some sixty years later, when an anonymous writer translated the same passage, he imagined ‘mixed language’ in the form that had come to dominate macaronic language, that is mixed English-Latin: Who so biddeth God with-oute deuocion of herte, he speketh to God a langage that men clepen patrolard, as who-so speketh half Englische and half Latyn (Francis 1942: 233).

At different times, at different places, and across different genres, then, language mixing took different forms. A properly historical exploration of multilingual texts and codices must therefore be attentive to diachronic changes in the co-occurrence of languages that could enter the multilingual mix. There is interesting research of this kind being done by scholars working on medieval Welsh texts and manuscripts (e.g. Lloyd-Morgan 2015; Russell 2019; Callander 2020; and see also Callander, this volume), and their findings similarly point to a combinatory shift in the fifteenth century. Welsh is commonly found together with Latin in manuscripts throughout the medieval  Fifteenth-century macaronic lyrics containing French known to us include items 137 (possibly two poems, both trilingual: ‘De amico ad amicam’ and ‘Responcio’), 138 (‘En Jesu Roy Soveraign’, French and English) and the trilingual drinking song ‘Fetys bel chere’ from the Selden carol book (Oxford, Bodleian Library, MS Selden B.26). 9

14  Textual and Codicological Manifestations of Multilingual Culture 

431

period, but it did not combine readily with the two other vernaculars of medieval Wales: French and English. There is only one surviving Welshlanguage manuscript which, in its present form, has a French text in it, and that is a medical miscellany (Cardiff, Central Library, MS 3.242, formerly Havod 16). As far as English is concerned, it is in the very late Middle Ages that it begins to be seen in the company of Welsh, again in both manuscript and on the smaller scale of the macaronic poem. For instance, while the mid-fifteenth-century compilations Aberystwyth, National Library of Wales, MSS Peniarth 26 and 50 are not dissimilar from earlier manuscripts in combining Latin and Welsh, they innovate by incorporating also texts in English, and that same innovative combination of Welsh and English is replicated in Tudur Penllyn’s macaronic Ymddiddan Rhwyng Cymro a Saesnes, a colloquy between a Welshman and an English woman who rebuffs the former’s advances (made in Welsh) with indignation (in English). The poem also dates from the midfifteenth century (Johnston 1991: 74–77).

5 Conclusions We have presented in this chapter an overview of the issues involved in the study of multilingual literary texts and manuscripts written in medieval England, using MS Harley 2253 as an illustration of the rich and complex interplay of linguistic resources and of the challenges and the rewards that are involved in unpacking and (re)interpreting the multilingual repertoires of scribes and authors. As we remarked at the beginning of this chapter, Middle English has not always been seen in the context of the larger ecology of languages in which it evolved. Editors have often presented Middle English texts in isolation from texts in other languages that often accompany them in the medieval codex. This has been true of the Middle English items in the Harley manuscript too. For the study of Middle English texts in a multilingual context, the recent trend of editing complete manuscripts (as in Fein’s edition of MS Harley 2253) is a welcome development. The availability of editions of complete multilingual manuscripts will facilitate a process that is well under way in historical linguistics just as it is in literary criticism: the move from curating Middle

432 

Ad. Putter et al.

English in isolation to an appreciation of its broader multilingual basis and context. Looking back at the scholarship of the past twenty years or so, we have seen research adopting a less Anglocentric perspective, even as there remains a tendency to reduce the multilingualism of medieval England to a question of two or three languages. The phenomenon of the trilingual Latin / French / English manuscript was put firmly on the scholarly agenda by, for instance, Hunt (1999) and Scahill (2003), but there also survive from medieval England manuscripts that are, for example, bilingual English / Dutch (London, Guildhall Library, MS 15838; dated 1501), trilingual Latin / French / Dutch (Cambridge, Trinity College, MS O.3.20; early fourteenth century) and quadrilingual Latin /French / English / Dutch (London, Mercers’ Company, Book of Privileges; c.1485, ed. Sutton and Visser-Fuchs 2009). These manuscripts remain neglected. A similar reductive tendency has been at work in the tendency to downplay the status of other languages in comparison to English. While no one denies the continuing importance of Latin as a language of record, it is still repeatedly said that it was a ‘dead’ language compared with the living language of English (Rothwell 1994; Crespo García 2000). In the case of French, too, it has been claimed that it became ‘artificially acquired’ after 1200 (Rothwell 1976; Kibbee 1991). We doubt that this was the experience of many speakers of Anglo-French and, yes, speakers of Latin in medieval England. The importance of French as a spoken vernacular in medieval England only declined after c.1350, but it is multilingual literature and manuscripts that allow us to track these important cultural changes. The structural nature of multilingual repertoires requires closer inspection as well. As we have demonstrated in this chapter, multilingual scribes drew on their multiple linguistic resources in a variety of ways, from switching between wholly monolingual texts in a Sammelhandschrift, to changing languages within a text. Here, the integration of two or more linguistic systems was driven and regulated by a variety of factors—textual, stylistic, grammatical. We are starting to appreciate the non-random character of linguistic mixture in the so-called macaronic texts and even in texts seemingly monolingual, where glimpses of other linguistic systems in a scribe’s repertoire can be seen in grammatical endings (Wright

14  Textual and Codicological Manifestations of Multilingual Culture 

433

2002), in discourse markers or at the level of spelling. These characteristics vary from text to text, from scribe to scribe, and should be incorporated in our study of multilingual repertoires. This brings us to another issue that needs further work: the diachronic fluctuations in the relations between the different languages in different texts and codices of the period, where (as we have emphasised) English mixed not just with French and Latin, but also with other languages. The texts and codices from medieval England never ceased to be multilingual but the possible combinations of the languages deployed in them did change. For example, while there exist poems and manuscripts that mix Welsh and English from the late medieval period, such poems and manuscripts do not exist before the fifteenth century. And, while macaronic poems with French and English (and Latin) were popular in the fourteenth century, French dropped out of the macaronic equation in the fifteenth. There is thus no single culture, albeit multilingual, and since French and Latin were international languages, there are important connections to be made with literature from across the North Sea. The Harley poet’s Latin code-switch in Ele est la flur in omni regis curia bears comparison with that in the famous medieval German Christmas carol In dulci jubilio, which uses a similar phrase Und die Schellen klingen / In regis curia (‘And the bells ring in the King’s court’; ed. and trans. Foster 1957: 74–75). The connection between the poems is not one of direct literary influence; rather, what connects them is the multilingualism of educated clerics and the international world of learning. ‘Dum ludis floribus’ was composed enmi la vile de Paris, according to the last stanza. Presumably, we should imagine an English student at the University of Paris, which drew many clerics from Herefordshire and elsewhere (Birkholz 2020: 79 and 99). One final point needs to be re-emphasised. If we take multilingualism in the Middle Ages to be the rule rather than the exception, at least amongst those trained to read and write, we also need to interrogate the supposed monolingual status that has been accorded to Middle English texts. The ‘ABC of Women’ in the Auchinleck Manuscript has an Anglo-­ French counterpart in MS Harley 2253; the latter is written in a stanza form that appears to have been borrowed from Early Middle English

434 

Ad. Putter et al.

poems. The field of literary production, in short, was multilingual. As a number of contributors to this volume show, more attention needs to be paid to the multilingual traditions underlying seemingly monolingual texts. Close inspection of scribal practice and morphology complicates the monolingual status of Middle English texts further. Titles of ‘English’ items in medieval manuscripts are not uncommonly rubricated in French or Latin (as we have seen in MS Harley 2253); sometimes proper names in English texts are inflected as they would be in Latin; instead of English ‘and’ Latin (or French?) et is sometimes used. To respond to the current research agenda, where multilingualism has become the new paradigm, there is a need to engage with close textual and codicological analysis and to be mindful of the multimodality of communication. In modern editions foreign-language phrases are marked in ways that are easily recognised (e.g. French phrases embedded in English prose are typically italicised). However, as illustrated in the sections ‘A Manuscript Miscellany as a Multi-layered Multilingual Text’ and ‘Dum ludis floribus’, more work is needed to understand the use of visual cues in a multilingual context by medieval scribes. We should pay attention to the layout choices, changes of ink colour, letter size and other visual and multimodal aspects of multilingual communication, and track the variations in practice and changing trends over time. Scholarship on the multilingual contexts of Middle English texts and manuscripts has come a long way and has much further to go.

References ‘ABC a femmes’. 2014–2015. In The Complete Harley 2253 Manuscript, ed. Susanna Fein, trans. David Raybin and Jan Ziolkowski. 3 vols. Kalamazoo, MI: Medieval Institute Publications. ‘ABC of Women’. https://auchinleck.nls.uk/mss/abc.html. Accessed 30 May 2023. Auer, Peter, and Raihan Muhamedova. 2005. ‘Embedded Language’ and ‘Matrix Language’ in Insertional Language Mixing: Some Problematic Cases. Journal of Italian Linguistics 17: 35–54.

14  Textual and Codicological Manifestations of Multilingual Culture 

435

Babington, C., and J.R. Lumby, eds. 1865–1886. Polychronicon Ranulphi Higden monachi Cestrensis. 9 vols. London: Longman. Bergner, Heinz. 1978. MILES: A Crux in MS. Harley 2253 f. 71v. Neuphilologische Mitteilungen 79: 354–358. Birkholz, Daniel. 2020. Harley Manuscript Geographies: Literary History and the Medieval Miscellany. Manchester: Manchester University Press. Böddeker, K., ed. 1878. Altenglische Dichtungen des MS. Harley 225. Berlin: Weidmannsche Buchhandlung. Brook, G.L. 1933. The Original Dialects of the Harley Lyrics. Leeds Studies in English 2: 38–61. ———, ed. 1978. The Harley Lyrics: The Middle English Lyrics of MS. Harley 2253. 4th ed. Manchester: Manchester University Press. Burnley, David, and Alison Wiggins. 2003. The Auchinleck Manuscript. Online facsimile edition. https://auchinleck.nls.uk/index.html. Accessed 30 May 2023. Callander, David. 2020. Bringing Medieval Welsh and English Literature Together. Kelten 84. https://kelten.vanhamel.nl/k84-­2020-­callander-­english-­ welsh-­literature-­multilingualism. Accessed 30 May 2023. Carmina Burana, poems 110 (‘Quis furor est in amore’) and 111 (‘O comes doloris, amor’). http://www.fh-­augsburg.de/~harsch/Chronologia/Lspost13/ CarminaBurana/bur_cam4.html#110. Accessed 30 May 2023. Carroll, Ruth, Matti Peikola, Hanna Salmi, Mari-Liisa Varila, Janne Skaffari, and Risto Hiltunen. 2013. Pragmatics on the Page: Visual Text in Late Medieval English Books. European Journal of English Studies 17: 54–71. Coseria, Eugenio, and Reinhard Meisterfeld. 2003. Geschichte der romanischen Sprachwissenschaft, von den Anfängen bis 1492. Tübingen: Gunter Narr. Crespo García, Begoña. 2000. Historical Background of Multilingualism and Its Impact on English. In Multilingualism in Later Medieval Britain, ed. David Trotter, 23–55. Cambridge: D. S. Brewer. Critten, Rory G. 2020. The Multilingual English Household in a European Perspective: London, British Library MS Harley 2253 and the Traffic of Texts. In Household Knowledges in Late-Medieval England and France, ed. Glenn D. Burger and Rory G. Critten, 219–242. Manchester: Manchester University Press. ———. 2022. The Exploitation of French-English Lexical Transfer in Early Middle English Poetry. Early Middle English 4: 31–50. Dove, Mary. 2000. Evading Textual Intimacy: The French Secular Verse. In Studies in the Harley Manuscript: The Scribes, Contents, and Social Contexts of

436 

Ad. Putter et al.

British Library MS Harley 2253, ed. Susanna Fein, 329–349. Kalamazoo, MI: Medieval Institute Publications. Fein, Susanna, ed. 2014–2015. The Complete Harley 2253 Manuscript. Trans. David Raybin and Jan Ziolkowski. 3 vols. Kalamazoo, MI: Medieval Institute Publications. ———, ed. 2019. Interpreting MS Digby 86: A Trilingual Book from Thirteenth-­ Century Worcestershire. York: York Medieval Press. Foster, Leonard, ed. 1957. Penguin Book of German Verse. Harmondsworth: Penguin. Francis, W.N., ed. 1942. The Book of Vices and Virtues, Early English Text Society, o.s. 217. London: Oxford University Press. Griffith, Gareth, and Ad Putter. 2014. Linguistic Boundaries in Multilingual Miscellanies. In Middle English Texts in Transition: A Festschrift Dedicated to Toshiyuki Takamiya on His 70th Birthday, ed. Simon Horobin and Linne Mooney, 116–124. York: York Medieval Press. Harvey, Carol J. 1978. Macaronic Techniques in Anglo-Norman Verse. L’Esprit Créateur 18: 70–81. Heuser, Wilhelm, ed. 1904. Die Kildare-Gedichte, Bonner Beiträge zur Anglistik, 14. Hunt, Tony. 1999. Insular Trilingual Compilations. In Codices miscellanearum: Brussels Van Hulthem Colloquium 1999, ed. R. Jansen Sieben and H. van Dijk, 51–70. Brussels: Archives et Bibliothèques de Belgique. Ingham, Richard. 2012. The Transmission of Anglo-Norman: Language History and Language Acquisition. Amsterdam: John Benjamins. Ingham, Richard, Louise Sylvester, and Imogen Marcus. 2021. Lone Other-­ Language Items in Later Medieval Texts. Journal of Historical Sociolinguistics 7: 179–205. Jefferson, Judith A. 2013. Scribal Responses to Latin in the Manuscripts of the B-Version of Piers Plowman. In Multilingualism in Medieval Britain, c. 1066–1529: Sources and Analysis, ed. Judith A. Jefferson and Ad Putter, 137–152. Turnhout: Brepols. Johnston, Dafyydd, ed. 1991. Canu Maswedd yr Oesoedd Canol / Medieval Welsh Erotic Poetry. Cardiff: Tafol. Kibbee, Douglas. 1991. For to speke Frenche trewely: The French Language in England, 1100–1600. Amsterdam: John Benjamins. Kopaczyk, Joanna. 2018. Administrative Multilingualism on the Page in Early-­ Modern Poland: In Search of a Framework for Written Code-Switching. In

14  Textual and Codicological Manifestations of Multilingual Culture 

437

Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skafari, and Laura Wright, 275–298. Berlin: De Gruyter Mouton. ———. 2022. The Challenges of Bringing Together Multilingualism and Multimodality: Unpacking the Structural Model of Multilingual Practice. In Multilingualism from Manuscript to 3D: Intersections of Modalities from Medieval to Modern Times, ed. Matylda Włodarczyk, Jukka Tyrkkö, and Elżbieta Adamczyk, 119–138. London: Routledge. Lerer, Seth. 2008. Dum ludis floribus: Language and Text in the Medieval English Lyric. Philological Quarterly 87: 237–255. Levelt, Sjoerd, and Ad Putter. 2021. North Sea Crossings: The Literary Heritage of Anglo-Dutch Relations. Oxford: Bodleian Library Publishing. Lloyd-Morgan, Ceridwen. 2015. Writing without Borders: Multilingual Content in Welsh Miscellanies from Wales, the Marches, and Beyond. In Insular Books: Vernacular Manuscripts in Late Medieval Britain, ed. Margaret Connolly and Raluca Radulescu, 175–192. Oxford: Oxford University Press. Lusignan, Serge. 1986. Parler vulgairement: Les intellectuels et la langue française aux XIIIe et XIV siècles. Paris: Vrin. Machan, Tim. 2011. The Visual Pragmatics of Code-Switching in Late Middle English Literature. In Code-Switching in Early English, ed. Herbert Schendl and Laura Wright, 303–333. Berlin: De Gruyter Mouton. Morris, Richard, ed. 1965. Dan Michel’s Ayenbite of Inwyt or Remorse of Conscience, revised by Pamela Gradon, Early English Text Society, o.s. 23. Oxford: Oxford University Press. Mustanoja, Tauno F. 1960. A Middle English Syntax, Part I: Parts of Speech, Mémoires de la Société Néophilologique de Helsinki, 23. Helsinki: Société Néophilologique. Myers-Scotton, Carol. 2002. Contact Linguistics. Oxford: Oxford University Press. Ormrod, Mark, Bart Lambert, and Jonathan Mackman. 2019. Immigrant England, 1300–1550. Manchester: Manchester University Press. Padel, Oliver J. 2017. Where was Middle Cornish Spoken? Cambrian Medieval Celtic Studies 74: 1–32. Pahta, Päivi, Janne Skaffari, and Laura Wright, eds. 2018. Multilingual Practices in Language History: English and Beyond. Berlin: Mouton de Gruyter. Pearsall, Derek. 1992. The Life of Geoffrey Chaucer. Oxford: Blackwell. Peikola, Matti, Aleksi Mäkilähde, Hanna Salmi, Mari-Liisa Varila, and Janne Skaffari, eds. 2017. Verbal and Visual Communication in Early English Texts. Turnhout: Brepols.

438 

Ad. Putter et al.

Pickering, Oliver. 2005. Stanzaic Verse in the Auchinleck Manuscript: The Alphabetical Praise of Women. In Studies in Late Medieval and Early Renaissance Texts in Honour of John Scattergood, ed. Anne Marie D’Arcy and Alan J. Fletcher, 287–204. Dublin: Four Courts Press. Putter, Ad. 2015. Anglo-French Letters and the Private Correspondence of an Abbot to His Mistress. In Language in Medieval Britain: Networks and Exchanges, ed. Mary Carruthers, 201–213. Donington: Shaun Tyas. ———. 2016. The Linguistic Repertoire of Medieval England, 1100–1500. In Imagining Medieval English, ed. T.W. Machan, 126–144. Cambridge: Cambridge University Press. ———. 2018. The Singing of Medieval Romance: Stanza Forms and Contrafacta. In The Transmission of Medieval Romance: Metres, Manuscripts and Early Prints, ed. Judith Jefferson and Ad Putter, 69–90. Cambridge: D.S. Brewer. Richter, Michael. 1979. Sprache und Gesellschaft im Mittelalter: Untersuchungen zur mündlichen Kommunikation in England von der Mitte des elften zum Beginn des vierzehnten Jahrhunderts. Stuttgart: Hiersemann. Robbins, R.H., ed. 1959. Historical Poems of the XIV and XV Centuries. New York: Columbia University Press. Rogos-Hebda, Justyna. 2020. Visual Pragmatics of Abbreviations and Otiose Strokes in John Lydgate’s Siege of Thebes. Journal of Historical Pragmatics 21: 1–27. Rothwell, William. 1976. The Role of French in Thirteenth-Century England. Bulletin of the John Rylands University of Manchester 58: 445–466. ———. 1994. The Trilingual England of Geoffrey Chaucer. Studies in the Age of Chaucer 16: 45–67. Russell, Paul. 2019. Bilingualisms and Multilingualisms in Medieval Wales: Evidence and Inference. Transactions of the Honourable Society of Cymmorodorion 25: 7–22. Scahill, John. 2003. Trilingualism in Early Middle English Miscellanies: Languages and Literature. Yearbook of English Studies 33: 18–32. Schendl, Herbert. 2002. Code-Choice and Code-Switching in Some Early Fifteenth-Century Letters. In Middle English from Tongue to Text: Selected Papers from the Third International Conference on Middle English, ed. P.J. Lucas and A.M. Lucas, 247–262. Frankfurt: Peter Lang. Schendl, Herbert, and Laura Wright, eds. 2011. Code-Switching in Early English. Berlin: De Gruyter Mouton. Schmidt, A.V.C., ed. 1995. The Vision of Piers Plowman. London: Dent.

14  Textual and Codicological Manifestations of Multilingual Culture 

439

Sebba, Mark. 2012. Researching and Theorising Multilingual Texts. In Language Mixing and Code-Switching in Writing: Approaches to Mixed-Language Written Discourse, ed. Mark Sebba, Shahrzad Mahootian, and Carla Jonsson, 1–26. London: Routledge. Sharpe, Richard. 1996. Latin in Everyday Life. In Medieval Latin: An Introduction and Bibliographical Guide, ed. F.A.C. Mantello and A.G. Rigg, 315–341. Washington: Catholic University of America Press. Smith, Llinos Beverley. 2000. The Welsh and English Languages in Late-­ Medieval Wales. In Multilingualism in Later Medieval Britain, ed. David Trotter, 7–24. Cambridge: Brewer. Sutton, Anne F., and Livia Visser-Fuchs, eds. 2009. The Book of Privileges of the Merchant Adventurers of England, 1296–1483. Oxford: Oxford University Press. Tiddeman, Megan. 2017. Early Anglo-Italian Contact: New Loanword Evidence from Two Mercantile Sources, 1440–1451. In Merchants of Innovation: The Languages of Traders, ed. Esther-Miriam Wagner, Bettina Beinhoff, and Ben Outhwaite, 217–234. Berlin: De Gruyter Mouton. ———. 2018. Lexical Exchange with Italian in the Textile and Wool Trades in the Thirteenth to Fifteenth Centuries. Medieval Clothing and Textiles 14: 113–140. ———. 2020. More Sugar and Spice: Revisiting Medieval Italian Influence on the Mercantile Lexis of England. In The Multilingual Origins of Standard English, ed. Laura Wright, 381–410. Berlin: De Gruyter Mouton. Trotter, David, ed. 2000. Multilingualism in Later Medieval Britain. Cambridge: Brewer. Turville-Petre, Thorlac. 1996. England the Nation: Language, Literature, and National Identity, 1290–1340. Oxford: Oxford University Press. ———, ed. 2015. Poems from BL Harley 913, ‘The Kildare Manuscript’, Early English Text Society, o.s. 345. Oxford: Oxford University Press. Wehrle, William Otto. 1933. The Macaronic Hymn Tradition. Washington: Catholic University of America Press. Wright, Laura. 2002. Code-Intermediate Phenomena in Medieval Mixed-­ Language Business Texts. Language Sciences 24: 471–489. ———, ed. 2021. The Multilingual Origins of Standard English. Berlin: Mouton De Gruyter.

15 Adapting Winefride in Welsh, Latin and English David Callander

1 Introduction: Welsh Saints and Trilingual Traditions Recent years have seen a great uptake in interest in the Lives of the Welsh saints. Whereas these texts were traditionally treated uncritically as witnesses to the Age of Saints in the sixth and seventh centuries, scholars since the latter part of the twentieth century have highlighted how they are more productively viewed as shedding light on the historical context and textual culture in the years in which they were composed, in the late eleventh and twelfth centuries (Dumville 1977; Evans and Wooding 2007). This more recent approach has culminated in two large projects funded by the UK’s Arts and Humanities Research Council (AHRC): Cwlt y Seintiau yng Nghymru / The Cult of Saints in Wales (2013–2017), which edited the saints’ Lives and poems to saints surviving in Welsh, and

D. Callander (*) Cardiff University, Cardiff, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_15

441

442 

D. Callander

Vitae Sanctorum Cambriae (2017–2019), which edited the Latin Lives of the Welsh saints.1 A number of these editions are now available on the project website (https://www.welshsaints.ac.uk/) with further editions forthcoming. Our improved understanding of these Lives and their transmission is highly relevant to the study of medieval English in a multilingual context: the Lives of Welsh saints provide an excellent opportunity for making cross-linguistic comparisons. These are texts which are almost always composed in Latin initially, as with so much hagiography of the eleventh and twelfth centuries, but subsequently a number of them are translated into Middle Welsh and, more surprisingly, Middle English. This is true, for example, of the most famous of Welsh saints, St David. His Latin vita was originally composed by Rhygyfarch of Llanbadarn Fawr near Aberystwyth in the late eleventh century, but survives in a wide variety of Latin versions (see James 1967; Sharpe 2007). It was also translated and adapted into Middle Welsh as Buchedd Dewi, perhaps in the fourteenth century, its earliest manuscript dating from 1346 (ed. Evans 1959). A Middle English Life of David by fifteenth-century prior Osbern Bokenham also recently came to light. This was not identified until 2005, when Simon Horobin discovered the Abbotsford manuscript (Abbotsford, Advocates Library, Abbotsford MS B3), containing 183 saints’ Lives by Bokenham (Horobin 2005, 2007). The Middle English Life of David, which is ultimately in the tradition of Rhygyfarch’s Life and concludes with Latin poetry, has now been published (ed. Horobin 2020: 339–345) and invites comparative research: how and why does it differ from the Latin and Middle Welsh versions? How did the Middle Welsh and Middle English authors adapt shared Latin sources? Teilo, a patron saint of Llandaf, also has a Middle English Life, although no extant Middle Welsh Life, and we can similarly examine how the Middle English adaptor changed the Latin vita and consider why they chose to translate a text from Wales (ed. Kooper and Callander 2016). While the Lives of David and Teilo may enable

 The Cult of Saints in Wales project was led by David Parsons of the University of Wales Centre for Advanced Welsh and Celtic Studies. The Vitae Sanctorum Cambriae project was led by Paul Russell of the University of Cambridge. 1

15  Adapting Winefride in Welsh, Latin and English 

443

Fig. 15.1  Stemma of the Lives of Winefride

comparative research, there is one border-crossing saint par excellence in medieval Wales, namely Winefride or, in Welsh, Gwenfrewy. Medieval texts from the twelfth century onwards normally present Winefride as a virgin martyr from North-East Wales, who was resurrected by St Beuno but retained the scar where she had been decapitated. The site of her martyrdom was identified with that of Winefride’s Well at Holywell, Flintshire, which is said to have burst forth after her head fell to the ground. The stemma in Fig. 15.1 provides the broad outline of how medieval Winefride texts relate to one another.2 As can be seen from Fig. 15.1, of pivotal importance are the two early Latin Lives of Winefride. These are twelfth-century works, Robert of Shrewsbury having composed his lengthy Life around 1137–1142 (Pepin and Feiss 2000; ed. Callander 2023b). Robert’s Life moves from Winefride’s lifetime to Robert’s present, describing how he and others translated her relics from Gwytherin to Shrewsbury in 1137/8. The anonymous Life, being much shorter, but with descriptions of many miracles at Winefride’s well site, appears to have been composed in the  For an excellent and detailed overview of the trilingual Winefride traditions, focusing more on broader textual changes than textual multilingualism and the translation of individual words and names as here, see Gregory (2012). The stemma presented here is in some ways likely to be a simplification. Sims-Williams (2018: 23–32) has argued that Robert of Shrewsbury’s Life influenced Buchedd Beuno in its presentation of Winefride’s martyrdom (whether directly or by having fed back into the tradition of the Latin Life of Beuno at an anterior stage), although this is unproven; cf. Callander (2019: 137). 2

444 

D. Callander

vicinity of Holywell (ed. Wade-Evans 1944: 288–308; ed. Callander 2023a). Robert of Shrewsbury’s Life was far more widely circulated in medieval Britain and forms the basis of the Welsh prose translation Buchedd Gwenfrewy, the earliest manuscript of which dates from the second half of the fifteenth century (ed. Cartwright Forthcoming-a; ed. Baring-Gould and Fisher 1907–1913, iv: 397–423) and the various Middle English adaptations. Robert’s Life also survives in a shortened version in the thirteenth-century manuscript Cambridge, Trinity College, MS O.4.42 (ed. Callander 2023c) and in an even shorter abridgement as part of John of Tynemouth’s Sanctilogium (ed. Horstmann 1901, ii: 415–421; ed. Marzella Forthcoming). Elements of both Lives are combined in the short compound version in the fourteenth-century codex London, British Library, MS Lansdowne 436 (ed. Gregory 2016; ed. Callander 2023d). All the manuscripts of the different versions of Winefride’s Latin Lives appear to have been copied in England, with no Latin Life of Winefride surviving in manuscripts copied in Wales. Winward (1999) has argued that the great similarities seen between the two twelfth-century Latin Lives of Winefride indicate they drew on a shared source, namely a lost Latin Life of Beuno. Buchedd Beuno, a Middle Welsh translation of the Life of Beuno, survives and contains the story of Winefride’s martyrdom (ed. Sims-Williams 2018). Five late medieval Welsh poems to Winefride and her well also survive, although due to the nature and brevity of the poetry it is difficult to connect them directly with either of the Lives (for discussion, see Callander 2020). As is clear from the stemma, the shape of Winefride’s traditions differs in Middle English, Middle Welsh, and Latin. Lewis has recently highlighted the role of language in determining the form of hagiographical traditions in Wales.3 He highlighted that in Latin, the material is primarily prose or prosimetric vitae produced in the late eleventh and twelfth centuries, with relatively little independent verse. In Welsh, the prose saints’ Lives are later and derivative, in almost every case being translations of Latin originals. However, celebratory poetic hagiography

 Barry Lewis made these contributions in his presentations at the Leeds International Medieval Congress (Leeds, 2019) and the Vitae Sanctorum Cambriae Conference (Cambridge, 2019). 3

15  Adapting Winefride in Welsh, Latin and English 

445

flourished in Welsh, particularly in the later Middle Ages, and it is clear that the Winefride tradition in Wales fits this model well. Yet Winefride is unique among Welsh saints in having a substantial tradition in Middle English as well as Middle Welsh and Latin. How do the Middle English texts compare? Middle English Winefride traditions flourished primarily from the fifteenth century onwards (Gregory 2012). Most of these texts are closely related to Robert of Shrewsbury’s Life, but with many unique innovations. The earliest extant Middle English Winefride text is John Mirk’s Festial sermon on Winefride, perhaps composed in the later 1380s, Mirk then being a canon of the Augustinian house at Lilleshall in Shropshire (ed. Powell 2009–2011). This relatively early interest in Winefride in Shropshire is not surprising, given the strength of her cult at Shrewsbury, where her relics had been translated. Following the elevation of Winefride’s feast by Canterbury in 1415, more Middle English Winefride texts appear. John Audelay the Blind (fl. c.1417–1431), a chantrist at the Augustinian Haughmond Abbey near Shrewsbury, addressed two Middle English poems to Winefride: a carol and a salutacio (ed. Whiting 1931; ed. Fein 2009). Both John Audelay and John Mirk can be associated with Shrewsbury or its vicinity and a Shrewsbury focus is clear in their texts too, but Winefride travelled further than this. There is a Life of Winefride in a single mid-fifteenth-­ century manuscript of the Gilte Legende (London, British Library, MS Add. 35298; ed. Hamer and Russell 2000) and an unrelated Life of Winefride in one fifteenth-century manuscript of the South English Legendary, Oxford, Bodleian Library, MS Bodley 779 (ed. Horstmann 1889). In the mid-fifteenth century, Osbern Bokenham, a friar of Clare Priory in Suffolk, produced a metrical English Life of Winefride, found, like the Middle English Life of David, in the newly discovered Abbotsford manuscript (ed. Gregory 2012: 412–420). Bokenham had made a pilgrimage to Winefride’s well at Holywell in 1448 (Gregory 2012: 285), illustrating its popularity as a destination for pilgrims in the later Middle Ages. As well as editing and printing the Gilte Legende Life of Winefride in 1483/4 as part of his Golden Legend (ed. Ellis 1900, vi: 127–132; Gregory 2012: 277), around 1484 Caxton translated and published Robert of Shrewsbury’s Life of Winefride as an individual pamphlet (ed. Horstmann 1880). In places he translated fairly closely and elsewhere

446 

D. Callander

(particularly in the description of Winefride’s life itself ) omitted large amounts of material. While Winefride traditions flourished in Latin, Welsh and English, no Winefride texts are known to survive in French, which may be a result of the late flourishing of the cult, primarily in the fifteenth century. The medieval traditions of Winefride are thus trilingual insofar as there is substantial material surviving in Latin, Welsh, and English. Yet it is striking when we turn to the texts themselves that they are most commonly monolingual. The Latin Lives are all in Latin only, although there is some interesting treatment of Welsh personal and place-names, discussed further below. The Welsh poems to Winefride are all in Welsh only. In the two earliest manuscripts of Buchedd Gwenfrewy (Aberystwyth, National Library of Wales, MS Peniarth 27ii; s. xv2; and Aberystwyth, National Library of Wales, MS Llanstephan 34, s. xvi ex.), all the material is in Welsh only, including the title for the Life. Aberystwyth, National Library of Wales MS Peniarth 252 (1594–1610), on the other hand, is a trilingual Latin-Welsh-English compilation and includes some Latin words and marginalia in its copy of Buchedd Gwenfrewy (Cartwright Forthcoming-b), although the text is still overwhelmingly in Welsh only. Latin incipits and explicits appear to be less common as a rule in Middle Welsh texts than Middle English texts, although this conclusion must remain tentative until a full study is undertaken. It is thus the primarily Middle English Winefride texts which are most multilingual. Beyond the use of personal and place-names, this multilingualism is always English / Latin, never English / Welsh. Bokenham’s Life and the Gilte Legende Life are entirely in Middle English. This is also true of the Bodley 779 Life, apart from the use of Deo Gracias in l. 62, rhymed with was, which refers to what the people in the church say when Winefride is resurrected. Mirk’s sermon is all in English but has a Latin title and the collection as a whole has a Latin incipit and explicit in London, British Library, MS Cotton Claudius A II. The Festial survives in whole or in part in forty-five manuscripts as well as early prints (Powell 2009–2011, i: xliii–xliv). Intriguingly, its textual tradition includes a sixteenth-century manuscript (Cardiff, Central Library, MS 2.632 (Havod 22)) which contains a translation of some of the temporale

15  Adapting Winefride in Welsh, Latin and English 

447

sermons into Welsh, but not of the Winefride sermon.4 In terms of the Middle English Winefride tradition, however, it is Audelay and Caxton who make the most extensive use of Latin. The extent of the use of Latin by Audelay and Caxton in treating of Winefride has not received sufficient attention. In both instances the Latin texts have been presented as peripheral through italicisation or, in the case of Horstmann’s edition of Caxton’s pamphlet Life, placement in footnotes. Audelay’s poems survive in a single fifteenth-century manuscript (Oxford, Bodleian Library, MS Douce 302), which employs significant amounts of both Latin and English. While the Winefride salutacio has a brief Latin incipit and explicit, the main text of carol and salutacio is in English only (unlike some of Audelay’s other English salutaciones, which repeat a Latin word or words at the start of successive lines). However, these Middle English texts are followed by a Latin rhymed antiphon and a Latin collect. The collect is also found in Caxton’s Winefride pamphlet, while a version of the same rhymed antiphon to Winefride occurs in John of Tynemouth’s Sanctilogium (Harper 2010). Thus the following Winefride collection is found in Douce 302, fols 26r–27r: 1 . ‘Wenefrede, thou swete may’ (Middle English Winefride carol) 2. ‘Salutacio Sancte Wenefrede’ (Middle English Winefride salutacio) 3. ‘Virgo pia Wynfryda’ (Latin rhymed antiphon to Winefride) 4. ‘Deus qui beatam virginem…’ (Latin collect for Winefride). What the codex provides here is a small bilingual collection of Winefride material. The Latin texts are not distinguished by script in the manuscript from the English as they are by italicisation in the modern editions of Audelay’s work and there is no reason to think they are of a lower status than the preceding Middle English texts, albeit they are far shorter. Indeed, one might expect the Latin liturgical texts to be of higher status, although the manuscript does not indicate this either.

 For an edition, see Lewis (1925). On translation of Middle English texts into Welsh and other languages, see Byrne (2020). 4

448 

D. Callander

The Latin texts in Audelay’s collection have clear liturgical associations, and this is also a relevant context for the Latin texts in Caxton’s Winefride pamphlet. Whereas the Life of Winefride itself and the story of the moving of her relics are in English only, including incipit and explicit (the latter noting Caxton as the translator), the English is followed by three and a quarter pages of Latin, which concludes the pamphlet. These are extensive Latin prayers for different Winefride festivals, including that of her passion but also that of her translation. Liturgical use, particularly at Shrewsbury, would provide a strong motivation for the inclusion of Latin here. Indeed, Harper (2007: 215) suggests that this liturgical collection ‘was probably assembled direct from material used at Shrewsbury’, with Lowry (1983: 116) arguing that ‘the commission and editorial advice came from Shrewsbury’, probably via Thomas Mynde, abbot of St Peter’s, Shrewsbury (cf. Sutton 2005). Some of the Latin material in Caxton’s pamphlet is unique, while other pieces are also found in other liturgical books (Harper 2007: 215–216). As noted, the Latin collect in John Audelay’s collection also appears in Caxton’s printing, although the rhymed antiphon ‘Virgo pia Wynfryda’ does not. Winefride’s liturgy has not received much attention and far more work is needed to edit and compare these texts, which are normally in Latin and survive in England rather than Wales. It is worth noting that the two words of Latin in the Bodley 779 Life, Deo gracias, are also strongly associated with a liturgical context, this being a response at mass, and, in the text, occurs in the context of attending mass at church. Looking at the Winefride texts as a whole, while the tradition of Winefride is trilingual, most individual works or groups of works are wholly or largely monolingual. The main exceptions to this are Audelay’s Winefride collection and Caxton’s Winefride pamphlet. In these cases, the Latin texts provided are liturgical. Beyond the use of personal and place-names, we do not find mixed-language English / Welsh texts. This is perhaps surprising as extant Winefride texts were often produced in or aimed at areas in England near Wales like Shrewsbury, where there would have been substantial numbers of Welsh speakers, or in North-East Wales, the main area of Winefride’s Welsh cult, where there would have been substantial numbers of English speakers (cf. the extensive use of English loanwords in Welsh poetry describing luxury goods in a Marcher context,

15  Adapting Winefride in Welsh, Latin and English 

449

discussed by Fulton, this volume). This is not reflected in the texts that we have and it is worth asking why this is the case. With a few creative exceptions (cf. Roberts 1958: 53–54), mixed-language texts in Welsh and English do not appear to have been common in the Middle Ages, although after 1400 we do find some examples of multilingual manuscripts which contain different Welsh and English texts, often alongside Latin, such as Aberystwyth, National Library of Wales, MSS Peniarth 26, 50 and 326 (Lloyd-Morgan 2015). The lack of direct influence between vernacular traditions in Welsh and English (which is much starker in the period before 1300) might push up against and complicate some of the work on multilingual exchange in medieval Britain which does not take Welsh as a focus. Nevertheless, the Welsh language does not disappear entirely in the Latin and Middle English Winefride traditions and its presence there will be the main focus of the remainder of this chapter.

2 The Welsh Language in Latin and Middle English Winefride Texts As noted, Winefride’s Latin Lives are almost entirely monolingual in all their versions, but Robert of Shrewsbury’s Life is interested in engaging with Welsh, and this is maintained and adapted in a number of later texts drawing on Robert’s work. Some of this may derive from the lost Latin Life of Beuno which Robert was using, although this is by no means certain. Whatever the explanation, Robert certainly had no problem in calling attention to the fact that people spoke a different language in the area where Winefride lived and died. One interesting part of this relevant to the study of how communication functioned between different linguistic groups is that Robert specifically refers to oral translation during the removal of Winefride’s relics in 1137/8. At Gwytherin in North-East Wales, where Winefride had been buried, Robert speaks to the locals, who are initially very hesitant and who only agree to hand her over data pecunia, a fact which Robert’s later adaptors are keen to omit. But Robert does not speak to the people directly in their language; he uses a local priest as an interpreter:

450 

D. Callander

Tunc prior, eodem presbitero mediante seu interprete, cateruam hominum quę astabat allocutus… ‘Then the prior, with that priest mediating and likewise being an interpreter, having addressed the company of people which was standing nearby…’ (§34; text and translation from Callander 2023b).

Robert is a good witness to these events: he was present and he himself was the prior speaking here. While he may have invented other elements of the Life and Translation, I see no motivation for inventing this translator, and thus this provides one possible model for communication between speakers of different languages in medieval Wales, where a local priest could act as a translator. Presumably, the priest is translating into Welsh, although this is not stated explicitly. Yet, it is unclear which language Robert was speaking here. Was it Latin, French, or English? This is not stated, and they are all possible options. Of all the Latin, Middle Welsh, and Middle English works, the only later text to maintain this reference to the translator is Caxton’s pamphlet Life of Winefride: Thenne the pryour by the moyen of the preest or by his interpretacion spacke vnto alle the companye of men that were there (Horstmann 1880: 310). Caxton thus maintains this acknowledgement that the people at Gwytherin spoke a different language to that of Prior Robert. This is in stark contrast to many of the other Middle English versions of the Life, some of which omit all reference to Wales. What can have motivated Caxton alone to retain this? While Caxton significantly reduces Robert’s description of Winefride’s Life itself, the story of the Translation of her relics is kept much more fully. As noted, this may be because the pamphlet was commissioned by clerics at Shrewsbury, who would have had a particular interest in the story of how Winefride’s relics were transferred to them. Caxton thus provides a full and detailed translation of this part of Robert’s work and the retention of this reference is part of this. This is the only reference to oral translation in the texts. Otherwise, the appearance of the Welsh language in the Latin and Middle English works is primarily associated with place-names and personal names. We can see an example of this in Robert of Shrewsbury’s provision of the Welsh

15  Adapting Winefride in Welsh, Latin and English 

451

epithet casulsych (‘dry-chasuble’) for Beuno. After Beuno departs from Holywell, Winefride sends him a chasuble every year, which is miraculously carried to him by the well-stream and in the sea, while remaining completely dry. As Robert writes, Ex hoc autem beato uiro additum cognomen est, quod apud Walie homines hodieque memoriale habetur. Nam Beunous Casulsech, id est, Beunous Casulasicca uocatur, eo quod per aquas ab aquis illesa ad eum sicca casula deferretur. ‘Now on account of this a nickname was given to the holy man, which is held to be worthy of remembrance today as well among the people of Wales. For he is called Beuno Gasulsych, that is, Beuno Dry-chasuble, because the dry chasuble would be conveyed to him through waters undamaged by waters’ (§11; text and translation from Callander 2023b).

This epithet is attested in other sources, like versions of the twelfth- or thirteenth-century Welsh genealogical collection Bonedd y Saint (‘The Pedigree of the Saints’; Bartrum 1966: 59; on dating the collection, see Lewis 2020). Robert provides both the Welsh form and a Latin translation, which indicates that he did not expect his readers to understand Welsh, but that he nevertheless thought the Welsh form was worth preserving. Casulsech would have been a possible Welsh spelling of the epithet at the time Robert was writing and may well derive from a written Latin source from Wales, such as the lost Latin Life of Beuno. This epithet is retained in both Welsh and Latin in the shortened versions of Robert’s Life in the Trinity O.4.42 recension (§11) and by John of Tynemouth (§7). The Welsh epithet is retained but the Latin removed in the Middle Welsh translation of Robert’s Life, Buchedd Gwenfrewy (§20). There is a similar section in the anonymous Latin Life, but with Beuno’s epithet in Latin only: Huic casule uirginis ob meritum tanta uirtus inerat quod ubicumque Beunonus foret hanc indutus, nec pluuia humectari nec pilus eius posset uento moueri. Talis rei euentu, Beunonus Casulam Siccus agnominatus est.

452 

D. Callander

‘This chasuble, on account of the virgin’s merit, contained such power that, wherever Beuno should be wearing it, it could neither be made wet by rain nor could its material be moved by wind. After that had happened, Beuno was distinguished by the nickname Dry-Chasuble’ (§17; text and translations from Callander 2023a).

The compound Lansdowne recension here follows the anonymous Life and retains the final sentence largely unchanged: ‘Tali euentu rei, Bennonius Casula Sicca congnominatus est’ (§4; text from Callander 2023d). The fact that this miracle is described in both Robert’s Life and the anonymous Life might indicate that it was found in the lost Latin Life of Beuno, although the Middle Welsh Buchedd Beuno contains no such miracle, perhaps having omitted it as part of the process of adapting, translating, and shortening the Life. While the epithets are largely retained in the Welsh and Latin texts drawing directly on Robert’s Life, it is notable that no epithet is kept in any of the Middle English texts. Of these, all but Mirk’s Festial describe the relevant miracle, but do not connect this with Beuno name, or name him as Casulsych, Casula Sicca, or a Middle English equivalent. As can be seen, Robert of Shrewsbury does not merely record Welsh names in his Latin Life: sometimes he explains and comments upon them. Thus he discusses Winefride’s Well, which at her martyrdom burst forth where her head touched the ground: Locus uero ubi sanguis illius effusus est, primitus Siccauallis dicebatur. Postquam autem capud uirginis abscisum terram tetigit et, ut prefati sumus, fons ibi aquę salientis emanauit, qui etiam manet usque in hodiernum diem, sanans omnes languores tam in hominibus quam inpecoribus, locus isdem de nomine puelle uocabulum sortitus est. Nam illorum lingua Fennan Wenefrede, nostra uero Fons Wenefrede nominatus est. Fennan enim latino sermone fontem sonat. ‘Now the place where her blood was spilled was called Dry Valley to begin with. But after the virgin’s severed head touched the ground and, as we mentioned earlier, a spring of gushing water flowed out there, which still remains up to this very day, healing all illnesses both in people and in animals, that place acquired a designation from the girl’s name. For it was

15  Adapting Winefride in Welsh, Latin and English 

453

called in their language Winefride’s Fennan, but in ours Fons Wenefrede [Winefride’s Well]. For fennan denotes fons [well] in Latin speech’ (§5; text and translation from Callander 2023b).

Robert here provides the Welsh and Latin names for Winefride’s well, although notably the second part of the Welsh form is simply the saint’s Latin name, rather than a form of Welsh Gwenfrewy. Robert’s spelling fennan merits discussion as the presumed basis for the use of the word in Middle English adaptation. It could derive from an oral or written form of Welsh ffynnon, which in the twelfth century is normally spelt , with being a possible form, representing /ə/. Robert’s form could thus derive from a miscopying of in a source or, given the form in the Trinity recension (the added by the scribe), could be an error for Robert’s original , which would reflect the spoken form of the word. Robert’s discussion here is paralleled by a similar section in Buchedd Beuno, which could indicate that this derives from the lost Latin Life of Beuno. Yn y lle y syrthyawd y gwaet ar y dayar, y kyvodes ffynnyawn odyno; a’r ffynnawn honno hyt hediw yssyd yn rodi yechyt y dynyon ac anyveileit oc eu heinyeu a’e clwyfeu. A’r ffynnawnn honno a enwit o enw y vorwyn, ac a elwit Ffynnawn Wenvrewy (§13; Sims-Williams 2018: 147). ‘In the place where her blood fell on the earth, a well rose up from there; and that well to this day gives health to people and animals from their diseases and wounds. And that well was named after the maiden’s name, and was called Winefride’s Well’ (my translation).

In Buchedd Beuno (as also in the Middle Welsh translation of Robert’s Life) the Welsh name is given without explanation, as this is unnecessary. But was there an explanation in the Latin Life of Beuno or was this added by Robert? If the latter, this raises the question of how Robert came to learn the meaning of the place-name and may be indicative of his interactions with locals in Wales or with Welsh monks at Shrewsbury. This explanation of the well-name is without equivalent in the anonymous

454 

D. Callander

Life and Lansdowne Life, although they do include the Welsh name Sechnant (Anonymous Life §7; Lansdowne Life §1) for the dry valley, which Robert includes only in Latin as Siccaualis. Robert’s statement is kept largely unchanged in the abbreviated version in Trinity O.4.42. John of Tynemouth retains the Latin name, and adds a reference to Basingwerk, but leaves out the Welsh: Locus vero vbi sanguis illius effusus est, non longe a monasterio nomine Basingwerk in Norwallia, primitus ‘Sicca uallis’ dicebatur; sed post eius mortem vsque in hodiernum diem Fons Wenefrede appelatur. ‘Now, the place where her blood was spilled, not far from the monastery called Basingwerk in North Wales, was originally called Dry Valley, but after her death until our days it has been called ‘‘Winefride’s Well”’ (§3; Text and translation from Marzella Forthcoming).

John of Tynemouth’s version of the Life is greatly shortened and clearly he did not consider it essential to retain this particular onomastic detail. In the Middle English texts, the Welsh name is overwhelmingly not retained, broadly following the pattern already seen with Beuno’s epithet. Thus, we find in Mirk’s Festial a fayre welle, a wonderful wel in Audelay’s carol, a fayre welle in the Gilte Legende Life, and a swythe fayr Welle in the Bodley 779 Life. Caxton, similarly to John of Tynemouth, presents the well’s name only in the language in which he is writing, as saint Wenefredes well. References to Wales and Welsh place and personal names are generally reduced in the Middle English texts to varying degrees of extremity. As Gregory (2012: 280) writes, the Bodleian 779 Life of Wenefred alters the names of several principal characters rather severely. Wenefred herself appears as ‘Uonefred’ or ‘Wonefred(e);’ Caradog is unnamed and instead appears variously as ‘the Kinges Sone,’ ‘this fool Child’ and ‘this ȝong Child;’ Eleri is recast as ‘Bulopius;’ and his mother, Tenoi, appears as ‘Eusebie’ and ‘Euzebye.’ Gwytherin has become ‘Veterat’ and ‘Syre Aleyn’ is ‘King of Ingelonde.’ Only ‘Beuno’ appears unchanged. Titled in the manuscript Seint Vonefrede þe Holi Virgine, this Life never mentions Wales or any place therein, and no comments are made upon the origin of Wenefred’s name.

15  Adapting Winefride in Welsh, Latin and English 

455

This is perhaps the most extreme example of the erasure of Welsh material in the Middle English texts, but it is paralleled to a degree by all of them. A reader looking at this text would have no reason to think that Winefride was not simply an English saint and the personal names, with the exception of Beuno, would not associate her with Wales. There is an obvious temptation to view this in the context of medieval England’s political domination of medieval Wales, with the appropriation of Winefride’s cult being one part of this. However, further questions require answers before moving to this conclusion. Do we see here an attempt to make the names less markedly Welsh (and thus more neutral?) so that the good example of the saint may be retained without connecting her with Wales? Or is this an instance of a far broader pattern of removing irrelevant material, not needed in a new English context, which could draw the minds of readers and listeners away from the text’s important lessons? Currently we do not know, but studying the Lives of Welsh saints and their treatment in Middle English could improve our understanding both of such processes of condensation and simplification and of attitudes towards Wales and the Welsh in medieval England. Such a study would need to assess how all Middle English Lives of Welsh saints (those of Teilo and David as well as those of Winefride) engage with and / or remove Welsh detail and proper nouns. This cannot be done effectively in a vacuum and would need to be viewed in the context of the broader treatment of local detail and proper nouns in the collections and by the authors noted. How do the writers of collections like the Gilte Legende treat Latin Lives of saints based in areas of less obvious political relevance to medieval England? Do we see the same type and scale of erasure as with the material concerning Wales, or is there a difference? If there is a difference, what is particular about the treatment of Wales and Welsh names? A further question is what differences can be perceived between changes made in the Middle English adaptations and later Latin collections from England, such as John of Tynemouth’s Sanctilogium. He often omits proper nouns, but in the case of Winefride he does not go as far as some of the Middle English texts: is this part of a wider pattern of more local details being retained in Latin texts from England than Middle English texts? The basic point which can be made currently is that the Middle English Winefride texts tend to erase local

456 

D. Callander

Welsh detail and proper nouns to a greater extent than the Latin or Middle Welsh texts. The precise significance of this erasure is not yet known. It was noted above that the Welsh name for Winefride’s well was overwhelmingly not retained by the Middle English adaptors of Robert’s Life, but there is one counterexample. In Bokenham’s Life, a form of the Welsh term is maintained: Gret wundyr yt was for the place befo Wher this newe plenteuous welle dede sprynge Was wone watyr to han neuere more Lesse gret reyn thedyr ony dede brynge Yet whanne yt ther cam wyth owte taryge Among the stonys yt wanysshd gret & smal Wherfore yt clepyd was the drye wale But nowe chawgyd ys wtowte drede The name therof as the legende doth spelle For in here language ./ Fenwey . Wenefrede Yt ys clepyd & in oure speche the welle So man & beste of euery maladye Wyth deuocyôn souht ys founde remedye (ll. 422–434; Gregory 2012: 416). ‘It was a great marvel, because the place where this new plentiful well sprang up previously did not usually have water, unless great rains brought any there. Yet when it came there, without delay it disappeared among the stones, big and small, for which reason it was called the Dry Valley. But now its name is changed, without doubt, as the saint’s life tells; for in their language it is called Fenwey Wenefrede, and in our speech the welle. Thus for man and beast a remedy of every disease, when sought with devotion, is found’.

We find here the expected reference to this newe plenteuous welle, but a version of the Welsh term Fenwey Wenefrede also occurs at l. 431. The form Fenwey can be explained as developing from Robert’s written form in the following manner: Fennan > Fenuan > Fenwan > Fenway > Fenwey. The letters and could be confused, particularly if the descender of the were unclear, and the ending may show the influence of other place-names ending in -wey, including anglicisations of Welsh -wy

15  Adapting Winefride in Welsh, Latin and English 

457

endings, such as Conwey. As with Robert’s Life, the Latin form Wenefrede is used in the Welsh name, rather than Middle Welsh Gwenfrewy. Bokenham thus retained something similar to Robert’s rendering of the Welsh form and does not seem to want to remove the Welsh elements. It is noteworthy how balanced his treatment is, referring to English and Welsh in a similar manner: in here language / in oure speche. In l. 431–432, the name Wenefrede appears to apply to both the Welsh Fenwey and the English the welle, and English readers can only access the English name Winefride’s well by piecing together the English noun in l. 432 with the use of the saint’s name in the Welsh place-name Fenwey Wenefrede in l. 431. Perhaps Bokenham’s local interest was sparked by his own pilgrimage to Holywell. He shows knowledge of Wales elsewhere in the text, utilising Higden’s Polychronicon and unique information about miracles which he claims to have learned by relacyon of men in that cuntre (l. 513; Gregory 2012: 289–290; Sanok 2018: 101, 106–108). Regardless of the general pattern, Bokenham highlights how the treatment of Wales in Middle English texts is not exclusively concerned with erasure.

3 Naming Winefride The fourth and most detailed instance where Robert engages with the Welsh language is in his attempt to explain Winefride’s name. Its origin remains somewhat uncertain. The form Wenefreda normally used in the Latin texts appears to be a Latinisation of Welsh Gwenfrewy, first attested in writing in the fifteenth century (Winward 1999), and the English forms draw upon the Latin, with the influence of such pre-existing English names as Winefrid. The first element of this is most likely gwen (the feminine form of the Welsh word for ‘white’), but the second element is of unknown meaning and poorly attested elsewhere (Winward 1999: 100–103). Like modern scholars, Robert was aware that there were two elements in her name, and explains how this came to be after Beuno resurrects her: Inde ferunt illius prouintie homines eam Wenefredam fuisse uocitatam, cum antea, ut ipsi asseuerant, Brewa nominata sit. Quod enim ipsi in sua

458 

D. Callander

lingua Wen dicunt, latine candidum uocant. Sicque ex occasione albedinis collum circumdantis, ex re nomen habere dicta est, adiecta particula, quę est Wen, atque eufonie causa pristini nominis duabus litteris transmutatis, compositum obtinere uocabulum quod Wenefreda dicitur. ‘People of that province say that she was called Winefride on account of that [i.e. her scar], when previously, as they assert, she was called Brewa. For what they call Wen in their language, they call in Latin candidus (white). And so, on account of the whiteness encircling her neck [i.e. the scar], she was said to have her name from that, a particle having been added, which is Wen, and, the two letters of her earlier name having changed on account of sounding better (eufonie causa), to have the composite name, which is called Wenefreda’ (§5; text and translation from Callander 2023b).

Robert here shows familiarity with Winefride’s Welsh name Gwenfrewy. As noted above, (g)wen is the feminine form of the Welsh adjective for white, and this is how Robert interprets the first part of her name. What follows is an attempt by Robert to explain the mutations. The Celtic languages are well known for their initial mutations. One instance of where mutation takes place in Welsh is at the start of the second element of compound words. Wen and Brewa formed a compound here and, as the second element of this, Brewa undergoes lenition to Frewa, resulting in the form ‘Gwenfrewa’. How exactly the was pronounced in Wenefreda is unclear. If it was /f/ rather than /v/, perhaps existing English names like Winefrid had already exerted an influence. Nevertheless, what we see here is someone from outside Wales looking at the mutations and trying to make sense of them, arguing that the phonological change takes place because it sounds better. Robert groups this with an unrelated change, namely of /w/ to /d/ in the second element of the compound (on this change, cf. Winward 1999: 102). Both Trinity (§5) and John of Tynemouth (§2) retain this explanation of the name, although Trinity cuts off the final section and thus does not attempt to explain the mutations and how the combination of wen and Brewa could lead to the form Wenefreda.

15  Adapting Winefride in Welsh, Latin and English 

459

A somewhat different explanation of her name is given in the anonymous and Lansdowne Lives (the latter here drawing on the anonymous Life), which add the Latin word so that no explanation of the Welsh is needed. This might indicate that the anonymous Life, despite having been composed in Wales, was aimed at a partly non-Welsh audience. Winefride’s scar is not described as white in the anonymous Life, and thus Robert’s explanation would not work here. Rather, her name is linked to her wisdom: Quanta scientie eloquentia hec uirgo prefulserit non est nostre facultatis edicere, sermones quippe elogii ipsius melle dulciores lacteque nitidiores in sensibus audientium reputabantur. Unde et ab omnibus publice Candida Wenefræda prenominata est, candore enim sophie eloquebatur et confirmiter degebat. ‘It is beyond our ability to tell with what eloquence of knowledge this virgin shone forth, for the words of her speech were considered to be sweeter than honey and whiter than milk in the thoughts of those listening. And for that reason she was named White Winefride by everyone openly, for she spoke with the whiteness of wisdom and she lived steadfastly’ (Anonymous Life §20; text and translation from Callander 2023a; cf. Lansdowne Life §5; Callander 2023d).

Candida Wenefreda is tautological in a Welsh context as gwen already means white, and thus the phrase would mean ‘White White-Brewa’. The forms of proper nouns elsewhere in the text would point towards an author with knowledge of Welsh, but this deployment of Candida may be aimed at a partially non-Welsh audience. It is worth remembering in this context that all the Latin versions of Winefride’s Lives survive only in manuscripts produced outside Wales. The addition of Candida here is much simpler than Robert’s linguistic explanation and would be immediately intelligible to any Latin reader. There is no explanation of Winefride’s name in Buchedd Beuno. Again it is uncertain whether this was present in the lost Latin Life of Beuno and has been removed as unnecessary in the vernacular rendering or was an innovation in Winefride’s Lives. In Buchedd Gwenfrewy (the Welsh

460 

D. Callander

translation of Robert’s Life), Robert’s explanation of the name is retained but greatly simplified: A phobyl y wlad honno a ddywad mae Brewy oedd i henw hi. Yna, ac o achaws yr edau wenn oedd ynghylch y mynwgyl, y gelwid hi o hynny allan Gwenn Vrewy (§13; Cartwright Forthcoming-a). ‘And the people of that land said that her name was Brewy. Then, and on account of the white thread which was around her neck, she was called thenceforth Gwen Frewy’ (my translation).

We can see how the vernacular context has affected the explanation here. There is no need to explain what gwenn means, as it is known to mean ‘white’, and thus her name clearly derives from the edau wenn (‘white thread’) around her neck. The translator does not need to explain the mutations and there is also far less distance to go in moving from gwen+Brewy to the Welsh Gwenn Vrewy than from wen+Brewa to the Latin Wenefreda. In most of the Middle English texts, Winefride’s name is unexplained. Audelay’s carol and salutacio, the Bodley 779 Life and Bokenham describe the scar as white, but do not link this to her name. The Gilte Legende Life and Caxton’s pamphlet Life present her scar as red, and again do not link it to her name. However, the Latin sequence Gaude virgo Wenefreda in Caxton’s pamphlet does appear to link Winefride’s name with the scar: Gaude quia collum cingit ictus ensis qui non fingit sed te probat martirem. Gaude Wenefrede [sic] hinc est dicta, martir constans et non ficta, seruans hunc carecterem (Horstmann 1880: 314–315). ‘Rejoice because the sword’s wound surrounds the neck, which is not feigned but proves you a martyr. Rejoice, Winefride was named because of this, a steadfast martyr and not contrived, preserving this distinguishing mark’ (my translation).

The sequence claims that Winefride was so named because of her scar but without explanation. Including a detailed etymological explanation like

15  Adapting Winefride in Welsh, Latin and English 

461

Robert’s would be incongruous with genre here, but perhaps this also indicates that the intended audience of this sequence was already familiar with the interpretation of Winefride’s name found in Robert’s Life. This would not be unexpected if this sequence derived from the liturgy in use at Shrewsbury, as the subsequent reference post translata ad vrbem Salopie would suggest. Interestingly, this etymology contradicts the red scar Winefride is said to have in Caxton’s Life, and illustrates how the sequence predates Caxton’s work and is drawing on Robert’s Latin Life rather than this Middle English translation. The only Middle English Winefride text to offer an explanation of Winefride’s name is the earliest, that is, Mirk’s Festial. This adapts Robert of Shrewsbury’s etymology and does so in such a way as to blend Welsh and English: eure whil scheo lyvyd, þer was a whyte serkul aboute hur nek þeras þe strok was, leke a whyte threde. Wherefore þereas scheo was before called Brewa, fro þis day forth men called hyr Wenefrede, þat is in englys a whyte threde (ll. 63–66; Powell 2009–2011, I: 163). ‘ever while she lived, there was a white circle around her neck where the blow had been, like a white thread. For which reason, whereas she was previously called Brewa, from this day forth people called her Wenefrede, that is in English a whyte threde’.

Mirk uses Robert’s explanation to translate wen to ‘white’, although the presence of another language is implicit rather than explicit. Mirk makes no attempt to explain the mutations. He may have been unconvinced by or could not understand Robert’s explanation on the basis of euphonia. He does not present a connection between the second element frede and her previous name Brewa. Instead he links frede to the more similar English word threde. Th-fronting in English appears to be an eighteenth-­century innovation (Beal 2012: 71) and therefore the initial consonants here must have been distinct from one another. Nevertheless, they are sufficiently similar for this to have offered writers like Mirk a tempting explanation for the name. Indeed, the monosyllabic frede in Middle English is closer to threde than the Latin freda would have been,

462 

D. Callander

the only difference being the initial consonants. Mirk apparently has no problem with Winefride’s name containing an English (or similar-toEnglish) element, although it is noteworthy that he describes frede > threde as an English translation. Winefride is presented as having lived in Wales in the sermon (although, compared to Robert’s Life, this is greatly reduced and confined to the story of the Translation in ll. 90 and 147), and the presence of another language is not made explicit. Thus, having a semi-­English element in her name may be more easily explicable. Of the English texts, only Mirk offers an actual explanation of the name, yet every version of the Latin Lives maintains some form of explanation, perhaps reflecting a greater tendency in medieval Latin writing towards learned etymology.

4 Conclusion Winefride’s Lives provide an opportunity to look at how Latin hagiography was adapted in a range of different contexts and areas by writers in Latin, Middle Welsh, and Middle English. Winefride is exceptional for a Welsh saint in the amount of material surviving in Middle English as well as Middle Welsh and Latin, reflecting the strength of her cult in both England and Wales in the fifteenth century. She is nevertheless not unique in having a Middle English Life. The Latin Lives of the Welsh saints are border-crossing texts which often survive in manuscripts in England and could be translated into Middle English. Notably, Winefride’s Middle Welsh and Middle English traditions do not appear to have much, if any, direct influence on one another. Presumably there was great mutual influence orally in bilingual areas like Oswestry hundred, but with the Winefride texts this is not reflected by bilingual Welsh-English texts or by substantial direct Welsh or English influence on one another. Nevertheless, the Latin Lives of the Welsh saints can engage richly with the Welsh language, as Robert of Shrewsbury’s Life testifies, and the adaptations of these in Welsh and English can give us an indication of differing priorities. In the majority of the Middle English texts, the clear pattern is erasure of the Welsh language, but there is also some creative engagement and adaptation, as highlighted by Mirk’s ingenious translation of

15  Adapting Winefride in Welsh, Latin and English 

463

Winefride’s name or Bokenham’s treatment of Ffynnon Wenfrewy. In scholarship on Welsh with reference to the multilingual context of medieval English, the focus has understandably been on the Marches and areas of direct linguistic contact (cf. Lloyd-Morgan 2015; Flood 2016). Yet this study highlights that there are also other paths we might pursue: a twelfth-­ century monk of Shrewsbury who wrote about a Welsh saint in Latin took an interest in the different language spoken in the land Winefride came from. Later, a fifteenth-century monk living in Suffolk, Osbern Bokenham, read this Latin Life and also, apparently, maintained a degree of interest in the Welsh language in composing his Middle English adaptation. The multilingual context of medieval English is not always where we expect to find it.

References Baring-Gould, S., and J. Fisher. 1907–13. The Lives of the British Saints. 4 vols. London: The Honourable Society of Cymmrodorion. Bartrum, P.C., ed. 1966. Early Welsh Genealogical Tracts. Cardiff: University of Wales Press. Beal, Joan C. 2012. Evidence from Sources after 1500. In The Oxford Handbook of the History of English, ed. T. Nevalainen and E.C. Traugott, 63–78. Oxford: Oxford University Press. Byrne, Aisling. 2020. From Hólar to Lisbon: Middle English Literature in Medieval Translation, c.1286–c.1550. Review of English Studies 71: 433–459. Callander, David. 2019. Review of P.  Sims-Williams, Buchedd Beuno. Studia Celtica Fennica 16: 136–138. ———. 2020. Y seintiau a thraddodiad llenyddol: Achos y canu i Wenfrewy. Studia Celtica 45: 99–114. ———, ed. 2023a. Vita Sanctae Wenefredae (Anonymous). In Vitae Sanctorum Cambriae Project, ed. Paul Russell, David Parsons, and Rosalind Love. https:// www.welshsaints.ac.uk/. Accessed 25 May 2023. ———, ed. 2023b. Vita Sanctae Wenefredae (Robert of Shrewsbury). In Vitae Sanctorum Cambriae Project, ed. Paul Russell, David Parsons, and Rosalind Love. https://www.welshsaints.ac.uk/. Accessed 25 May 2023. ———, ed. 2023c. Vita Sanctae Wenefredae (Robert of Shrewsbury: Trinity Recension). In Vitae Sanctorum Cambriae Project, ed. Paul Russell, David

464 

D. Callander

Parsons, and Rosalind Love. https://www.welshsaints.ac.uk/. Accessed 25 May 2023. ———, ed. 2023d. Vita Sanctae Wenefredae (Lansdowne Recension). In Vitae Sanctorum Cambriae Project, ed. Paul Russell, David Parsons, and Rosalind Love. https://www.welshsaints.ac.uk/. Accessed 25 May 2023. Cartwright, Jane, ed. Forthcoming-a. Buchedd Gwenfrewy. In Prosiect Cwlt y Seintiau yng Nghymru / The Cult of Saints in Wales Project, ed. David Parsons, Jane Cartwright, Ann Parry Owen, and Barry Lewis. https://www.welshsaints.ac.uk/. Accessed 25 May 2023. ———, ed. Forthcoming-b. The Welsh Versions of the Life of Gwenfrewy. In Seintiau Cymru, Sancti Cambrenses: Astudiaethau ar Seintiau Cymru / Studies in the Saints of Wales, ed. David Parsons and Paul Russell. Aberystwyth: CAWCS. Dumville, David N. 1977. Sub-Roman Britain: History and Legend. History 62: 173–192. Ellis, F.S., ed. 1900. The Golden Legend or Lives of the Saints as Englished by William Caxton. 7 vols. London: J. M. Dent. Evans, D. Simon, ed. 1959. Buchedd Dewi. Caerdydd: Gwasg Prifysgol Cymru. Evans, J. Wyn, and Jonathan M. Wooding, eds. 2007. St David of Wales: Cult, Church and Nation. Woodbridge: Boydell Press. Fein, Susanna Greer, ed. 2009. John the Blind Audelay: Poems and Carols (Oxford, Bodleian Library MS Douce 302). Kalamazoo, MI: Medieval Institute Publications. Flood, Victoria. 2016. Prophecy, Politics and Place in Medieval England: From Geoffrey of Monmouth to Thomas of Erceldoune. Cambridge: D. S. Brewer. Gregory, James Ryan. 2012. A Welsh Saint in England: Translation, Orality, and National Identity in the Cult of St. Gwenfrewy, 1138–1512. Unpublished doctoral thesis, University of Georgia. ———. 2016. The Life of St. Winifred: The Vita S.  Wenefrede from BL Lansdowne MS 436. Medieval Feminist Forum Subsidia 4: 1–57. Hamer, R., and V. Russell, eds. 2000. Supplementary Lives in Some Manuscripts of the Gilte Legende, Early English Text Society, o.s. 315. Oxford: Oxford University Press. Harper, Sally. 2007. Music in Welsh Culture Before 1650: A Study of the Principal Sources. London: Routledge. ———. 2010. Traces of Lost Late Medieval Offices? The Sanctilogium Angliae, Walliae, Scotiae, et Hiberniae of John of Tynemouth (fl. 1350). In Essays on the History of English Music in Honour of John Caldwell: Sources, Style,

15  Adapting Winefride in Welsh, Latin and English 

465

Performance, Historiography, ed. E. Hornby and D. Maw, 1–21. Woodbridge: Boydell Press. Horobin, Simon. 2005. ‘The Angle of Oblivioun’: A Lost Medieval Manuscript Discovered in Walter Scott’s Collection. Times Literary Supplement, November 11: 12–13. ———. 2007. A Manuscript Found in the Library of Abbotsford House and the Lost Legendary of Osbern Bokenham. In English Manuscript Studies, 1100–1700, ed. A.S.G. Edwards, 132–164. Oxford: Blackwell. ———, ed. 2020. Bokenham’s Lives of Saints, I, Early English Text Society, o.s. 356. Oxford: Oxford University Press. Horstmann, Carl. 1880. Prosalegenden. Anglia 3: 293–360. ———. 1889. Des ms. Bodl. 779 jüngere Zusatzlegenden zur südlichen Legendensammlung. Archiv für das Studium der neueren Sprachen und Literaturen 82: 307–422. ———. 1901. Nova Legenda Anglie. 2 vols. Oxford: Clarendon Press. James, J.W., ed. 1967. Rhigyfarch’s Life of St. David. Cardiff: University of Wales Press. Kooper, E., and D. Callander. 2016. The Middle English Life of St Teilo. The Mediaeval Journal 6: 29–72. Lewis, Barry. 2020. Bonedd y Saint, Brenhinedd y Saesson, and Historical Scholarship at Valle Crucis Abbey. In The Chronicles of Medieval Wales and the March: New Contexts, Studies, and Texts, ed. B. Guy, G. Henley, O.W. Jones, and R. Thomas, 139–154. Turnhout: Brepols. Lewis, Henry, ed. 1925. Darn o’r Ffestifal, Supplement to the Transactions of the Honourable Society of Cymmrodorion, Session 1923–24. London: The Honourable Society of Cymmrodorion. Lloyd-Morgan, Ceridwen. 2015. Writing without Borders: Multilingual Content in Welsh Miscellanies from Wales, the Marches, and Beyond. In Insular Books: Vernacular Manuscript Miscellanies in Late Medieval Britain, ed. M. Connolly and R.L. Radulescu, 175–192. Oxford: Oxford University Press. Lowry, M.J.C. 1983. Caxton, St Winifred and the Lady Margaret Beaufort. The Library, 6th ser. 5: 101–117. Marzella, Francesco, ed. Forthcoming. Vita Sanctae Wenefredae (John of Tynemouth). In Vitae Sanctorum Cambriae Project, ed. Paul Russell, David Parsons, and Rosalind Love. https://www.welshsaints.ac.uk/. Accessed 25 May 2023. Pepin, Ronald, and Hugh Feiss, trans. 2000. Two Mediaeval Lives of Saint Winefride. Toronto: Peregrina Publishing Co.

466 

D. Callander

Powell, S., ed. 2009–2011. John Mirk’s Festial: Edited from the British Library MS Cotton Claudius A.II, Early English Text Society, o.s. 335. 2 vols. Oxford: Oxford University Press. Roberts, Thomas, ed. 1958. Gwaith Tudur Penllyn ac Ieuan ap Tudur Penllyn. Caerdydd: Gwasg Prifysgol Cymru. Sanok, C. 2018. New Legends of England: Forms of Community in Late Medieval Saints’ Lives. Philadelphia: University of Pennsylvania Press. Sharpe, Richard. 2007. Which Text is Rhygyfarch’s Life of St David? In St David of Wales: Cult, Church and Nation, ed. J.  Wyn Evans and Jonathan M. Wooding, 90–105. Woodbridge: Boydell Press. Sims-Williams, Patrick, ed. 2018. Buchedd Beuno: The Middle Welsh Life of St Beuno. Dublin: Dublin Institute for Advanced Studies. Sutton, Anne F. 2005. Caxton, the Cult of St. Winifred, and Shrewsbury. In The Fifteenth Century, V: Of Mice and Men: Image, Belief and Regulation in Late Medieval England, ed. Linda Clark, 109–126. Woodbridge: Boydell Press. Wade-Evans, A.W., ed. 1944. Vitae Sanctorum Britanniae et Genealogiae. Cardiff: University of Wales Press. Whiting, E.K., ed. 1931. The Poems of John Audelay, Early English Text Society, o.s. 184. London: Oxford University Press. Winward, Fiona. 1999. The Lives of St Wenefred (BHL 8847–8851). Analecta Bollandiana 117: 89–132.

16 Let Each One Tell its Own Story: Language Mixing in Four Copies of Amore Langueo Mareike Keller and Annina Seiler

1 Introduction The Lenten sermon Amore langueo (‘I languish with love’) is a bilingual, or ‘macaronic’ sermon;1 its base language Latin is interspersed with English insertions of varying size, from single words, through phrases and We would like to thank Sara Pons-Sanz and Louise Sylvester for their support as well as for getting us to join forces for this project in the first place. We are grateful to Marigold Norbye and Orietta Da Rold for their generous help with palaeographical details. Our sincere thanks also go to Daryl Green, Anne Chesher and David Rundle (Magdalen College); Amy Boylan (Balliol College); and Frank Bowles (Cambridge University Library) for hosting visits, assisting us with their expertise while inspecting the manuscripts and allowing us to take digital images; and to the imaging service of Trinity College Dublin for supplying us with high-quality images.  Macaronic literature, especially macaronic poetry, is characterised by conscious language mixing on the word level which is used to parody the morphology of classical Latin (see e.g. Demo 2014, 2018: 34; Schendl 2015: 236). Even though this is not the kind of language mixing found in Amore langueo, we have decided to use the term alongside code-switching and language mixing, as it is the most widely known label for the text type under investigation. 1

M. Keller (*) University of Mannheim, Mannheim, Germany e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_16

467

468 

M. Keller and A. Seiler

clauses, to entire English prayer verses. Some of the English elements are syntactically integrated into Latin clauses. Such intra-sentential code-­ switching is formally and functionally different from inter-sentential code-switching as found, for example, in London, British Library, MS Harley 2268 (O’Mara 2002), where Latin Bible quotes occur in English-­ language sermons without any syntactic integration. In contrast to heavily mixed sermons like the ones in Oxford, Bodleian Library, MS Bodley 649 (Horner 2006), Amore langueo has one clear base language, Latin, throughout the entire sermon. Hence, the focus of our study is not language mixing per se but the forms and functions of syntactically independent and syntactically integrated vernacular insertions in a Latin text. Amore langueo survives in four manuscripts dating from the fourteenth and fifteenth centuries. Despite significant differences in the Latin parts of the text, the copies show a surprising overlap regarding the English insertions, suggesting that these insertions served a function that was to be kept intact despite other adaptations. In the following pages, we want to show what the four copies can tell us about the purpose of the bilingual structure and about the possibility of actual bilingual performance. To our knowledge the issue has not been discussed on the basis of different copies of the same bilingual text, and our contribution aims to address this gap, focusing on the functions of the vernacular insertions and on structural differences and similarities concerning language distribution, including aspects of mise en page. The questions guiding our analysis are: (a) What is the function of the different types of English insertions in the sermon? (b) What can the differences between copies concerning English insertions tell us about the reception of the different copies? In Sect. 2, we discuss preaching in medieval England and the roles of English and Latin in the religious domain. In Sect. 3, we introduce the sermon and its transmission. Section 4 contains the analysis of our data, beginning with the rhetorical macrostructure of the sermon and then zooming in on the microstructure of intrasentential code-switches and

A. Seiler University of Zurich, Zürich, Switzerland e-mail: [email protected]

16  Let Each One Tell its Own Story 

469

related language contact phenomena. In Sect. 5, we interpret our findings with regard to the broader question of how we can determine the mode in which a bilingual sermon was presented and / or consumed.

2 Preaching in Medieval England Preaching must have been a very common occurrence in late medieval England, taking place on numerous occasions in parish churches, at the universities, in monastic contexts, as well as in other settings.2 Sermons are essentially a multilingual genre: traditionally, Latin was the established language of the religious domain, but in order to reach a lay congregation, the vernacular was needed, especially for the sermon, which is the part of the mass that is—ideally—intended for teaching the congregation about Christian virtues.3 Medieval Latin was open to innovations (Latin derivations and compounds to denote new professions, objects, etc.) and loanwords (from Greek, Hebrew, Arabic, local vernaculars). Thus, the co-occurrence of words and expressions from several languages in one text was nothing exceptional per se. Multilingual sermons, which contain more language contact phenomena than the occasional use of loanwords, just take this general openness to language mixing one step further. While many written sermons from the period survive, there is little concrete evidence on how sermons were actually delivered. Or, as Wenzel puts it, ‘we can only guess at what was actually said in the pulpit and heard by the congregation’ (Wenzel 2005: 16).4 This uncertainty also applies to the use of languages in preaching. Generally, Latin was used for preaching to clerical audiences while English was reserved for lay audiences, though, according to Wenzel (2005: 10), this distinction may have broken down by the fifteenth century. Yet sermons—at least the orthodox ones—were usually written down in Latin even when they were preached in the vernacular. One famous description of language mixing  On preaching in medieval England, see Owst (1926) and Wenzel (1994: 105–129; 2005: 227–310). 3  On multilingualism in the religious domain, see Pahta and Nurmi (2011). 4  For theoretical and methodological issues concerning multilingual medieval sermons, see Odstrčilík (2020). 2

470 

M. Keller and A. Seiler

in a medieval sermon is provided by Chaucer’s Pardoner in the Canterbury Tales. In his Prologue, the Pardoner describes how he spikes his preaching with Latin expressions to stir people to devotion (cf. Fletcher 2013: 139): And in Latyn I speke a wordes fewe, / To saffron with my predicacioun / And for to stire hem to devocioun (Benson 2008: 194, ll. 344–346). The Pardoner himself provides one such example, when he cites the theme of all his sermons, Radix malorum est Cupiditas (l. 334: ‘Greed is the root of all evil’), a quotation from I Timothy 6.10. This use of Latin for adding authority to a statement in an otherwise vernacular text is well attested in sermons (e.g. sermons in MS Harley 2268) and in a wide range of other medieval text genres (e.g. Schendl 2015: 238, 241, on medieval English poetry and drama; Hunt 2000 on medical texts). In Amore langueo, however, as in other surviving sermons (see handlist in Wenzel 1994), the overall distribution of English and Latin is reversed: the sermon is written in Latin and features different types of English insertions. As the use of Middle English in a Latin context certainly cannot be assumed to convey authority, this raises interesting questions concerning the functions of code-switches into the vernacular. It has been a matter of some debate whether such ‘macaronic’ sermons were actually preached in a bilingual way. Whereas Wenzel (1994) and Johnson (2012) conclude that this may well have been the case, Fletcher (2013) maintains that this type of language mixing in sermons is more likely a written phenomenon (see also Schendl 2013: 167). In his view, it is unlikely ‘that medieval preachers would have introduced willingly a possible stumbling block in the way of comprehension into their form of delivery’ (Fletcher 2013: 142). Fletcher (2013: 147) argues that the written bilingualism is owed to the fact that the author of a sermon had to be ‘linguistically bifocal’ in as far as he was working on a text intended for oral delivery in English but, at the same time, producing a Latin written text that could be read by literate audiences. While the question cannot be answered generally for all existing mixed-language sermons, Amore langueo might well have been intended for bilingual preaching, as it contains several features suggestive of conceptual orality.5 On the one hand,  On ‘orality in text’, see, for example, Oesterreicher (1997) and Schaefer (2012).

5

16  Let Each One Tell its Own Story 

471

the function of a large majority of code-switches is primarily rhetorical and thus forms an intrinsic part of the artful composition of the sermon. On the other hand, the structural characteristics of the mixed clauses conform to patterns commonly found in fluent oral code-switching (Keller 2020: 79–83). Moreover, Amore langueo includes other features which Wenzel considers ‘marks of orality’ (2005: 19), such as addresses to the audience or the use of inclusive we. There are also two instances in which the preacher remarks that he is running out of time and needs to move on or come to a conclusion.6 To complement existing research on the potential oral delivery of mixed-language sermons, in this contribution we focus on the parallel transmission of a single sermon in order to investigate whether the differences between copies provide any evidence on the intended ‘consumption’ of the sermon. To set the scene for the analysis, we now give a brief overview of the sermon’s structure and content, followed by information on the existing copies.

3 Amore Langueo Amore langueo is a sermon for Good Friday based on an excerpt from Canticles 2.5: fulcite me floribus stipate me malis quia amore langueo (‘stay me up with flowers, compass me about with apples: because I languish with love’; Vulgate with Douay-Rheims translation). As a scholastic sermon, it does not provide an exegesis of the entire Bible passage but only uses the expression amore langueo as a theme (thema), which is developed throughout the sermon into a division with further subdivisions according to a fixed scheme.7 Based on a pseudo-etymological interpretation of the different syllables of the theme, the sermon introduces Good Friday as both a sorrowful and a blissful day: it is sorrowful because Jesus died on the cross but blissful since Jesus redeemed humankind. Subsequently, seven signs of love and mourning from different Bible passages are  W644 and W761; references are to line numbers in Wenzel’s edition (1994; = W).  For an overview of the structure of Amore langueo, see Wenzel (1994: 213); for a discussion of the form of sermons, see Wenzel (2005: 11–16); for a more detailed summary and interpretation of the sermon, see Johnson (2012: 61–66). 6 7

472 

M. Keller and A. Seiler

identified (protheme). This is followed by the introduction, a narrative of a knight and his unfaithful lady, who later repents of her sins and gives her husband a golden truelove—the plant Paris quadrifolia. An allegorical reading of the story identifies the knight as Christ and equates the lady with the renegade believer who repents and returns to faith. The sermon moves on to the division, which introduces the seven signs of languishing love. In the development, the sermon branches out into a series of subdivisions: each sign—loss of strength, sighing, solitude, listening for one’s beloved, speaking lovingly to one’s beloved, suffering bitter pain, giving gifts—is ‘proved’ (introduced with hoc probant) with another set of seven items from the Bible, for example the seven torments of Christ. Each elaboration of a sign of languishing love ends with a quote from a psalm, which is then paraphrased as a short verse in English. As a last step, each sign is taken up once more in the context of a Bible story (figura), also ending with an English prayer verse. The development, with its clear structure and repetitive pattern, is the core of the scholastic sermon. In Amore langueo, the beginning is elaborated in detail. Towards the end the Latin parts are noticeably shorter and less systematically developed, and the frequency of English insertions drops significantly. We attribute this to the final parts of the sermon division not being written out in full, rather than to an intentional change in language distribution in favour of Latin. The sermon topic was fairly popular at the time, but we have no record of who wrote the original sermon and can only guess where, by whom and for what reason it was copied. Amore langueo is transmitted in four manuscripts dating from the late fourteenth to the mid-fifteenth century: • Oxford, Balliol College, MS 149, fols 31r–38v, s. xiv ex. (= Bal.)8 • Cambridge, University Library, MS Kk.4.24, fols 143v–150r, s. xv in. (= Cam.)9  For a description of Bal., see Mynors (1963: 130–135) and Wenzel (1994: 177–180; 2005: 121–124). Digital images are available at https://www.flickr.com/photos/balliolarchivist/ sets/72157627107696303/ [accessed on 25 May 2023]. 9  For a description of Cam., see Wenzel (1994: 140–156; 2005: 136–139).

8

16  Let Each One Tell its Own Story 

473

• Oxford, Magdalen College, MS lat. 93, fols 152r–157r, s. xv2/4 (= Magd.)10 • Dublin, Trinity College, MS 277, pp. 185–198, s. xv med. (= Trin.).11 None of the copies appears to be an autograph, nor is any copy clearly dependent on any of the others. The copies are transmitted in different textual environments: in a collection of primarily Latin popular sermons presumably originating from Oxford (Bal.); in a random collection of sermons, which includes references to villages where the sermons were preached (Cam.); in a sermon notebook from the Carthusian John Dygon, who studied in Oxford (Magd.); and in a miscellany of thematically grouped religious texts (Trin.).12 Neither the text of Amore langueo nor any of the manuscripts provide any clear evidence concerning the intended audience of the sermon (Wenzel 1994: 32). The different contexts suggest that it may have been preached to clerical as well as lay—or mixed—audiences. The composition of at least three of the manuscripts strengthens the impression that Amore langueo is a ‘real’ sermon intended for oral delivery (Wenzel 2005: 3). Cam. in particular, in which the marginalia preserve a record of actual preaching, points in that direction. The odd one out is Trin., whose composition suggests that the manuscript was intended for private study. As we will argue, this interpretation is borne out by its treatment of code-switching. Wenzel (1994: 212–267) provides a semi-critical edition of Amore langueo based on Bal. with variants in the English material in the apparatus. In the edition, English elements are printed in bold, though the two languages are not distinguished visually in the manuscripts.13 The English  For a description of Magd., see Wenzel (1994: 180–182; 1996: 16–17), Hanna (2003), Hanna and Rundle (in preparation) and Coxe (1973: ii, 49–51). 11  For a description of Trin., see Wenzel (1994: 156–157; 1996: 19–21), Colker (1991–2008: i, 509–532) and Scattergood et al. (2021). 12  For a classification of the different types of sermon collections, see Wenzel (1996). 13  Occasionally, the word Anglice or a symbol resembling the @ sign is inserted at the beginning of Middle English elements. Some medieval English texts use rubrication to highlight code-switching, for example Piers Plowman (Machan 2012; Halmari 2021; see also Putter, Kopaczyk and Bridges, this volume, esp. Sect. 4.2); the use of different scripts for different languages is common in early medieval English sources but does not occur in later medieval English texts (see Skaffari 2018: 128–129). 10

474 

M. Keller and A. Seiler

verses included in Amore langueo have also been edited separately (Digital Index of Middle English Verse; Wenzel 1986; Scattergood et al. 2021, verses in Trin.) and discussed in several publications (e.g. Butterfield 2013: 457–462; Killian 2019).

4 Analysis For our analysis of Amore langueo, we have used Wenzel’s (1994) text and apparatus as a starting point; all vernacular elements have been collated from digital images of the four manuscripts. Examples are cited directly from the manuscripts with abbreviations expanded in square brackets.14 The copies mostly agree in content and wording, though there are some differences in both the Latin and Middle English parts. The variation in the Latin text concerns spelling variants, the use of tense forms and word order preferences, and intentional omission / reduction of text. Typical abbreviations are used to varying degrees, mostly for the Latin material, but occasionally also for English. Many English insertions occur in at least three of the four copies, but the size of the insertions and the exact lexical content vary. Overall, syntactically independent English insertions, which are almost exclusively verses, appear consistently across copies. Syntactically dependent English units are well integrated into the structure of the clause in which they occur, and most of them occur in the same places in different copies, even though their exact form may vary. Trin. differs significantly from the other three copies in that it replaces many of the vernacular insertions by Latin translation equivalents in the running text and supplies the English material in the form of interlinear glosses. Overall, the sermon (as attested in the Bal. manuscript) consists of approximately 7500 orthographic words. Of these, some 580 are English (about 400 in verses and other syntactically independent structures,  Expuncted or crossed-out letters are marked as such, English items are underlined. Glosses are indicated by slashes. Punctuation is gently normalised. References are given to folio or page numbers; cross-references to Wenzel (1994) are also provided. Translations are by Wenzel unless indicated otherwise. As Bal. can be viewed online (see Footnote 8), we have made it the primary reference point for most of our examples, inviting readers to look at details for themselves. 14

475

16  Let Each One Tell its Own Story 

about 180 in mixed clauses), including a handful of ambiguous tokens (e.g. proper names, Tironian ‘and’) at switch sites. The proportion of vernacular elements (7.5 percent) is comparatively low. Wenzel (1994: 81) reports figures between 2.28 and 33.36 for the forty-three macaronic sermons analysed in his study. The English elements constitute some seventy switches (cf. Wenzel 1994: 347). The English insertions in Amore langueo fall into a number of categories regarding their size and how they are integrated into the Latin text, from (a) verses that constitute their own little texts, via (b) clauses and (c) phrases, to (d) single-word insertions. The overview in Table 16.1 is based on Bal., but the same types also occur in all other copies.15 Twenty of the Middle English insertions in Amore langueo are syntactically not integrated (Table 16.1, (a)). One of these switches consists of a list of terms (the seven torments, (5) below); one switch is a single sentence (2); the remaining eighteen are (groups of ) rhyming couplets (1). The distribution and function of these verses will be discussed below. We have counted complex sentences in which the main clause is Latin and Table 16.1  Overview of code-switching in Amore langueo according to linguistic level of English insertion Linguistic level Example

No. of switches

(a) Textual

for loue of l[es]u my swete herte y morne 7 seke 20 wyþ teres smert. (‘For the love of Jesus, my sweet heart, I morne and sigh with painful tears’; Bal. fol. 31v, W73)

(b) Clausal

Secundum signum est quod he syket as a sorful man (‘The second sign is that he sighs like a grieving person’; Bal. fol. 34v, W367)

4

(c) Phrasal

Tercium tormentum fuit gret trauayle (‘The third torment is great physical strain’; Bal. fol. 33v, W260)

27

(d) Lexical

de corde isto p[ro]cedat vnu[m] trewloue (‘from the heart a truelove shall appear’; Bal. fol. 32v, W170)

18

 There are slight discrepancies in the figures between the manuscripts since some of the differences affect the level and number of code-switches (see details below). 15

476 

M. Keller and A. Seiler

the subordinate clause English as switches on a clausal level (Table 16.1, (b)). Such switches are fairly infrequent in the text. Phrase-level insertions in English occur throughout the text, though they are more common in the first half of Amore langueo, which is elaborated in greater detail. Many of the phrasal insertions consist of prepositional phrases (e.g. of gret mornyng, Bal. fol. 32r, W102), noun phrases (Table 16.1, (c)) or coordinated constituents (e.g. Set maledi[i]c[t]i optulebant ei aisell and galle, ‘His cursed tormentors offered him vinegar and gall’; Magd. fol. 155v, W692). In examples with coordinated constituents, there is variation across the manuscripts as to how they render the conjunction and, which is sometimes spelled out but sometimes abbreviated with Tironian ‘and’ (⁊). Such abbreviation symbols often function as ‘visual diamorphs’, a term that describes items that potentially belong to more than one language.16 In this sense, coordinated constituents can also be interpreted as separate single-word switches. Single-word borrowings17 from English are all lexical words from one of the main word classes, that is, adjectives (thrall ‘enslaved’, free, sorfful, hungry, sad ‘firm’), nouns and nominalisations (trelow ‘Paris quadrifolium’ [2x]; actoun ‘padded jacket’; hauberk ‘chain mail’; babel ‘a scourge with spiked balls on thongs’; abobete ‘a game; probably blind man’s buff’; cornwaxand ‘growing of grain’; drury ‘beloved’) and verbs (to-borstoun ‘shattered’ [2x], to-cleuen ‘split’ [2x], blowen). We begin our analysis with the English verses and related elements (Sect. 4.1). Some of these are complemented in the sermon by smaller insertions, which show strong formal and functional overlaps with the verse passages (seven signs of love, seven torments of Christ) and form part of the rhetorical backbone of the sermon. Then we move on to English insertions that are motivated less by the rhetorical structure of the sermon than by lexical need (Sect. 4.2). Finally, we point out structures that contain clues to the competition of languages during the copying process (Sect. 4.3).

 The term was coined by Wright (2011: 203) in analogy to the concept of ‘homophonous diamorph’, which describes elements with ambiguous language affiliation at switch sites in oral code-­ switching (cf. Muysken 2000: 133). 17  See Jake et al. (2002) for arguments against and Poplack (2018: 141–157) for arguments in favour of a distinction between code-switching and borrowing. 16

16  Let Each One Tell its Own Story 

477

4.1 English Verses and Related Elements The most systematically vernacular elements in Amore langueo are a number of short pieces of verse. They are completely monolingual English in themselves and thus not to be confused with medieval macaronic poetry (see Footnote 2). Vernacular verses like these are a very common element in sermons of the time (Butterfield 2013: 455). They constitute their own little texts within the sermons and are thus structurally distinct from vernacular elements that are syntactically integrated. Despite being syntactically independent, the verses are well integrated into the sermon in terms of content. Like all other vernacular insertions, they are not distinguished visually from their surrounding Latin text in any of the copies. The most prominent of the verse pieces consist of pairs of corresponding lyrics, the first spoken by Jesus, as it were, and the second as a prayer to be spoken by the congregation, as in (1): (1) nam [Christus] pot[er]it hodie veraciter di[cer]e generi hum[an]o amore langueo q.d. y morne for loue þu may see þ[a]t makide me deye for þe ista ecia[m] verba p[oss]u[n]t e[ss]e verba generis hu[ma]ni ad [Christu]m sponsum suum quia quilibet n[ost]r[u]m tenet[ur] hodie dicere [Christ]o; amore langueo, q.d. for loue of ih[es]u my swete herte y morne ⁊ seke wyþ teres smert, nam ista […] ‘For Christ could truly say to mankind today “I languish with love,” as if he were saying: I mourn for love, as thou canst see, That made me die because of thee. And these words can be considered as being spoken by mankind to Christ, its spouse, for each one of us is held to say to Christ today, “I languish with love,” as if to say: For Jesus’ love, my sweetheart dear, I mourn and sigh with painful tear’ (Bal. fol. 31v; cf. Cam. fol. 144rb, Magd. fol. 152v, Trin. p. 186; W66). As the sermon explains, these two verse pieces paraphrase the sermon theme ‘I languish with love’ from two different perspectives. In both instances, the language switch is flagged by the abbreviation q.d. (or quasi diceret, as written out in one instance in Magd.). As Killian (2019: 216–223) has convincingly argued, the change of language here signals a

478 

M. Keller and A. Seiler

change of speaker voice. The words outlined for Jesus and the congregation cited above expand on the vernacular translation of the theme, which occurs a few lines earlier: (2)  ⁊ ista sunt illa duo verba, que dixi vobis in p[r]i[nc]i[p]io ⁊ sunt scripta, in libro amoris, amore langueo. Angelice Y morne fore loue ‘And these are the two words I told you in the beginning, and they are written in the Book of Love: amore langueo. In English: “I mourn for love”’ (Bal. fol. 31v; cf. Trin. p. 186, omitted in Cam. and Magd.; W64). The code-switched theme is singular insofar as it constitutes the only example of an inter-sentential code-switch that is not expanded into a verse. However, as an excerpt from Canticles, it ties in with other multilingual sections of the sermon in which citations from the psalms or sayings by saints are first cited in Latin and then provided with a verse paraphrase or extended translation in English. Like the prayers, the citations from the psalms and similar elements are introduced by verba dicendi, for example quod ipse dicit per prophetam in Psalmo (‘what he himself said through the prophet in a psalm’; W56), potest Christus […] dicere cum Psalmista (‘Christ may […] say with the psalmist’; W334) and so on. Therefore, Kilian’s reading of the code-switches as changes in speaker voice also applies here. Interestingly, the use of the vernacular to mark Bible quotations mirrors the use of Latin by Chaucer’s Pardoner (or of many religious texts in which the main language is English). A switch into the vernacular may not have lent authority to a quotation in the same way that a switch into Latin does. However, repeating it in both languages would certainly draw attention to it and render it more memorable.18 One central verse that does not fit into the pattern of prayer verse pairs described above contains a description of the seven signs of love (3). It consists of four rhyming couplets in Middle English (though written as  See Butterfield (2013), who points out the aesthetic function of the prayer verses and the affective meaning associated with the vernacular when used for devotional purposes. 18

16  Let Each One Tell its Own Story 

479

running text in all manuscripts), and it marks the beginning of the division of the sermon. At almost fifty words in length, it is the longest passage in the vernacular. Contrary to the prayer verses, this element does not involve a change of voice. However, the passage fulfils a key rhetorical and mnemonic function: it provides a preview of the division and draws attention to the sevenfold structure along which the principal section of the sermon is organised, namely the seven signs by which we recognise languishing love: (3)  ⁊ sunt ista .7. he losys his myght and wexyth al wan, he syketh as a sorwfull man. Allone he drawes fro company and eu[er] herkenes of his drewry. Louely he spekes to his hert, for hym that suffres peynes smert, [Thurgh tokynys gret of gyftes gyfing,]19 y swete yn hert for mornyng. Ista sunt septem signa […] ‘And there are seven: He loses his strength and grows all wan, he sighs like a sorrowful man. Alone he withdraws from company and always listens for his lover. Lovingly he speaks to his heart, for which he suffers bitter pain. Through great tokens of gift giving he showed the mourning of his heart. These are the seven signs […]’ (Our translation; Magd. fol. 153r–v; cf. Bal. fol. 32v–33r, Cam. fol. 145rb, Trin. p. 188; W194). This passage is relevant to many of the shorter, syntactically integrated English elements, since chunks from this verse piece are repeated as intra-­ sentential code-switches later on. On the text level, the division of Amore langueo lays out the seven signs according to a fixed pattern: in each subdivision, the sign is introduced in Latin and in English, and repeated in English at the end. For example, the first sign is that Jesus’s body was weakened. The running text first mentions this in Latin in the preview of the sermon: Nam primo fuit Chr[istus] corp[or]alit[er] extenuatus (‘For first Christ was weakened in his body’; Bal. fol. 31r; W32). The  The line on gift giving is missing in Magd. and Cam.; in Bal., it is added in a marginal gloss. In (3) the reading is supplied from Trin., which omits the verse in the main text but adds it in the top margin of p. 188 (starting at Alone). As Wenzel (1994: 228, n. 64) points out, the distribution of seven signs across eight verses may have confused the scribes. 19

480 

M. Keller and A. Seiler

subdivision ends with a repetition of the sign in English integrated into a Latin clause, which more or less faithfully repeats the phrase from the verse: Primu[m] [er]go signu[m] amoris languentis est q[uod] ille qui sic languet lesis is myth ⁊ waxis wan (‘Thus, the first sign of languishing love is that he who thus languishes loses his strength and fades away’; Bal. fol. 34v; W365). While all four manuscripts preserve this basic structure, there are some differences in how they treat individual switches (see Figs. 16.1, 16.2, 16.3, and 16.4). The code-switch discussed as an example of a switch at the clausal level in Table 16.1, namely, that one sighs like a sorrowful man, is rendered as follows: (4) a. 2m signu[m] est q[uod] he syket as a sorful man (Bal. fol. 34v; W367) b. 2m signu[m] e[st] þ[at] he sychyth as a sory man (Cam. fol. 146vb) c. 2m sig[num] he that sykes a sorowfull man (Magd. fol. 154r)              \A[nglice] he syghys as a sorofull man/ d. S[e]c[un]d[u]m sig[num] e[st] q[uod] ip[se] suspirat quasi ho[mo] doloros[us] (Trin. p. 191).

Fig. 16.1  Bal. fol. 34v (Oxford, Balliol College, MS 149): The code-switch discussed as ex. 4a occurs on the second line after the pilcrow. Reproduced by kind permission of the Master and Fellows of Balliol College

Fig. 16.2  Cam. fol. 146v. (Cambridge, University Library, MS Kk.4.24): Code-switch 4b is located in column b; it starts in the line of the first red pilcrow. Reproduced by kind permission of the Syndics of Cambridge University Library

Fig. 16.3  Magd. fol. 154r (Oxford, Magdalen College, MS lat. 93): Code-switch 4c occurs five lines from the bottom. Reproduced by kind permission of the President and Fellows of Magdalen College, Oxford

Fig. 16.4  Trin. p. 191 (Dublin, Trinity College, MS 277): Ex. 4d is the second interlinear gloss, starting close to the right-hand margin and continuing on the next line. Reproduced by kind permission of the Board of Trinity College Dublin

482 

M. Keller and A. Seiler

Bal. and Cam. render the sentence in the same way, except for one structural detail: the complementiser that introduces the subordinate clause is given in Latin in Bal. (4a) and in English in Cam. (4b). With respect to structural code-switching theory, this is interesting because both variants are observed regularly in modern oral data, and there has been a long-standing theoretical dispute, mainly entertained by proponents of different versions of generative grammar, as to whether the complementiser as a functional head must be realised in the matrix language of the main clause or not (Belazi et al. 1994; Myers-Scotton and Jake 2016). Bal. and Cam. give evidence that both options have co-existed for a long time. The English insertion in Magd. (4c) appears to be syntactically incomplete and does not convey the same sense as the parallel clauses in the other copies. This is not surprising given that the entire manuscript contains a much higher number of copying errors and scribal oversights than the other three copies. Trin. (4d) shows a form of bilingualism that is fundamentally different from what we see in the other manuscripts. The main text is presented entirely in Latin. The part that constitutes an English insertion in the other copies is translated into Latin; however, the English text is reinserted in the shape of an interlinear gloss flagged A(nglice) but using the same wording as the other copies. This is typical of Trin., and Wenzel (1994: 212–213) pointed out that this manuscript ‘frequently renders English words and phrases that are part of the divisions […] in Latin, though here and there this rendition is in turn supplemented with an interlinear gloss in English, which therefore returns to what I take to be the original English’. In fact, Trin. replaces almost twenty of the code-­ switches found in the other manuscripts. In about half of those instances, there is a Middle English gloss, which translates the Latin elements partially or completely. The wording normally corresponds to the code-­ switching in the other manuscripts, though occasionally further synonyms are provided. Conversely, there are four instances in which a code-switch that Trin. has preserved is glossed in Latin, and four code-switches are omitted. The glossing in Trin. reflects a common type of medieval bilingual text, employing a parallel (or paradigmatic) rather than linear (or syntagmatic) combination of languages. The relationship to the other versions

16  Let Each One Tell its Own Story 

483

highlights the connections between the two types of bilingualism. As Trin. is the youngest of the four manuscripts and furthermore deviates most significantly from the others, the linear code-switching is clearly primary in this sermon. This means that we are not dealing with glosses that were integrated into the Latin text in the course of written transmission. Rather, the Trin. copyist removed the code-switching from the main text and placed the vernacular elements between the lines or in the margins. Glossing is, of course, a very common feature of medieval texts and represents an essential source of bilingual material surviving from the Middle Ages. However, glossing is typical of written texts and necessitates reading ‘in stereo’—a term coined by Sarkonak and Hodgson (1993) for the reading process demanded by bilingual parallel texts. Code-switching, on the other hand, is commonly associated with spoken language. The more ‘written’ character of the bilingualism in Trin. ties in with the overall composition of the manuscript: unlike the others, Trin. is not a sermon collection; its make-up implies that it was intended for private devotion. A structuring function similar to the seven signs of languishing love attaches to the following list of seven torments, which form a subdivision that serves as proof of the first sign: (5) Et fuerunt vii. blodwetyng, hard byndyng, gret traualyng, smert betyng, longwakyng, croys beryng, scherp prikyng ‘And there were seven: blood-drenching,20 hard binding, great labour, painful beating, long waking, cross-bearing, sharp pricking’ (Our translation; Bal. fol. 33r; cf. Cam. fol. 145rb–va, Magd. fol. 153v, Trin. p. 188; W211). The list occurs in English in all four copies, though only Trin. flags it as A[nglice]. In the subsequent sections of the sermon, each of the seven torments is discussed in turn. Like the seven signs of languishing love, the  The other manuscripts read swetyng at this point, which corresponds to Latin sudatio ‘sweating’; however, ME weting ‘the act of moistening or wetting’ (Middle English Dictionary, hereafter MED, s.v. wetinge) also makes sense. 20

484 

M. Keller and A. Seiler

torments are reintroduced in Latin (e.g. Primum tormentum fuit sanguinis sudacio, ‘The first torment was sweating of blood’; Bal. fol. 33r, W218) or with a code-switch (e.g. Septi[mu]m torme[nt]u[m] fu[i]t scharp prykki[n]g, ‘The seventh torment was sharp pricking’; Cam. fol. 146rb, W325). In these repetitions, there is some variation in the use of the vernacular across the four manuscripts. Most of the terms are given in Latin; however, Bal. and Magd. use a phrase-level code-switch into English for the third and seventh torment. For the five other torments, which are discussed in the main text only in Latin, Bal. provides bilingual glosses in the margins, which are structured as follows: (6) \4 torme[n]t[um] . dura etc. An[glice] . smart betyng/ ‘4th torment—hard etc. En(glish)—painful beating’ (Bal. fol. 33v; W275). Such marginal elements flag the structure of the subdivision of the sermon. However, numerals and a reference to the Latin term for the torments would have sufficed. The vernacular elements in the Bal. glosses are clearly repetitions of the English phrases from the list of torments at the beginning of the subdivision (5). Why did the scribe repeat the English terms? One interpretation is that the English phrases are, as it were, reminders of code-switches that a preacher might have used during an oral delivery of the sermon, similar to the seven signs. Cam. uses code-switching only for the last torment; it also includes a highly abbreviated gloss in the top margin of fol. 145v, which reads, ­sud. li.la.flag.vi.ba.pric.Mr.Davi, which can be expanded into sudacio, ligacio, laboratio, flagellacio, vigilacio, baiulacio, pricking, Miraculum Danielis VI. Interestingly, the marginal list repeats the language distribution as used in the text and thus also switches to English for the abbreviation of ‘pricking’. This gloss in Cam. is one of many such entries, all located in the top margin. They summarise the elements of the sermon discussed on the respective page and point to the allegorical interpretation as described in the sermon. In this way, they fulfil a function similar to marginal elements in Bal. and Cam. As in other instances, the Trin. copyist substituted Latin for the code-switches that he must have found in his exemplar—even for scharp pryking, which appears in the vernacular in the three other manuscripts. However, an English rendering is provided

16  Let Each One Tell its Own Story 

485

by a triple gloss in the margin: prykky[n]g or naylyng or brody[n]g (‘pricking or nailing or tacking’). The subdivision on the seven torments shows the ease with which all four scribes move between the two languages. The overall consistency implies that they generally copied the linguistic distribution as they found it. Yet, they were perfectly capable of replacing a Latin form with an English one or adding an English translation equivalent. This implies that an oral delivery of the sermon might have included rather more code-switching, possibly through the reintroduction of additional vernacular terms to highlight the structure of the sermon. As we mentioned in our summary of Amore langueo (Sect. 2), the other signs of languishing love are also illustrated by sevenfold sets of events from the Bible; however, none of those are elaborated in as much detail, and consequently they do not display the same clear pattern of rhetorical code-switching.

4.2 Single-Word Borrowings: Rhetorical Function or Lexical Gap? As discussed, English in the verses and related elements is used to highlight the structure of the sermon. A primarily rhetorical intent also suggests itself for some of the single-word insertions, especially the ones occurring in pairs. Such switches are often concomitant with a recurrent rhetorical figure, namely the use of semantic and syntactic parallelisms for highlighting key aspects of the sermon topic.21 Below are two examples; (7) occurs across all copies, whereas (8) is reduced to a single insertion (rewful) in Cam. and Magd: (7) iste q[u]i fuit thrall fact[us] fuit free (‘he who was enslaved became free’; Trin p. 185; cf. Bal. fol. 31v, Cam. fol. 144ra, Magd. fol. 152r; W48) (8) Quintu[m] verbum erat verbum rewful ⁊ blissful (‘The fifth word was a sorrowful and a blissful word’; Bal. fol. 38r and Trin. p. 197; reduced clause in Cam. fol. 149rb and Magd. fol.155v; W688). 21

 Magd. underlines these parallel insertions, thus enhancing their visibility on the page.

486 

M. Keller and A. Seiler

(7) has a Latin syntactic frame into which English antonymous adjectives (thrall and free) are inserted. The sentence refers to the major transformation of humanity through the Passion of Christ: all people were slaves to their sins, and Christ’s sacrifice set them free. The switch into English adds rhetorical weight to the rhematic information. This kind of rhetorical refinement marks this text as a piece of art rather than a reflection of spontaneous speech. (8) refers to the paradoxical nature of Good Friday as a day of bliss (blissful) and at the same time a day of great sorrow (rewful). The omission of blissful in Cam. and Magd. suggests that the copyists amended the text according to their interpretation—in this case framing Good Friday as a prototypical day of sorrow—waiving the rhetorical figure but still keeping the vernacular insertion for the rhematic element in order to highlight the semantic core of the message. The English insertions discussed so far appear to have been included to support the rhetorical structure of the sermon. However, there are some English insertions, mostly single words, that appear to have been selected for reasons other than adding rhetorical weight via the use of the vernacular. Among them are a number of one-word insertions spread out over the sermon which seem to be triggered by a need for a specific lexeme, including its connotation. We illustrate this along with some observations on details of morphosyntactic integration. The lexeme truelove in (9) refers to the plant Paris quadrifolia, or herb paris, which in medieval English literature serves as a symbol of love (Fein 1991; Johnson 2012: 102; MED, s.v. treu-love, n., sense 4d). It locates the sermon not only linguistically but also culturally in an English context: (9) de corde isto p[ro]cedat vnu[m] trewloue (‘from this heart a truelove must grow’; our translation; Bal. fol. 21v; cf. Cam. fol. 145ra, Magd. fol. 153r, Trin. p. 187; W170). Two sentences later, all copies write about [q]uatuor folia istius fidelis amoris (‘the four leaves of this truelove’; Bal. fol. 32v, Cam. fol. 145ra, Magd. fol. 153r, Trin. p. 187; W172). Trin. glosses fidelis amoris with trewluff, suggesting that the two expressions have the same referent in the given text. However, fidelis amoris appears to be a spontaneous loan translation of the vernacular term, rather than an established Latin term; as attested

16  Let Each One Tell its Own Story 

487

in the Harley Nominale (c.1500; cited from the manuscript in MED), the regular Latin name for the plant is quadrifolium, which does not lend itself to the wordplay needed in the sermon. With respect to the structural integration of trewloue into the Latin clause, the numeral vnum is in Latin, which at the time the sermon was copied was already used as an indefinite article in medieval Latin (Beeson 1925: 19). Integrating a noun from one language into a syntactic frame of another language by marking the grammatical categories required by the host language (here, number and gender) on the determiner, in our case vnum, is very common in modern oral code-switching (Myers-Scotton and Jake 2016: 16–17; see also Poplack’s 2018: 122–140 discussion of nonce borrowings). The following examples, (10) and (11), appear to be motivated by the need for a specific technical term. They occur in the same sentence, and both concern pieces of protective clothing, namely actoun ‘padded jacket’ and hauberk ‘coat of mail’; furthermore, both are loanwords that entered Middle English via Old French.22 They figure in the interpretation of the story of the knight and his unfaithful lady. Christ is depicted as the knight, not clad in shining armour but instead bedecked with wounds. The clause containing the word actoun is rendered the same in all four copies. (10) Primo [e]n[im] h[ab]uit suu[m] actou[n] (‘For first he had his padded jacket’; Bal. fol. 32v; cf. Cam. fol. 144vb, Magd. fol. 153r, Trin. p. 187; W153). Contrary to actoun, the word hauberk shows significant differences between copies (11): (11) a. ⁊ p[ro] haberk q[uod] est ful of holes h[ab]uit corpus suum plenum vulneribus (‘and as his coat of mail, which is full of holes, he had his body full of wounds’; Bal. fol. 32v; W153)     \haburiciu[m]/ (addition in left margin by different hand)  Ultimately, Old French auqueton goes back to a Spanish-Arabic word al-quṭūn ‘cotton’ (with prefixed article; Oxford English Dictionary, hereafter OED, s.v. acton; Französisches Etymologisches Wörterbuch, hereafter FEW, s.v. quṭun), while hauberc and its variant forms are of Germanic origin (OED, s.v. hauberk; FEW, s.v. *halsberg). 22

488 

M. Keller and A. Seiler

b.  ⁊ p[ro] suo hauberk quod e[st] ful of holys h[ab]uit suu[m] corp[us] vuln[er]atu[m] plenu[m] fora[min]ib[us] (Cam. fol. 144vb–145ra) c. p[ro] habuticu[m] que est full of holys h[ab]uit suu[m] corpus vuln[er]atu[m] plenu[m] vuln[er]ib[us] (Magd. fol. 153r) d. et p[ro] suo hawberk q[uo]d est plenu[m] foraminib[us] h[ab]uit corp[us] suu[m] plenu[m] vuln[er]ib[us] (Trin. fol. 187)      \lorica haburgu[n]/ (interlinear gloss above hawberk). Bal. and Cam. insert both the piece of armour (hauberk) and its pitiful state (ful of holes) in English. A later hand added an otherwise unrecorded Latinised haburiciu[m] in the left margin of Bal. Another unattested Latinised form, habuticu[m], is used in Magd. in the running text, and only the state of the garment is inserted in English. Conversely, Trin. uses the English form hauberk but adds the fact that it is perforated in Latin. The scribe also added an interlinear gloss, giving the Classical Latin translation equivalent lorica, as well as an expuncted second vernacular variant, spelled haburgu[n]. In the same passage, Wenzel identifies one further piece of armour as a vernacular term, namely plate in the phrase cirothecis de plate ‘iron gloves’ (W155). Given that the collocation is attested a few times in fourteenth-century Latin sources (see Dictionary of Medieval Latin from British Sources, henceforth DMLBS, s.v. chirotheca) and that plate is ultimately of Romance origin, we have not counted this example as a code-switch. However, all three terms belong to the same semantic field of body armour, and they seem to represent ‘international’ terminology, as it were, which transcends language boundaries.23 Latinised variants that correspond to the first two items are well attested, for example auketonus, aketona and so on, and haubercus, haubergonum (DMLBS, s.vv. alcoto, haubercus and haubergo), while the collocation gloves of plate is used in Middle English (e.g., in Sir Gawain and the Green Knight;  Sylvester (2018: 92–94) describes similar instances of shared vocabulary in the field of clothing terms. Her conclusion is that such terms form a language that is both precise and universal and represents ‘something approaching a technolect’ (93). From a theoretical perspective, Sylvester argues that neither code-switching nor borrowing describes the use of such terms adequately. 23

16  Let Each One Tell its Own Story 

489

MED, s.v. plate, sense 3c). In all three cases, complex processes of borrowing resulted in establishing corresponding variants in Middle English, medieval Latin, as well as French. The scribes of the different copies of Amore langueo made use of some of those variants—or maybe even created some of their own. Only the Trin. scribe opts for a Classical Latin form; however, even here the addition of a second vernacular variant (though subsequently expuncted) shows that the specific meaning of this particular lexeme was probably not captured by the Classical Latin lorica. The use of hauberk / habergoun/haburicium is perhaps also reinforced by the fact that this lexeme collocates frequently with aketoun (e.g. in Chaucer’s Tale of Sir Thopas; Benson 2008: 215, ll. 860–861). To sum up, the primary purpose of most vernacular elements in Amore langueo appears to be rhetorical; however, the single-word insertions in (9)–(11) illustrate that some code-switches might be better explained in terms of lexical need—despite the fact that Latin equivalents would have been available: in the case of trewluf, only the vernacular plant name allows for the wordplay on true love. The words referring to body armour, on the other hand, represent specific terminology that cannot appropriately be rendered by Classical Latin equivalents.

4.3 The Fuzzy Borders between Languages Readers acquainted with the topic of multilingualism in medieval times are right to point out that the clear separation of codified and standardised languages cannot be applied to the historical setting we are discussing here. Nevertheless, the vast majority of mixed structures in Amore langueo consist of items which can clearly be assigned to one language. But we also find cases that reveal something about the fuzzy borders between languages. We finish our analysis with a look at two examples that illustrate the intricate interplay of languages in the bilingual mind and the constant competition that may explain some of the more unobtrusive differences between copies. Studies of oral bilingual discourse describe the phenomenon of delayed lexicalisation (Muysken 2000: 256), which means that the syntax can switch earlier than the lexicon. There are traces of this phenomenon in

490 

M. Keller and A. Seiler

Amore langueo. In (12), three copies switch to Latin after an English prayer verse (12a), but Cam. keeps using English lexical items (of þ[ys] torme[n]t[is] ‘about these torments’) for the beginning of the prose text (12b): (12) a. Y am disseset ⁊ al for schende sori ⁊ sykande ⁊ alle to rent De \istis/ vii . tormentis que Chr[istus] hodie p[er]tulit h[ab]emus figuram in sac[ra] scriptura (‘I am distressed and all broken, aching and sighing and all torn up. About those seven torments, which Christ suffered today, we have a figura in sacred scripture’; Bal. 34r–v; cf. Magd. fol. 154r, Trin. p. 190; W336) b.  Y am desesyt ⁊ al for schend . sor[e] ⁊ syke ⁊ alto rend of þ[ys] torme[n]t[is] que Ch[ristus] hodie p[er]tulit h[ab]em[us] figura[m] in sac[ra] sc[ri]pt[ur]a (Cam. fol. 146va). This leaking of English into the subsequent Latin sentence suggests simultaneous activation of languages (Kroll et al. 2006). It is facilitated through the visual diamorph tormentis, which can be Latin or English and occurs at the switch site in Cam. Another instance that points to the interaction of languages in the bilingual mind is the semantic parallelism at the end of (13): (13) a. terra in se est mag[is] sad , ⁊ stabilis (‘the earth itself is very firm and stable’; Bal. fol. 36v; W579) b. t[er]ra e[st] mag[is] sadde ⁊ stab[i]lis (Trin. p. 195) c. t[er]ra est in se magis sad and stabill (Magd. fol. 156v) d. t[er]ra e[st] i[n] se mag[is] sadde ⁊ stabil[?] (Cam. fol. 148va). In the parallelisms discussed earlier ((7) and (8)), both conjuncts are clearly English. The situation in (13) is more complex. In Bal. (13a) and Trin. (13b), only sad is inserted clearly in English. In contrast, in Magd. (13c) the whole string sad and stabill is English. The language affiliation of the items in Cam. (13d) is not as easy to determine. The first conjunct, sadde, is English. The conjunction is given as ⁊. The second adjective is

16  Let Each One Tell its Own Story 

491

spelled stabil, which suggests an English word form (MED, s.v. stable, adj.), but there is a horizontal stroke at the end of the word which functions as an abbreviation sign elsewhere (e.g. instabil[is] two lines further down). Summing up, (12) and (13) show once more that the copyists did not merely reproduce the language distribution from the exemplars but that their competence in both English and Latin as well as the interaction of languages in the bilingual mind shaped the individual form of each copy.

5 Interpretation and Conclusion What does our analysis tell us about the function of the different types of English insertions in the sermon? Clearly, most code-switches fulfil a distinct rhetorical function and constitute key elements of the sermon text. The content and distribution of most of the switches follows a clear pattern which supports the message of the sermon on different levels. The use of English in verses paraphrasing psalms and so on, as well as the prayers for the congregation, highlight a change in speaker voice and mark a different discourse level. The insertions are not motivated by lexical need, as in all cases a Latin quotation is given first, which is then loosely translated into English verse. The verses are distributed throughout the entire sermon. Many of the key concepts introduced in verse form, like the couplets describing the ‘seven signs of love’, reappear as vernacular clauses and phrases at strategic places in the sermon. The strong formal and semantic link between the verses and the clausal and phrasal insertions shows that the vernacular insertions are by no means random but serve a distinct rhetorical purpose—namely that of outlining the structure of the division and subdivisions. Another group of vernacular insertions on the phrasal and lexical level constitute syntactic and semantic parallelisms, which add rhetorical flourish to particularly prominent passages of the sermon, supporting its hortative character. A further motivation for vernacular insertions, which is, of course, not specific to artfully composed texts like sermons, is the need for culturally specific terms or connotations (e.g. truelove and hauberk). The ease with which such cultural borrowings are inserted into Latin clauses, mostly without

492 

M. Keller and A. Seiler

any verbal or visual flagging, suggests that vernacular insertions were nothing exceptional. Overall, the language mixing structures do not differ from those reported most frequently for modern oral code-switching, which also suggests that the bilingual composition of the sermon would not have been an obstacle to oral delivery. The fact that the function of the code-switching in Amore langueo is primarily rhetorical may well be one of the reasons why the language distribution is comparatively uniform across the four copies analysed in this chapter. Yet, despite the similarity in the macrostructure, the four copies display significant differences with respect to the frequency and presentation of the vernacular material. This brings us back to our second research question: What can the differences concerning English insertions between copies tell us about the reception of the different copies? Our analysis reveals that there is no uniform answer holding for all four copies, as the textual and paratextual elements point in different directions: Bal. contains structural notes, including some vernacular glosses in the margins. Cam. also lays out the sermon structure in the margins, but uses hardly any flags or glosses, which suggests that the copyist did not feel a need to highlight the vernacular insertions as such. Magd. mostly agrees with Bal. and Cam. in terms of content; however, the text is less carefully copied, and there are no later corrections or emendations, nor any traces of use. While there are no specific linguistic clues to distinguish it from Bal. and Cam. in terms of its bilingualism, the copy gives the impression of a manuscript that was produced with a view to preserving rather than preaching the sermons it contains. Trin., on the other hand, differs significantly from the other three copies. As shown in our analysis, about half of the items that are rendered as code-switches in the running text in the other copies are ‘outsourced’ by the Trin. scribe into glosses and replaced by Latin translations in the main text, thus changing the type of bilingualism from the syntagmatic to the paradigmatic level. This type of bilingualism demands a reader who can take a detour before returning to the main text, rather than a listener who is bound to the linear flow of speech. Trin. also uses the language marker Anglice much more frequently than any of the other copies. While this language label is not very prominent, it is one of the few visual cues to code-switching in the sermon.

16  Let Each One Tell its Own Story 

493

The careful rhetorical use of the vernacular, in combination with a general lack of visual highlighting, leads us to assume that the code-­ switches fully unfold their rhetorical potential only during oral performance.24 A purely Latin or English preaching mode would have rendered useless the obvious effort of placing the vernacular insertions at rhetorically relevant points. Because Amore langueo uses language contrast to draw attention to the central message, we regard the code-switching in Amore langueo as one of the ‘oral’ features displayed by this text. However, even if the older surviving versions were indeed intended for oral performance, we need to take into account that the intended mode of delivery and / or consumption can be changed by a copyist. Trin., for example, adapts a primarily oral text genre to a written mode with glosses and verbal flagging of vernacular insertions—both are features of written bilingualism well known from other medieval text genres. In conclusion, we maintain that a uniform answer to whether macaronic sermon manuscripts could actually have been preached in a bilingual manner cannot be given. Instead, exploring the possible mode of delivery must take into account the linguistic details of a specific manuscript, and it needs to assess all different textual and paratextual clues against what is known about bilingual recording and performance, both in the past and in the present.

References Beeson, Charles Henry. 1925. A Primer of Mediaeval Latin: An Anthology of Prose and Verse. Chicago, IL: Scott, Foresman and Company. Belazi, Hedi M., Edward J. Rubin, and Almeida Jacqueline Toribio. 1994. Code Switching and X-Bar Theory: The Functional Head Constraint. Linguistic Inquiry 25: 221–237. Benson, Larry D. 2008. The Riverside Chaucer, 3rd edn. Oxford: Oxford University Press.

 Halmari (2021: footnote 34) notes that the absence of visual marking of code-switches might point to a text being intended for oral delivery rather than for silent reading. 24

494 

M. Keller and A. Seiler

Butterfield, Ardis. 2013. Fuzziness and Perceptions of Language in the Middle Ages: Part 3: Translating Fuzziness: Countertexts. Common Knowledge 19: 446–473. Colker, Marvin L. 1991–2008. Trinity College Library Dublin Descriptive Catalogue of the Mediaeval and Renaissance Latin Manuscripts. 3 vols. Dublin: Four Courts Press / Aldershot: Scolar Press. Coxe, Henry Octavius. 1973. Bodleian Library, Quarto Catalogues II: Laudian Manuscripts, reprinted from the ed. of 1858–1885, with corrections and additions, and an historical introduction by R.W. Hunt. Oxford: Bodleian Library. Demo, Šime. 2014. Towards a Unified Definition of Macaronics. Humanistica Lovaniensia 63: 83–106. ———. 2018. Macaronic Latin: When Language Goes Wild. Cursor 14: 34–37. The Digital Index of Middle English Verse. Based on the Index of Middle English Verse (1943) and Its Supplement (1965). Ed. Linne R. Mooney, Daniel W. Mosser, and Elizabeth Solopova with Deborah Thorpe, David Hill Radcliffe, and Len Hatfield. https://www.dimev.net. Accessed 25 May 2023. [DMLBS =] Dictionary of Medieval Latin from British Sources. 2018. Ed. Richard Ashdowne, David Howlett, and Ronald Latham. Oxford: Oxford University Press. http://clt.brepolis.net/dmlbs/. Accessed 25 May 2023. Fein, Susanna Greer. 1991. Why Did Absolon Put a ‘Trewelove’ under His Tongue? Herb Paris as a Healing ‘Grace’ in Middle English Literature. The Chaucer Review 25: 302–317. [FEW =] Französisches etymologisches Wörterbuch: Eine Darstellung des galloromanischen Sprachschatzes. 1928–2003. Ed. Walther von Wartburg, Jean-­ Pierre Chambon, Margarete Hoffert, and Yan Greub. 25 vols. Bonn: Klopp. https://apps.atilf.fr/lecteurFEW/index.php/. Accessed 25 May 2023. Fletcher, Alan J. 2013. Written Versus Spoken Macaronic Discourse in Late Medieval England: The View from a Pulpit. In Multilingualism in Medieval Britain c. 1066–1520, ed. Judith A. Jefferson and Ad Putter, 137–152. Turnhout: Brepols. Halmari, Helena. 2021. Language Mixing as a Persuasive Strategy in Oxford, MS Bodley 649. Medieval Worlds 13: 177–194. Hanna, Ralph. 2003. Producing Magdalen College MS lat. 93. The Yearbook of English Studies 33: 142–155. Hanna, Ralph, and David Rundle. In preparation. A Descriptive Catalogue of the Medieval Manuscripts of Magdalen College, Oxford. Oxford: Oxford Bibliographical Society.

16  Let Each One Tell its Own Story 

495

Horner, Patrick. 2006. A Macaronic Sermon Collection from Late Medieval England: Oxford, MS Bodley 649. Toronto: Pontifical Institute of Mediaeval Studies. Hunt, Tony. 2000. Code-Switching in Medical Texts. In Multilingualism in Later Medieval Britain, ed. David A. Trotter, 131–147. Woodbridge: D. S. Brewer. Jake, Janice L., Carol Myers-Scotton, and Steven Gross. 2002. Making a Minimalist Approach to Codeswitching Work: Adding the Matrix Language. Bilingualism: Language and Cognition 5: 69–91. Johnson, Holly. 2012. The Grammar of Good Friday: Macaronic Sermons of Late Medieval England. Turnhout: Brepols. Keller, Mareike L. 2020. Codeswitching: Unifying Contemporary and Historical Perspectives. New York: Palgrave Macmillan. Killian, Ann. 2019. Listening for Lyric Voice in Sermon Verses and The Book of Margery Kempe. Studies in the Age of Chaucer 41: 211–237. Kroll, Judith F., Susan C. Bobb, and Zofia Wodniecka. 2006. Language Selectivity is the Exception, Not the Rule: Arguments Against a Fixed Locus of Language Selection in Bilingual Speech. Bilingualism: Language and Cognition 9: 119–135. Machan, Tim William. 2012. The Visual Pragmatics of Code-Switching in Late Middle English Literature. In Code-Switching in Early English, ed. Herbert Schendl and Laura Wright, 303–333. Berlin: Mouton de Gruyter. [MED =] Middle English Dictionary. 1952–2001. Ed. Robert E. Lewis. Online Edition in Middle English Compendium. 2000–2018. Ed. Frances McSparran and Paul Schaffner. Ann Arbor: University of Michigan Press. https://quod.lib.umich.edu/m/med/. Accessed 25 May 2023. Muysken, Pieter. 2000. Bilingual Speech: A Typology of Code-Mixing. Cambridge: Cambridge University Press. Myers-Scotton, Carol M., and Janice L. Jake. 2016. Revisiting the 4-M Model: Codeswitching and Morpheme Election at the Abstract Level. International Journal of Bilingualism 21: 340–366. Mynors, Roger. 1963. Catalogue of the Manuscripts of Balliol College Oxford. Oxford: Clarendon Press. O’Mara, Veronica M. 2002. Four Middle English Sermons: Edited from British Library MS Harley 2268. Heidelberg: Carl Winter. Odstrčilík, Jan. 2020. Multilingual Medieval Sermons: Sources, Theories and Methods. Medieval Worlds 12: 140–147.

496 

M. Keller and A. Seiler

[OED =] The Oxford English Dictionary. 1884–. 1st edn (1884–1928), ed. by James Murray, Henry Bradley, William Craigie and Charles Onions; Supplement (1972–1976), ed. by Robert Burchfield; 2nd edn (1989), ed. by John Simpson and Edmund Weiner; 3rd edn, OED Online, ed. by John Simpson (–2013) and Michael Proffitt (2013–). Oxford: Oxford University Press. https://www.oed.com. Accessed 25 May 2023. Oesterreicher, Wulf. 1997. Types of Orality in Text. In Written Voices, Spoken Signs: Tradition, Performance, and the Epic Text, ed. Egbert J. Bakker and Ahuvia Kahane, 190–214. Cambridge, MA: Harvard University Press. Owst, Gerald Robert. 1926. Preaching in Medieval England: An Introduction to Sermon Manuscripts of the Period c.1350–1450. New York: Russell & Russell. Pahta, Päivi, and Arja Nurmi. 2011. Multilingual Discourse in the Domain of Religion in Medieval and Early Modern England: A Corpus Approach to Research on Historical Code-Switching. In Code-Switching in Early English, ed. Herbert Schendl and Laura Wright, 219–251. Berlin: De Gruyter Mouton. Poplack, Shana. 2018. Borrowing: Loanwords in the Speech Community and in the Grammar. Oxford: Oxford University Press. Sarkonak, Ralph, and Richard Hodgson. 1993. Seeing in Depth: The Practice of Bilingual Writing. Visible Language 27: 6–39. Scattergood, John, Niamh Pattwell, and Emma Williams. (2021). Trinity College Dublin: A Catalogue of Manuscripts Containing Middle English and Some Old English. Dublin: Four Courts Press. Schaefer, Ursula. 2012. Spoken and Written English—Orality and Literacy. In English Historical Linguistics: An International Handbook, ed. Alexander Bergs and Laurel J. Brinton, 1274–1288. Berlin: de Gruyter Mouton. Schendl, Herbert. 2013. Code-Switching in Late Medieval Macaronic Sermons. In Multilingualism in Medieval Britain (c.1066–1520), ed. Judith F. Jefferson and Ad Putter, 153–169. Turnhout: Brepols. ———. 2015. Code-switching in Early English Literature. Language and Literature 24: 233–248. Skaffari, Janne. 2018. Code-Switching in the Long Twelfth Century. In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 121–142. Boston: De Gruyter Mouton. Sylvester, Louise. 2018. A Semantic Field and Text-Type Approach to Late-­ Medieval Multilingualism. In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 77–96. Boston: De Gruyter Mouton.

16  Let Each One Tell its Own Story 

497

Wenzel, Siegfried. 1986. Preachers, Poets, and the Early English Lyric. Princeton, NJ: Princeton University Press. ———. 1994. Macaronic Sermons: Bilingualism and Preaching in Late-Medieval England. Ann Arbor, MI: University of Michigan Press. ———. 1996. Sermon Collections and Their Taxonomy. In The Whole Book: Cultural Perspectives on the Medieval Miscellany, ed. Stephen G. Nichols and Siegfried Wenzel, 7–21. Ann Arbor, MI: University of Michigan Press. ———. 2005. Latin Sermon Collections from Later Medieval England: Orthodox Preaching in the Age of Wyclif. Cambridge: Cambridge University Press. Wright, Laura. 2011. On Variation in Medieval Mixed-Language Business Writing. In Code-Switching in Early English, ed. Herbert Schendl and Laura Wright, 191–218. Berlin: De Gruyter Mouton.

17 The Materiality of the Manières de langage Emily Reed

1 Introduction Or l’écriture médiévale ne produit pas de variantes, elle est variance (Cerquiglini 1989: 111). ‘Medieval writing does not produce variants, it is variance’. Go, litel boke, go, litel myn tragedye […] And for ther is so gret diversite In Englissh and in writyng of oure tonge, So prey I God that non myswrite [‘miscopies’] the, Ne the mysmetre [‘ruins your meter’] for defaute [‘failure’] of tonge (Troilus and Criseyde, V, ll. 1786–1796; Benson 1988: 584).

E. Reed (*) Independent Scholar, Sheffield, UK © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_17

499

500 

E. Reed

Cerquiglini (1989) stated that medieval writing is variance, a fact of which Chaucer was evidently aware when he wished that nobody would ‘miswrite’ his little book. When we speak of medieval texts as variance, we recognise the fact that manuscripts are the product of a multitude of people: author, scribe, illuminator, copyists, binders, and readers. This sometimes overlooked aspect of studying medieval texts has been the focus of a scholarly endeavour that takes several names: bibliography, new, material or pragmaphilology, visual pragmatics, and simply manuscript studies. Whatever label scholars work under, they are all united by a common goal: understanding medieval texts through their material preservation. Indeed, the tenet behind such approaches is that we cannot meaningfully discuss medieval texts without paying attention to the manuscripts in which they are preserved. The number of manuscript copies, their contents, their size, the way the texts are laid out on the page, the scripts and hands, all can serve to enlighten us about who engaged with the texts and how. This mode of study forces us to recognise each manuscript as an ‘edition’ in its own right. Modern editions can give the illusion of a stable text. In traditional editorial models, a medieval text is ‘lifted’ from a manuscript that has been deemed the ‘preferable’ witness. Variation from other manuscripts might appear in footnotes or endnotes, and the whole text is printed in a beautifully bound book, with copyright assertions, publication date and a dedication from the editor to their partner and children who made the editorial process more bearable. One could be forgiven for forgetting the other surviving texts lying on pages of vellum, in different manuscripts, in the stacks of archives around the world. Such editions, in replacing the text(s), risk giving us a false sense of security: Larry Benson is Chaucer himself. If relied upon, a single edition precludes difficult questions of the text: how was it used? By whom? In what contexts? Did this change? Medieval scribes had little sense of our modern notions of intellectual property, and indeed, there is a less demarcated boundary between author and scribe. Instead, in the process of copying, a scribe might insert, delete or reorder words and passages. A particular scribe might use their own dialectal forms (cf. Cole and Pons-Sanz, this volume; see also the papers by Putter, Kopaczyk and Bridges, and Keller and Seiler). They might gloss certain segments of text. A ‘singular text’ may thus experience several textual realisations. This variation is borne out in the following

17  The Materiality of the Manières de langage 

501

extracts, taken from the texts I discuss in the present study: a group of dialogues that aimed to teach French to English speakers in Britain at the turn of the fifteenth century (points for discussion underlined): (1)          

 —Dame, Dieu soit oveque vous. —Sire, vous soiez le bien venu. —Dame, ou est le seigneur de ciens? —Sire, il est ale hors de la ville, mais il vendra tantost, se croy je. Ditez moy, sire, vostre bousoigne.    —Dame, je vouldroye parler a luy.    —Vrayement, il ne y est pas.    —Dame, vendra il point tost?    —Par ma alme, sire, je ne scay. Sire, ditez moy vostre message.    —Nemy, dame, je le luy vouldroie dire moy mesmes (OA 1399;                      Kristol 1990: 334).                           

‘—Lady, God be with you. —Sir, be you welcome. —Lady, where is the lord of here? —Sir, he has gone out of town, but he will come soon, so I believe. Tell me, sir, your need. —Lady, I would like to speak with him. —Truly, he is not in. —Lady, will he come at all soon? —By my soul, I don’t know. Sir, tell me your message. —No, lady, I would like to tell it to him myself ’.

(2)   —Dame, Dieu soit ovesque vous. Dame, ou est le seignur de cieinz?    —Sir, il vendra tost come jeo quide. Ditez a moy, sir, que est vostre volunte?    —Dame, jeo vottroy parler ovesque luy.    —Verrament, sir, il n’est pas al hostiell.    —Dame, vendra il point tost?    —Par m’alme, sir jeo ne say, mays jeo crie q’il ne tiendra longement hors. Sir, voilles vous seier et attendre sa venue?    —Nony, dame, qar j’ay graundement affaire endementiers.   —Sir, voillez moy dire vostre message?    —Nony, dame, je le voille dire mesmes (LL 1399;               Kristol 1990: 334).

502 

E. Reed

                 

‘—Lady, God be with you. Lady, where is the lord of here? —Sir, he will come soon I believe. Tell me, sir, what is your will? —Lady, I would like to speak with him. —Truly, he is not at the hostel. —Lady, will he come at all soon? —By my soul, I don’t know, but I believe he will not be delayed long. Sir, do you want to sit and await his arrival?    —No, lady, for I have much to do in the meantime.    —Sir, would you like to tell me your message?    —No, lady, I would like to say it myself ’. (3)   —Dame, ou est le sir de ciens?   —Aves vous ascune coignoissance de lui?   —Oyl dea, je lui coignois tres bien.    —Sir, il viendra tost come je quide. Dites a moy: Que est vostre volunte?    —Dame, je vouldray parler a lui.    —Vraiment, sir, il n’est pas ciens ore.    —Dame, viendra il tost?    —Par ma foy, sir, je ne scai. Vous purres a moy dire vostre volunte.    —Nonil, je lui vouldroy dire moy mesmes (CD 1396;                 Kristol 1990: 335).                           

‘—Lady, where is the lord of here? —Sir, do you know him? —Yes, by God, I know him very well. —Sir, he will come soon I believe. Tell me, what is your will? —Lady, I would like to speak with him. —Truly, sir, he is not in now. —Lady, will he come soon? —By my faith, I don’t know. You could tell me your will. —No, I would like to tell it to him myself ’.

Between these three extracts, all taken from the same point in the texts, we encounter three versions of the same suggestion (Sire, ditez moy vostre message (1) ‘Sir, tell me your message’; Sir, voillez moy dire vostre message? (2) ‘Sir, do you wish to say to me your message?’; (il vendra (tan)tost ‘he will come (very) soon’ and Vous purres a moy dire vostre volunte (3) ‘You could say your will to me’). Moreover, the scribe in extract (3) adds two sentences not found in either of the two other manuscripts (Aves vous

17  The Materiality of the Manières de langage 

503

ascune coignoissance de lui? Oyl dea, je lui coignois tres bien). However, some elements are stable: the general structure of the conversation, some key phrases (ou est le seignur de cieinz ‘where is the lord of here’, il vendra (tan)tost ‘he will come (very) soon’), and (pro)nominal address terms. Furthermore, each manuscript represents a dialogue between scribes and readers. A reader may leave their own marks on a manuscript in the form of marginalia, glosses and signs of wear and tear. In short, each manuscript can tell us about contexts for reading and use. In order to redefine the concept of the medieval text, scholars such as Zumthor and Cerquiglini developed terms to capture this process of writing, rewriting and adaptation. Basing his work within the relationship between oral and written cultures, Zumthor (1972) conceptualised this as mouvance to describe the movement of text between people. Cerquiglini (1989), meanwhile, preferred the term variance, emphasising the relationship between the differing written manifestations of vernacular texts, and the subsequent consequences for the medieval concept of textual authority. Scholarship that grounds textual study within study of the manuscript, whatever terminology is used, draws on this foundational work. The focal points of such scholarship have been summed up as follows by Driscoll (2010: 90–91), using the term material philology: Literary works do not exist independently of their material embodiments, and the physical form of the text is an integral part of its meaning; one needs therefore to look at ‘the whole book’, and the relationships between the text and such features as form and layout, illumination, rubrics and other paratextual features, and, not least, the surrounding texts. These physical objects come into being through a series of processes in which a (potentially large) number of people are involved; and they come into being at particular times, in particular places and for particular ­purposes, all of which are socially, economically and intellectually determined; these factors influence the form the text takes and are thus also part of its meaning. These physical objects continue to exist through time, and are disseminated and consumed in ways which are also socially, economically and intellectually determined, and of which they bear traces.1  See also Moore (2016), who uses the phrase visual pragmatics in her study of visual markers for organising verbal discourse on the manuscript page. See also Machan (2011), and Carroll et al. (2013), who engage with the methodology of using material clues to inform linguistic argument. 1

504 

E. Reed

Such issues are relevant, and indeed central, to the group of dialogues at the centre of this study. As a group they have been given the title Manières de langage (henceforth Manières).2 The Manières sought to teach students to parlere, bien sonere, et parfitement escriere doux francés, that is, ‘to speak, to pronounce well and to perfectly write in beautiful French’ (CD 1396: 3).3 This small piece of publicity explicitly states that the goal was near-­fluency in French. Following the methodology of material philology, I use material elements of the Manières dialogues to make arguments about how learners may have physically engaged with the texts, and thereby the process of learning French. These dialogues were written largely in French although there are also two bilingual French-English fragments. They also address a striking range of topics: one could expect to learn how to ask for directions, how to secure a room at an inn, how to haggle at the marketplace and how to ask for news. One could also learn how to console a crying child, how to engage in a slanging match, how to preach to a sick person, how to refuse beggars, how to flirt and how to solicit the services of a prostitute. Some of the exchanges have a somewhat comedic feel. For example, after having spent the previous night partying, a lord attempts to rouse his servant: (4)   Et quant il venra au matinée, il soi levera sus bien matin et appellera tantost son chambrer par nom ainsi:     —Janyn, dors tu?     —Non il, mon signeur.     —Que fais tu doncques?     —Mon signeur, s’il vous plaist, je sounge.  I use this term for convenience rather than strict accuracy. We have little reason to presume that they were conceptualised as an indivisible and singular entity, or monolithic ‘text’, a position adopted by the editor (Kristol 1995: xvii). 3  Citations from the Manières consist of a siglum, followed by the year attributed to the Manière in question (1396, 1399, or 1415) and then a page reference from Kristol’s (1995) edition of the dialogues. I use these sigla when referring to manuscripts: Cambridge, University Library, MS Add. 8870 (CA); Cambridge, University Library, MS Dd.12.23 (CD); Cambridge, University Library, MS Ii.6.17 (CI); Cambridge, Trinity College, MS B.14.40 (CT); Lincoln, Lincolnshire Archives Office, Formulary Book 23 (LL); London, British Library, MS Add. 17716 (LA); London, British Library, MS Harley 3988 (LH); Oxford, All Souls College, MS 182 (OA); Oxford, Bodleian Library, MS Lat. Misc. e. 93 (OB); Paris, Bibliothèque Nationale, MS Nouv. Acq. Lat. 699 (PN). All translations are my own unless otherwise stated. 2

17  The Materiality of the Manières de langage 

505

    —Reveille toi de par le deable et de par sa mere ov tout. Quey ne m’as tu reveillé bien matin, comme je te comandoi hier soir?     —Mon signeur, par mon serement, si fesoi je.    — Hé, tu mens fausement parmy la gorge. Quelle heure est il maintenant?     —Mon signeur, il n’est que bien matin encore (LH 1396: 43).     ‘And when the morning comes, he will arise early and immediately call his servant by name thus:     —Janyn, are you sleeping?     —No, my lord.     —What are you doing then?     —My lord, if you please, I’m dreaming.     —Wake up, by the devil and his mother and the whole lot! Why did you not wake me at an early hour as I ordered you last night?     —My lord, by my oath, I did.     —Ah, you’re lying through your throat! What time is it now?     —My lord, it is still but early in the morning’.

Moreover, there are differences in the textual organisation of the dialogues. Some dialogues are largely organised by a narratorial voice that describes events: (5) Doncques le signeur se monte a chival et baise la fillete sa compaigne et li baille trent francs a paier pour ses despens et li dit courtoisement ainsi (LH 1396: 44).   ‘Then the lord mounts his horse and kisses the girl and her friend and gives her thirty franks to pay for her expenses, and courteously says this to her’.

or speaks as a teacher: (6)  Ore je vous moustray la manere de parler a un enfant (CD 1396: 24).    ‘Now I will show you the way to speak to a child’.

506 

E. Reed

Conversely, other dialogues are presented as lists of short utterances with no strict conversational progression or indication of who speaks, although headings are still used. For example, CA fol. 34v (7)                

Autre manere de parler de langgage De vendre et de aschatez touz chose Sir ditz moy coment serra ceo vendit Ditz coment avera jeo pur ceo Sir vous me durrez taunt Nanyl vous le faitz trep cher (Baker 1989: 100).

                 

‘Another was to speak of language To sell and to buy all things Sir, tell me how this will be sold Say for how much will I have this Sir, you will give me X [cf. Sect. 5] No, you make it too expensive’.

In the above extract, there is no indication which lines belong to the buyer and which to the seller. This question of paratext will form part of my discussion about how learners engaged with the dialogues. Indeed, the focus of this study is to examine some of the material features of the Manières manuscripts and ask whether these may be brought to bear on the language learning experience. Would manuscript size make a difference to how the learner engaged with the text? What do script, page layout, marginalia or glosses have to tell us about the language learning process? What about other signs of use? Although the editorial efforts of Kristol (1990, 1995) and Södergård (1953) have drawn attention to the textual variation between manuscripts, such questions have rarely come into consideration in Manières scholarship, which has hitherto focused on linguistic (Denoyelle 2013; Lagorgette 2013) or historical (Kibbee 1991; Critten 2015, 2019) commentary on the texts. I first describe the corpus in more detail, before discussing the social context for the dialogues. This discussion will bring to light the diversity of the learners who used these dialogues: adults and children who were

17  The Materiality of the Manières de langage 

507

geographically dispersed across England and who had different professional aspirations. Having established this background, I discuss the size of the manuscripts and their paratext to investigate the pedagogical approaches at play in the manuscripts. This research seeks to add to our understanding of how the Manières were used to learn French in England, although as will be discovered, this is an exceedingly diverse picture and much more work is needed in this area.

2 The Corpus There are three groups of the Manières which correspond to presumed dates of composition and revision: the 1396 family, the 1399 family and the 1415 family. The 1396 and 1415 dialogues can each be further divided into two families on the basis of content and transmission (Kristol 1995). Extracts from these dialogues may be found in ten manuscripts. Internal evidence informs the attribution of these dates. For example, the 1396 family is named after a note in LH fol. 23r that it was escript a Bury Saint Esmon en la veille de Pentecost lan du grace mil trois cenz quare vinz et sesze (‘written in Bury Saint Edmunds on the Eve of Pentecost in the year of grace 1396’). The date supplied for the 1399 family ne tient qu’à un fil (‘hangs by a thread’; Kristol 1995: xxxiii). The basis for the 1399 date is the last episode in this collection, which Kristol (1995: xxxiii) argues may represent an addition, which makes reference to the deposition of Richard II. Similarly, the date of 1415 for the third group is given on the basis that it mentions the Battle of Agincourt. It should be noted that some manuscripts contain fragments from different groups. The manuscripts present varying scope for the texts. While CD contains fragments of the 1396 and 1415 Manières (forty folios at 11,726 words and twelve folios at 3073 words, respectively), LL contains a short bilingual fragment of the 1399 Manières (two folios at 766 words). The relationship of dialogue to manuscript is represented graphically in Fig. 17.1:

508 

E. Reed

1396 B

1415 C

1396 A CT

PN LH

LA 1399 OA CA CD CI LL

1415 D OB

Kristol’s (1995) Manières families: 1396 A and B 1399 1415 C and D Fig. 17.1  Distribution of manuscripts across Manières families (see fn. 3 for full shelfmarks)

3 The Context for the Manières Before surveying aspects of materiality and their possible effects on the language learners, it is worth further contextualising the Manières. In the early fourteenth century, thus pre-dating the Manières, French was regarded as an important language to know in England, although in some instances it had to be taught, rather than naturally acquired (Putter and Busby 2010; Ingham 2012; Critten 2015). Until the second half of the fourteenth century, French was also the main language for teaching in schools (Iglesias-Rabade 1995: 187; Ingham 2012; see also Walkden et al., this volume) and in domestic pedagogy (Putter 2019). So, what circumstances engendered the Manières? While there are many social and political factors that contributed to the decline of L1 French competence in Britain, of most relevance to the mid-fourteenth century was the Black Death ‘whose incidence was particularly high in towns and religious communities where A-N [Anglo-Norman] was used’ (Ingham 2012: 162). The devastation that the Black Death wrought on

17  The Materiality of the Manières de langage 

509

the clergy was a significant factor in the break of transmission of French from one generation to the next, since it was the clergy that largely maintained schools where French was the main language used for teaching Latin (Ingham 2012: 33). As a consequence, ‘the plague depleted the educated population of the monasteries and convents, opening the door to poorer, less well-educated English monolinguals’ (Kibbee 1991: 59). This is corroborated with evidence from Piers Plowman (c.1370–90), in which the character Anima bemoans the lack of any clerks who can understand French: (8)  Grammer, the ground of al, bigileth now children:    For is noon of thise newe clerkes—whoso nymeth hede—    That kan versifye faire ne formaliche enditen,    Ne naught oon among an hundred that an auctour kan construwe,   Ne rede a lettre in any langage but in Latyn or in Englissh (Piers Plowman B.15, ll.370–374; Schmidt 1978).               

‘Grammar, the foundation of everything, now confuses children: For none of these new clerks—whoever pays attention— Can compose verse nor write formally, Nor can one in a hundred understand an author, Nor read a letter in any language but in Latin or English’.

Meanwhile, Kibbee argues that during the decline in French engendered by multiple plague outbreaks, the comparative prosperity caused by the Hundred Years’ War ‘furnished the middle classes and lower classes, again English monolinguals, with the opportunity and wealth to gain access to power’ (1991: 59). While the middle classes were able to make profits from a booming trade in wine and wool, the lower classes were able to demand better conditions due to fewer workers. New colleges at Oxford and Cambridge were subsequently founded to ‘replenish the number of priests, clerks, lawyers and businessmen well-educated enough to carry on the business of the day’, and along with these sprang up professional training schools that included French-language teaching in their pedagogy (Kibbee 1991: 60).

510 

E. Reed

It is within this context that the Manières dialogues are most discussed and understood. Two Oxford-based pedagogues are particularly important in this story: William Kingsmill and his predecessor, Thomas Sampson, neither of whom was attached to the university, but who operated professional training courses with a ‘full program in French’ (Kibbee 1991: 60) on the fringes of the institution. The business course taught by Sampson, a ‘prolific’ and ‘successful’ (Richardson 1941: 259) private teacher whose long career flourished between the 1360s and c.1409, included instruction in writing and witnessing charters, model letters and epistolary greetings and the Orthographia Gallica (a treatise on spelling and grammar). Cornelius (2010: 306), who examines the link between career ambitions and the study of this subject matter, states that [t]he aims of this curriculum are clearly stated in the teaching materials of Thomas Sampson […] In addition to conveyancing and letter-writing, the business course taught French grammar, common-law pleading, accountancy, and perhaps even heraldry. Teaching materials were in French and Latin, reflecting the use of both languages in contemporary domains of law, commerce, and estates management.

The evidence from the manuscripts supports this environment for many of the dialogues. For instance, CT, CD, LA and OA contain treatises on French grammar and orthography that can be connected to the teaching courses of Kingsmill and Sampson. Indeed, the manuscripts also appear to be related to students, particularly students in law, as half the manuscripts also contain either legal treatises or documents related to the legal profession. Consider also this note in LA (fol. 13v): This is edwarde dotheroles boke of chichester proficiens in chichurgia (‘This is Edward Dotherole’s book, advancing in surgery’). The later course imparted by Kingsmill (fl. 1415–1430) incorporated similar materials to Sampson’s: he composed the 1415 Manières, in which are included rhymes to instruct in elementary French, treatises on French numerals and days of the week, compositional treatises for letters and charters and a treatise on legal pleading in French. It is here that we find internal textual evidence that adds a little more to the pedagogical diversity addressed by the Manières. In the following extract, a boy (named

17  The Materiality of the Manières de langage 

511

Jehan Boun Enfant ‘John Good Boy’), addresses an unnamed knight in the hopes that he will be taken on as his apprentice: (9)                      

 —Moun fiz, avez vous esté a l’escole? —Oy, syre, par vostre congé. —A quel lieu? —Syr, a l’ostelle de Will[iam] Kyngesmylle Escriven. —Beau fyz, comben de temps avez vous demurrez ovesque luy? —Sire, forsque un quart de l’an. —Cella n’est que un poi temps, mes qu’avez vous apriz la en ycel terme? —Syr, moun maystre m’ad enseigné pur escrire, enditer, acompter et fraunceys parler (1415: CT, 76).

                    

‘—My boy, have you been to school? —Yes sir, by your leave. —At which place? —Sir, at the hostel of the scrivener William Kingsmill. —Good lad, how long have you stayed with him? —Sir, for only a quarter of a year. —That is only a short while, but what have you learned there in that time?    —Sir, my master has taught me how to write, compose, take accounts, and speak French’.

This piece of publicity demonstrates a pedagogy that combines the skills that will secure the boy employment in business administration or the legal profession, both of which required proficiency in French. Moreover, as Kristol remarks (1995: xliii), this extract adds to our understanding of who would undertake such a course: This section allows us to see that the school of William Kingsmill must have been open to different categories of student: aside from classes for adolescent or adult learners and business, secretarial or legal apprentices, there had to be an elementary course in which children could acquire their first scholarly understanding.4  ‘Cette partie permet de voir que l’école de William Kingsmill devait être ouverte à différentes catégories d’étudiants: en dehors des classes pour étudiants adolescents ou adultes, apprentis commerçants, secrétaires ou juridiques, il devait y avoir un cours élémentaire dans lequel des enfants pouvaient acquérir leurs premières connaissances scolaires’ (my translation). 4

512 

E. Reed

Such a picture is corroborated by the following note in PN fol. 114r (italics my own): (10) Ici comence un tretis de douls franceis que enformera aussi bien lez petiz comme lez granz a parler bien et perfectement beau franceis    ‘Here begins a treatise of charming French which will teach children as well as adults to speak well and perfectly good French’.

Despite the strong link between the Manières and Oxford, there is also evidence that they were known elsewhere, and ‘might have originated elsewhere’ (Critten 2015: 934). Indeed, not only are there multiple places alluded to in the Manières (Dunbar, York, Essex, Winchester and Windsor), but also, as stated in Sect. 2, LH was written in Bury Saint Edmunds. There are moreover striking anomalies in manuscripts that contain Manières, suggesting a milieu away from the classrooms of the Oxford teachers Sampson and Kingsmill, such as LL (a legal formulary book) and CI (a legal notebook, perhaps belonging to a student of the Inns of Court or Inns of Chancery in London; see Baker 1989: 81). Hitherto unnoticed in Manières scholarship is the dialect analysis of the Middle English portions of the CT manuscript, which place it in the region of Somerset (Linguistic Atlas of Late Mediaeval English, LP 5250; Benskin et al. 2013–). In addition to the linguistic analysis a previously undocumented gloss on fol. 180v (in very faded ink) reads berkelay [Berkeley],5 a town twenty-five miles north of the Somerset border, in Gloucestershire. Productive analysis in this area could be done on the LL manuscript, with dialect variations such as keve for ‘give’ (a remarkable de-voiced velar). Thus, four of the ten manuscripts (i.e. almost half the corpus) can reasonably be located away from Oxford, potentially in the hands of different types of learners. The Manières are valuable texts for study because they contribute to our understanding of the ways in which French and English influenced  This word appears above a fabliau. It glosses above the phrase nous purrons faire tout nostre desire en tens a venir (‘we will be able to do all our desire [i.e. have sex] in the future’). 5

17  The Materiality of the Manières de langage 

513

each other, or in some cases did not, in late medieval England. The texts oftentimes present idealistic attitudes towards (certain types of ) French. The apparent goal of the dialogues was to teach English-native students to parlere, bien sonere, et parfitement escriere doux francés, qu’est la plus beale et la plus gracious langage et la plus noble parlere aprés latyn de scole, that is, ‘to speak, to pronounce well and to perfectly write in beautiful French, which is the most beautiful and gracious language and the most noble to speak after scholarly Latin’ (CD: 1396: 3). In OA, this becomes doulz franceoys selon l’usage et la coustume de France (‘beautiful French according to the usage and custom of France’; OA: 1396, 81), thereby prioritising a specifically continental French. In PN, this is more specific, selon l’usage et la manere de Paris et Aurilians (‘according to the useage and custom of Paris and Orleans’; PN: 1396, 81). It is perhaps significant that this latter statement aligns the French within the Manières to that spoken within continental university cities. It is noteworthy that, while there appears to be a keenness to impart a ‘superior’ continental French, some of these conversations are set in England. (11)  Sire, cest vile est appelle Windessore (CI: 1399, 53).     ‘Sir, this town is called Windsor’. (12)  q’est le droit chymyn vers [Loundrez]? (LA: 1396, 8)     ‘What is the right way to London?’

Moreover, despite the advertised quality of the French in the Manières, some of the French in the dialogues has a markedly English character (Kristol 1995: xxxi, xxxix-xl, xlix).

4 Manuscript Size With this background in place, we know that there was a diverse age profile of learners, who wanted to learn French for professional advancement, but how are they likely to have interacted with these dialogues? I

514 

E. Reed

focus primarily on size in this section, arguing that a smaller manuscript like CI would have been likely passed around among a group of students, while larger manuscripts such as LL may have been used differently, as part of an individual’s private collection or as a teacher’s copy to be read from a lectern. Size is a noticeable variation within the corpus that would have likely changed how a person engaged with the text. However, there has not been any scholarly interest in this aspect of the Manières corpus. This omission is surprising given that the learning process would be different if the texts were read aloud to a large group of learners, or if they were to be read by an individual privately. Although pursuing a conclusive answer to this question would be an ill-judged venture, it is nonetheless possible to make informed suppositions. The following figure visually captures the extent of the difference of manuscript dimensions within the Manières corpus (Fig. 17.2). There are three groups of manuscripts within the corpus: three small manuscripts (CI, CD and LA), four medium (OB, CT, CA and LH) and three large (OA, PN and LL). At the large end of the scale is LL, at 29.5 cm × 21 cm, while at the other end, CI is the smallest manuscript, at 14.5 cm × 10 cm (a quarter of the size of LL). LL is the only manuscript for which we can name an owner. It was compiled c.1415 for John Kemp, an ecclesiastical lawyer, when he was made Dean of Arches. The rest of the manuscript contains legal formularies and the Manières appears as a fragment within the manuscript. The hand appears to be different (using more anglicana forms than are found in the other surrounding texts). The fragment concerned includes two conversations from the 1399 family: delivering a message to a man who is not at home and booking a room in an inn. It is unclear why this fragment has been bound up within this particular volume. Either this is an administrative / archival error, or it is deliberate. In the latter case, the manuscript’s owner, a practising lawyer in the ecclesiastical courts, would either have used the manuscript for personal study or as an aid if he had teaching duties. Some schoolmasters were known to personally own their own instructional books (Orme 2006: 454). It is also possible that pedagogical manuscripts were compiled by students themselves for private ownership, as is suspected of CA, a mid-sized manuscript (Baker 1989).

17  The Materiality of the Manières de langage 

515

Fig. 17.2  Manuscript dimensions (cms)

In the post-Black Death period, the cost of manuscript production had risen sharply due to the absence of available scribal labour, although around the time of the Manières it had started to decrease again as the book production trade countered increased demand with greater efficiency (Overty 2008: 12). It is nevertheless more likely that wealthier students and schoolmasters were more likely to have owned books, while less well-off students in a schoolroom would have had to share books.

516 

E. Reed

It is likely that CI was one such ‘shared’ book. CI, the smallest manuscript, is a miscellany of medical and legal treatises and documents. It contains a bilingual fragment from the 1399 family. I discuss the mise en page of CI later in this chapter, but one striking feature is the manuscript page having worn away at the point where the thumb naturally lies when holding the book. This wear is intriguing, as the damage is more extensive than what could have been produced by a single, intermittent reader. This suggests a more heavily used book. The same manuscript begins with a treatise of urines that begins thus: (13) This boke that we haue now in hande to turne in to englissh is of the craft of vrines. but who it is that werketh [‘translated’]) it we knowe not sauf onely god. Nathelesse we suppose and vs thenketh therby (CI fol.1r).

The phrase we have in hande is noteworthy. The first-person plural we may suggest a group of readers and the words in hande refer to the handheld size of the volume. At the introduction to some of the extracts, the Manières texts are also discursively positioned as a text for a group. The 1396 dialogue begins as follows: (14) Ici a nostre commencement de cesti tretis nous dirrons ainsi: En noun de Pier et Filz et Seint Espirit, amen (CD 1396: 3).    ‘Here at our beginning of this treatise we will say this: In the name of the Father and Son and Holy Spirit, amen’.

The small size of CI could alternatively indicate a private study model associated with small books, an analysis that is corroborated by these often-quoted lines from Chaucer’s Prioress’s Tale: (15)  This litel child, his litel book lernynge,     As he sat in the scole [‘school’] at his prymer    (Prioress’s Tale ll.503–536; Benson 1988: 210).

However, there is both internal textual evidence and contextual evidence that this smaller Manières manuscript (and others) might have

17  The Materiality of the Manières de langage 

517

been shared and performed aloud in a group classroom setting. In terms of evidence yielded by the Manières themselves, there is an exchange in the 1415 Manière in which a young boy recites his lessons from the hostel of William Kingsmill to an unidentified knight in order to secure an apprenticeship. As he recites his vocabulary lists, the boy utters these words: (16)  Auxi, beal fyz, je toy enseigne de comune langage et d’autre maner de parlance (CT 1415: 78; emphasis my own).     ‘Also, good lad, I will teach you the common language and another way of speaking’.

Here, not only does the boy inhabit the role of his teacher, but he replicates a verbal exchange between master and student. Moreover, throughout all the dialogues we encounter various marks of orality (such as multiple exclamations of que dea and hé). While emulating spoken exchanges between various characters across medieval society, these also serve as a possible indicator that the texts could have been performed aloud. Regarding evidence for oral performance of the dialogues within the classroom setting, Orme’s (2006) study on medieval schools provides a wealth of practical information on their organisation and teaching practice. Orme (2006: 147) states that pedagogical practice within the grammar schools comprised memorisation (either dictated by the master or read aloud by pupils) and questions and answers, which was the format used by influential fourth-century grammarian Aelius Donatus and his medieval imitators. This indicates that masters would have asked a question in class and trained the pupils to make the appropriate responses. It is therefore compelling to imagine that students may have been asked comprehension questions about their dialogues after performing them. The internal textual evidence, the pedagogical context of Donatus and the worn edge of the CI pages indicate that CI was a well-used text. I would argue that the damage to the vellum is likely not caused by a single reader, or several successive single readers. Rather, this book was likely to be shared in a group of learners in the process of learning French. Moreover, the presence of larger as well as smaller manuscripts in the Manières corpus calls to mind medieval depictions of classrooms such

518 

E. Reed

that found in Castres, Bibliothèque Municipale, MS 3, fol. 277r. In this image, the teacher sits at a lectern, reading from a large tome, while the students sit on the floor sharing a smaller volume. One can imagine larger manuscripts such as LL in the hands of the teacher, and smaller manuscripts like CI in the hands of the student.

5  Mise en page and Paratext Another source of considerable variation within the Manières corpus is the visual presentation of the dialogues across manuscripts. The dialogues are mostly presented in continuous prose, but also as two columns and, in CI, a single column of phrases in French with interlinear translations into English. It is remarkable, although I am not suggesting correlation, that the largest (**) and smallest (*) manuscripts surveyed also have the least common layout types. This could suggest that they are material outliers (Fig. 17.3).

CI*

CA, LL**, OA

Continuous text (6)

CD, CT, LA, LH, OB, PN

Two columns (3)

Fig. 17.3  Layout types in the Manières corpus

Single column (1)

17  The Materiality of the Manières de langage 

519

For this section I compare the mercantile dialogues of OA, CA and CI as they contain similar types of language, owing to the subject matter. Some of these conversations, especially those found in the 1396 family, are rather embellished and contain many rhetorical flourishes: (17) (Vendor) Savez vous que vous ferez? Vous me donnrez pour ces trois madlardes de rivere noef deniers, car ils sont bien bons, gros et gras, et je me fais fort que vous ne mangastes, sentistes ne maniastez au deux ans passez du millours qu’ils ne sont. Ore regardez, biau sire, comment ils sont rampliz du sain.      (Buyer) Oil dea, je le veoi bien, mais vrayement vous demandez trop    […]     (Buyer) Si Dieu m’ait, vous estes le plus cher homme ov qui j’ay marchandee cest annee, car j’en avrai aillours trois aussi bons anes comme ils sont pour sept deniers, mais il ne peut chaloir, car un autre fois je m’aviserai mieulx.     (Vendor) Par saint Pol, je sai bien que non avrez si bons de la price en toute ce ville, car j’en sui certain que vous ne vistes aucques mais du millours anes qu’ils ne sont (LH 1396: 38).     ‘ (Vendor) Do you know what you’ll do? You will give me nine pence for these ducks, since they are so good, large and fat, and I am assured that you have never eaten, tasted or handled in these past two years any better than these. Now look, good sir, how full of fat they are.     (Buyer) Yes by God, I see that, but you are really demanding too much    […]     Buyer) So help me God, you are the most expensive man from whom I’ve bought this year, because elsewhere I would have had three ducks of equal quality for seven pence, but no matter, because I will be better advised next time.     (Vendor) By Saint Paul, I know well that you would not have such good ducks for this price in all this town, because I’m certain that you have never seen better ducks than these’.

In the above extract, the interlocutors engage in several post-­elementary pragmatic and syntactic behaviours using a wide range of constructions:

520 

E. Reed

rhetorical questions (Savez vous que vous ferez? ‘Do you know what you will do?’), a mixture of verb tenses (je m’aviserai ‘I will be advised’; vous ne mangastes ‘you have not eaten (subjunctive)’) and accusations (vous estes le plus cher homme ov qui j’ay marchandee cest annee ‘you are the most expensive man from whom I’ve bought this year’). Conversely, the conversations of the 1399 and 1415 dialogues are characterised by much shorter utterances, and less detail (if any) about the interlocutors. (18)  (Vendor) Ditez, vueillez ce la?     Saie, wil ye that there?     (Buyer) Oy sir, il est bon pour moy.     Yee sir, it is good for me.     (Vendor) Ditez, coment le vueilliez doner? Sir, vous durrez moy tant.     Saie, howe wil ye yeve it? Sir, ye shal yeve so moche.     (Buyer) Nony sir, ne vault pas tant. Ditez a un mot coment l’avera.     Nay sir, it is not worth so moch. Saie at one worde, how shal I have it.     (Vendor) Sir, coment le vueilliez vous aver?     Sir, how wil ye have hit?     (Buyer) Ne vous durray tant d’argent.     I ne shal yeve so moch silver (CI; Södergård 1953: 213–214).6     ‘(Vendor) Say, do you want that there?     (Buyer) Yes sir, it is good for me.     (Vendor) Say, how [much] will you give [for] it? Sir, you will give X much.     (Buyer) No sir, it is not worth X much. Say in one word how I shall have it.     (Vendor) Sir, how will you have it?     (Buyer) I will not give you so much silver’.

There is a noticeable difference between the single-sentence utterances above, and the longer utterances to be found over the course of the other mercantile conversations. This suggests two different pedagogical  Labels indicating vendor and buyer in (17) and (18) are my editorial additions.

6

17  The Materiality of the Manières de langage 

521

approaches: one that anticipates a competent learner who already has some knowledge of French, and who will also be able to take in the behavioural didactics on display; and another that imagines a less able learner and presents simple sentences to be faithfully memorised and recycled. Because some manuscripts contain both elaborate and simpler styles of conversations, it is important to state that this is not an evolution of pedagogy, with one type being replaced by another. Rather, the two styles appear to coexist. While the more embellished conversations of the 1396 texts include more conversational twists and turns with more information about the speakers, in these shorter conversations, utterances seem to be condensed around salient phrases useful for haggling, such as key request forms (vous me donnerez ‘you will give me’; moustrez moi ‘show me’), contestation (il ne vault pas le argent ‘it is not worth the money’) and pressing the interlocutor for a sum (mot draine ‘last word’, a un mot ‘at one word’), among other things. Another feature of these dialogues is the presence of ‘placeholder’ language: instead of a specific sum of money, as we saw in the earlier dialogues, the word tant (translated as so moch in CI) is used. Likewise, instead of naming the product being bought and sold, the phrase bone chose ‘good thing’ is used. The modern equivalent to this placeholder language would be to use the letter X instead of a number or name (‘for X amount of time’) while giving an example. In the context of the mercantile dialogues, one can imagine that the learner who merely wants to know a few useful phrases could insert their own prices and products in the places of tant and bone chose. There is little investment in observing the complex interplay between two developed and believable characters. Adding to this impression is the lack of paratextual features, such as narrative interludes, or features that enable a reader to make sense of conversational turns (i.e. who the speaker is, and separation of clauses into possible things one could say). Below is an example of the mise en page of the more embellished 1396 draper’s apprentice conversation as it appears in OA fol. 317r (altered to highlight features of interest; Fig. 17.4). Highlighted above, we can see paratextual features such as narrative frame (Et lapprentiz leur dit tout courtoisement en ce manere, ‘and the

522 

E. Reed

Fig. 17.4  Oxford, All Souls College, MS 182 (OA), fol. 317r (close-up)

apprentice says to them very courteously in this way’) and vel sic (‘or this’) which makes evident that there are two choices of utterance that the apprentice might say. This is underlined, which visually marks it out from the direct speech. Moreover, small dots are used to separate clauses and the scribe also uses large capital letters at the beginning of clauses to make clear the changes between speakers and between direct speech and narrative frame. Somewhat more subtly, clause-initial v has an ascender (lines 4, 9 and 10), whereas the non-clause-initial v has descenders (lines 4, 9 and 19). Overall, the reader is given a lot of support in reading this dialogue.

17  The Materiality of the Manières de langage 

523

Fig. 17.5  Cambridge, University Library, MS Add. 8870 (CA), fol. 34v (close-up)

The picture changes when we turn to the simpler dialogues, such as this one below from CA fol. 34v (Fig.  17.5). While the utterances are separated line-by-line, there is no narrative frame beyond the simple title autre manere de parler de langgage de vendre et de aschatez touz chose (‘another way of speaking to sell and to buy all things’) and no obvious change of speaker. When we look at ll. 4–7 (highlighted in Fig. 17.5), the question of who is speaking becomes obscured: (19)  Sir ditz moy coment serra ceo vendit     Ditz coment avera jeo pur ceo     Sir vous me durrez taunt     Nanyl vous le faitz trep cher (Baker 1989).

524 

E. Reed

Fig. 17.6  Cambridge, University Library, MS Ii.6.17 (CI), fol.103r (close-up)

    ‘Sir, tell me how this will be sold     Say for how much will I have this     Sir, you will give me X     No, you make it too expensive’.

Indeed, there is no indication (beyond the reader’s inference) that the first two lines are the buyer, the third line is the vendor, and the fourth line is the buyer again. CI, discussed above and shown in Fig. 17.6, shows this same tendency to omit the paratextual framework. In the CI fragment, there are lines to indicate translation pairs, and while phrases are separated from each other, no speakers are identified, and no title is provided to separate this conversation from the previous

17  The Materiality of the Manières de langage 

525

one. It is furthermore unclear which of the traveller’s two utterances are being responded to by the innkeeper when he asks at l. 10 Sir, coment le vueilliez vous aver? (‘Sir, how will you have it?’). On the basis of this style of presentation, it would appear that the separate phrases and their translation into English is much more important than developing characters and demonstrating the more complex discursive manoeuvres of the 1396 dialogues. Rather, what we see in the examples from CA and CI are phrases that one might expect to hear in the marketplace, loosely grouped together into a bargaining structure but not embedded in a particular narrative or representation of reality. This speaks to a ‘stock phrase’ style of learning that may be found in modern day phrasebooks or indeed, in Early Modern vocabularies such as those found in William Wey’s itineraries from the mid-fifteenth century: (20)  Good morrow, calomare    Welcome, calosertys     Tel me the way, dixiximo strata     Wyth goyd wyl, mitta karas     Com hethyr, elado     Sytte, catase     Gyff me that, doys me tutt     Goo thy way, ame     Anone, lygora     Bryng me, fer me     What seyst thou, the leys     I vnderstond the not, apopon kystys     God be wyth the, Theos metasana (Itineraries of William Wey; Williams 1857: 102).     ‘Good morrow    Welcome     Tell me the way     With good will     Come here    Sit     Give me that     Go on your way

526 

E. Reed

   Anon     Bring me     What are you saying?     I don’t understand you     God be with you’.

Gallagher (2014: 39) states that such manuals ‘filled a niche for those who needed basic vocabulary and phrases appropriate to a variety of situations, many to do with travel and commerce’. But even earlier, in the late fourteenth century, some of Chaucer’s pilgrims can be seen to have adopted an ad hoc approach to language learning, recycling snippets of language. For instance, the Pardoner admits to seasoning his sermons with choice phrases of Latin in order to encourage devotion in his audience: (21)  And after that thanne telle I forth my tales;     Bulles of popes and of cardynales,     Of patriarkes and bishopes I shewe     And in latyn I speke a wordes fewe,     To saffron with my predicacioun,      And for to stire hem to devocioun (Pardoner’s Prologue, VI, ll.341–346; Benson 1988: 194).                        

‘And after that then I tell my tales; Letters of popes and cardinals, Of patriarchs and bishops I show. And I speak a few words of Latin, To season my preaching, And to move them to devotion’.

The friar of the Summoner’s Tale moreover uses some French phrases in his speech: (22)             

O thomas, je vous dy, thomas! thomas! This maketh the feend; this moste ben amended. Ire is a thyng that hye God defended, And therof wol I speke a word or two.

17  The Materiality of the Manières de langage 

527

                               

Now, maister, quod the wyf, er that I go, What wol ye dyne? I wol go theraboute. Now dame, quod he, now je vous dy sanz doute, Have I nat of a capon but the lyvere, And of youre softe breed nat but a shyvere, And after that a rosted pigges heed— But that I nolde no beest for me were deed— Thanne hadde I with yow hoomly suffisaunce (Summoner’s Tale, VI, ll.1832–1843; Benson 1988: 130).

                       

‘“Oh Thomas, I tell you, Thomas! Thomas! This makes the fiend; this must be amended. Ire is something that high God forbade, And on that will I speak a word or two.” “Now master,” said the wife,” before I go, What do you want to eat? I will make it.” “Now lady,” he said, “now I will tell you without doubt I’ll have only the liver of a capon, And of your soft bread nothing but a sliver, And after that a roasted pig’s head, But I wish that no beast died for me, Then I will have had with you enough plain food”’.

                   

While the Manières are part of a more intentional French teaching programme, the presence of code-switching between English and French at the level of phrases reinforces the possibility of this style of learning in operation for some learners.

6 Conclusion Material philology allows us to fully consider the implications of the diversity of the manuscripts of the Manières corpus. Indeed, the material aspects of the texts can illuminate the various ways in which people learned French at the turn of the fifteenth century. A heterogenous corpus, the Manières are worth further exploration, and there is much more ground to cover. This chapter has sought to demonstrate how aspects of

528 

E. Reed

size and paratext can inform our understanding of how students used the Manières to learn French. We know that learners varied in age, prior knowledge of French and profession. This diversity is reflected in the variety of the manuscripts in the Manières corpus. The manuscripts vary greatly in size, layout, textual content, hand and more besides. I have argued that the portable dimensions of some of the Manières manuscripts, like CI, make it possible that these dialogues may have been passed around in a classroom setting, in order to be read aloud (in a manner similar to the traveller’s modern phrasebook). This reflects the question-­and-response pedagogical method inspired by Donatus, resulting in a classroom with ‘plenty of oral interchange’ (Orme 2006: 148). On the other end of the scale, large manuscripts like LL, for which we have an owner, point towards a potentially different style of use. Analysis of glosses could add to this picture. If a manuscript was shared, and not a personal possession, would this affect the nature of the glosses? For example, would students be less able to gloss a shared book than a student for whom the book was a personal possession? What would multilingual glosses add to the picture? For example in LH, glosses are in both Anglo-­ French and Middle English (fol. 6v poucies: ordures ‘scraps of filth or dirt’, and pourretes: poudres ‘dust’) while in PN glosses are in both Middle English and Latin (fol.119r pourettez: dustes ‘dust’ and soriz: muribus ‘mice’). Mise en page and textual arrangement can also inform our understanding of these dialogues. While the shorter dialogues likely drew the attention of the learner towards internalising short sentences, other manuscripts supplied a lot of visual support to a reader navigating the more complex and discursive 1396 dialogues.

References Baker, J.H. 1989. A French Vocabulary and Conversation-Guide in a Fifteenth-­ Century Legal Notebook. Medium Aevum 58: 80–102. Benskin, Michael, Margaret Laing, Vasilios Karaiskos, and Keith Williamson. 2013–. An Electronic Version of A Linguistic Atlas of Late Mediaeval English.

17  The Materiality of the Manières de langage 

529

Edinburgh: University of Edinburgh. http://www.lel.ed.ac.uk/ihd/elalme/ elalme.html. Accessed 28 May 2023. Benson, Larry, ed. 1988. The Riverside Chaucer. Oxford: Oxford University Press. Carroll, Ruth, Matti Peikola, Hanna Salmi, Mari-Liisa Varila, Janne Skaffari, and Risto Hiltunen. 2013. Pragmatics on the Page. European Journal of English Studies 17: 54–71. Cerquiglini, Bernard. 1989. Éloge de la variante: Histoire critique de la philologie. Paris: Seuil. Cornelius, Ian. 2010. The Rhetoric of Advancement: Ars Dictaminis, Cursus, and Clerical Careerism in Late Medieval England. In New Medieval Literatures 12, ed. Rita Copeland, David Lawton, and Wendy Scase, 289–330. Turnhout: Brepols. Critten, Rory. 2015. Practicing French Conversation in Fifteenth-Century England. The Modern Language Review 110: 927–945. ———. 2019. The Manières de Langage as Evidence for the Use of Spoken French within Fifteenth-Century England. Forum for Modern Language Studies 55: 121–137. Denoyelle, Corinne. 2013. Les realisations des actes de langage directifs dans les Manières de langage. In Marques d’Oralité en Français Médiéval, Diachroniques, III, 149–175. Paris: Presses de l’université Paris-Sorbonne. Driscoll, M.J. 2010. The Words on the Page: Thoughts on Philology, Old and New. In Creating the Medieval Saga: Versions, Variability, and Editorial Interpretations of Old Norse Saga Literature, ed. Judy Quinn and Emily Lethbridge, 85–102. Odense: Syddansk Universitetsforlag. Gallagher, John. 2014. Vernacular Language-Learning in Early Modern England. Unpublished doctoral thesis, University of Cambridge. Iglesias-Rabade, Luis. 1995. Multilingual Education in England 1200-1500. Studia Neophilologica 67: 185–195. Ingham, Richard, ed. 2012. The Transmission of Anglo-Norman: Language History and Language Acquisition. Amsterdam: John Benjamins. Kibbee, Douglas. 1991. For to Speke Frenche Trewely: The French Language in England, 1000–1600: Its Status, Description and Instruction. Amsterdam: John Benjamins. Kristol, Andres. 1990. L’enseignement du Français en Angleterre (XIIIe–XVe siècles): Les sources manuscrites. Romania 111: 289–330. ———, ed. 1995. Manières de langage (1396, 1399, 1415), Anglo-Norman Text Society, 53. London: Anglo-Norman Text Society. Lagorgette, Dominique. 2013. Insultes et registres de langue dans les Manières de langage: Transgression et effets d’oralité. In Marques d’Oralité en Français

530 

E. Reed

Médiéval, Diachroniques 3, 119–147. Paris: Presses de l’université Paris-Sorbonne. Machan, Tim. 2011. The Visual Pragmatics of Code-Switching in Late Middle English Literature. In Code-Switching in Early English, ed. Herbert Schendl and Laura Wright, 303–333. Berlin: Mouton de Gruyter. Moore, Colette. 2016. Visual Pragmatics: Speech Presentation and Middle English Manuscripts. In The Cambridge Handbook of English Historical Linguistics, ed. Merja Kytö and Päivi Pahta, 481–496. Cambridge: Cambridge University Press. Orme, Nicholas. 2006. Medieval Schools. New Haven, CT: Yale University Press. Overty, Joanne. 2008. The Cost of Doing Scribal Business: Prices of Manuscript Books in England, 1300–1483. Book History 11: 1–32. Putter, Ad. 2019. The Linguistic Repertoire of Medieval England, 1100–1500. In Imagining Medieval English: Language Structures and Theories, 500–1500, ed. Tim William Machan, 126–144. Cambridge: Cambridge University Press. Putter, Ad, and Keith Busby. 2010. Introduction: Medieval Francophonia. In Medieval Multilingualism: The Francophone World and Its Neighbours, ed. Christopher Kleinhenz and Keith Busby, 1–13. Turnhout: Brepols. Richardson, H.G. 1941. Business Training in Medieval Oxford. American Historical Review 6: 259–280. Schmidt, A., ed. 1978. The Vision of Piers Plowman: A Critical Edition of the B-Text. London: J. M. Dent and E. P. Dutton. Södergård, Östen. 1953. Une ‘Maniere de parler’: Ms. Cambridge, Bibliothèque de l’Université, I i. 6.17. Neuphilologische Mitteilungen 54: 201–225. Williams, George, ed. 1857. The Itineraries of William Wey, Fellow of Eton College, to Jerusalem, A.D. 1458 and A.D. 1462; and to Saint James of Compostella, A.D. 1456. London: Roxburghe Club. Zumthor, Paul. 1972. Essai de poétique médiévale. Paris: Seuil.

18 Afterword Sara M. Pons-Sanz and Louise Sylvester

As we noted in the Introduction to this volume, in the 1990s, when scholars first began to focus on the variety of languages in the administrative records of medieval Britain, the theoretical approaches that were available were those derived from modern studies of bilingualism and code-switching. However, by the time of the publication of Pahta, Skaffari and Wright’s collection in 2018, questions about definitions of code-­ switching, and whether the term adequately described all the multilingual phenomena found in medieval texts, were under consideration. The introduction to that volume is titled ‘From Historical Code-Switching to Multilingual Practices in the Past’. The editors discuss the range of definitions that exist for the term code-switching and note other approaches to

S. M. Pons-Sanz (*) Cardiff University, Cardiff, UK e-mail: [email protected] L. Sylvester University of Westminster, London, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2_18

531

532 

S. M. Pons-Sanz and L. Sylvester

multilingualism that ‘take a different starting point for viewing the phenomenon and disregard the notion of codes in the sense of separate languages between which speakers switch’ (2018: 4). Moreover, significantly, the first essay in that collection addresses ‘Historical and Modern Studies of Code-Switching’ and defines the research contained in the volume as representing ‘the most recent stage in research carried out by a group of historical sociolinguists’ (Gardner-Chloros 2018: 20). In planning this volume, it seemed to us that the label applied by Gardner-Chloros is now too narrow. The researchers whose work is collected in our volume have participated in the continued opening-up of a wider range of approaches with which to investigate multilingualism in medieval British textual culture and its effects on the development of English. We may note that where code-switching continues to be used as a model for understanding multilingual phenomena in medieval texts, it may now be combined with other approaches, as shown in Keller and Seiler’s chapter (this volume), which examines variation between copies on the lexical and the structural level and their discourse-pragmatic functions on the textual level against the backdrop of models for modern oral and written code-switching. We may also note the chapters by Putter et al., which considers the role of visual pragmatics in the creation of the text’s meaning in its social context; and by Reed, which takes account of the multitude of people involved in the production of manuscripts as well as those who used them, through an exploration of the manuscripts’ materiality. We have continued to see research grounded in sociolinguistic approaches; these now include the exploration of networks and communities of practice: for example, Pons-Sanz’s (2004) study of the Norse-­ derived terms recorded in the Old Northumbrian glosses to the Lindisfarne and Rushworth Gospels; Timofeeva’s (2013) exploration of bilingual communities of practice in England in the Old English period; Ingham’s (2018) article investigating the role of the clergy in the diffusion of words of French origin; and a co-edited special issue of the journal English Language and Linguistics on mechanisms of French contact influence in Middle English (Timofeeva and Ingham 2018). The distinction between borrowing and imposition suggested by van Coetsem (2000) has also led to fruitful discussions. For instance, it has

18 Afterword 

533

been invoked in order to try to account for the differences in the semantic fields represented by the Norse-derived terms first attested in the Old English period and those associated with the terms first recorded during the Middle English period; these differences are taken to be indicative of disparities in terms of types and intensity of sociolinguistic interactions and speakers’ linguistic competence in Old English and Old Norse between the Scandinavianised and non-Scandinavianised areas (Townend 2002). A project led by Sylvester, The Semantics of Word Borrowing in Late Medieval English, also seeks to make use of van Coetsem’s distinction by applying it to the range of senses of French words that were borrowed into Middle English, on the basis that borrowing only one sense or the full range of senses may inform us about the levels of competence in French of speakers in late medieval Britain.1 Further work based in historical sociolinguistics is represented by Kinn and Walkden’s (this volume) application of the methods and findings of modern heritage language research to their work on possessive constructions in Norn. Moreover, research on Multicultural London English by Fox (Fox and Sharma 2017; Fox and Torgersen 2018) and the new town koine discussed by Kerswill and Williams (2000) has the potential to shed light on the interaction of closely related languages, such as Norse and English (see Fox et al., this volume). Much work has also been done to further our understanding of the impact that contact with Old Norse, French and Latin had on the development of English, and this has led, in turn, to a reassessment of very well-established views in scholarly debates. Thus, the Gersum project team, on the basis of piloting work by Dance (2019), have reconsidered the nature and reliability of the different types of evidence that one needs to take into consideration in order to identify Norse-derived terms in English, and hence the extent of Norse impact on the English lexicon (Dance et al. 2019; see also Pons-Sanz 2021). However, as Verloot’s, and Cole and Pons-Sanz’s contributions to this volume make clear, this is an area where much more work is still needed (see also Cole 2018, Forthcoming). Work by Sylvester et al. (2020, 2022a, 2022b, 2023,;  Richard Ingham and Kathryn Allan are co-investigators, and Megan Tiddeman research fellow. The project team is grateful to the Leverhulme Trust for funding across 2021–2024. 1

534 

S. M. Pons-Sanz and L. Sylvester

Sylvester and Tiddeman, this volume) strongly challenges the notion that the incoming loanwords led to the native terms undergoing semantic shifts, such as broadening or narrowing, or dropping out of usage within the period (see further Dance et al., this volume). New research is also being undertaken on morphosyntactic effects of language contact in the medieval period and the interface between morphosyntax and semantics (see the chapters by García García and Ingham, and by Adamczyk, this volume). Research on multilingualism in the medieval period has also led to a thorough reassessment of development of Standard English; it is now clear that its origins and spread are a result of the multilingual ecology of medieval Britain (see Wright 2020; Sylvester Forthcoming). All of this work is being facilitated by the use of new technologies, in work undertaken by Stokes, Trips and others, and discussed in detail in Trips and Stokes’s chapter in this volume. The support offered by electronic tools will continue as the field of humanities computing evolves. Another development in the field is the analysis of the influence of languages other than Latin, French and Norse on English. Trotter’s (2000) edited volume contains papers investigating the relationship and contact between English and Welsh (Smith 2000), and the (scant) evidence of testimony in Welsh (Richter 2000); and Pahta et al.’s collection includes a chapter that opens up the question of possible contact influence from Welsh on the manuscript of Sir Gawain and the Green Knight (Meecham-Jones 2018). Research on medieval Welsh that contextualises it in the study of multilingualism has continued to thrive, as we see in various contributions to this volume: Callander’s chapter looks at English, Welsh and Latin; while Fulton’s focuses on English, Welsh and French, particularly the latter two languages. Research on the role of Dutch has gained momentum with a project led by Putter and van Houts: The Literary Heritage of Anglo-Dutch Relations c.1050–c.1600 (see Levelt and Putter 2021; Putter 2021; McSheffrey and Putter 2023),2 but see also Roig-Marín (2021). Given the proximity between the two languages, the question of Dutch loans in English has the potential to be explored through the Gersum methodology. Tiddeman’s work (2017, 2018, 2020,  Sjoerd Levelt and Moreed Arbabzadah are research assistants, and the project is funded by the Leverhulme Trust. 2

18 Afterword 

535

2023) analyses the influence of Italian through trade in the Middle English period, and further work in this area is being undertaken as part of the Progetto di Ricerca di Interesse Nazionale (PRIN) Writing Expertise as a Dynamic Sociolinguistic Force: The Emergence and Development of Italian Communities of Discourse in Late Antiquity and the Middle Ages and Their Impact on Languages and Societies.3 A further area of focus rooted in the medieval literary culture that remains to us is the work on texts and manuscripts; for example a neglected Dutch source for the Middle English text Solomon and Marcolf has recently been suggested by Putter (2023). Work on administrative and business texts such as the Corpus of Middle English Local Documents (MELD) project led by Stenroos (Stenroos et al. 2022; and the resulting volume edited by Stenroos and Thengs 2020) and the Emerging Standards: Urbanisation and the Development of Standard English c.1400–1700 project led by Auer (Auer et al. 2022) have extended our understanding of the languages in which the bureaucracy was conducted in Britain and how the use and balance of the different languages of record shifted across the medieval period. All these advances and the work that still remains to be done are a testament to the thriving character of this research field and the important contribution that it can make to, besides benefiting from, neighbouring areas.

References Auer, Anita, Moragh Gordon, Femke van Hilton, Mike Olson, and Tino Oudesluijs. 2022. Emerging Standards: Urbanisation and the Development of Standard English c.1400–1700. https://wp.unil.ch/emst/anita-­auer/. Accessed 23 May 2023. Cole, Marcelle. 2018. A Native Origin for Present-Day English they, their, them. Diachronica 35: 165–209. ———. Forthcoming. A History of They. New Approaches to English Historical Linguistics. London: Palgrave Macmillan.  National coordinator Piera Molinelli, University of Bergamo; Associated universities: Bergamo, Cagliari, Chieti-Pescara “G. d’Annunzio”, Naples “Federico II”. 3

536 

S. M. Pons-Sanz and L. Sylvester

Dance, Richard. 2019. Words Derived from Old Norse in Sir Gawain and the Green Knight: An Etymological Survey, Publications of the Philological Society, 50. 2 vols. Chichester: Wiley-Blackwell. Dance, Richard, Sara M. Pons-Sanz, and Brittany Schorn. 2019. The Gersum Project: The Scandinavian Influence on English Vocabulary. Cambridge, Cardiff and Sheffield. https://gersum.org. Accessed 23 May 2023. Fox, Susan, and Devyani Sharma. 2017. The Language of London and Londoners. In Urban Sociolinguistics—The City as a Linguistic Process and Experience, ed. Patrick Heinrich and Dick Smakman, 115–129. Abingdon: Routledge. Fox, Susan, and Eivind Torgersen. 2018. Language Change and Innovation in London: Multicultural London English. In Sociolinguistics in England, ed. Natalie Braber and Sandra Jansen, 189–213. London: Palgrave Macmillan. Gardner-Chloros, Penelope. 2018. Historical and Modern Studies of Codeswitching: A Tale of Mutual Enrichment. In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 19–36. Berlin: Mouton de Gruyter. Ingham, Richard. 2018. The Diffusion of Higher-Status Lexis in Medieval England: The Role of the Clergy. English Language and Linguistics 22: 207–224. Kerswill, Paul, and Ann Williams. 2000. Creating a New Town Koine: Children and Language Change in Milton Keynes. Language in Society 29: 65–115. Levelt, Sjoerd, and Ad Putter. 2021. North Sea Crossings: The Literary Heritage of Anglo-Dutch Relations, 1066–1688. Oxford: Bodleian Library. McSheffrey, Shannon, and Ad Putter. 2023. The Dutch Hatmakers of Late Medieval London with an Edition of Their Bilingual Guild Ordinances. Boydell and Brewer. Meecham-Jones, Simon. 2018. Code-Switching and Contact Influence in Middle English Manuscripts from the Welsh Penumbra—Should We Re-Interpret the Evidence from Sir Gawain and the Green Knight? In Multilingual Practices in Language History: English and Beyond, ed. Päivi Pahta, Janne Skaffari, and Laura Wright, 97–119. Berlin: Mouton de Gruyter. Pahta, Päivi, Janne Skaffari, and Laura Wright, eds. 2018. Multilingual Practices in Language History: English and Beyond. Berlin: Mouton de Gruyter. Pons-Sanz, Sara M. 2004. A Sociolinguistic Approach to the Norse-Derived Words in the Glosses to the Lindisfarne and Rushworth Gospels. In New Perspectives on English Historical Linguistics: Selected Papers from 12 ICEHL, Glasgow, 21–26 August 2002, II, Lexis and Transmission, ed. Christian Kay,

18 Afterword 

537

Carole Hough and Irené Wotherspoon, 177–192. Amsterdam: John Benjamins. ———. 2021. Norse-Derived Vocabulary in La Estorie del Evangelie. Folia Linguistica Historica 42: 461–491. Putter, Ad. 2021. Materials for a Social History of the Dutch Language in Medieval Britain: Three Case Studies from Wales, Scotland, and England. Dutch Crossing 45: 97–111. ———. 2023. Multilingualism. In The Routledge Companion to Medieval English Literature, ed. Raluca Radulescu and Sif Rikhardsdottir, 107–119. London: Routledge. Richter, Michael. 2000. Collecting Miracles Along the Anglo-Welsh Border in the Early Fourteenth Century. In Multilingualism in Later Medieval Britain, ed. David A. Trotter, 53–61. Cambridge: D. S. Brewer. Roig-Marín, Amanda. 2021. Middle Dutch-Origin Lexis in the Durham Account Rolls: The Cinderella of Medieval Multilingual Texts. Studia Neophilologica 93: 92–106. Smith, Llinos Beverley. 2000. The Welsh and English Languages in Late-­ Medieval Wales. In Multilingualism in Later Medieval Britain, ed. David A. Trotter, 7–21. Cambridge: D. S. Brewer. Stenroos, Merja, and Kjetil Thengs, eds. 2020. Records of Real People: Linguistic Variation in Middle English Local Documents. Amsterdam: John Benjamins. Stenroos, Merja, Kjetil Vikhamar Thengs, and Kenneth Solberg-Harestad. 2022. A Corpus of Middle English Local Documents (MELD). https://www.uis. no/en/corpus-­middle-­english-­local-­documents-­meld. Accessed 18 May 2023. Sylvester, Louise. Forthcoming. Early Standardisation. In New Cambridge History of English, I: Context, Contact and Development, ed. Laura Wright. Cambridge: Cambridge University Press. Sylvester, Louise, Megan Tiddeman, and Richard Ingham. 2020. An Analysis of French Borrowings at the Hypernymic and Hyponymic Levels of Middle English. Lexis: Journal of English Lexicology 16. ———. 2022a. Lexical Borrowing in the Middle English Period: A Multi-­ domain Analysis of Semantic Outcomes. English Language and Linguistics 26: 237–261. ———. 2022b. Semantic Shift in Middle English: Farming and Trade as Test Cases. Transactions of the Philological Society 120: 427‒446. ———. 2023. Lexical Replacement, Retention and Borrowing in Middle English: A Case Study. In Language Contact and the History of English: Processes and Effects on Specific Text-Types, ed. Gabriella Mazzon, 165‒18. Frankfurt am Mein: Peter Lang.

538 

S. M. Pons-Sanz and L. Sylvester

Tiddeman, Megan. 2017. Early Anglo-Italian Contact: New Loanword Evidence from Two Mercantile Sources, 1440–1451. In Merchants of Innovation: The Languages of Traders, ed. Esther-Miriam Wagner, Bettina Beinhoff, and Ben Outhwaite, 217–234. Berlin: De Gruyter Mouton. ———. 2018. Lexical Exchange with Italian in the Textile and Wool Trades in the Thirteenth to Fifteenth Centuries. Medieval Clothing and Textiles 14: 113–140. ———. 2020. More Sugar and Spice: Revisiting Medieval Italian Influence on the Mercantile Lexis of England. In The Multilingual Origins of Standard English, ed. Laura Wright, 381–410. Berlin: De Gruyter Mouton. ———. 2023. The Cantelowe Accounts: Multilingual Merchant Records from Tuscany, 1450–1451, Records of Social and Economic History. Oxford: Oxford University Press for The British Academy. Timofeeva, Olga. 2013. Of ledenum bocum to engliscum gereorde: Bilingual Communities of Practice in Anglo-Saxon England. In Communities of Practice in the History of English, ed. Joanna Kopaczyk and Andreas H. Jucker, 201–224. Amsterdam: John Benjamins. Timofeeva, Olga, and Richard Ingham, eds. 2018. Mechanisms of French Contact Influence in Middle English: Diffusion and Maintenance. English Language and Linguistics 22 (2). 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. Trotter, David A., ed. 2000. Multilingualism in Later Medieval Britain. Cambridge: D. S. Brewer. van Coetsem, Frans. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter. Wright, Laura, ed. 2020. The Multilingual Origins of Standard English. Berlin: De Gruyter Mouton.

Index1

A

Abbotsford, Advocates Library, Abbotsford MS B3, 442, 445 ABC a femmes, 425, 426, 433 ABC of Women, 425, 426, 433 Aberystwyth, National Library of Wales, MS Llanstephan 34, 446 Aberystwyth, National Library of Wales, MS Peniarth 26, 431, 449 Aberystwyth, National Library of Wales, MS Peniarth 27ii, 446 Aberystwyth, National Library of Wales, MS Peniarth 50, 431, 449

Aberystwyth, National Library of Wales, MS Peniarth 252, 446 Aberystwyth, National Library of Wales, MS Peniarth 326, 449 Ælfric Catholic Homilies, 259, 355, 357 Heptateuch, 74 Letter to Sigeweard, 74 Against the King’s Taxes, see ‘Dieu, roy de magesté Aled, Tudor, 204–206 Alliterative Revival, 162, 183 Piers Plowman, 315, 316, 322, 323, 428, 473n13, 509 Sir Gawain and the Green Knight, 488

 Note: Page numbers followed by ‘n’ refer to notes.

1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 S. M. Pons-Sanz, L. Sylvester (eds.), Medieval English in a Multilingual Context, New Approaches to English Historical Linguistics, https://doi.org/10.1007/978-3-031-30947-2

539

540 Index

Amis et Amile, 429 Amore langueo, 467, 467n1, 468, 470–480, 471n7, 482–493 Ancrene Riwle/Wisse, 112, 315, 321 Anglo-Saxon Chronicle, 245, 350 Arabic, 197n14, 469, 487n22 Assumption of our Lady, 316 Audelay, John (the Blind), 445, 447, 448, 454, 460 Ayenbite of Inwyt, 166, 265, 268

Brittonic languages, 21, 97, 100–102, 246, 248, 249, 379n3 (see also Cornish; Welsh) Gaelic languages, 21, 97, 101, 379, 379n3 (see also Irish; Scottish Gaelic) Pictish, 379, 379n3 The Buke of the Howlat, 387 C

B

Bacon, Roger, 409 Bede Historia Ecclesiastica Gentis Anglorum, 245 Beowulf, 355, 357 Bible, 203, 221, 305, 470–472, 478, 485 English Prose Psalter, 267 Wycliffite Bible, 166, 173, 203, 264 Binomials, 223, 224 Black Death, 262, 428, 508, 515 Boccaccio, Giovanni Genealogia Deorum Gentilium, 220, 221 Boethius, Anicius Manlius Severinus Consolatio Philosophiae, 357 Bokenham, Osbern, 442, 445, 446, 456, 457, 460, 463 Bonedd y Saint (‘The Pedigree of the Saints’), 451 Book of Margery Kempe, 71 Bren, Llywelyn, 189–194, 196 British Celtic, 22, 26, 241, 243n5, 244–252, 262, 269, 283n2

Cambridge, Corpus Christi College, MS 140, see West Saxon Gospels Cambridge, Trinity College, MS B.14.40, 504n3, 508, 510–512, 514, 515, 517, 518 Cambridge, Trinity College, MS O.2.41, 56 Cambridge, Trinity College, MS O.3.20, 432 Cambridge, Trinity College, MS O.4.42, 444, 454, 458 Cambridge, Trinity College, MS R.3.8, 289n4, 299–303 Cambridge, University Library, MS Add. 8870, 504n3, 506, 508, 514, 515, 518, 519, 523–525 Cambridge, University Library, MS Dd.12.23, 502, 504, 504n3, 507, 508, 510, 513–516, 518 Cambridge, University Library, MS Ii.6.17, 504n3, 508, 512–521, 524, 525, 528

 Index 

Cambridge, University Library, MS Kk.4.24, 472, 472n9, 473, 477–488, 479n19, 490, 492 Cardiff, Central Library, MS 2.632 (Havod 22), 446 Cardiff, Central Library, MS 3.242, 431 Carmina Burana, 424 Castres, Bibliothèque Municipale, MS 3, 518 Caxton, William, 213–231, 409 Blanchardyn & Eglantine, 223 Book of Curtesye, 225n4 Canterbury Tales, 225n4 Charles the Grete, 218, 225n4 Dialogues in French and English, 219, 229 Eneydos, 223, 225n4 Game and Playe of the Chesse, 231 Golden Legend, 445 History of Jason, 214, 216–231 Life of St Roch, 219 Life of St Winefride, 445, 447, 448, 450, 454, 460, 461 Ovide Moralisé, 231 Recuyell of the Historyes of Troye, 215, 217, 218, 220, 225, 226, 228, 229 Reynard the Fox, 219, 225, 226, 228, 229, 231 Chartier, Alain Treatise of Hope, 351 Chaucer, Geoffrey, 111, 112, 264, 265, 409, 427, 478, 500 Canterbury Tales, 191, 267, 352, 470, 489, 516, 526 Troilus and Criseyde, 499 Cheapside, 203, 204, 206

541

Chrétien de Troyes, 195 Erec et Enide, 192 Code-switching, 113–118 Cornish, 102, 196n10, 196n12, 245, 410 Creolization, 256 Cursor Mundi, 280, 281, 289, 289n4, 298–306 D

Dafydd ab Edmwnd, 203, 203n18, 203n19 Dafydd ap Gwilym, 201, 202 Daniel’s Dreams, 317 Danish, 105, 127, 130, 132, 134, 135, 391, 392n13, 393, 395, 397 Dares Phrygius De Excidio Trojae Historia, 220 De amico ad amicam, 430n9 De Coninck, Pieter, 410 De ludis floribus, 433 Definiteness, 387, 388, 393, 394, 397n14 Delle Colonne, Guido Historia Destructionis Troiae, 220 Dialectal variation, 22, 26, 36, 37, 71, 72, 80, 105, 106, 109, 124, 126, 128, 129n7, 130–153, 240, 254, 256–260, 264, 279–281, 285–290, 289n4, 292–306, 312–316, 318–322, 324, 325, 328, 335–337, 364, 500, 512, 513 Mischsprache, 315, 320 Dieu, roy de magesté, 429

542 Index

Digital tools, 49, 50, 52–55, 57, 59–63, 65–80, 82, 83 handwritten text recognition (HTR), 58, 58n6, 58n7, 61, 62 International Image Interoperability Framework (IIIF), 53, 54, 58, 59, 62, 82 linguistically annotated corpora, 62–80, 264, 267, 268, 313, 314, 319, 319n3, 320n4, 321, 327, 337 machine learning, 59–62 metadata, 52–54, 79–81 optical character recognition (OCR), 58, 58n6, 58n7, 60–62 Text Encoding Initiative (TEI), 57, 82 Diplomatarium Norvegicum, 63, 380, 389–391, 394, 395 Donatus, Aurelius, 517, 528 Ars Minor, 218 Dublin, Trinity College, MS 277, 473, 474, 477–490, 479n19, 492, 493 Dulwich College, MS XXII, 314, 318, 319, 319n3, 323, 327, 328, 337 Dum ludis floribus, 419–421, 423–425, 429 Durham, Cathedral Library, MS A. IV.19, see Durham Collectar/Ritual Durham Collectar/Ritual, 254, 288–290, 292–298, 304–306

Dutch, 104, 128, 129n7, 130, 132, 134, 139, 140n15, 141, 149, 150, 152, 153, 379, 383 Middle Dutch, 21, 23, 30, 103, 103n8, 106, 107, 110, 215, 216, 218, 219, 225–231, 343, 344, 346, 379, 409–411, 432 Old Dutch, 149, 153 Dygon, John, 473 Dynamic Model, 67, 71 E

Edinburgh, National Library of Scotland, MS Advocates MS 19.2.1, 325, 425, 426, 433 Edinburgh, Royal College of Physicians, MS of Cursor Mundi, 289n4 Edward I, 200, 201 Edward III, 189 Edward IV, 216, 217 Edward V, 216, 217 Elizabeth of Ulster, 427 En Jesu Roy Soveraign, 430n9 Épinal Glossary, 136n13 Epistle a claudie lemperour, 414 Exultemus et letemus, 422n6 F

Falkvor Lommansson, 389, 390, 393, 394 Faroese, 135 Flemish, 21, 23, 219, 229, 410 See also Dutch

 Index 

The Flemish Insurrection, 410, 411 Floris and Blanchefleur, 429 Frankish, 103 French, 23, 24, 26, 37, 51, 74, 97, 98, 265 Middle French, 52, 53, 187–191, 208, 215–219, 222–224, 226, 227, 229–231, 446, 501, 504–514, 516–521, 523–525, 527, 528 Old French, 62, 64, 71–74, 76–80, 103, 106, 111, 159–161, 164, 164n5, 165, 167, 168, 171, 172, 174, 175, 178, 179, 183, 187, 188, 190–192, 194, 196–199, 208, 241, 251n11, 252, 260, 264, 265, 267, 269, 346–349, 358, 359, 446, 450, 487, 489 Anglo-French, 21–23, 27, 35–39, 41, 64–67, 103, 107–109, 111–117, 164, 165, 171, 174, 175, 195, 196n10, 197, 199, 260–269, 349, 351n6, 353, 354, 359–364, 367, 371–373, 408–415, 417–421, 423–434 vs. Anglo-Norman, 67–71, 260n20 Frisian, 126–135, 129n7, 137–139, 140n15, 141, 143, 149, 151–153 Old Frisian, 128, 129n7, 130, 132, 134, 140n15, 149, 152, 153, 346

543

G

Gaelic languages, 100 German Old High German, 346 Gildas De Excidio et Conquestu Britanniae, 245 Gilte Legende, 445, 446, 454, 455, 460 Glyn Cothi, Lewys, 207 Glyn, Guto’r, 204, 205 Gothic, 134, 249 Göttingen University Library, MS theol. 107, 289n4, 299, 301 Greek, 23, 26, 27, 102, 115, 191, 196, 469 Gregory the Great Dialogues, 355 Pastoral Care, 74 H

Harley lyric no. 28, 421 Harley Nominale, 487 Hebrew, 203, 469 Heimskringla, 387 Heritage language, 378, 382–385, 388, 394, 396–399 Higden, Ranulf Polychronicon, 428, 457 High German, 127, 128, 132, 134, 137, 149, 153, 219 Old High German, 134, 249 Hildina, 380–382, 388n9, 389, 390n11, 391, 393–395, 397, 399 The Historie of ane Nobil and Wailʒeand Squyer, 387

544 Index

Holcot, Robert, 428 Hundred Years’ War, 509 Hystorie van Reynaert die vos, 219, 226, 229 I

Icelandic, 130, 134, 135, 152, 153 Old Icelandic (see Old Norse) In dulci jubilio, 433 Intrascribal variation, 313, 315, 319–338 Irish, 102, 106, 107, 116 Old Irish, 250 Italian, 193, 410 J

John of Tynemouth Sanctilogium, 444, 447, 454, 455, 458 K

Kemp, John, 514 King Horn, 429 Kingsmill, William, 510–512, 511n4, 517 Koineization, 19, 28, 131, 132, 137–140, 140n15, 143, 144, 149, 151, 254–258, 260, 269, 306, 324 L

La estorie del evangelie, 63, 256, 313–317, 319–338 La Queste del Saint Graal, 264

La Somme le Roi, 267, 268 Laȝamon Brut, 321, 353 Lambeth Homilies, 357 Langland, William, 428 See also Alliterative Revival Language mixing, 35–37, 39, 41, 410, 411, 417, 419, 423, 424, 427, 429, 430, 432–434 code-switching, 25, 40, 410, 411, 417, 419–421, 423, 424, 429–434, 467, 468, 470–480, 482–493 Language shift, 19, 21 Latin, 21–24, 26, 35, 37, 38, 40, 51, 53, 55, 56, 60–62, 102, 103, 107–109, 111–117, 151, 159, 160, 164n5, 165, 166, 168, 171, 173, 174, 178, 191, 196, 197, 199, 215, 218, 219, 241, 242, 242n2, 244, 261, 263, 379, 408–415, 417–421, 423, 424, 427, 429–434, 442–463, 467–472, 474–480, 482–493, 510, 513, 528 Laurent d’Orléans La somme le roi, 265, 430 Le jongleur d’Ely e le roi d’Angleterre, 429 Lefèvre, Raoul Histoire de Jason, 216–222, 222n2, 224, 230 Receuil des histoires de Troyes, 217, 218, 220 Legende doree, 218

 Index 

Lexical replacement, 160, 162, 168, 172, 173, 179–183 Life of St Chad, 265 Lincoln, Lincolnshire Archives Office, Formulary Book 23, 501, 504n3, 507, 508, 512, 514, 515, 518, 528 Lindisfarne Gospels, 254, 258, 288–290, 292–298, 304–306 Linguistic transfer adstratum vs. substratum vs. superstratum, 19, 20, 22, 26, 27 borrowing vs. copying, 67n12 borrowing vs. imposition, 20, 241–243, 247, 247n8, 253, 255, 256, 258, 262, 263, 269, 283, 284 borrowing scale, 22, 23, 255, 281, 282, 284 lexical, 18, 19, 21–27, 37, 38, 63–80, 99–101, 103, 104, 106–118, 123–153, 159–183, 187, 188, 190–199, 202, 205–208, 216, 222, 223, 226–228, 231, 240, 247, 253, 255, 256, 258, 261, 262, 265–269, 280, 282, 284, 288, 294, 300, 303, 305, 311, 312, 324, 337, 344, 358, 359n7, 371–373, 410, 417, 476, 476n17, 485–489, 488n23, 491 morphosyntactic, 19, 21–26, 63, 125, 228, 229, 239–241, 243, 244, 248, 252,

545

254–256, 260, 262, 266, 267, 269, 311, 312, 324, 337, 344 argument structure, 64–69, 261, 266–268 genitive/of-objects, 67, 69–71 lability, 66, 266, 267, 267n21, 269, 344–367, 369–373 reflexivity, 71–80 secondary predicates, 67, 68, 71 cleft constructions, 248 comparative nor, 244, 248, 250–252 complexification vs. simplification, 241, 243, 244, 252, 255, 257, 258, 283, 306 copulas, 239, 248–250, 257 do-periphrasis, 248 nominal morphology, 254, 256, 257, 279–282, 284–289, 292–306, 419 Northern Subject Rule, 248, 258 possessive constructions, 63, 248, 378, 385–391, 393–398 progressive aspect, 248 relativiser sum, 26 verbal morphology, 229, 256, 258, 259 word order, 240, 259–266 orthographic, 224, 225, 419 phonological, 19, 21 Lithuanian, 107 Liturgy, 445–448, 461 Livre des mestiers, 219

546 Index

Llywarch ap Llywelyn, 195, 197 Llywelyn ap Gruffudd, 200 London, British Library, MS Add. 17716, 504n3, 510, 513 London, British Library, MS Add. 35298, 445 London, British Library, MS Add. 46919, 413n3 London, British Library, MS Cotton Claudius A II, 446 London, British Library, MS Cotton Caligula A IX, 321 London, British Library, MS Cotton Nero D IV, see Lindisfarne Gospels London, British Library, MS Cotton Otho C XIII, 321 London, British Library, MS Cotton Tiberius A VI, 56 London, British Library, MS Cotton Vespasian A III, 289n4, 299–303 London, British Library, MS Cotton Vitellius D III, 429 London, British Library, MS Harley 913, 408, 409 London, British Library, MS Harley 2253, 407, 409–415, 417–421, 423–427, 429, 431–434 London, British Library, MS Harley 2268, 468, 470 London, British Library, MS Harley 3988, 504n3, 505, 507, 508, 512, 514, 515, 518, 519, 528 London, British Library, MS Lansdowne 388, 317–319, 323, 327, 329–332

London, British Library, MS Lansdowne 436, 444, 452, 454, 459 London, British Library, MS Royal 17 C XVII, 317–319, 319n3 London, Guildhall Library, MS 15838, 432 London, Mercers’ Company, Book of Privileges, 432 London, University of London Library, MS Sterling V. 17, 316–318, 319n3, 320, 320n5, 322–326, 325n6, 328–338, 330n7 Low German, 103, 107, 127, 130, 131, 149, 150, 152, 344, 379 Middle Low German, 21, 23, 130, 137, 153, 228 Old Saxon, 134, 135, 149, 150, 249, 346 M

Mabinogion, 195–198, 199n17 Macaronic texts, see Language mixing Malory, Thomas Le Morte Darthur, 263 Mandeville’s Travels, 316 Manières de langage, 429, 501, 502, 504–508, 504n3, 510–514, 516–521, 523–525, 527, 528 Mannyng, Robert Handlyng Synne, 315, 316, 322 Mayden, moder milde, 429 Meditations on the Supper of our Lord, 316

 Index 

Michel, Dan, of Northgate, 430 Minot, Laurence Siege of Tournay, 426 Mirk, John, 445 Festial, 445–447, 452, 454, 461–463 Mirror of St Edmund, 265 Mise en page, 39, 412, 414, 423, 424, 434, 468, 473, 479n19, 483–485, 492, 493, 500, 503, 503n1, 506, 516, 518–521, 523–525, 527, 528 Morphosyntax nominal, 258 Mynde, Thomas, 448 N

New dialect formation, 27–29, 34–39, 41 Multicultural London English, 30–33 Norn, 63, 377–399 Northern Homily Cycle, 199 Northern Passion, 314, 319 Norwegian, 63, 99, 106, 135, 378, 386, 389, 390, 393, 396–398, 397n14 American Norwegian, 383, 385, 396, 397, 399 O

Of Arthur and of Merlin, 351 Óláfs saga ins helga, 386 Old English laws, 350 Old English Martyrology, 73

547

Old Norse, 21–24, 26, 27, 29, 35, 97–99, 103–107, 103n8, 109, 123–153, 160, 164, 166, 171, 199n16, 241, 243n5, 244, 249, 252–260, 262, 264, 269, 279–282, 284–289, 292–306, 311, 312, 324, 337, 359n7, 377–379, 381, 385–388, 390, 393, 395–399, 397n14 Older Scots, 196n10, 199 Ormulum, 26, 260, 319, 327, 330, 334, 337 Orthographia Gallica, 510 Owain, Gutun, 203 Oxford, All Souls College, MS 182, 501, 504n3, 508, 510, 513–515, 518, 519, 521, 522 Oxford, Balliol College, MS 149, 472–480, 472n8, 474n14, 479n19, 482–488, 490, 492 Oxford, Bodleian Library, MS Additional C 38, 316–318, 319n3, 320, 320n5, 322 Oxford, Bodleian Library, MS Auct. D. 2. 19, see Rushworth Gospels Oxford, Bodleian Library, MS Bodley 649, 468 Oxford, Bodleian Library, MS Bodley 779, 445, 446, 448, 454, 460 Oxford, Bodleian Library, MS Digby 86, 413 Oxford, Bodleian Library, MS Douce 302, 447

548 Index

Oxford, Bodleian Library, MS Eng. Poet a.1, 313, 315–318, 319n3, 320, 322–324, 338 Oxford, Bodleian Library, MS Fairfax 14, 289n4, 299, 301, 302 Oxford, Bodleian Library, MS lat. misc. e.93, 504n3, 508, 514, 515, 518 Oxford, Bodleian Library, MS Laud misc. 416, 301 Oxford, Bodleian Library, MS Rawlinson C 655, 314, 315, 318, 322, 323, 325–332, 334, 335, 337 Oxford, Bodleian Library, MS Selden B.26, 430n9 Oxford, Magdalen College, MS lat. 93, 473, 473n10, 476–488, 479n19, 490, 492 P

Paris, Bibliothèque Nationale, MS fr. 331, 217 Paris, Bibliothèque Nationale, MS fr. 12148, 222n2 Paris, Bibliothèque Nationale, MS nouv. acq. lat. 699, 504n3, 508, 512–515, 518, 528 Paston Letters, 116, 326, 328, 330 Persian, 197n14 Place-names, 95–101, 128, 446, 448, 450, 452–454, 456, 457 Poema Morale, 199 Prayer for Protection, 429 Prayer on the Five Joys of Our Lady, 429

Preaching, 467–473, 475–480, 482–487, 491–493 Princeton University Library, R. H. Taylor MS 10, 316 Pronouns, 239 demonstrative, 255, 312, 321, 324, 326, 327, 329, 333, 337 personal, 26, 63, 125, 151, 254–256, 311–313, 317–338, 319n3, 386, 386n8, 387, 389, 395–398, 428 reflexive, 71–80, 248, 344, 357, 358, 361, 362, 371, 372, 387, 388, 394, 395, 397, 398 relative, 22, 248 R

Responcio, 430n9 Rhygyfarch of Llanbadarn Fawr, 442 Richard II, 507 Robert of Shrewsbury, 443–445, 443n2, 449–454, 456–463 Rolle, Richard, 315 Ego Dormio, 322 Roman de la Rose, 190, 190n3 Romaunt of the Rose, 352 Rule of St Benet (prose), 259, 260 Rushworth Gospels, 254, 288–290, 292–298, 304–306 S

St Beuno, 451, 454, 455, 457 St David, 442 St Teilo, 442 St Winefride, 443, 445, 446, 455, 462

 Index 

Saints’ Lives, 441, 442, 455 St Beuno, 443n2, 444, 449, 451–453, 459 St David, 442, 445, 455 St Teilo, 442, 455 St Winefride, 443–463 Salesbury, William, 203n19 Sampson, Thomas, 510, 512 Scots, 97–101, 130, 137, 149, 378, 380, 383, 384, 393, 395 Older Scots, 98, 99, 105, 244, 248, 250–252, 378–380, 383, 385, 387, 388, 393, 395–398 Scottish Gaelic, 106 Semantic hierarchy, 161–183, 164n5 Semantic shift, 37, 99, 100, 104, 105, 131, 139, 144, 150, 160–164, 161n1, 170, 172, 174, 176, 177, 179–181, 223, 226–228 Sir Tristrem, 426 Slavonic languages, 107 South English Legendary, 316, 320, 321, 445 Spanish, 487n22 Standardisation, 21, 33–39, 41, 124, 162, 163, 182, 183, 213–215, 222, 230 Swedish, 105, 135 T

Technical language, 38–40, 100, 112, 114, 117, 160–183, 187–199, 202, 205–208 Trade, 36, 100, 163–177, 182, 188, 193, 200, 204, 205, 410, 509 cloth, 110, 165, 169, 187–199, 201–203, 205–208, 214, 219, 509

549

Translation, 79, 80, 203, 216–231, 264, 265, 268, 426, 442–463 Trevisa, John, 215, 428 U

Utrecht, University Library, MS Hattem C5, 227n8 V

Vegetius, Flavius De Re Militari, 351 Vitas patrum, 414 W

The Walling of New Ross, 408 Wars of Alexander, 352 Washington, D. C., Folger Library, MS V.b.236, 316 Welsh, 101, 102, 110, 111, 245, 246, 249, 250 Middle Welsh, 187–202, 204–208, 251, 252, 409, 410, 430–434, 441–463 Old Welsh, 249, 251, 252 Welsh Marches, 97, 187, 189–194, 196, 199, 200, 202, 205–208, 410, 463 West Saxon Gospels, 288–290, 292–298, 304, 305 Wey, William, 525 William of Shoreham, 357 Y

Ypotis, 320