Loss and Renewal: Australian Languages Since Colonisation 9781614518792, 9781614518877

Citation preview

Felicity Meakins and Carmel O’Shannessy (Eds.) Loss and Renewal

Language Contact and Bilingualism

Editor Yaron Matras

Volume 13

Loss and Renewal

Australian Languages Since Colonisation

Edited by Felicity Meakins and Carmel O’Shannessy

ISBN 978-1-61451-887-7 e-ISBN (PDF) 978-1-61451-879-2 e-ISBN (EPUB) 978-1-5015-0103-6 ISSN 2190-698X Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. 6 2016 Walter de Gruyter Inc., Boston/Berlin Cover image: Anette Linnea Rasmus/Fotolia Typesetting: RoyalStandard, Hong Kong Printing and binding: CPI books GmbH, Leck ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Pukaka-wu For Patrick

Acknowledgments This volume stems from the Language Contact Symposium held at the Australian National University 6–7 March 2014. We would like to thank Jeff Siegel and Harold Koch who acted as discussants at the symposium. They provided feedback to all of the presenters before they submitted papers for review. We are also immensely grateful to all of the reviewers (in alphabetical order) who provided detailed comments on the papers in the volume: Judith Bishop, Grev Corbett, Alan Dench, Nick Evans, Alice Gaby, Diana Guillemin, Luise Hercus, Harold Koch, Rachel Nordlinger, Erich Round, Eva Schultze-Berndt, Jeff Siegel, Ruth Singer, Jeffrey Steele, Marine Vuillermet, and Don Winford. We also thank Brenda Thornley for her work on many of the maps, Cameron Bothner, Elizabeth Gall and Leah Zekelman for formatting work and Nahyun Kwon for the indexing and proof checking. Finally, we thank the contributors, for their own papers and comments on other parts of the volume.

Table of contents Acknowledgements List of contributors xii Maps xiv List of figures xvii List of tables xx Preface

I

v xi

Introduction Carmel O’Shannessy and Felicity Meakins Australian language contact in historical and synchronic perspective

3

II Transfer of form: Structure 1

Nicholas Evans As intimate as it gets? Paradigm borrowing in Marrku and its implications 29 for the emergence of mixed languages

2

Ilana Mushin and Janet Watts Identifying the grammars of Queensland ex-government reserve varieties: 57 The case of Woorie Talk

III Transfer of form: Lexical 3

4

5

Patrick McConvell Kinship loanwords in Indigenous Australia, before and after 89 colonisation David Nash Placenames from NSW Pidgin: Bulga, Nyrang

113

Greg Dickson Rethinking the substrate languages of Roper Kriol: The case of 145 Marra

x

Table of contents

IV Transfer of form: Phonological 6

Rikke Bundgaard-Nielsen and Brett Baker The continuum in Kriol: Fact or furphy?

7

Carmel O’Shannessy Entrenchment of Light Warlpiri morphology

177

217

V Transfer of function, structure, distribution and semantics 8

Denise Angelo and Eva Schultze-Berndt Beware of ‘bambai’ – soon it may turn apprehensive

9

Maïa Ponsonnet Reflexive, reciprocal and emphatic functions in Barunga Kriol

255

297

Sophie Nicholls 10 Grammaticalization and interactional pragmatics: A description of the 333 recognitional determiner det in Roper River Kriol

VI (Further) Development of new structures 11

Felicity Meakins No fixed address: The grammaticalisation of the Gurindji locative as a 367 progressive suffix

John Mansfield 12 Light verb structure in Murrinh-Patha

397

Felicity Meakins and Rob Pensalfini 13 Gender Bender: Superclassing in Jingulu gender marking

Index

453

425

List of contributors Denise Angelo Australian National University ARC Centre of Excellence for the Dynamics of Language [email protected]

Ilana Mushin University of Queensland ARC Centre of Excellence for the Dynamics of Language [email protected]

Brett Baker ARC Centre of Excellence for the Dynamics of Language University of Melbourne [email protected]

David Nash Australian National University [email protected]

Rikke Bundgaard-Nielsen La Trobe University [email protected]

Sophie Nicholls University of Western Sydney ARC Centre of Excellence for the Dynamics of Language [email protected]

Greg Dickson University of Queensland ARC Centre of Excellence for the Dynamics of Language [email protected]

Carmel O’Shannessy University of Michigan ARC Centre of Excellence for the Dynamics of Language [email protected]

Nicholas Evans Australian National University ARC Centre of Excellence for the Dynamics of Language [email protected]

Rob Pensalfini University of Queensland r.pensalfi[email protected]

John Mansfield University of Melbourne ARC Centre of Excellence for the Dynamics of Language jbmansfi[email protected] Patrick McConvell Australian National University [email protected] Felicity Meakins University of Queensland ARC Centre of Excellence for the Dynamics of Language [email protected]

Maïa Ponsonnet French National Centre for Scientific Research (Dynamique du Langage) [email protected] Eva Schultze-Berndt University of Manchester [email protected] Janet Watts University of Queensland janet.watts@griffithuni.edu.au

xii

List of contributors

Figure 1: Languages of Australia discussed in the volume (Brenda Thornley 2015)

List of contributors

Figure 2: Languages of northern Australia discussed in the volume (Brenda Thornley 2015)

xiii

List of figures Figure 1 Figure 2

Languages of Australia discussed in the volume xii Languages of northern Australia discussed in the volume

xiii

Preface Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6

xx Patrick McConvell with his parents as a child Pincher Nyurrmiari and Patrick McConvell at Daguragu in the mid xxi 1970s xxii McConvell teaching at SAL in the 1980s McConvell in 1984 discussion the establishment of the xxiii KLRC xxiv McConvell in the mid 1980s working at Turkey Creek McConvell at the Gurindji dictionary launch in 2013 at Kalkaringi with Violet Wadrill, Topsy Dodd Ngarnjal, Marie Japanngali and xxv others

Chapter 1 Figure 1 Figure 2 Figure 3

32 Languages of the Cobourg Peninsula region Proposed relations between languages of the Iwaidjan 33 family Monolingual and interlingual constructional elements in a 50 multilingual setting

Chapter 2 Figure 1

Locations of Australian languages and communities mentioned in 60 this chapter

Chapter 3 Figure 1 Figure 2

Map of spread of Wanderwort ramparr > lamparr(a) 95 Tyamuny

94

Chapter 4 Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

121 Distribution of Cabonne and Nyrang etc placenames 123 Placenames involving bulga Looking south to The Bulga ~ Black Bulga ~ Bulga Mountain ~ 124 Black Bulga Mountain ~ Moolagundi 126 Gullen placenames 128 Path of pastoral industry and NSW Pidgin

List of figures

Figure Figure Figure Figure Figure Figure Figure

6 7 8 9 10 11 12

xv

Placenames with budgery etc. and merrijig 129 130 Distribution of yarraman and nantu ‘horse’ 131 Distribution of placenames with yarraman 132 Distribution of placenames with blackfellow 133 Placenames with black gin ~ blackgin, and lubra 134 Placenames with piccaninny, piccaninnie, picaninny 135 Placenames with sugarbag

Chapter 5 Figure 1

Language of origin of 60 commonest non-English based verbs in 166 Roper Kriol

Chapter 6 Figure Figure Figure Figure Figure Figure

1 2 3 4 5 6

Figure 7 Figure 8 Figure 9

194 Voice Onset Time (VOT) in Kriol speaker PT 194 Intervocalic Constriction Duration (CD) in Kriol speaker PT 196 VOT in naturally occurring speech 196 Constriction Duration (CD) in naturally occurring speech 199 Word-initial Kriol VOT from a Gen1 speaker Correct medial stop discrimination scores for Kriol 204 speakers Correct initial fricative discrimination scores for Kriol 205 speakers Mean word initial and word-medial voiceless and voiced VOT across 208 stop place of articulation Intervocalic constriction duration (CD) across place of 209 articulation

Chapter 7 Figure 1 Figure 2

224

Locations of Warlpiri communities, Northern Territory 247 The Monster Story

Chapter 8 Figure 1

Apprehensive subtypes identified by Lichtenberk (1995)

Chapter 9 Figure 1

Picture used in elicitation

313

Chapter 11 Figure 1

Traditional languages of the Victoria River District

371

258

List of figures

xvi

Chapter 12 Figure 1 Figure 2

Simple model of the phrasal/synthetic diachronic cycle 406 The Murrinhpatha verb template

402

Chapter 13 Figure 1 Figure Figure Figure Figure Figure

2 3 4 5 6

Figure 7

Hierarchy of gender disagreement across Australian 427 languages 430 Languages and their usage in the Barkly Region 439 Gender superclassing in Jingulu 440 Gender superclassing in Jingulu 441 Number superclassing in Jingulu Hierarchy of gender agreement superclassing across Australian 446 languages 447 Hierarchy of gender agreement superclassing in Gaagudju

List of tables Chapter 1 Table 1 Table 2 Table 3

Body-part terms with singular possessors 45 45 Reflexive pronouns, singular forms Verbal subparadigms and excerpted pronominal prefixes for 48 five Marrku verbs

Chapter 3 Table 1 Table 2

Father’s father in Ngumpin-Yapa Nyulnyul grandparental terms

95 100

Chapter 4 Table 1 Table 2

114 Australian placename sets after European Contact Australian placename sets with Pidgin placename sets shaded light 116 grey

Chapter 5 Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7

Active and passive fluency in Ngalakgan and Ngandi in 151 2001 Local Aboriginal language food names occurring in Joynt 157 (1918) 158 Kriol kinterms derived from languages other than English 159 Kriol kinterms and equivalents in Marra and Alawa Occurrence of Kriol kinterms in traditional languages of the Roper 160 Region Summary of etymologies of sixty common non-English based Kriol 165 verbs Etymology of non-English based names of most salient bush 168 medicines known to young Kriol speakers

Chapter 6 Table 1 Table 2 Table 3 Table 4

181 Phoneme inventories in the Roper substrate languages 189 The obstruent inventory of Roper Kriol 190 Examples of correspondences between Kriol and English (American) English stop consonant Constriction Duration and VOT 193 durations, and standard deviations, in milliseconds

xviii Table 5 Table 6

List of tables

Number of tokens, mean VOT, standard deviation and p values for 197 voiced-voiceless comparisons Speaker GN’s wordlist, with standard Kriol and GN’s pronunciation 215 variants (in a phonemic representation)

Chapter 7 Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9

228 Case-marker forms in classic Warlpiri 229 Case-marker forms in contemporary Warlpiri Number of ergative case marker tokens in Light Warlpiri texts, 2005 236 and 2010, AV and VA word orders Fisher’s exact test: Ergative case marking in Light Warlpiri, ages 7 237 and 9, 2005 and 2010 Number of ergative case marker tokens in Warlpiri texts, 2005 and 238 2010, AV and VA word orders 239 Ergative allomorphs in each language, 2005 and 2010 Number of prepositional and suffix forms of dative case in Light 239 Warlpiri, 2005 and 2010 Frequencies of dative case forms in each language, 2005 and 2010, 240 with and without a final vowel Transitive verbs appearing in the Light Warlpiri texts in 2005 and 249 2010

Chapter 8 Table 1 Table 2

Bambai in temporal function: Range of meaning and clause posi272 tion according to dialect Summary of apprehensive and temporal functions in clause initial 276 and final position

Chapter 9 Table 1 Table 2 Table 3 Table 4 Table 5

302 Participants in the study 306 Some verbs found in mijelp reflexive constructions 307 Some verbs found in mijelp reciprocal constructions Some verbs found in gija reciprocal constructions (non-exhaustive 309 list) 319 Realignment of reflexive and reciprocal categories

Chapter 10 Table 1

Kriol demonstratives

344

List of tables

Chapter 11 Table 1 Table 2

Summary of continuative marking in Gurindji Kriol 385 387 Summary of the verb template in Gurindji Kriol

Chapter 12 Table 1 Table 2 Table 3

411 Inherited coverbs Verb structure types in older and younger speaker 420 samples 420 Individual verb usage by three young men

Chapter 13 Table 1

Patterns of gender agreement superclassing across Australian 445 languages

xix

Preface

Figure 1: Patrick McConvell with his parents as a child

We are very pleased to dedicate this volume to Patrick McConvell who has contributed enormously to our understanding of how Australian languages have changed, both historically and since colonisation, as a result of language contact processes. This overview of McConvell’s work life makes reference to the exhaustive bibliography at the end of the preface. Particular papers and their contributions to language contact and change in Australia are discussed before the bibliography. McConvell was born at Isleworth Hospital in western London on 11 April 1947. He was the only child of Betty (Mason) and William McConvell of Twickenham. McConvell was always interested in languages and, as a child, he would invent his own languages. McConvell also always loved travel, encountering different cultures and learning different languages. He was enthralled as a boy by the stories his aunt Peg and uncle Harry told him when they would return to England for holidays from their home in Nigeria which perhaps prompted his later PhD work on Hausa. In McConvell’s high school holidays, he twice travelled with school friends through France, once as a cycling holiday and once as a hitchhiking holiday. McConvell received a (high school) scholarship to study at Hampton Grammar School where he completed A levels in Latin, French and German. McConvell was offered a scholarship to study Arabic at Cambridge University but he declined the offer in favour of studying at the University of London (SOAS) where he wanted to study linguistics and anthropology. He was awarded his BA (Hons) in 1969 and later his PhD in 1973 for his thesis Cleft Sentences in Hausa? A Syntactic Study of Focus which was supervised by David Arnott and Neil Smith. This was a time in linguistics when the Chomskian ‘turn’ was taking

Preface

xxi

Figure 2: Pincher Nyurrmiari and Patrick McConvell at Daguragu in the mid 1970s

hold and linguistics was forgetting its anthropological roots. Throughout this shift, McConvell remained an anthropological linguist but one who also engaged with formal grammatical theory. Although he did not continue his study of Chadic languages, his PhD work on information structure provided the foundation for early foundational work in Australian languages, specifically on pronominal clitic placement (1980a, 1996a) and the restrictive clitic =rni (1983a) in NgumpinYapa languages. McConvell was considered for a job at SOAS but some of his political activities disqualified him and so he successfully applied for a job at the Australian Institute of Aboriginal Studies (AIAS, now AIATSIS) in Canberra. From 1973–77, McConvell worked as a Research Fellow in Linguistics where he first encountered Australian languages. He was one of six Research Fellows attracted from overseas, the others being Jeffrey Heath, Paul Black, Alan Rumsey, Francesca Merlan, Frank Wordick and later Michael Silverstein. They joined Peter Sutton who was the Linguistics Research Officer at AIATSIS (before Michael Walsh replaced him in 1975) and who also combined an interest in linguistics and anthropology. During this time, McConvell conducted two years of fieldwork at Daguragu in northern Australia on Gurindji, Mudburra and Bilinarra. His recordings and transcriptions which consist of around 60 hours of elicitation, narratives, conversations and procedural texts are the earliest comprehensive documentation of

xxii

Preface

Figure 3: McConvell teaching at SAL in the 1980s

these languages, with their value to language communities and linguists immeasurable. They also form the basis of a number of dictionaries, the most substantial being Gurindji to English Dictionary (2013), and numerous papers on language contact, historical linguistics and kinship (see ‘Ngumpin-Yapa languages’ below). During his time at Daguragu, McConvell became involved the Aboriginal land rights movement which was centred around a number of groups, including the Gurindji. The Gurindji had walked off a Wave Hill Station in 1966 in protest against their treatment in the cattle industry. This protest, now referred to as the Gurindji Walk-Off, transformed into a bid to regain control of their traditional lands. In 1975, the then Australian Prime Minister, Gough Whitlam went to Daguragu and made a commitment to return their land to the protest leader, Vincent Lingiari, and a crowd of Gurindji activists, their families and non-Indigenous supporters from the North Australian Workers Union, other unions, the Communist Party of Australia and the Australian Labor Party. Vincent Lingiari’s response (1986 [1975]), which remains one of the important political speeches of this era, was delivered in Gurindji and McConvell’s translation, which has appeared in many activist contexts, gave the broader Australian public access to Lingiari’s words. Subsequent to Whitlam’s visit to Daguragu, the Aboriginal Land Rights (Northern Territory) Act (1976), which provides the basis upon which Aboriginal people can claim rights to traditional land, was legislated by the Fraser government. McConvell worked on the first land claim conducted by the Central Land Council (CLC) as the anthropologist for the Gurindji claimants to areas of their traditional lands (Peterson et al. 1978). He was also the principal anthropologist on seven other land claims and native title cases (see ‘Native title and land claims’ below) in the Victoria River District (VRD) which together form a comprehensive and coherent ethnography of Gurindji, Malngin, Nyininy, Bilinarra,

Preface

xxiii

Figure 4: McConvell in 1984 discussion the establishment of the KLRC

Ngarinyman and Mudburra people including their interconnected Dreamings, and shared land tenure and kinship systems. McConvell, referred to as Pukaka1 by Aboriginal people in the VRD, remains highly respected in this region for his contribution to the repatriation of traditional lands. McConvell was also supportive of programs aimed at revitalising Aboriginal languages and cultural practices. From 1978–79, he worked at Strelley Community (Pilbara, WA) as a linguist supporting the school bilingual program. From 1979 to 1986, he was a Lecturer at the School of Australian Linguistics (SAL) (Batchelor, NT) where he worked with Aboriginal students from widespread communities, facilitating their study of many aspects of their languages, from developing and teaching writing systems to analysing the linguistic workings of their languages. He also taught NAATI courses to qualify Aboriginal people to interpret in courts, hospitals etc. In 1984, he and Joyce Hudson wrote the seminal report Keeping Language Strong which sounded the alarm about language endangerment in Australia. He went on leave from SAL at this time and, together with Joyce Hudson and Peter Yu (Kimberley Land Council), helped establish the Kimberley Language Resource Centre which continues to service the language needs of Gija, Walmajarri and Jaru people (among others). From 1985 to 1986, McConvell worked at the Turkey Creek School (WA) in its bilingual program developing materials, training staff and preparing a Gija dictionary and a working grammar. During this time, he produced many papers advocating 1 Pukaka is the name of an important site for Ngarinyman people on the Wickham River in Judbarra/Gregory National Park.

xxiv

Preface

Figure 5: McConvell in the mid 1980s working at Turkey Creek

for the value of bilingual education (listed below in ‘Language documentation, endangerment and education’). When SAL was merged with Batchelor College in 1987, McConvell began lecturing in anthropology and linguistics, first at the University of Queensland’s Darwin branch, the University College Northern Territory (UCNT) (1987–1989), then at the Northern Territory University (now Charles Darwin University) (1989– 1999) and finally at Griffith University in Brisbane (1998–1999). In 1991 during his appointment at UCNT, he held the first ARCLING (Archaeology and Linguistics) conference from which he and Nick Evans edited a volume of works in 1997. This interest in archaeology and language continued, as detailed in ‘Linguistics and archaeology’ below. McConvell moved back to Canberra in January 2000 where he was a Research Fellow at AIATSIS (2000–2008) and most recently at ANU (2008–2012). During his time at AIATSIS, he was involved in setting up and supporting the Australian Research Council-funded Aboriginal Child Language project and associated PhD projects, ultimately leading to some of the research in this volume, notably Meakins and O’Shannessy’s work. Later at ANU, he returned to his first love of anthropological linguistics, particularly the social processes underlying historical language change and tracing change in kinship organisation using linguistic evidence. Numerous papers and collaborations have stemmed from this work, as detailed below. Over his career, McConvell has written a number of seminal papers. Here we focus on those which have influenced our understanding of how Australian languages have evolved, particularly as a result of contact with other languages.

Preface

xxv

Figure 6: McConvell at the Gurindji dictionary launch in 2013 at Kalkaringi with Violet Wadrill, Topsy Dodd Ngarnjal, Marie Japanngali and others

McConvell brought his background in anthropological linguistics to the task of language description and was one of the first people in Australia to begin documenting community language practices in his work on code-switching between Gurindji dialects and Kriol (1985b; 1988c). This research provided the basis for McConvell and Meakins (2005) who claimed that the mixed language, Gurindji Kriol, is derived from these code-switching practices, refuting previous claims that mixed languages could not find their origins in normal language contact processes. McConvell’s work on Gurindji and language contact also looks back into the historical past. His 2009a paper is a careful lexico-semantic case study of Gurindji which shows an unusually high borrowing rate of 45.6% which he details by word class, semantic domain, lexical semantics and structure. Despite the fact that high rates of borrowing have been claimed by R. M. W. Dixon to create too much noise in the historical signal, McConvell and Laughren (2004) convincingly justify the Ngumpin-Yapa language family. This interest in language contact extends to discoveries of how complex kinship systems have evolved in Australia. McConvell (1985a, b) shows how the eight term subsection system, which classifies members into socio-centric categories necessary for understanding descent and marriage relations, developed from contact between two linguistic systems across northern Australia. This work is an excellent example of how bilingualism and contact scenarios can lead to the complexification of pre-existing linguistic and kinship systems.

xxvi

Preface

McConvell’s work on language change in Australia extends beyond NgumpinYapa languages. His 1981 paper on how Lardil became accusative is significant because it argued that Ken Hale’s analysis about the direction of syntactic change from Active > Passive > Ergative could not be sustained given the Lardil data which showed ERG-ABS > NOM-ACC > Passive. This paper also laid the foundations for some of Nick Evans’ work on language change in a related language, Kayardild, including how insubordination worked to generate different structures. McConvell’s 1976 paper on nominal hierarchies in Yukulta also made an important contribution to our understanding of the role of feature hierarchies in determining voice (cf. Silverstein’s animacy hierarchy in the same volume).

Ngumpin-Yapa languages 1980a. Hierarchical variation in pronominal clitic attachment in the Eastern Ngumpin languages. In Bruce Rigsby & Peter Sutton (eds.), Papers in Australian linguistics no. 13, 31–117. Canberra: Pacific Linguistics. 1983a. “Only” and related concepts in Gurindji. Batchelor: Unpublished manuscript. 1986 (1975). Translation of Lingiari, Vincent. Vincent Lingiari’s Speech. In Luis Hercus & Peter Sutton (eds.), This is what happened: Historical narratives by Aborigines, 313–315. Canberra: AIAS Press. 1986 (1975). Translation of Kijngayarri, Long Johnny. The Wave Hill strike. In Luis Hercus & Peter Sutton (eds.), This is what happened: Historical narratives by Aborigines, 305–311. Canberra: AIAS Press. 1988a. Nasal cluster dissimilation and constraints on phonological variables in Gurindji and related languages. Aboriginal Linguistics 1. 135–165. 1996a. The functions of split-Wackernagel clitic systems: Pronominal clitics in the Ngumpin languages. In Aaron Halpern & Arnold Zwicky (eds.), Approaching second: Second position clitics and related phenomena, 299–332. Stanford CA: CSLI Publications. 1999. with Mary Laughren. Down under in central Australia. Paper presented at the American Anthropological Association Meeting, Chicago. 2002a. Mix-im-up speech and emergent mixed languages in Indigenous Australia. Texas Linguistic Forum (Proceedings from the 9th Annual Symposium about Language and Society) 44(2). 328–349. 2004. with Mary Laughren. Ngumpin-Yapa Languages. In Harold Koch & Claire Bowern (eds.), Australian languages: Reconstruction and subgrouping, 151–177. Amsterdam: Benjamins. 2005a. Gurindji grammar. Canberra: Unpublished manuscript. 2006a. Grammaticalization of demonstratives as subordinate complementizers in NgumpinYapa. Australian Journal of Linguistics 26(1). 107–137. 2009a. Loanwords in Gurindji, a Pama-Nyungan language of Australia. In Martin Haspelmath & Uri Tadmor (eds.), Loanwords in the world’s languages: A comparative handbook, 790– 822. Berlin: Mouton de Gruyter. 2009b. Where the spear sticks up: The variety of locatives in placenames in the Victoria River District, Northern Territory. In Harold Koch & Luise Hercus (eds.), Aboriginal place names: Naming and re-naming the Australian landscape, 359–402. Canberra: ANU Research School of Pacific Studies.

Preface

xxvii

2010a. Mood swings: Imperative verbs attract pronominal enclitics in Ngumpin-Yapa (Australian) and Southern European languages. In Rachel Hendery & Jennifer Hendriks (eds.), Grammatical change: Theory and description, 123–156. Canberra: Pacific Linguistics. 2012. with Jane Simpson. Fictive motion down under: The locative-allative case alternation in some Australian Indigenous languages. In Diana Santos, Krister Lindén and Wanjiku Ng’ang’a (eds.), Shall we play the Festschrift game?: Essays on the occasion of Lauri Carlson’s 60th birthday, 159–180. Berlin: Springer. 2013. with Felicity Meakins, Erika Charola, Norm McNair, Helen McNair & Lauren Campbell (compilers). Gurindji to English dictionary. Batchelor, Australia: Batchelor Press.

Linguistics (other than Ngumpin-Yapa) 1976. Nominal hierarchies in Yukulta. In R.M.W. Dixon (ed.), Grammatical categories in Australian languages, 191–200. Canberra: AIAS. 1977. Relativisation and the ordering of cross-reference rules in Hausa. Studies in African Linguistics 8. 1–31. 1980b. Manjiljarra wangka: Manjiljarra-English dictionary. Darwin: Strelley Literacy Centre/SAL. 1981. How Lardil became accusative. Lingua 55. 141–179. 1983. with Ron Day & Paul Black. Making a Meriam Mir dictionary. In Peter Austin (ed.), Papers in Australian linguistics no 15: Australian Aboriginal lexicography, 19–30. Canberra: Pacific Linguistics. 1985a. The role of Aboriginal languages in story. Australian Aboriginal Studies 2. 74–76. 1988b. To be or double be: changes in the English copula. Australian Journal of Linguistics 8(2). 287–305.

Language documentation, endangerment and education 1982a. Supporting the two-way school. In Jeanie Bell (ed.), Language planning for Australia Aboriginal languages, 60–76. Alice Springs: IAD Press. 1983b. Domains and domination. Aboriginal Languages Association Newsletter 5. 8–14. 1983c. Aboriginal interpreter services: The present crisis. ITEAA Newsletter. 1984a. Domains and domination. NT Bilingual Education Newsletter 1(2). 48–52. 1986. Aboriginal language programmes and language maintenance in the Kimberley. Australian Review of Applied Linguistics 3. 108–122. 1987. Kija Sounds and Spelling. Broome: Jawa Print. 1987. with Catholic Education language team. Two-way learning. Broome: Jawa. 1991a. Understanding language shift: A step towards language maintenance. In Susanne Romaine (ed.), Language in Australia, 143–155. Melbourne: Cambridge University Press. 1991b. Cultural domain separation: Two-way street or blind alley? Stephen Harris and the NeoWhorfians on Aboriginal education. Australian Aboriginal Studies, 1. 13–24. 1994a. Language shift and maintenance in the Asia-Pacific region. Australian Aboriginal Studies 1994. 86–89 1994b. Two-way exchange and language maintenance in Aboriginal schools. In Deborah Hartman & John Henderson (eds.), Aboriginal languages in education, 235–256. Alice Springs: IAD Press.

xxviii

Preface

1994c. Oral proficiency assessment for Aboriginal languages. In Deborah Hartman & John Henderson (eds.), Aboriginal languages in education, 301–315. Alice Springs: IAD Press. 2010. with Claire Hill. Emergency language documentation teams: The Cape York Peninsula experience. In John Hobson, Kevin Lowe, Susan Poetsch and Michael Walsh (eds.), Reawakening languages: Theory and practice in the revitalisation of Australia’s Indigenous languages, 418–432. Sydney: Sydney University Press. 2001a. Looking for the two-way street. Cultural Survival Quarterly. Summer 2001. 18–21. 2002b. Changing places: European and Aboriginal styles. In Luise Hercus, Flavia Hodges & Jane Simpson (eds.), The land is a map: Placenames of Indigenous origin in Australia, 51–61. Canberra: Pandanus Books. 2003. with Nicholas Thieberger. Language data assessment at the national level: learning from the state of the environment process in Australia. In Joe Blythe and Robert McKenna Brown (eds.), Maintaining the links, 51–57. Bath, UK: Foundation for Endangered Languages. 2004. Understanding language shift: A step towards language maintenance. In Susanne Romaine (ed.), Language in Australia, 143–155. Cambridge: Cambridge University Press. 2005. with Nicholas Thieberger. Languages Past and Present. In Bill Arthur & Frances Morphy (ed.), Macquarie atlas of Indigenous Australia, 78–87. Sydney: Macquarie University Press. 2006. with Nicholas Thieberger. Keeping track of Indigenous language endangerment in Australia. In Denis Cunningham, D.E. Ingram and Kenneth Sumbuk (eds.), Language diversity in the Pacific: Endangerment and survival, 54–84. UK: Multilingual Matters Ltd. 2007a. Video: a linguist’s perspective: reply to Nathan. Language Archives Newsletter 10. 2–3. 2010. with Jane Simpson & Jo Caffery. Maintaining languages, maintaining identities: What bilingual education offers. In Brett Baker, Ilana Mushin, Mark Harvey and Rod Gardner (eds.), Indigenous language and social identity: Papers in honour of Michael Walsh, 409–429. Canberra: Pacific Linguistics.

Language contact 1985b. Domains and codeswitching among bilingual Aborigines. In Michael Clyne (ed.), Australia, meeting place of languages, 95–125. Canberra: Pacific Linguistics. 1988c. Mix-im-up: Aboriginal codeswitching old and new. In Monica Heller (ed.), Codeswitching: Anthropological and sociolinguistic perspectives, 97–124. Berlin: Mouton de Gruyter. 1994e. Discourse frame analysis of code-switching. In Durk Gorter & Anna Christien Piebenga (eds.), Code-switching: Papers from the Leeuwarden Summer School. 1995. with Lorraine Dalton, Sandra Edwards, Ros Farquharson & Sarah Oscar. Gurindji children’s language and language maintenance. International Journal of the Sociology of Language 113. 83–98. 2005b. Language contact interaction and possessive variation. Monash University Linguistics Papers 4(2). 87–105. 2005. with Margaret Florey. Introduction: Language shift, code-mixing and variation. Australian Journal of Linguistics 25(1). 1–7. 2005. with Felicity Meakins. Gurindji Kriol: A mixed language emerges from code-switching. Australian Journal of Linguistics 25(1). 9–30.

Preface

xxix

2007b. Language ecology as determinant of language shift or language hybridity: Some Australian Aboriginal cases. Paper presented at the International Symposium on Bilingualism, Hamburg. 2008a. Language mixing and language shift in Indigenous Australia. In Jane Simpson & Gillian Wigglesworth (eds.), Children’s language and multilingualism: Indigenous language use at home and school, 205–227. New York: Continuum. 2008b. Mixed Languages as outcomes of code-switching: Recent examples from Australia and their implications. Journal of Language Contact 2. 187–212. 2010b. Language contact and Indigenous languages in Australia. In Ray Hickey (ed.), Handbook of language contact, 770–794. Oxford: Blackwell.

Historical linguistics 1996b. Backtracking to Babel: the chronology of Pama-Nyungan expansion in Australia. Archaeology in Oceania 31(3). 125–144. 1997a. Semantic shifts between fish and meat and the prehistory of Pama-Nyungan. In Darrell Tryon & Michael Walsh (eds.), Boundary rider: Essays in honour of Geoffrey O’Grady, 303– 325. Canberra: Pacific Linguistics. 1998. with Nicholas Evans. The enigma of Pama-Nyungan expansion in Australia. In Roger Blench, Roger and Matthew Spriggs. Archaeology and language, volume 2: Correlating archaeological and linguistic hypotheses, 174–191. London: Routledge. 2001b. Language shift and language spread among hunter-gatherers. In Catherine Panter-Brick, Robert H. Layton, Peter Rowley-Conwy (eds.), Hunter-gatherers: Social and biological perspectives, 143–169. Cambridge: Cambridge University Press. 2003a. Headward migration: a Kimberley counter-example. In Nicholas Evans (ed.), The nonPama-Nyungan languages of northern Australia: Comparative studies of the continent’s most linguistically complex region, 75–92. Canberra: Pacific Linguistics. 2006b. Shibbolethnonyms, ex-exonyms and eco-ethnonyms in Aboriginal Australia: The pragmatics of oxymization and archaism. Onoma 41. 185–214. 2006c. Comment on Clendon ‘Reassessing Australia’s Linguistic Prehistory’. Current Anthropology. 47(1). 53–54. 2010c. Getting the constraints right. Behavioral and Brain Sciences. 33(5). 394–395. 2011. Lexical contact phenomena in Australian Linguistic Prehistory: Substrates and Wanderwörter. Paper presented to the symposium Methodology in Linguistic Prehistory. Berlin, October 2011. http://www2.hu-berlin.de/kba/events/mlp/mcconvell.pdf 2011. with Claire Bowern, Patience Epps, Russell Grey et al. Does lateral transmission obscure inheritance in hunter-gatherer languages? PLoS ONE 6.9: 1–9. 2011. with Claire Bowern. The prehistory and internal relationships of Australian languages. Language and Linguistics Compass 5(1). 19–32. 2013a. Australia: Linguistic History. In I. Ness & P. Bellwood (eds.), The Encyclopaedia of global human migration. pp. 329–332. Oxford: Wiley-Blackwell. 2015. Language and culture history: The contribution of linguistic prehistory. In Farzad Sharifian (ed.), The Routledge handbook of language and culture, 209–224. Abingdon and New York: Routledge.

xxx

Preface

2014. with Claire Bowern, Hannah Haynie, Catherine Sheard et al. Loan and inheritance patterns in hunter-gatherer ethnobiological systems. Journal of Ethnobiology 34(2). 195–227. 2014. with Hannah Haynie, Claire Bowern, Patience Epps et al. Wanderwörter in languages of the Americas and Australia. Ampersand: An International Journal of General and Applied Linguistics. 1. 1–18.

Kinship 1982b. Neutralisation and degrees of respect in Gurindji. In Jeffrey Heath, Francesca Merlan & Alan Rumsey (eds.), Languages of kinship in Aboriginal Australia, 86–106. Sydney: University of Sydney. 1985c. The origin of subsections in northern Australia. Oceania 56. 1–33. 1985d. Time perspective in Aboriginal Australian culture: Two approaches to the origin of subsections. Aboriginal History 9(1). 53–79. 1997b. Long lost relations: Pama-Nyungan and northern kinship. In McConvell, Patrick & Nicholas Evans (eds.), Archaeology and linguistics: Aboriginal Australia in global perspective, 207–235. Melbourne: Oxford UP. 1998. ‘Born is nothing’: Roots, family trees and other attachments to land in the Victoria River District and the Kimberleys. Aboriginal History. 22. 180–202. 2002. with Barry Alpher. On the Omaha trail in Australia: Tracking skewing from east to west. Anthropological Forum 12(2). 159–175. 2008c. Grand-daddy morphs: The importance of suffixes in reconstructing Pama-Nyungan kinship. In Claire Bowern, Bethwyn Evans & Luisa Miceli (eds.), Morphology and language history: In honour of Harold Koch, 313–327. Amsterdam: John Benjamins. 2010. with Laurent Dousset, Rachel Hendery, Claire Bowern et al. Developing a database for Australian Indigenous kinship terminology: The Austkin project. Australian Aboriginal Studies 1. 42–56. 2012. Omaha skewing in Australia: Overlays, dynamism and change. In Thomas R Trautmann and Peter M Whiteley (eds.), Crow-Omaha: New light on a classic problem of kinship analysis, 243–260. Tucson: University of Arizona Press. 2012. with Laurent Dousset. Tracking the dynamics of kinship and social category terms with AustKin II. Joint Workshop of LINGVIS & UNCLH for Computational Linguistics (EACL 2012), 98–107. Stroudsburg, US: Curran Associates. 2013b. Granny got cross: Semantic change of kami ‘mother’s mother’ to ‘father’s mother’ in Pama-Nyungan. In Robert Mailhammer (ed.), Lexical and structural etymology: Beyond word histories, 147–184. Boston: Mouton de Gruyter. 2013. with Heather Gardner. The descent of Morgan in Australia: Kinship representation from the Australian colonies. Structure and Dynamics: eJournal of Anthropological and Related Sciences 6(1). 1–23. 2013c. Comment on Denham’s beyond fictions of closure in Australian Aboriginal kinship. Mathematical Anthropology and Cultural Theory: An International Journal 5(3). 1–6. 2013d. Proto-Pama-Nyungan kinship and the AustKin project: Reconstructing proto-terms for ‘Mother’s Father’ and Their Transformations. In Patrick McConvell, Ian Keen, Rachel Hendery (ed.), Kinship systems: Change and reconstruction, 192–216. Salt Lake City: University of Utah Press.

Preface

xxxi

2013. with Rachel Hendery & Patrick McConvell. Mama and Papa in Indigenous Australia. In Patrick McConvell, Ian Keen, Rachel Hendery (eds.), Kinship systems: Change and reconstruction, 217–238. Salt Lake City: University of Utah Press. 2013. with Ian Keen, Rachel Hendery (eds.), Kinship systems: Change and reconstruction. Salt Lake City: University of Utah Press. 2013e. Introduction: Kinship change in anthropology and linguistics. In Patrick McConvell, Ian Keen, Rachel Hendery (eds.), Kinship systems: Change and reconstruction, 1–18. Salt Lake City: University of Utah Press. 2013. with Ian Keen. The transition from Kariera to an asymmetrical system: Cape York Peninsula to North-east Arnhem Land. In Doug Jones & Bojka Milicic (eds.), Kinship, language and prehistory: Per Hage and the Renaissance in Kinship Studies, 99–132. Salt Lake City: University of Utah Press. 2015. with Helen Gardner. The Unwritten Kamilaroi and Kurnai: Unpublished kinship schedules collected by Fison and Howitt. Sydney: Palgrave Macmillan. in press-a. Enhancing The Kinship Anthropology Of Scheffler With Diachronic Linguistics And Centricity. In W. Shapiro (ed.), [Title TBA] Proceedings of the American Philosophical Society. in press-b. Long-Distance Diffusion Of Affinal Kinship Terms As Evidence Of Late Holocene Change In Marriage Systems In Aboriginal Australia. In P. Toner (ed.), [Title TBA] Canberra: ANU Press. forthcoming. with Piers Kelly and Sebastien Lacrampe (eds.), Skin, kin and clan: The dynamics of social categories in Indigenous Australia. forthcoming. The birds and the bees: origins of sections in Queensland. In Patrick McConvell, Piers Kelly and Sebastien Lacrampe (eds.), Skin, kin and clan: The dynamics of social categories in Indigenous Australia. forthcoming. with Maia Ponsonnet. Generic Terms For Subsections (‘Skins’) In Australia: Sources And Semantic Networks. In Patrick McConvell, Piers Kelly and Sebastien Lacrampe (eds.), Skin, kin and clan: The dynamics of social categories in Indigenous Australia.

Linguistics and archaeology 1990. The linguistic prehistory of Australia: opportunities for dialogue with archaeology. Australian Archaeology 31. 3–27. 1997. with Nicholas Evans (eds.) Archaeology and linguistics: Aboriginal Australia in global perspective. Melbourne: Oxford UP. 1997. with Nicholas Evans. Clues to Australia’s human past: pulling together the strands. In McConvell, Patrick & Nicholas Evans (eds.), Archaeology and linguistics: Aboriginal Australia in global perspective, 1–16. Melbourne: Oxford University Press. 2003. with Michael Smith. Millers and Mullers. The Archaeo-Linguistic Stratigraphy of Technological Change in Holocene Australia. In Henning Anderson (ed.), Language contacts in prehistory: Studies in stratigraphy, 177–200. Amsterdam: John Benjamins. 2010d. The archaeo-linguistics of migration. In Jan Lucassen, Leo Lucassen & Patrick Manning (eds.), Migration history in world history: Multidisciplinary approaches, 155–188. Leiden: Brill.

xxxii

Preface

2014. with Thomas Saunders & Stef Spronck. Linguistic prehistory of the Australian boab. In. Lauren Gawne & Jill Vaughan (eds.), Proceedings of the 44th conference of the Australian linguistic society, 295–310. Melbourne: University of Melbourne. 2015. with Rangan, Haripriya, Karen Bell, David Baum et al. New Genetic and Linguistic Analyses Show Ancient Human Influence on Baobab Evolution and Distribution in Australia. PLOS ONE 10(4). 1–18.

Other linguistic anthropology 2001c. People, countries and the Rainbow Serpent: systems of classification among the Lardil of Mornington Island. Journal of Linguistic Anthropology 11(2). 311–313. 2005. with Nicholas Peterson, Heather McDonald et al. Social and cultural life. In Bill Arthur & Frances Morphy (eds.), Macquarie atlas of Indigenous Australia, 88–107. Sydney: Macquarie University Press.

Language Reports 1983d. Language needs of the Torres Strait. Submission to Senate National Language Policy Enquiry. 1984. with Joyce Hudson. Keeping language strong: Report of the pilot study for the Kimberley Language Resource Centre. Broome, WA: Kimberley Language Resource Centre. 2001. with Nicholas Thieberger. The state of Indigenous languages in Australia. Australia state of the environment technical paper series (natural and cultural heritage). Canberra: Department of Environment and Heritage. 2002. with Robert Amery, Mary-Anne Gale, Christine Nicholls, Jonathan Nicholls, Lester Rigney, Tur Irabinna, Ulalka Simione. “Keep that language going!”: A needs-based review of the status of Indigenous languages in South Australia. Adelaide: Aboriginal and Torres Strait Islander Commission. 2009. with Jane Simpson, Jo Caffery. Gaps in Australia’s Indigenous language policy: Dismantling bilingual education in the Northern Territory. Canberra: AIATSIS.

Native title and land claims 1975. A claim for land at Victoria River Downs by the Yarralin community. Darwin: Northern Land Council. 1978. with Nicolas Peterson, Stephen Wild & Rod Hagen. A claim to areas of traditional land by the Warlpiri and Kartangarurru-Kurintji. Alice Springs: Central Land Council. 1979. with Arthur B. Palmer. Yingawunarri Mudbura Land Claim. Darwin: Northern Land Council. 1986. with Rod Hagen. A traditional land claim by the Gurindji to Daguragu Station. Alice Springs: Central Land Council. 1993. Malngin and Nyinin Claim to Mistake Creek: Anthropologist’s report. Alice Springs: Central Land Council.

Preface

xxxiii

2002c. Linguistic stratigraphy and Native Title: The case of ethnonyms. In John Henderson, David Nash, (eds.), Language in Native Title, 259–290. Canberra: Aboriginal Studies Press. 2010d. A short ride on a time machine: linguistics, culture history and native title. In Sandy Toussaint (ed.), Crossing boundaries: Anthropology, linguistics, history and law in Native Title, 151-175. Melbourne: Melbourne University Press. 2010e. Kalkarindji Native Title consent determination. Alice Springs: Central Land Council.

Reviews 1982c. Review of ‘Brandenstein, C G von, 1980 Ngadjumaja: An Aboriginal language of southeast Western Australia’. Oceania 53. 192–193. 1984b. Review of ‘Donaldson, Tamsin, 1980 Ngiyambaa: the language of the Wangaaybuwan’. Oceania 54. 265. 1992. Review of Fishman 1991. Australian Journal of Linguistics 12(1). 209–220. 2001e. Review of ‘D. Nettle Language Diversity’. Language in Society 30. 97–100. 2001f. Review of ‘D. McKnight People, Land and the Rainbow Serpent’. Journal of Linguistic Anthropology. 11(2). 311–313. 2003b. Book Review: Forty Years on: Ken Hale and Australian Languages. Oceanic Linguistics 42(1). 252–258. 2013. with Maia Ponsonnet. Results and Prospects in the Study of Semantic Change: A Review of From Polysemy to Semantic Change (2008). Journal of Language Contact. 6(1). 180–196.

Thanks Many thanks to Mary Laughren, David Nash and Harold Koch for help with the bibliography, Adrienne and Billy McConvell for information about Patrick’s early life and photos; Harold Koch for a timeline of his work career in Australia; Philip Jagger, Graham Furniss and Ian Maddieson for details of his time at SOAS; Peter Sutton and Michael Walsh for stories about the early years at AIAS (mostly unpublishable!); and finally Nick Evans, Mary Laughren, David Nash, Jane Simpson and Erich Round for direction on his seminal work.

I Introduction

Carmel O’Shannessy and Felicity Meakins

Australian language contact in historical and synchronic perspective 1 Introduction This volume is the first collection of research dedicated to the effects of recent language contact processes on Australian languages. Multilingualism and language contact have always been pervasive in Australia (Bowern & Koch 2004; Koch 1997; McConvell & Bowern 2011), but have often been discussed in the context of identifying genetic relationships between languages. At the time of British colonisation, there were approximately 250 languages spoken in Australia, many with several dialects. Colonisation brought the extensive diffusion of English and with it a dramatically different configuration of languages in contact, including the emergence of pidgins, creoles and mixed languages and a range of English-lexified varieties and dialects, such as Aboriginal English. Now relatively few traditional languages are spoken day-to-day or are being transmitted to children. Yet notions of simplification and loss do not adequately capture the complexity and dynamics of the contemporary contexts. Indigenous people have developed complex linguistic repertoires, often including other traditional languages and varieties of English and/or Kriol (an English-lexified creole), or a mixed language. Many of the contact languages co-existed for periods of time with traditional languages, and in some cases, still do, raising questions of continuing and bidirectional contact influences. Indigenous speakers have shifted, or are shifting away from traditional languages in many locations, but in some places traditional languages remain the primary languages spoken, with English or Kriol included in speakers’ repertoires. These constantly evolving scenarios raise questions of what kinds of language contact mechanisms and outcomes are at play in contemporary language-in-use, and this volume collates research at the vanguard of that exploration. The research presented in this volume marks a new era of linguistic work on Australian languages. The last 40 years have seen a concerted effort to describe traditional Australian languages rather than contact varieties. The focus is largely the result of the urgency of documenting these endangered languages. However, in the 1970s through to the mid-1980s, attention was given to the English-based pidgin and Kriol. The pidgin developed in the Sydney colony and diffused into the Pacific and northern Australia, and transformed into north Australian Kriol, that developed as a result of interactions between speakers of the pidgin, traditional languages and English. The interest in Australian pidgin

4

Carmel O’Shannessy and Felicity Meakins

and creole languages abated until recently, when there has been a resurgence of interest in Australian contact languages, including Kriol varieties, newlyidentified mixed languages and restructured traditional Australian languages. It is likely that, as more traditional Australian languages lose their first language learners, more attention will be drawn to these contact languages. The volume contributes new data and theoretical analyses to discussions of contact languages and language contact processes. Theories of language contact need descriptions of mechanisms and outcomes from all parts of the world and from many language types, yet until recently they have been informed by relatively little data from Australia, which is rich in typological diversity and known for extensive contact historically. For example, the theoretical literature on pidgin and creole languages is largely Atlantic-focussed, with Australian varieties not included in the debates which have preoccupied creolists over the last two decades, such as whether creolisation is the result of an innate bioprogram (Bickerton 1984), relexification of substrate languages by colonial languages (Lefebvre 1998) or the result of more general second language learning processes (Siegel 2008). A related issue that Australian pidgins and creoles can contribute to greatly is the relative roles of the substrates and superstrates in creole formation and the role of pidgins as a predecessors of creoles (Chaudenson 1992; Mufwene 2000). The development of Kriol is relatively recent, and detailed analyses of many of the substrate languages are available or within reach. Chapters 5, 6, 8, 9 and 10 take perspectives on factors in the formation of Kriol and continuing influences on it where it remains in contact with traditional languages. Recent research on the Australian mixed languages, Gurindji Kriol (McConvell (2008), Meakins (2011)) and Light Warlpiri (O’Shannessy (2012)), has contributed new data and analyses to debates such as whether common contact processes such as code-switching can lead to the development of mixed languages and whether inflectional morphology can be borrowed. Further developments of these mixed languages will be discussed in Chapters 7 and 11. Additionally, Chapter 1 will examine the borrowing of a verb paradigm in Marrku, a process similar to those seen in the formation of mixed languages. Work on Australian restructured varieties has informed the larger field of language contact. In particular, Schmidt’s (1985) work on the restructured variety of Dyirbal spoken by younger generations has become a classic case study often cited in the literature on language shift. Nonetheless much of Schmidt’s work and others in the field of language obsolescence frame the language of younger generations as language loss. This characterisation will be challenged in Chapters 12 and 13 which focus on new varieties of Murrinh-Patha and Jingulu, describing recent grammatical changes as newly emergent systems rather than language loss.

Australian language contact in historical and synchronic perspective

5

The volume stems from papers given at the Australian Languages Workshop (ALW) Language Contact Symposium (6–7 March 2014) run by Felicity Meakins and Carmel O’Shannessy at the Australian National University.

2 Contact between Australian languages before 1788 Extensive language contact has been a feature of the Australian languages context for thousands of years, making the role of language contact an important issue historically (e.g. Dixon 1980, 2002; Koch 1997; McConvell & Bowern 2011). One claim is that language contact effects in Australia are so extensive, and involve such great time-depth, that they obscure genetic relationships (Dixon 2001: 88; 2002: 38), and that all Australian languages form one large linguistic area, created approximately 40,000–50,000 years ago (Dixon 2001: 25, 39; 2002). But processes of language contact versus genetic inheritance have been distinguished using historical-comparative methods, providing substantial evidence for genetic subgrouping within Australia (e.g. Alpher 2004; Bowern & Atkinson 2012; Bowern et al. 2011; Dench 2001; Evans 1988; Evans & McConvell 1998; Heath 1978; L. A. Hercus 1979; McConvell & Bowern 2011; McConvell & Laughren 2004; O’Grady Voegelin & Voegelin 1966). The major genetic distinction is between the Pama-Nyungan family, with sub-groups covering approximately 90% of the continent, and the group labelled Non-Pama-Nyungan, which is not a genetic family, but a group of approximately 27 families in the north (Evans 1988; Evans & McConvell 1998; McConvell 1996). There were approximately 250 languages spoken before colonisation, and the sizes of language groups ranged from 100–200 people, to about 3,000 people (including dialects) (McConvell 2010b: 1). The small numbers of speakers promote multilingualism as groups communicate with each other. Hunter-gatherer economies were maintained until soon after colonisation, and multilingualism was common (e.g. Evans 2010; McConvell & Bowern 2011). It had been thought that hunter gatherer populations borrow proportionally more words than agriculturalist populations (Bowern et al. 2011 and references therein), and this claim has been made often for Australian languages (Koch 2004), but recent work by Bowern, et al (2011) shows that this is not the case. In a survey of 122 languages from hunter-gatherer and agriculturalist societies in three continents, Australia (31 languages), and North and South America, Bowern, et al (2011) found that the mean proportion of borrowed words in basic vocabulary is 9.4% – far less than often supposed. However, the Australian languages sampled have

6

Carmel O’Shannessy and Felicity Meakins

more loans and show more variation than those on the other two continents. The reason is that some languages have loan rates of up to 48% (e.g. Gurindji, Pama-Nyungan), but these languages are outliers in terms of the amount of borrowing, and the high rates are the results of particular population movements (e.g. McConvell 1996, 2009). When the outlier languages are removed, the mean proportion of loans is 6.6%. Languages with the highest rates of loans have one or more of the following properties – speakers are relatively few in number, have a lower density of population in any one location, are relatively mobile, and practise linguistic exogamy. In a study of highly borrowed items in 53 Australian languages in the north west, it was found that basic vocabulary is borrowed least often, and terms for flora and fauna and items of material culture, including technologies, are borrowed more often (Bowern et al. 2014). Words for some highly traded items, for instance ‘spearthrower’ and ‘pearl shell’, diffused across great distances, such as the word for ‘pearl shell’ travelling from the Kimberley coast to central Australia and the South Australian Bight (Haynie, Bowern, Epps, Hill, & McConvell 2014: 13–14). Some very early language contact events can be inferred from combining anthropological and archaeological information with current and reconstructed linguistic patterns. For instance, it is likely that as speakers of Pama-Nyungan languages migrated across the continent, perhaps around 5,000–2,000 BP, in some places their languages replaced those of the peoples already living in those areas, after a phase of mutual bilingualism (Evans & McConvell 1998: 125–7; McConvell 1996). The language shift may be linked to a change in social organisation from isolated, endogamous groups to that of increased “alliance, exchange and exogamy” (Evans & McConvell 1998: 184) including, for example, ceremonial events with speakers of other languages, increased trade exchanges, and a view of the incoming language(s) as high status. Changes in climate, availability of natural resources and changes in technology would have played a role in people’s movements. In McConvell’s (2010a) “upstream-downstream” model, people may have recolonised areas when their technology allowed for greater use of the resources in the area, or moved to more richly resourced areas when needed. In some instances these movements would have brought about language shift, in others, transfer of words and structure. For example, when speakers of Eastern Ngumpin (Pama-Nyungan) languages moved from the semidesert area into the riverine region, perhaps 2000 years ago, they borrowed many words and likely some verbal structure from non-Pama-Nyungan languages into their own languages (McConvell 1996: 133, 2009: 811, 2010b: 776). Lexical semantic patterns also diffused among these languages (McConvell 2010b: 776). Extensive diffusion of features is found in areas where speakers of different languages interacted frequently, for example, in the linguistic area in south east

Australian language contact in historical and synchronic perspective

7

Arnhem Land, where large amounts of vocabulary and morphological structure were transferred between Pama-Nyungan and non-Pama-Nyungan languages, especially Ritharrngu (Pama-Nyungan) and Ngandi (non-Pama-Nyungan) (Heath 1978, 1981). The intensive diffusion is attributed to small group size, linguistic exogamy and type of internal social organisation. A dramatic example of language contact in Australia is the spread of section and subsection terms (known colloquially as ‘skin names’). Subsection terms are used along with kinship terms and personal names, and identify the relationship of the speaker to others in the community. McConvell’s (1985) reconstruction of the origin of Australian subsection systems shows that earlier four-section systems, originating in the south-west, were expanded into eight subsection systems through marriages between different language groups in the lower Victoria River Basin. Four terms belonging to each of two language groups were merged, resulting in a system of eight terms. The subsection system then spread south and west through further language contact. McConvell suggests that the diffusion took place over hundreds of years, perhaps beginning around 1,000 BP, with more recent changes being completed in the mid-nineteenth century (McConvell 1985: fn6 23, 1996: 130–6). Several cultural-linguistic conventions that favour multilingualism and therefore, language contact, have been practised for long time periods. In East Arnhem Land linguistic exogamy is the norm. Men and women continue to maintain their own languages, and children grow up learning both, ensuring constant multilingualism (Evans 2010). In many areas, when people travel through lands belonging to speakers of other languages, they speak the language belonging to the land they are on (e.g. Evans 2010). Another wide-spread practice is placing a taboo on the names of recently deceased persons, along with other words that begin with similar sounding syllables. Borrowings from other languages can be a source of replacement words. This can lead to high levels of change in vocabulary (Dixon 1980: 28, 2002: 27), including in basic vocabulary (McConvell 2009: 30), which is usually less amenable to replacement than nonbasic vocabulary, but not so much that genetic lines have been obscured (Koch 1997: 41 and references therein).

3 Contact between Australian languages post 1788 Over the last two hundred years, English has left an indelible footprint on the linguistic landscape of Indigenous Australia. Of the approximately 250 languages which were spoken at first contact, only around 18 remain strong. The

8

Carmel O’Shannessy and Felicity Meakins

shift to English has seen a decline in the use of Indigenous languages and the emergence of a number of contact varieties of English and Indigenous languages. The main language now spoken in many Indigenous communities across northern Australia is Kriol, a creole language which uses English vocabulary while preserving the some of the sound system, semantics and grammatical features common to many Indigenous languages (Sandefur 1979; SchultzeBerndt, Meakins, & Angelo 2013) (§3.1). In other areas of Australia, Aboriginal English varieties have come to dominate (Eades 2014) (§3.2). Other Indigenous languages have formed mixed languages by combining with Kriol or English to create a new language, the best known examples being Gurindji Kriol (Meakins 2013a) and Light Warlpiri (O’Shannessy 2005, 2013) (§3.3). Where traditional Indigenous languages are still spoken, many have undergone grammatical restructuring under the influence of English, such as Lardil (Richards 2001), Areyonga Teenage Pitjantjatjara (Langlois 2004) and Young People’s Dyirbal (Schmidt 1985) (§3.4). The volume provides much-needed descriptions and explanations of contemporary language contact processes in Australia, including work on each of the main types of contact outcomes since colonisation – NSW Pidgin (Ch. 3 McConvell, Ch. 4 Nash), north Australian Kriol (Ch. 5 Dickson, Ch. 6 BundgaardNielsen and Baker; Ch. 8 Angelo & Schultze-Berndt; Ch. 9 Ponsonnet; Ch. 10 Nicholls), Aboriginal English (Ch. 2 Mushin & Watts), mixed languages (Ch. 7 O’Shannessy; Ch. 11 Meakins), and restructured traditional languages (Ch. 1 Evans; Ch. 12 Mansfield; Ch. 13 Meakins and Pensalfini). This section gives an introduction to the linguistic shift that Australia, as a continent, has undergone since 1788 when British colonists first arrived. More substantial reviews of Australian contact varieties can be found in McConvell (2010c) and Meakins (2014) and specific reviews of the pidgin and creole literature are provided by Mühlhäusler (2008) and can be found in the Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas (Wurm, Mühlhäusler, & Tryon 1996).

3.1 NSW Pidgin and north Australian Kriol Most of the work which has been carried out on Australian contact languages has focused on pidgin and creole languages. A number of non-English-based pidgins were used in the past between Indigenous and non-Indigenous people including some based on Australian languages such as Jargon Kaurna spoken in Adelaide (Simpson 1996), Pidgin Ngarluma spoken in the Pilbara region (Dench 1998), and those based on non-Australian languages other than English

Australian language contact in historical and synchronic perspective

9

such as Pidgin Macassan (Evans 1992; Urry & Walsh 1981), and Broome Pearling Lugger Pidgin which was based on Kupang Malay (Hosokawa 1987). However the majority of pidgins spoken in Australia were English-based. Of particular interest are New South Wales (NSW) Pidgin and Queensland Pidgin English due to their influence on the development of north Australian Kriol, estimated to be spoken by up to 30,0001 Aboriginal people across the Top End of Australia from western Cape York (Queensland) to Broome (Western Australia) and south to Tennant Creek (Northern Territory) (see map). NSW Pidgin originated in the Sydney area when Australia was colonised in 1788 by the British. It probably has its roots in South Seas Jargon which was brought to Australia by early sailing crews. A small number of lexical items from this contact language are still found in north Australian Kriol including too much ‘plenty’, piccaninny ‘child, small’, sabi ‘know, understand’ and catch ‘get’ (Baker 1993; R. Clark 1979). Based on historical sources, Troy (1990) provides a detailed account of the linguistic context in which NSW Pidgin arose which involved South Seas Jargon, English and local Indigenous languages. Communication between Aboriginal and non-Aboriginal people largely involved the settlers obtaining information which would help the colony, but communication also occurred on a social level with mixed drinking parties. NSW Pidgin further developed with more formal attempts at teaching the local Aboriginal people English. These attempts involved kidnapping Aboriginal men or orphaned children. They were kept captive as English language students, with the expectation that they would become interpreters. The captive students also attempted to teach the colonists some of the Sydney Language which is how a number of Dharuk words entered the developing pidgin. Troy (1990: 47) suggests that the developing pidgin may have also been a lingua franca among Aboriginal people early on. A likely scenario is that large numbers of Aboriginal people within individual language groups had been decimated by introduced diseases such as small pox and at the hand of the Europeans in killing sprees, and the survivors had then formed new allegiances, using the developing pidgin as a lingua franca. People were also displaced as their land was gradually taken over by the colonists. As a result different groups of Aboriginal people were forced into a closer proximity than had been the case traditionally, with the pidgin becoming their main means of communication. Amery and Mühlhäusler (1996: 38) observe that by 1804 NSW Pidgin was wellestablished as the language of communication between Aboriginal and nonAboriginal people, and between Aboriginal people themselves. 1 Current, accurate published data on numbers of speakers is not available. See, e.g. Angelo and McIntosh (2014) on issues with census data and English-lexified varieties.

10

Carmel O’Shannessy and Felicity Meakins

The diffusion of NSW Pidgin took a number of routes, including into the Pacific through sea trade (Troy 1990). Varieties of Melanesian Pidgin originate in NSW Pidgin (Baker 1993). NSW Pidgin also spread north to Queensland through both inland and coastal routes (Dutton 1983). Creole varieties subsequently developed in some areas. In the Northern Territory, the pidgin developed into varieties of Kriol, and in the far north Queensland region it developed into Torres Strait Creole (aka Broken, Yumplatok) (Shnukal 1988; Sellwood & Angelo 2013) and its dialect Cape York Creole (Crowley and Rigsby 1979). In addition, a mosaic of English-lexified varieties emerged in Queensland (e.g. Angelo & Carter 2015; Mushin, Angelo & Munro forthcoming). By the early 1900s, reports from a mission at Roper River (now Ngukurr) in the north suggested that pidgin was no longer just a lingua franca used between Aboriginal and non-Aboriginal people, but was being learnt as the first language of many Aboriginal children. This is the earliest indication of the emergence of a creole language in northern Australia. Its link to NSW Pidgin is indisputable due to the few words which remain in Kriol from the Sydney Language including bogi ‘swim, bathe’, gabarra ‘head’ and binji ‘stomach’ (Harris 1986: 286 onwards; Troy 2003: 47). The most detailed account of the emergence of a north Australian Kriol variety comes from Ngukurr (Harris 1986). Ngukurr was originally established as a mission on the banks of the Roper River in 1908. There may have been competing motivations for its establishment, including Christianisation and the creation of a refuge for Aboriginal people escaping massacres at the hands of the Eastern and African Cold Storage Company, which was seizing land in the area for pastoral leases. The mission provided sanctuary for Aboriginal people from nine different language groups. Most Aboriginal people were fluent in two or more of these languages and would have spoken the pidgin English that had spread across northern Australia. For many Aboriginal people, the pidgin became their lingua franca at the mission, with traditional languages reserved for in-group communication. English also had a strong influence on children through schooling. Two theories exist for the diffusion of Kriol, (i) the monogenesis hypothesis and (ii) the multi-regional hypothesis. The monogenesis hypothesis suggests that Kriol originated at Ngukurr and then spread across northern Australia as a fullyfledged creole in a process of language shift and expansion (Munro 2000). The multi-regional theory proposes that pidgin English spread throughout the north of Australia through the pastoral industry (via imported Aboriginal labour) and creolised in different places (Sandefur & Harris 1986). The monogenesis hypothesis is popularly espoused, however this is probably the result of the disproportionally large amount of work which has focussed on Ngukurr Kriol. It is likely

Australian language contact in historical and synchronic perspective

11

that dialects of north Australian Kriol developed independently and their similarity is the result of their shared origins in the pidgin spoken on cattle stations, and similarities in substrate languages (see Meakins 2013b: 381 onwards for arguments). This volume provides domain specific studies of NSW Pidgin, specifically in the area of kinship (Ch. 3) and placenames (Ch. 4). It also examines a number of structures in north Australian Kriol more closely and in comparison with its Indigenous substrates, including lexical influence in the domain of verbs and kinship (Ch. 5), phonology and the question of a Creole Continuum (Ch. 6), the development of apprehensives (Ch. 8), distinctions between reciprocal and reflexive marking (Ch. 9) and the semantics of determiners (Ch. 10).

3.2 Aboriginal English The name Aboriginal English is often used as an umbrella term to describe English-based contact varieties spoken by Aboriginal people which are neither Standard Australian English nor described (yet) as a creole. These range across a spectrum of varieties from those which are at a surface level very similar to English to those which are quite dissimilar (e.g. Angelo 2013; Mushin, Angelo & Munro in press; Eades 2013, 2014; Malcolm & Kaldor 1991; Sellwood & Angelo 2013). These varieties are spoken across much of the densely populated areas of Australia, particularly in southern Queensland and Western Australia, and in the southern states, New South Wales, Victoria, Tasmania and South Australia. Varieties of Aboriginal English show features of traditional Australian languages such as zero copula, no gender distinction in pronouns, and possession expressed by apposition. Some varieties are the result of an acrolectalisation of NSW Pidgin or Queensland Pidgin English (I. Clark, Mühlhäusler, & Amery 1996; Foster, Monaghan, & Mühlhäusler 2003) and others have developed as second language learner varieties (Malcolm 2008). Most of the research which has been undertaken on Aboriginal English has occurred within the context of legal and educational systems. Eades’ (2008) pioneering work within an ethnography of communication framework has demonstrated the different ways in which Aboriginal people speak English and how miscommunication between speakers of SAE and Aboriginal English disadvantage Aboriginal people within the Australian legal system. Work on Aboriginal English within the education system has similarly examined differences between SAE and Aboriginal English. In early work, these differences were framed as deficits, for example the Queensland Speech Survey undertaken by Elwyn Flint and his students in 29 Aboriginal communities in the 1960s (Flint 1968). The

12

Carmel O’Shannessy and Felicity Meakins

deficit approach made way for a bidialectal approach under the direction of Margaret Sharpe whereby children were encouraged to learn “school talk” but not at the expense of “home talk” (Teasdale & Whitelaw 1981). The bidialectal approach has been expanded, for example, as the program Fostering English Language in Kimberley Schools (FELIKS) (Catholic Education Office, W.A). More recent approaches recognise that Aboriginal and Torres Strait Islander students become multilingual within complex language ecologies consisting of traditional languages, contact varieties and the national standard English (Angelo & Carter 2015). Educational responses include an augmented approach to Siegel’s 2010 Critical Language Awareness, a celebration of students’ ‘translanguaging’ practices (e.g. Garcia 2009) and targeted teaching of English as an Additional Language/Dialect. Much descriptive work on Aboriginal English has studied the varieties within educational contexts (as shown above), notably also Harkins’ (1994) work in Alice Springs and Malcolm’s (2000) work in Perth, with some exceptions being Koch’s (2000, 2011) work with adult speakers of Central Australian Aboriginal English. This volume extends this descriptive work with a contribution from Mushin & Watts (Ch. 2) which re-examines one of the communities which participated the Queensland Speech Survey (Alexander, 1968).

3.3 Mixed languages Mixed languages are the result of the fusion of two identifiable source languages, normally in situations of community bilingualism (Matras & Bakker 2003; Meakins 2013b, to appear; Thomason & Kaufman 1988). This dual linguistic parentage means that mixed languages cannot be classified according to standard historical comparative methods (Thomason & Kaufman 1988: 108). They emerge in situations of severe social upheaval, serving as an expression of a new identity or the maintenance of an older identity. Thus they differ from pidgin and creole languages in that their genesis is a product of expressive rather than communicative needs (Golovko 2003: 191; Muysken 1997b: 375). Pidgin and creole languages such as Kriol are born out of the need for communication between people of a number of language groups, whereas mixed languages are created in situations where a common language already exists and communication is not at issue. Three mixed languages have been documented in Australia: Gurindji Kriol spoken by Gurindji people at Kalkaringi (Victoria River District, Northern Territory) (Meakins, 2013a), Light Warlpiri spoken by Warlpiri people at Lajamanu (North Tanami, Northern Territory) (O’Shannessy 2013), and New Tiwi spoken by Tiwi

Australian language contact in historical and synchronic perspective

13

people on Melville and Bathurst Islands (off Darwin, Northern Territory) (Lee 1987; McConvell 2002). These mixed languages fuse elements of a traditional Australian language with a contact variety of English. Unlike other mixed languages documented in other parts of the world, they are a recent development, probably only 30–40 years old. They can also be the main language spoken in their respective communities, although they continue to co-exist with their source languages. As a result there is good socio-historical and linguistic information available about the structure and use of these languages and the factors which contributed to their formation. The work stemming from the documentation of these Australian mixed languages has made an important contribution to a number of debates in the language contact literature surrounding the genesis of mixed languages, including specific questions of whether mixed languages result from the fossilisation of code-switching (McConvell & Meakins 2005; Meakins 2011b; O’Shannessy 2012), and the structural possibilities of language contact, for example the borrowability of inflectional morphology (Meakins 2011a) and the role of typology in shaping the mixed languages (McConvell 2002; Meakins & O’Shannessy 2012). Mixed languages which demonstrate a high degree of grammatical mixing are quite rare, with only five identified. Three of these are the Australian mixed languages. New Tiwi sources its NP structure from English and its VP structure from Tiwi. Gurindji Kriol and Light Warlpiri mirror this pattern, with Aboriginal English/Kriol providing TAM auxiliaries, and transitive and aspectual suffixes; and Gurindji or Warlpiri providing the nominal frame including derivational morphology and ergative, dative, comitative, locative, allative and ablative case marking. The following example schematically demonstrates a simple transitive clause in Gurindji Kriol in comparison with its source languages. Gurindji elements are italicised, Kriol elements are in plain font and bolded elements indicate the source of forms. Brackets indicate optional elements. The structure of Light Warlpiri is similar but has clear differences, especially in verbal structure (O’Shannessy, 2013). What is remarkable about the clause structure of Gurindji Kriol and Light Warlpiri is the combination of inflectional categories from both source languages given that the transfer of inflectional morphology from one language to another is exceptional in its rarity. It always occupies the lowest rung on borrowability hierarchies (Gardani, 2008; Matras, 2007) and is most generally derived from the more dominant language in code-switching (Muysken 2000, Myers-Scotton 2002). Yet it is this very structure which characterises the composite morphosyntactic structure of and Gurindji Kriol and Light Warlpiri. Indeed Matras (2003: 158) suggests that a particular feature of mixed languages is the seemingly unconstrained borrowing of grammatical elements, which in the past have been labelled as ‘loan proof’.

14

Carmel O’Shannessy and Felicity Meakins

(1) (Meakins & O’Shannessy 2010: 1697) det man im=in spiya-im (im) the man 3SG=PST spear-TR 3GS ↓ man(-tu) i bin jarrwaj (im) man-ERG 3SG PST spear 3SG

det guana gat jik the goanna PROP stick dat guana karnti-yawung the goanna stick-PROP

↑

(K)

(GK)

↑

ngumpit-tu ngu-ø-ø jarrwaj pa-ni kirrawa karnti-yawung (G) man-ERG CAT-3SG-3SG spear hit-PST goanna spear-PROP The man speared the goanna with a stick. This volume provides further studies of the case systems in Gurindji Kriol and Light Warlpiri including the grammaticalisation of the locative marker into a present tense suffix through a process of insubordination in Gurindji Kriol (Ch 11) and the entrenchment of case allomorphy in Light Warlpiri (Ch 7).

3.4 Restructured languages English has had a less obvious impact on a number of other Australian languages. These languages continue to be spoken but have undergone restructuring under the influence of English grammar. A number of studies have examined restructured varieties of Warlpiri (Bavin & Shopen 2004), Dyirbal (Schmidt 1985a, 1985b), Jingulu (Pensalfini 1999b), New Lardil (Richards 2001), Pitjantjatjara (Langlois 2004), and Arabana and Paakantyi (Hercus 2005). Most of these studies have focussed on changes in the case system resulting from contact with English, in particular (i) the solidification of word order coupled with the loss of structural case marking, and (ii) the loss of the in/alienable distinction in possessive constructions. One of the most commonly observed changes to Australian languages is the optional use of the ergative case marker. Optional ergativity has most commonly been observed as an internal feature of some Australian languages (see the special 2010 issue of Lingua 120.7), but can also be attributed to language contact, specifically, the adoption of the English/Kriol SVO word order system of indicating arguments, and the decreasing use of ergative case-marking (see Meakins & O’Shannessy 2010 for an overview). For example, Schmidt (1985), in her examination of structural change in Dyirbal, describes optional ergativity in terms of the incremental replacement of the case marking system. In Dyirbal the loss of the case marking system corresponds to a gradual increase in the use of

Australian language contact in historical and synchronic perspective

15

English word order and prepositions. In this in-between stage of language loss, ergative marking has become optional. Her predicted end point is the complete replacement of the Dyirbal system of argument marking with the English word order system. Optional ergativity is also a feature of the mixed languages Gurindji Kriol and Light Warlpiri. Changes in possessive constructions is another commonly observed change in languages undergoing restructuring. A number of languages including Arabana, Paakantyi and Areyonga Teenage Pitjantjatjara have undergone the erosion of the alienable-inalienable distinction. Hercus (2005) attributes this structural change in Arabana (northern South Australia) and Paakantyi (Darling River, New South Wales) to contact with English. In Arabana, inalienable and alienable nominals continue to be differentiated in possessive constructions which contain two nominals, however this distinction is not being maintained in constructions which relate possessive pronouns to nominals. Hercus also gives a similar account of in/alienability in Paakantyi, and claims that the change in both languages is an outcome of contact with English, which does not distinguish inalienable nominals in attributive possessive constructions. In the case of Areyonga Teenage Pitjantjatjara, both alienable and inalienable nominals are being marked genitive, where inalienables were previously unmarked (Langlois 2004: 84). This volume provides of further cases of restructuring beyond case systems such as the expansion of a marginal complex predicate structure in Murrinhpatha (Ch 12) and changes in the gender system of Jingulu (Ch 13).

4 Structure of the volume The structure of the book is given below. After the introduction, the book is broadly divided into three sections: II Transfer of form: Structure (1) As intimate as it gets? Paradigm borrowing in Marrku and its implications for the emergence of mixed languages (Evans) (2) Identifying the grammars of Queensland ex-Government reserve varieties: The case of Woorie Talk (Mushin & Watts) III Transfer of form: Lexical (3) Kinship loanwords in Indigenous Australia, before and after colonisation (McConvell)

16

Carmel O’Shannessy and Felicity Meakins

(4) Placenames from NSW Pidgin: Bulga, Nyrang (Nash) (5) Rethinking the substrate Languages of Roper Kriol: The case of Marra (Dickson) IV Transfer of form: Phonological (6) The continuum in Kriol: fact or furphy? (Bundgaard-Nielsen & Baker) (7) Entrenchment of Light Warlpiri morphology (O’Shannessy) V Transfer of function, structure, distribution and semantics (8) Beware of ‘bambai’ – soon it may turn apprehensive (Angelo & SchultzeBerndt) (9) Reflexive, reciprocal and emphatic functions in Barunga Kriol (Ponsonnet) (10) Grammaticalization and interactional pragmatics: A description of the recognitional determiner det in Roper River Kriol (Nicholls) VI (Further) Development of new structures (11) No fixed address: The grammaticalisation of the Gurindji locative as a progressive suffix (Meakins) (12) Light verb structure in Murrinh-Patha (Mansfield) (13) Gender bender: Superclassing in Jingulu gender marking (Meakins & Pensalfini) The first two sections follow Matras and Sakel’s (2007) distinction of MAT(ter) versus PAT(tern) borrowing. This distinction captures the difference between the transfer of surface level linguistic material such as lexemes, morphemes and phonological material versus the transfer of categories such as functional distributions, paradigmatic structures and lexical semantics. Section 2 focuses on MAT(ter) borrowing, examining the transfer of structural material such as verb paradigms (Ch. 1) and structures for temporal reference (Ch. 2); the transfer of lexical items such as kinship terms (Ch. 3), placenames (Ch. 4); and verbs (Ch. 5); and finally the transfer of phonological forms and processes such as phonemes (Ch. 6) and case markers and their allomorphs (Ch. 7). Section 3 then turns the reader’s attention to the PAT(tern) borrowings, specifically the transfer of Australian language semantics in the domains of a modal category, the apprehensive (Ch. 8), expressions of reciprocity and reflexivity (Ch. 9), and Australian language pragmatics in the domain of determiners (Ch. 10). Finally Section 4 considers the development of new structures in Australian contact languages, specifically in the functions of locative case (Ch. 11), complex verb structures (Ch. 12) and gender realignment (Ch. 13). The volume begins with Paradigm borrowing in Marrku by Nick Evans. This paper is a case study of language contact which occurred before the colonisation of Australia, nonetheless it gives an insight into the kinds of contact processes which were underway in Australia before English was added to the linguistic

Australian language contact in historical and synchronic perspective

17

landscape. Evans argues that Marrku (non-Pama-Nyungan), spoken in the highly multilingual context of the Croker Peninsular Region, can be regarded as a precursor to a mixed language (§3.3). Marrku has borrowed complete paradigms for subject/object/TAM prefixing and TAM suffixing for some borrowed verbs, as well as the pronominal paradigm for reflexives. In particular, the borrowing of entire paradigms of verb inflections has implications for the emergence of mixed languages such as Gurindji Kriol and Light Warlpiri which exhibit verb structures from a language different from that which provides the nominal structure. The complex layering of influences on a contact variety of English spoken in Woorabinda, a small community in Queensland, is investigated by Ilana Mushin and Janet Watts in Identifying the Grammars of Queensland ex-Government Reserve Varieties: The Case of Woorie Talk. Focusing on expressions of future time in a contemporary variety called “Woorie Talk”, the chapter traces influences on its formation from regional varieties of English, varieties of Pidgin and traditional languages. A strength of the indepth study is that it collates historical and sociolinguistic information about the forced relocation of speakers of many Australian languages, moved to government-controlled Reserves in the early twentieth century, and aligns this with structural analysis. In Kinship loanwords in Indigenous Australia, McConvell demonstrates a continuity of borrowing practices with respect to kinterms in Australia before and after colonisation. He observes that Matras’ (2009) borrowability hierarchy of kinship terms – collateral terms (e.g. uncle, aunt, cousin) > affinal terms (e.g. spouse, in-laws) > lineal terms (e.g. parents, siblings) – applies well to the Australian context in both borrowings between traditional Australian languages and also borrowings from Aboriginal English or Kriol in traditional languages. McConvell notes that, in the latter cases, English/Kriol words are mapped onto Indigenous kinship structures in a process of relexification. For example, granny has been borrowed into many Indigenous languages to refer to maternal grandchildren of either gender as well as a maternal grandmother which reflects the use of the comparable Indigenous terms. Most curious is the case of cousin which has come to mean mother-in-law. Throughout the paper McConvell details the considerable complexity of identifying contact influence versus inheritance in the domain of kinship terms, and shows that the kinds of processes that take place elsewhere in the world with kinterms have taken place in Australian languages also. Placenames supplement the rare insights into the use of nineteenth century Pidgin in Placenames from NSW Pidgin: Bulga, Nyrang by David Nash. The Pidgin placenames were probably co-constructed by locally resident speakers of traditional languages and of English, in contrast to those introduced by either of those groups of speakers independently. Nash details how the Pidgin names

18

Carmel O’Shannessy and Felicity Meakins

show a mix of the formal and semantic patterns of Indigenous and introduced placenames. For instance, Indigenous placenames are rarely modified by a descriptive adjective, but some Pidgin names incorporate the words for ‘big’ gabun and ‘small’ ngarang from the Sydney Language into the name. The chapter also shows how placenames can assist in tracing the spread of Pidgin beyond its origins in Sydney and New South Wales. The lexicon provides new data about the formation of Kriol, as Greg Dickson offers a nuanced view of how substrate languages may have contributed differentially to Kriol as it formed, in his chapter, Rethinking the Substrate Languages of Roper Kriol: The Case of Marra. Dickson analyses historical evidence, kin and ethnobiological terms, and verbs from traditional languages, arguing that rather than viewing all of the relevant languages as having made approximately equal contributions to grammar and lexicon, it is likely that some languages contributed more than others, specifically those in the Marran family. An under-researched area in language contact in Australia, that of Kriol phonology, is addressed in The continuum in Kriol: fact or furphy? by Rikke Bundgaard-Nielsen and Brett Baker. The authors use production and perception studies to investigate early claims that the phonological systems of Kriol speakers are best characterised as a continuum ranging from basilectal styles (more like the systems of Australian languages) to acrolectal styles (more like the system of Australian English) (Sandefur 1979, 1984, 1986), and that obstruents in Kriol do not have a voicing distinction. They find that speakers of Kriol as a first language produce and perceive Kriol obstruents in a manner consistent with a single, invariable phonological system. This system has elements from the Australian substrate languages, and from English, but is identical to neither. Phonology and morphology interact in Carmel O’Shannessy’s study of the dynamics Light Warlpiri ergative and dative case marking, in Entrenchment of Light Warlpiri Morphology. A longitudinal study of ergative and dative case morphology in children’s texts is examined at two time points, five years apart. A surprising finding is that the frequency of ergative case marking has increased in that time period, from occurring on an average of 59% of overt subjects of transitive verbs in 2005 data, to occurring on an average of 79% of overt transitive subjects in 2010 data, bringing the amount of marking in line with that in Warlpiri texts by the same cohort of speakers. The unexpected increase is attributed to continuing contact with Warlpiri. An examination of the allomorphy of the two cases shows that an earlier trend of reduction in the number of allomorphs (O’Shannessy, To Appear) had become entrenched over a five year period. The findings show both dynamism and regularity in the two systems of case morphology.

Australian language contact in historical and synchronic perspective

19

Section 3 turns to transfer of function, structure, distribution and semantics, where varieties of Kriol are influenced at multiple levels, by other Kriol varieties and by Australian languages. Denise Angelo and Eva Schultze-Berndt begin the section with their chapter, Beware of ‘bambai’ – soon it may turn apprehensive. In the study they present evidence for a semantic extension of a temporal marker bambai (‘soon’, ‘later’, ‘then’) which is attested in English-lexified contact languages of Australia and the Pacific, to the function of an apprehensive modal, in at least some varieties of Kriol. Through very detailed tracking of available data they show that substrate influence and internal grammaticalisation are both likely to have played roles in the semantic change. Australian languages typically have a category for expressing apprehension, providing a semantic motivation. This is coupled with crosslinguistic evidence that a temporal category can conventionalise to create an apprehensive reading. Structural transfer as a likely result of contact with other Kriol varieties as well as Australian languages, including substrate languages, is seen in Barunga Kriol, documented by Maïa Ponsonnet in Reflexive, reciprocal and emphatic functions in Barunga Kriol. Early descriptions of reflexive, reciprocal and emphatic functions in this variety show that they patterned similarly to English, in that reflexive and reciprocal functions were distinguished by different words. Now the word that described reflexive situations is also used to describe some reciprocal situations; and the word that formerly described all reciprocal situations now applies only in some of those contexts. The choice is made according to degree of transitivity – one word is used in semi-transitive clauses, and the other in transitive clauses. In addition, the emphatic marker, which used to be segmentally identical to the reflexive marker, has now developped as a distinct form. It appears that multiple levels of language contact have contributed to the innovations. Nicholls’ paper continues the examination of substrate influences of Australian languages on Kriol. In this study, Nicholls provides evidence for stages of the grammaticalisation of the English distal demonstrative that into a recognitional article det in Roper Kriol and suggests pragmatic motivations for such a change. Nicholls describes the use of det in discourse to introduce familiar referents to the discourse while allowing the speaker to maintain a high level of circumspection, particularly in relation to place and people reference. She argues that this function is not found in article systems in other creole languages, such as Melanesian pidgins, but rather is the result of influence from Australian languages which have articles that precede proper names, generic terms and mark topics in the same manner as Roper Kriol. Nicholls suggests that the transfer of this feature from Australian languages has created a continuity of discourse practices from traditional Aboriginal languages into Roper Kriol.

20

Carmel O’Shannessy and Felicity Meakins

Meakins’ chapter on the grammaticalisation of the locative case marker in Gurindji Kriol into a present tense suffix begins Section Four which focuses on (further) developments of new structures. Gurindji Kriol is a mixed language spoken in northern Australia which is a lexical and structural fusion of Gurindji, a Pama-Nyungan language, and Kriol, an English-lexifier creole language (§3.3). Gurindji Kriol has maintained the case system of Gurindji, albeit with some innovations. One of those innovations is the use of the locative case suffix -ja/-ngka to mark the present (progressive) in Gurindji Kriol in presentative constructions. Cross-linguistically, it is not unusual for locative constructions to grammaticalise into progressive markers and later present tense markers (Bybee, Perkins, & Pagliuca, 1994). This process occurs as the locative undergoes a metaphorical extension from indicating space to indicating time. Insubordination involving case markers such as the locative seems to be one path which contributes to this grammaticalisation process and has been historically reconstructed for a number of Australian languages (Evans 2007). In the case of Gurindji Kriol, the process has occurred within a single generation of Gurindji people and seems to have been accelerated by language contact. Thus Meakins suggests that situations such as these provide a window on language change processes such as grammaticalisation. In other language contact situations, an expansion in a minor use pattern occurs as a result of accommodating material from another language (Heine & Kuteva 2005). Mansfield demonstrates such a change in The rise of phrasal verbs in Murrinhpatha. Murrinhpatha is a polysynthetic language of the Daly River region in the Northern Territory which is one of the few Australian languages still being learnt by children. Mansfield argues that a recent and substantial increase in the use of a phrasal verb structure in Murrinhpatha, alongside the existing morphologically complex verb, is largely due to an influx of verb borrowings from English/Kriol. This structure is widespread in its usage amongst younger Murrinhpatha speakers and is one of many linguistic differences now apparent between generations of Murrinhpatha speakers. The volume ends with Gender Bender: Realignment of Gender in Jingulu. Meakins and Pensalfini argue that obsolescence situations where languages are undergoing rapid change may provide valuable data for understanding change in languages with stronger speaker bases. They claim that, although obsolescing languages are often characterised as displaying high levels of variation and optionality, changes are nonetheless highly rule governed. Meakins and Pensalfini use the noun class system of Jingulu, a highly endangered non-Pama-Nyungan language, as an example of such principled change. Jingulu was documented by Chadwick in the 1970s and later by Pensalfini in the 1990s. During this time, Jingulu had undergone a number of changes attributable to language obsoles-

Australian language contact in historical and synchronic perspective

21

cence. One change occurred in the noun class system. Jingulu distinguishes four genders: masculine (I), feminine (II), vegetable (III) and neuter (IV). NP modifiers such as adjectives and demonstratives generally show agreement in gender with the head noun, however Pensalfini (1999a) also observed that modifiers can optionally ‘disagree’ with their head. Disagreement is principled and hierarchical with masculine-marked modifiers optionally found with heads of all four genders and neuter-marked modifiers optionally found with heads of the vegetable gender. This system was not in place when Chadwick first documented the language 20 years prior, suggesting the phenomenon described is the result of language change. However similar patterns have also been observed in other languages, for example Wambaya (Nordlinger 1998) and a number of Gunwinyguan languages (Evans, Brown, & Corbett 2002) but are not considered the result of obsolescence, suggesting that the health of a language is irrelevant as an explanatory mechanism for gender disagreement.

References Alexander, Diane. 1968. Woorabinda Aboriginal English. (MA thesis), University of Queensland, Brisbane. Alpher, Barry. 2004. Pama-Nyungan: phonological reconstruction and status as a phylogenetic group. In Claire Bowern & Harold Koch (eds.), Australian languages: Classification and the comparative method, 93–126. Amsterdam: John Benjamins. Amery, Rob, & Peter Mühlhäusler. 1996. Pidgin English in New South Wales. In Stephen Wurm, Peter Mühlhäusler & Darrell Tryon (eds.), Atlas of languages of intercultural communication in the Pacific, Asia and the Americas Vol. 2.1. 33–52. Berlin: Mouton de Gruyter. Angelo, Denise. 2013. Identification and assessment contexts of Aboriginal and Torres Strait Islander learners of Standard Australian English: Challenges for the language testing community. Papers in Language Testing and Assessment (PLTA), 2(2). 67–102. Angelo, Denise & Nina Carter. 2015. Schooling within shifting langscapes: Educational responses in complex language contact ecologies. In: Androula Yiakoumetti (ed.), Multilingualism and language in education. Sociolinguistic and pedagogical perspectives from Commonwealth countries, 119–140. Cambridge: Cambridge University Press. Angelo, Denise & Sophie McIntosh. 2012. The (In)visibility of “language” within Australian educational documentation: Differentiating language from literacy and exploring particular ramifications for a group of “hidden” ESL/D Learners. In Richard Baldauf, (ed.), Future directions in Applied Linguistics: Local and global perspectives – 35th Applied Linguistics Association Australia (ALAA) Congress. 447–468. Brisbane: The University of Queensland Press. Baker, Peter. 1993. Australian influence on Melanesian Pidgin English. Te Reo, 36. 3–67. Bavin, Edith, & Tim Shopen. 2004. Warlpiri in the 80s: An overview of research into language variation and child language. In Suzanne Romaine (ed.), Language in Australia. Cambridge: Cambridge University Press.

22

Carmel O’Shannessy and Felicity Meakins

Bowern, Claire, & Harold Koch. 2004. Introduction: Subgrouping and methodology in historical linguistics. In Claire Bowern & Harold Koch (eds.), Australian languages: Classification and the comparative method. 1–16. Amsterdam: John Benjamins Bowern, Claire, & Quentin Atkinson. 2012. Computational phylogenetics and the internal structure of Pama-Nyungan. Language, 88(4). 817–845. Bowern, Claire, Patience Epps, Russell D. Gray, Jane Hill, Keith Hunley & Patrick McConvell. 2011. Does lateral transmission obscure inheritance in hunter-gatherer languages? PloS One 6.9.e25195. Bowern, Claire, Hannah Haynie, Catherine Sheard, Barry Alpher, Patience Epps, & Jane Hill. 2014. Loan and inheritance patterns in hunter-gatherer ethnobiological systems. Journal of Ethnobiology, 34(1). 195–227. Bybee, Joan, Rever Perkins & William Pagliuca. 1994. The Evolution of Grammar: Tense, aspect and modality in the languages of the world. Chicago: The University of Chicago Press. Clark, Ian, Peter Mühlhäusler & Rob Amery. 1996. Language contacts and Pidgin English in Victoria. In Stephen Wurm, Peter Mühlhäusler & Darrell Tryon (eds.), Atlas of languages of intercultural communication in the Pacific, Asia and the Americas Vol. 2.1. 53–68. Berlin: Mouton de Gruyter. Clark, Ross. 1979. In search of Beach-la-Mar. Towards a history of Pacific Pidgin English. Te Reo, 22. 3–64. Crowley, Terry, & Bruce Rigsby. 1979. Cape York Creole. In Tim Shopen (ed.), Languages and their status. Cambridge, Mass.: Winthrop Publishers. Dench, Alan. 1998. Pidgin Ngarluma: An Indigenous contact language in north western Australia. Journal of Pidgin and Creole languages, 13. 1–61. Dench, Alan. (001. Descent and diffusion: The complexity of the Pilbara situation. In A. Aikhenvald & Robert M. W.Dixon (eds.), Areal diffusion and genetic inheritance: Problems in comparative linguistics, 105–133. Oxford: Oxford University Press. Dixon, Robert M. W. 1980. The languages of Australia. Cambridge: Cambridge University Press. Dixon, Robert M. W. 2001. The Australian linguistic area. In A. Aikhenvald & Robert M. W.Dixon (eds.), Areal diffusion and genetic inheritance, 64–103. Oxford: Oxford University Press. Dixon, Robert M. W. 2002. Australian languages: Their nature and development. Cambridge: Cambridge University Press. Dutton, Tom. 1983. The origin and spread of Aboriginal Pidgin English in Queensland: A preliminary account. Aboriginal History, 7. 90–120. Eades, Diana. 2008. Courtroom talk and neocolonial control. Berlin: Mouton de Gruyter. Eades, Diana. 2013. Aboriginal ways of using English. Canberra: Aboriginal Studies Press. Eades, Diana. 2014. Aboriginal English. In H. Koch & R. Nordlinger (eds.), The languages and linguistics of Australia: A comprehensive guide, 417–447. Berlin: Mouton de Gruyter. Evans, Nicholas. 1988. Arguments for Pama-Nyungan as a genetic subgroup, with particular reference to initial laminalization. Aboriginal Linguistics 1. 91–110. Evans, Nicholas. 1992. Macassan loan words in Top End languages. Australian Journal of Linguistics, 12. 45–91. Evans, Nicholas. 2007. Insubordination and its uses. In I. Nikolaeva (ed.), Finiteness: Theoretical and Empirical Foundations, 366–431. Oxford: Oxford University Press. Evans, Nicholas. 2010. Dying words. Malden/Oxford: Wiley-Blackwell. Evans, Nicholas, Dunstan Brown & Greville Corbett. 2002. The semantics of gender in Mayali: Partially parallel systems and formal implementation. Language, 78(1). 111–155.

Australian language contact in historical and synchronic perspective

23

Evans, Nicholas, & Patrick McConvell. 1998. The enigma of Pama-Nyungan expansion in Australia. In Roger Blench & Matthew. Spriggs (eds.), Archeology and language II: Correlating archeological and linguistic hypotheses, 174–191. London/New York: Routledge. Flint, Elwyn. 1968. Aboriginal English: Linguistic description as an aid to teaching. English in Australia, 6.3–21. Foster, Robert, Paul Monaghan & Peter Mühlhäusler. 2003. Early Forms of Aboriginal English in South Australia, 1840s–1920s. Canberra: Pacific Linguistics. García, Ofelia. 2009. Education, multilingualism and translanguaging in the 21st century. In A. Mohanty, M. Panda, R. Phillipson & T. Skutnabb-Kangas (eds.), Multilingual Education for Social Justice: Globalising the local, 140-158. New Delhi: Orient Blackswan. Gardani, Francesco. 2008. Borrowing of inflectional morphemes in language contact. Frankfurt: Peter Lang. Harkins, Jean. 1994. Bridging two worlds: Aboriginal English and cross-cultural understanding. Brisbane: University of Queensland Press. Harris, Jean. 1986. Northern Territory pidgins and the origin of Kriol. Canberra: Pacific Linguistics. Haynie, Hannah, Claire Bowern, Patience Epps, Jane Hill & Patrick McConvell. 2014. Wanderwörter in languages of the Americas and Australia. Ampersand, 1. 1–18. Heath, Jeffrey. 1978. Linguistic diffusion in Arnhem Land. Canberra: Australian Institute of Aboriginal Studies. Heath, Jeffrey. 1981. A case of intensive lexical diffusion: Arnhem Land, Australia. Language, 57. 335–367. Heine, Berndt, & Tania Kuteva. 2005. Language contact and grammatical change. Cambridge: Cambridge University Press. Hercus, Luise. 2005. The influence of English on possessive systems as shown in two Aboriginal languages, Arabana (northern SA) and Paakantyi (Darling River, NSW). Monash University Linguistics Papers, 4(2). 29–42. Hercus, Luise. 1979. In the margins of an Arabana-Wanganguru dictionary: The loss of initial consonants. In Stephen A. Wurm (ed.), Australian Linguistic Studies (Pacific linguistics C-54), 621–651. CAnberra: Austalian National University. Hosokawa, Komei. 1987. Malay talk on boat: An account of Broome Pearling Lugger Pidgin. In Don Laycock & Werner Winter (eds.), A World of Language: Papers presented to Professor S. A. Wurm on his 65th birthday, Vol. C-100. 287–296. Canberra: Pacific Linguistics. Koch, Harold. 1997. Comparative linguistics and Australian prehistory. In Patrick McConvell & Nicholas Evans (eds.), Archaeology and linguistics: Aboriginal Australia in global perspective, 27–43. Oxford: Oxford University Press. Koch, Harold. 2000. Central Australian Aboriginal English: In comparison with the morphosyntactic categories of Kaytetye. Asian Englishes: An international journal of the sociolinguistics of English in Asia/Pacific, 3(2). 32–58. Koch, Harold. 2004. A methodological history of Australian linguistic classification. In Claire Bowern & Harold. Koch (eds.), Australian languages: Classification and the comparative method, 17–66. Amsterdam: John Benjamins. Koch, Harold. 2011. The influence of Arandic languages on Central Australian Aboriginal English. In Claire Lefebvre (ed.), Creoles, their substrates and language typology, 437– 460. Amsterdam: John Benjamins. Langlois, Annie. 2004. Alive and kicking: Areyonga Teenage Pitjantjatjara. Canberra: Pacific Linguistics.

24

Carmel O’Shannessy and Felicity Meakins

Lee, Jennifer. 1987. Tiwi today: A study of language change in a contact situation. Canberra: Pacific Linguistics. Malcolm, Ian and Susan Kaldor. 1991. Aboriginal English: An overview. In Romaine, Suzanne (ed.) Language in Australia, 67–83. Cambridge, Cambridge University Press. Malcolm, Ian. 2000. English and inclusivity in education for Indigenous students. Annual review of Applied Linguistics, 22(2). 51–66. Malcolm, Ian. 2008. Australian creoles and Aboriginal English: Phonetics and phonology. In Kate Burridge & Berndt Kortmann (eds.), Varieties of English 3: The Pacific and Australiasia, 124–141. Berlin: Mouton de Gruyter. Matras, Yaron. 2003. Mixed languages: Re-examining the structural prototype. In Y. Matras & P. Bakker (eds.), The mixed language debate: Theoretical and empirical advances, 151– 176. Berlin: Mouton de Gruyter. Matras, Yaron. 2007. The borrowability of structural categories. In Y. Matras & J. Sakel (eds.), Grammatical borrowing in cross-linguistic perspective, 31–73. Berlin: Mouton de Gruyter. Matras, Yaron. 2009. Language Contact. Cambridge: Cambridge University Press. Matras, Yaron & Peter Bakker. (eds.), 2003. The mixed language debate: Theoretical and empirical advances. Berlin: Mouton de Gruyter. Matras, Yaron & Jeanette Sakel. 2007. Grammatical borrowing in cross-linguistic perspective. Berlin: Mouton de Gruyter. McConvell, Patrick. 1985. The origin of subsections in northern Australia. Oceania, 56(1 (Special Aboriginal Issue)), 1–33. McConvell, Patrick. 1996. Backtracking to Babel: The Chronology of Pama-Nyungan Expansion in Australia. Archeology in Oceania, 31(3). 125–144. McConvell, Patrick. 2002. Mix-im-up speech and emergent mixed languages in Indigenous Australia. Texas Linguistic Forum (Proceedings from the 9th Annual Symposium about Language and Society), 44(2). 328–349. McConvell, Patrick. 2009. Loanwords in Gurindji, a Pama-Nyungan language of Australia. In Martin Haspelmath & Uri Tadmor (eds.), Loanwords in the world’s languages: A comparative handbook, 790–822. Berlin: Mouton de Gruyter. McConvell, Patrick. 2010a. The archaeolinguistics of migration. In Jan Lucassen, Leo Lucassen & Patrick Manning (eds.), Migration history in world history: Multidisciplinary approaches, 155–190. Leiden: Brill. McConvell, Patrick. 2010b. Contact and Indigenous languages in Australia. In Ray Hickey (ed.), The handbook of language contact, 770–794. Malden: Wiley Blackwell. McConvell, Patrick & Claire Bowern. 2011. The prehistory and internal relationships of Australian languages. Language and Linguistics Compass, 5(1), 19–32. McConvell, Patrick & Mary Laughren. 2004. The Ngumpin-Yapa subgroup. In C. Bowern & H. Koch (eds.), Australian Languages: Classification and the comparative method, 151–177. Amsterdam: Benjamins. McConvell, Patrick & Felicity Meakins. 2005. Gurindji Kriol: A mixed language emerges from code-switching. Australian Journal of Linguistics, 25(1). 9–30. Meakins, Felicity. 2011a. Borrowing contextual inflection: Evidence from northern Australia. Morphology, 21(1). 57–87. Meakins, Felicity. 2011b. Case marking in contact: The development and function of case morphology in Gurindji Kriol. Amsterdam: John Benjamins.

Australian language contact in historical and synchronic perspective

25

Meakins, Felicity. 2013a. Gurindji Kriol. In Susanne Michaelis, Phillippe Maurer, Martin Haspelmath & Magnus Huber (eds.), The survey of pidgin and creole languages, Vol III, 131–139. Oxford: Oxford University Press. Meakins, Felicity. 2013b. Mixed languages. In Yaron Matras & Peter Bakker (eds.), Contact Languages: A Comprehensive Guide, 159–228. Berlin: Mouton. Meakins, Felicity. 2014. Language contact varieties. In H. Koch & R. Nordlinger (eds.), The Languages and Linguistics of Australia: A comprehensive guide, 365–416. Berlin: Mouton de Gruyter. Meakins, Felicity. To appear. Mixed languages. In Mark Aronoff (ed.), Oxford Research Encyclopedia of Linguistics. Oxford: Oxford University Press. Meakins, Felicity & Carmel O’Shannessy. 2010. Ordering arguments about: Word order and discourse motivations in the development and use of the ergative marker in two Australian mixed languages. Lingua, 120(7). 1693–1713. Meakins, Felicity & Carmel O’Shannessy. 2012. Typological constraints on verb integration in two Australian mixed languages. Journal of Language Contact, 5(2). 216–246. Mühlhäusler, Peter. 1996a. Post contact languages in mainland Australia after 1788. In Stephen Wurm, Peter Mühlhäusler & Darrell Tryon (eds.), Atlas of Languages of Intercultural Communication in the Pacific, Asia and the Americas Vol. 2.1. 11–16. Berlin: Mouton de Gruyter. Mühlhäusler, Peter. 2008. History of research into Australian pidgins and creoles. In W. McGregor (ed.), Encountering Aboriginal Languages: Studies in the History of Australian Linguistics, 437–457. Canberra: Pacific Linguistics. Munro, Jennifer. 2000. Kriol on the move: A case of language spread and shift in Northern Australia. In Jeff. Siegel (ed.), Processes of language contact: Studies from Australia and the South Pacific, 245–270. Saint-Laurent (Quebec): Fides. Mushin, Ilana, Denise Angelo & Jen Munro. In press. Same but different: Understanding language contact in Queensland Indigenous Settlements. In Jean-Christophe Verstraete and Di Hafner (eds.), Land and Language in Cape York Peninsula and the Gulf Country. Nordlinger, Rachel. 1998. A grammar of Wambaya, Northern Territory (Australia). Canberra: Pacific Linguistics. O’Grady, Geoff., Charles F.Voegelin & Florence M. Voegelin. 1966. Languages of the world: IndoPacific fascicle 6. Anthropological Linguistics, 8(2). 1–199. O’Shannessy, Carmel. 2005. Light Warlpiri – A new language. Australian Journal of Linguistics, 25(1). 31–57. O’Shannessy, Carmel. 2012. The role of code-switched input to children in the origin of a new mixed language. Linguistics, 50(2). 305–340. O’Shannessy, Carmel. 2013. The role of multiple sources in the formation of an innovative auxiliary category in Light Warlpiri, a new Australian mixed language. Language, 89(2). 328–353. O’Shannessy, Carmel. To appear. Distributions of case allomorphy by multilingual children speaking Warlpiri and Light Warlpiri. Linguistic Variation. Special issue on Child Language Variation edited by Véronique Lacoste & Lisa Green. Pensalfini, Robert. 1999a. Optional disagreement and the case for feature hierarchies. Chicago Linguistics Society, 35. 343–353. Pensalfini, Robert. 1999b. The rise of case suffixes as discourse markers in Jingulu – a case study of innovation in an obsolescent language. Australian Journal of Linguistics, 19(2). 225–239.

26

Carmel O’Shannessy and Felicity Meakins

Richards, Norval. 2001. Leerdil Yuujmen bana Yanangarr (Old and New Lardil). In Jane Simpson, David Nash, Peter Austin & Barry Alpher (eds.), Forty years on: Ken Hale and Australian languages, 231–245. Canberra: Pacific Linguistics. Sandefur, John. 1979. An Australian creole in the Northern Territory: A description of NgukurrBamyili dialects (Part 1). Darwin: SIL. Sandefur, John & John Harris. 1986. Variation in Australian Kriol. In Joshua Fishman (ed.), The Fergusonian impact, 179–190. Berlin: Mouton de Gruyter. Schmidt, Annette. 1985. Young people’s Dyirbal: An example of language death from Australia. Cambridge: Cambridge University Press. Schultze-Berndt, Eva, Felicity Meakins & Denise Angelo. 2013. Kriol. In Susanne Michaelis, Phillippe Maurer, Martin Haspelmath & Magnus Huber (eds.), The survey of pidgin and creole languages, Vol I, 241–251. Oxford: Oxford University Press. Sellwood, Juanita & Denise Angelo. 2013. Everywhere and Nowhere: Invisibility of Aboriginal and Torres Strait Islander contact languages in education and Indigenous language contexts. Australian Review of Applied Linguistics, 36(3). 250–266. Shnukal, Anna. 1988. Broken: an introduction to the creole language of Torres Strait (Pacific Linguistics C-107). Canberra: Research School of Pacific Linguistics, Australian National University. Simpson, Jane. 1996. Early language contact varieties in South Australia. Australian Journal of Linguistics, 16(2). 169–207. Teasdale, Robert & Albert Whitelaw. 1981. The early childhood education of Aboriginal Australians: Australian Council for Educational Research. Thomason, Sarah & Terrence Kaufman. 1988. Language contact, creolization, and genetic linguistics. Berkeley: University of California Press. Troy, Jaky. 1990. Australian Aboriginal Contact with the English Language in New South Wales: 1788 to 1845. Canberra: Pacific Linguistics. Troy, Jaky. 2003. Language contact in early colonial New South Wales. In M. Walsh & C. Yallop (eds.), Language and Culture in Aboriginal Australia, 33–50. Canberra: Aboriginal Studies Press. Urry, James & Michael Walsh. 1981. The lost ‘Macassar language’ of northern Australia. Aboriginal History, 5. 91–108. Wurm, Stephen, Peter Mühlhäusler & Darrell Tryon (eds.), 1996. Atlas of languages of intercultural communication in the Pacific, Asia and the Americas Vol. 2.1. Berlin: Mouton de Gruyter.

II Transfer of form: Structure

Nicholas Evans

1 As intimate as it gets? Paradigm borrowing in Marrku and its implications for the emergence of mixed languages1 Abstract: Marrku, now close to extinct, is the language of Croker Island in the Northern Territory. Existing classifications of Australian languages have assigned Marrku to the same family as Iwaidja, Mawng and Amurdak in what is most commonly known as the Iwaidjan family (Schmidt 1919, O’Grady et al. 1966, Evans 2000, Dixon 2002). In fact the level of shared cognacy between Marrku and other languages of this putative family is quite low, so that what has appeared to be the best evidence for genetic relatedness comes from what appear to be shared patterns of prefixal morphology. Though Marrku verbs in particular have highly irregular morphological paradigms, with a large number of quite distinct patterns according to the lexical item involved, some show significant paradigmatic resemblances to verbs in Iwaidja or Mawng. Recent work on Marrku has allowed us to extend the analysis of Marrku grammar, by enlisting the aid of two ‘last hearers’ (Joy Williams and Khaki Marrala) to transcribe and translate hitherto unanalysed recordings made in the 1960s. Though our understanding is still fragmentary, it appears increasingly likely that Marrku is less close to the other Iwaidjan languages than was previously believed. Rather, there appears to have been borrowing of entire paradigms of at least two inflected verbs from other Iwaidjan languages. (Though it is unclear whether such extreme cases of paradigm borrowing reflect code-mixing in a language death situation, or resulted from more ‘normal’ areal influence facilitated by the fact that most verbs had their own distinct paradigms anyway.) In this paper I will re-evaluate the genetic position of Marrku, focussing on the very distinctive morphological structure of the Marrku verb and why it suggests that at least two verb paradigms have been borrowed wholesale – as well as the prefixal paradigm of reflexive pronouns. A reconsideration of the evidence, I will argue, points to Marrku being a family-level isolate, rather than 1 It is a great pleasure to include this paper in a volume dedicated to Patrick McConvell, whose friendship and collaboration have enriched my work and life in so many ways. His many works on language contact and change have been merely one aspect of this, and almost every conversation I have ever had with him has left me enlightened and challenged across a range of other problems from kinship to fine semantics to linguistic anthropology to the quest to put together the clues from language, archaeology, genetics, anthropology, palynology and other fields.

30

Nicholas Evans

a member of the Iwaidjan family – though belonging, at a deeper level, to the Australian phylum. At the same time, Marrku may be considered as an incipient mixed language of older vintage than more recently identified mixed languages such as Gurindji Kriol and Light Warlpiri.

1 Introduction Von aussen gesehen, bot uns eine Sprache nicht das Bild einer abgeschlossenen Einheit dar; nun zeigt sie sich auch ihrem innern Bau nach nicht als eine solche, sondern als eine Zusammensetzung aus Tatsachen, die zwar miteinander in mehr oder minder festem Verband stehen, aber doch nicht in unlösbarem – sonst wäre ja Mischung unmöglich.2 (Schuchardt 1928: 195) The transfer of entire inflectional paradigms has long been regarded as the last challenge to morphological borrowability (Gardani, Arkadiev & Amiridze 2014:11)

The existence of mixed languages in Australia has been shown by a spate of recent publications (Charola 2002; McConvell 2008; Meakins 2010, 2011, 2012; Meakins and O’Shannessy 2010; O’Shannessy 2012, 2013), as well as Meakins (this volume) and O’Shannessy (this volume). In each case, these derive from a combination of a traditional language (Gurindji and Warlpiri respectively) with a creolised form of English. So far, however, we have no reported examples of mixed languages forming by the interaction of two traditional Australian languages. In this chapter I discuss the case of Marrku, which appears to have borrowed a number of complex verb and nominal paradigms from neighbouring languages (Iwaidja, Ilgar/Garig) to which it is at best distantly related. Unfortunately the limited nature of our documentation of Marrku, and the fact that it is no longer spoken,3 places limits on what we can say about its analysis. Nonetheless, we have enough data to identify the sorts of wholesale mixing of complex morphology which characterise such well-known mixed languages as Michif (Bakker 1997) and Mednyj Aleut (Thomason 1997). While I will not be going so far as to characterise Marrku as a mixed language, the 2 “Seen from outside, language does not present us with the image of an enclosed unit; nor does it so appear according to its inner construction, but rather as a collection of facts, which stand in more or less tight connection to each other, though not indissolubly connected – otherwise mixture would be impossible.” (Translation NE) 3 For details see Evans (2001), Evans et al. (2006) and Evans (2009); since the situation described in those publications a key “last hearer”, Joy Williams Malwagag, has passed away, leaving only Khaki Marrala as an extremely frail last hearer and partial speaker.

Paradigm borrowing in Marrku and its implications

31

wholesale borrowing of complete paradigms resembles what Seifart (2012: 498) has said about Resígaro, as constituting “an intermediate case, a ‘missing link’ between borrowing and language mixing that may eventually help to bridge the perceived gap between ‘normal’ contact-induced changes and mixed languages”. Nonetheless, Seifart (2012: 473) signals a criterion which is only met by a subset of the Marrku material to be considered here: “Derivational and inflectional morphemes are considered as e1ectively borrowed only when they are used on at least some native stems, i.e. when their use is not restricted to equally borrowed stems.” Now it might be expected that a first stage in the borrowing of inflectional systems would be ‘trojan horse borrowings’ (Meakins 2011), in the form of borrowed words that host the relevant paradigms, with the extension to native stems being a later process. In that sense, we need to be on the lookout for both phenomena if we are to understand the full sequence of steps accompanying the borrowing of inflectional systems.4 In my view, we therefore need to include both situations in our purview, and it is helpful to have terms to distinguish them. For the case where borrowed paradigmatic material is confined to loanword hosts I will employ the terms ‘hosted inflectional borrowing’ (or ‘hosted paradigmatic borrowing’ when dealing with paradigms), while for situations where the borrowed material has been extended to native hosts I will use the term ‘recombinant inflectional borrowing’ (or ‘recombinant paradigmatic borrowing’).5 As we shall see, both types of borrowing are found in Marrku: hosted paradigmatic borrowing in the case of a couple of verbs, and recombinant paradigmatic borrowing in the case of pronominal prefixes on reflexive pronouns. The recognition that a portion of Marrku’s paradigmatic morphology – at least two verb paradigms and the prefixal paradigm of reflexive pronouns – is borrowed rather than inherited leads us to re-evaluate its claimed genetic position within the Iwaidjan family (as proposed in Evans 2000). A reconsideration of the

4 This point was already made by Weinreich (1968: 31): “Thus the plural ending -im in Yiddish pójerim ‘peasants,’ doktórjim ‘doctors’ is only ultimately, but not directly, of Hebrew origin; it is rather an analogical extension of the -im-plural from such Yiddish couples as min–mínim ‘sort’, gíber–gibójrim ‘strong man’, etc., etc. – free morphemes borrowed in pairs from Hebrew.” 5 Two deviations between my terminology and those used elsewhere in the literature should be noted here. First, some others (e.g. Gardani 2008) only regard morphology as borrowed once it has appeared on native material (i.e. my recombinant borrowing), so my use of the term is broader (and motivated by the fact that I regard Gardani’s terminology as unduly restrictive). Secondly, my terms ‘hosted inflectional borrowing’ and ‘recombinant inflectional borrowing’ correspond to the opposition ‘oikoclitic’ vs ‘xenoclitic’ in Elšík & Matras (2006: 324); I do not use these terms because of the potential for the ‘clitic’ component to be misleading in metalinguistic terminology. See also the discussion in Kossmann (2010) on ‘parallel system borrowing’.

32

Nicholas Evans

Figure 1: Languages of the Cobourg Peninsula region

evidence, I will argue, points to Marrku being a family-level isolate, rather than a member of the Iwaidjan family – though belonging, at a deeper level, to the Australian phylum.

2 Marrku and the Iwaidjan languages Marrku (also sometimes spelled Marrgu; ISO 639-3 mhg) was traditionally spoken on Croker Island in the Northern Territory of Australia. There are no fluent speakers left, though a couple of surviving people remember small amounts of the language. In addition to written material collected by Capell (1963) or one of his research assistants (mostly word-lists and questionnaire-type elicitations), some recordings were made in the 1960s by Heather Hinch (a missionary linguist who knew Mawng and was based on Goulburn Island) and Bernhard Schebeck (a linguist primarily working on Yolngu). These latter sources included some textual material recorded from a number of speakers (Alf Brown, Jumbo Jambululu, Dickie Malwagu and Hazel Mamiyarr), which I was able to transcribe in 2003–2005 with the assistance of the late Joy Williams, daughter of Hazel

Paradigm borrowing in Marrku and its implications

33

Mamiyarr. I was also able to record a limited amount of material during the 1990s from two senior men, Mick Yarmirr and Charlie Wardaga, both of whom spoke some Marrku (though not fluently). In the 1960s, when Hinch and Schebeck made their recordings, there appear to have been perhaps six Marrku speakers still alive; we know for certain that some of them spoke other languages (e.g. that Alf Brown also spoke Garig) and it is likely from typical language portfolios of senior Croker Islanders that they would in general have spoken Ilgar/Garig (two very closely related variants, with Garig spoken around Port Essington and Ilgar on the small islands to the east of Croker, such as Grant Island and New Year Island), Iwaidja and possibly Kunwinjku in addition to Marrku. Most existing classifications of Marrku and its neighbours place Marrku, along with the Popham Bay language Wurruku, in a group with Iwaidja, Garig/ Ilgar, Mawng (aka Maung) and Amurdak (essentially the same in Schmidt 1919; O’Grady et al. 1966; Walsh 1981; Dixon 20026; Evans 2000, 2003b). There is perhaps 20% shared vocabulary between Marrku and Iwaidja, though it is difficult to assess how much is loaned, and the figure drops to less than 5% for verbs. The family tree proposed in Evans (2000) is given below.

Figure 2: Proposed relations between languages of the Iwaidjan family (Evans 2000)

The earlier classifications (Schmidt 1919; O’Grady et al. 1966) predominantly drew on the not insignificant overall proportion of shared vocabulary (without giving extra weighting to the evidence from the number of cognate verbs, which would have pulled the figures down). The main evidence drawn upon in Evans (2000) came from certain paradigmatic similarities in the reflexive pronoun and some inflected verbs, predicated on the assumption that these would not have 6 Though Dixon (2002: 668) expresses skepticism about the evidence for including Marrku, Amurdak and the Popham Bay language), pointing out that the evidence for including them in this family is minimal.

34

Nicholas Evans

been borrowed. However, analytic further work on Marrku since then, including a more systematic approach to its verb paradigms (see Evans, Malwagag, and Marrala 2006), makes a paradigm-borrowing scenario more likely, as I will argue below.

3 “Typical” Marrku verb paradigms Though all the above languages are head-marking languages with considerable verb morphology, there are significant differences between Marrku and the other Iwaidjan languages when it comes to the forms, ordering and semantics of its verbal affixes.7 The basic structure of the Marrku verb, close to maximally expanded,8 is illustrated in (1); (2–4) give simpler and more typical examples. Forms with no overt marker of tense, e.g. (2), are construed as present, and those with no overt marker of number are construed as singular. Note that, in the practical Marrku orthography used here, h = /ɰ/ (only between two vowels), ng = /ŋ/, nh = /ṋ/, ny = /ɲ/, r = /ɻ /, rn = /ɳ /, rr = /r/, rt = /ɖ /, th = /ṱ/. (1)

Tense Subject mangapast 1st person mangawulkayi ‘We

Aspect Plural Stem wulkayi durative plural be, sit used to stay, used to live’

(2)

Subject Stem ngalayi 1st person be, sit ngalayi ‘I am, I remain’

(3)

Tense Subject Stem mangalayi past 1st person be, sit mangalayi ‘I was living’

7 In my comparisons I focus on Ilgar/Garig and Iwaidja, the languages which are geographically nearest and most likely to have formed part of Marrku speakers’ language portfolio. For information on Amurdak see Mailhammer (2009) and references therein, and on Mawng see Capell & Hinch (1970) and Singer (2011, 2015). 8 Some Marrku verbs have an overt tense suffix, e.g. past tense -yi in (5); in this example the final yi is part of the stem, as shown by its recurrence in the present form in (2).

Paradigm borrowing in Marrku and its implications

(4)

35

Tense Subject Aspect Stem mangawulayi past 1st person durative be, sit mangawulayi ‘I used to stay, used to live’

A first wrinkle to the above structure comes from the fact that some verbs have suppletive roots. For example, ‘come, arrive’ has the past root -urtyi and the nonpast root jahan (5): (5)

Tense Subject Plural Stem Tense maØ lkurt -yi past 3rd person plural arrive past malkurtyi ‘they arrived, they came’ (cf mangurtyi ‘I arrived’)

(6)

Subject Prefix Stem+Tense ngajahan 1st person arrive:NPst ngajahan ‘I come, arrive’ (cf. kirrijahan ‘I will arrive’)

A second wrinkle is the presence of certain subject prefix suppletions, in that different verbs select quite different forms for the same person/number combinations across the future vs non-future contrast (7). (7)

a.

kuthirri ‘I went’

kirrithirri ‘I will go’

b.

ngajahan ‘I came’

kirrijahan ‘I will come’

c.

ngalawuthi ~ thawuthi ‘I talked’

kirrilawuthi ‘I will talk’

It is possible that some of this may result from portmanteaux for subject + direction: a possible analysis is that ku- means ‘1sg present, away’ and nga- ‘1sg present, towards/neutral’. A comparable phenomenon is also found in Iwaidja and Ilgar, e.g. Ilgar ja- ‘1sgSubj:away’ vs nya- ‘1sgSubj:towards’. We certainly have some third person prefixes which show this contrast (8–10), but do not have clinching examples with the 1st singular, and note that neither the ‘towards/neutral’ vowel /a/ nor the ‘away’ vowel /i(yi)/ from (8–10) appear in the examples in (7).

36

Nicholas Evans

(8)

makaladbany ‘it went down, descended, set’ vs mikaladbany ‘it went down, away’9

(9)

malkurtyi ‘they came, arrived, appeared, turned up (here)’ vs miyilkurtyi ‘they turned up there, they arrived there’

(10)

manangayi ‘they got it/him/her’ vs minangayi ‘they took him away’

4 Comparison of Marrku verb morphology with Iwaidja and Ilgar/Garig If we turn to Marrku’s neighbours, Iwaidja and Ilgar/Garig, we see significant contrasts in how their verb morphology is organised. Note first that Iwaidja and Ilgar/Garig are very close, almost sister dialects. The two most important differences are that (a) Ilgar/Garig retains the proto-Iwaidjan masculine and feminine gender prefixes, for (some) nouns, (some) adjectives and in verbal agreement (both subject and object), whereas Iwaidja has almost entirely jettisoned them: the only exception is for the 3sg>3sg prefix combinations on verbs, which distinguish masculine from feminine subjects (ri- vs ka- respectively) as long as the referents are human. (b) Linked with its jettisoning of ancestral gender prefixes, Iwaidja has generalised an originally obscure fifth gender (the miscellaneous) aK-10, which had the morphophonemic effect of hardening following consonants (semivowels and nasals to stops at the same point of articulation) and eventually producing initial mutation in the singular forms and stop-initial singular forms in Iwaidja corresponding to sequences of gender prefix plus stop in Ilgar/Garig: cf. Ilgar/Garig imawurr ‘his arm’, inymawurr ‘her arm’, amawurr ‘their arms’; Iwaidja bawurr ‘his/her/its arm’, amawurr ‘their arms’, Ilgar/ Garig iwani ‘he sits’, inybani ‘she sits’, awani ‘they sit’; Iwaidja bani ‘he/she sits’, awani ‘they sit’. 9 As Iwaidja translations for this pair, Joy Williams offered bulakuny ‘it descended’ (direction neutral) for makaladbany and ijuwulakuny (direction away) ‘it went down / away’ for mikaladbany. 10 Where K- is a morphophoneme producing hardening in following semi-vowels and nasals (e.g. w > b, ng > k, m > b); the ancestral a, still attested in Mawng, is in Iwaidja only still found if the total prefixed word does not exceed two syllables, e.g. aK-yu [3sgS-lie] > aju ‘it lies’, and elsewhere has been lost.

Paradigm borrowing in Marrku and its implications

37

See Evans (1998, 2000) for further details. These changes will be relevant, below, to our deciding which language some loans into Marrku are sourced from. However, in terms of structural comparisons relevant here, all differences between Ilgar/Garig on the one hand and Iwaidja and the other can be reduced to a contrast in their 3sg forms for subject and/or object, including the knock-on effects of the Iwaidja general 3sg morpheme sometimes represented as K-, which produces a hardening of the following consonantal segment. Looking now more specifically at the differences in morphological organisation, in Iwaidja and Ilgar/Garig there is no tense prefix before subject, no aspect prefix, and no separate plural prefix (instead, there are distinct pronominal prefix forms for singular and plural). Iwaidja and Ilgar/Garig have a future/ potential prefix -(w)ana- between subject+direction and verb, which lacks a counterpart in Marrku. Examples (11) and (12) compare the Marrku and Iwaidja/ Ilgar/Garig forms for ‘we used to stay/live’, and (13) and (14) the corresponding future forms. Marrku, past imperfective: (11)

Tense Subject manga past 1st person mangawulkayi ‘We

Aspect Plural Stem wulkayi imperfective plural be, sit used to stay, used to live’

Iwaidja, past imperfective: (12)

Subject + Direction Tense/Modality ngadø 1st person pl. neutral realis ngadbaningan ‘We used to stay, used to

Stem11 bani be, sit live’

Tense/Aspect/Mood -ngan12 past imperfective

Marrku, future: (13)

Tense + Subject Stem kirrildayi 1st:FUT be, stay kirrildayi ‘I will be/stay’

11 With a distinct pattern of partial right-reduplication for duals and iteratives, not illustrated here, and some suppletive dual roots. 12 Strangely, this suffix marks perfective in Mawng with this root. It must be emphasised that the aspectual semantics of Iwaidja remains poorly understood.

38

Nicholas Evans

Iwaidja/Ilgar/Garig, future: (14)

Subject + Direction nga1st person sg.neutral ‘I will be/stay’

Tense/Modality nafuture/irrealis

Stem wani be, sit

To sum up the differences: (a) Marrku has an initial tense slot for past marking which Iwaidja lacks (b) In its subject marking, Marrku separates person from number marking (plural generally –lk-) into distinct slots (5, 10), whereas Iwaidja and Ilgar/Garig have pronominal prefixes where the number marking directly follows the person marking (1sg nga-, 1nsg ngarr- ~ ngad-) or is not related to it at all: 3sg i-/iny- (Ilg/Ga), K- (Iw), 3pl a-; 2sg intrans subject a(n)ng-, 2pl intrans subject kurr-). (c) Marrku signals aspect by prefix, between the subject marker and the plural marker or stem, whereas Iwaidja and Ilgar/Garig signal it by suffix after the stem (d) Iwaidja and Ilgar/Garig have a future/irrealis prefix –(wa)na, directly before the verb stem, while Marrku has a fused subject + future prefix, typically portmanteau in form These do not exhaust the differences between Marrku and the other languages in how they organise their verbs, but will be enough to show how different the systems normally are.

5 Some shared paradigms We now turn to three verbs in which the normal differences in verb morphology outlined in Section 4 are not observed. Instead, the Marrku forms closely resemble those found in Ilgar/Garig. The relevant verbs are: yama ‘work’; miyardma ‘want, like’, and wurdan ‘be from, come from’. For expository purposes I draw all examples from a single text – told by Jumbo Jamburlurlu about working for the missionaries gathering trepang (sea-cucumbers), and transcribed by myself with the assistance of Joy Williams Malwagag.13 However, there are other exam13 In the examples that follow, JJ are the speakers’ initials, followed by a line number in the Elan transcript of the tier, followed by the time code at the beginning of the line. The full Elan transcript and original sound file can be found in the DoBeS archive. The original recording was made by Bernhard Schebeck in 1966 and is archived in AIATSIS as Archive Tape 644, Track B.

Paradigm borrowing in Marrku and its implications

39

ples of these verbs (especially ‘want’) in the recorded materials, as will be mentioned below. In presenting these excerpts, over every word I write an abbreviation for the original source language: Iw = Iwaidja, G = Garig/Ilgar, M = Marrku, Mkr = Makassarese, an Austronesian language of Sulawesi from which scores of words were borrowed into languages of the region (Evans 1992). The symbol ≈ indicates that the resemblance of an inflected verb to the indicated language is only approximate, while ≠ indicates that the form is different from the designated language, and < indicates a loan source. (15)

G (≠Iw) M < Mkr M M a-miyardma-ng tharriba many-ma-yi thak 1SG.A3SG.o-want-NPST trepang PST:1>3-get-PST all ‘Do you want trepang? I’ve got it all.’ (JJ 19; 00:01:04)

The line in (15) concludes with the two clearly Marrku words manymayi thak ‘I’ve got it all’. Tharriba ‘trepang’, originally from Makassarese taripaŋ (Cense 1979) can be regarded as an assimilated loan (sporadic loss of final -ŋ is common – see Walker and Zorc 1981) in a form distinctive to Marrku, since other languages of the region lack interdentals and borrow the form as darriba rather than tharriba (Evans 1992), or else assimilate the initial to a palatal, as in Mawng jarripang. The word amiyardmang ‘I want it’ is distinctive to Garig and Ilgar, but different from the Iwaidja form which would be abiyardmang.14,15 The next line of interest, eight seconds further into the text, is: (16)

=Iw/G =Iw/G a-, whereas the Iwaidja form is historically a + K- where K- is the (etymologically) miscellaneous gender, generalised as the regular 3SG absolutive marker in Iwaidja. 15 Capell’s materials (1963: 6) give the following forms for ‘want’: amiyarrma ‘I like (him), imiyarrma ‘he likes (it)’, kunmiyarrma ‘I like you’, yanmiyarrma ‘you like me’ and nganmiyarrma ‘she likes me’. These are very close to the attested Ilgar/Garig forms, except that inimiyarrma would be expected for ‘he likes it’ and nganngamiyarrma for ‘she likes me’; the final –ng appears to be variably present, and the phonetic variation between rd and rr can be hard to hear in Iwaidja and Ilgar/Garig and may not even be significant in Marrku.

40

Nicholas Evans

Joy Williams translated this line into Iwaidja as iyi, nganayamang nuyi: the interjection iyi is identical across the languages of the Cobourg region (regionally shared affirmative interjections are nothing unusual in Australia), and the two languages employ different second person pronouns as expected. What is of interest in this line is the inflected verb nganayamang ‘I will work’, identical in Iwaidja and Marrku. This is based on the root yama, from Makassarese jáma ‘do, work, handle, touch’ or possibly Malay jamah ‘handle’, with lenition of the initial consonant (Evans 1997) and addition of -ng, which is stereotypical of Macassan loans (Evans 1992). The 1st singular future prefix ngana- is completely regular for intransitive verbs in Iwaidja (and Ilgar/Garig), but is completely different from the normal prefixes in Marrku, so this is a clear case where an inflected verb has been borrowed lock stock and barrel. The next relevant line is reproduced in (17); Joy Williams translated it into Iwaidja as jamangu ba darriba. Here the word jamangu has a distinctively Iwaidja form, showing the initial hardening distinguishing Iwaidja from Garig or Ilgar (both of which would have i-yamangu); the -(ng)u past imperfective suffix is also distinctive to Iwaidja, Garig and Ilgar but not normally found in Marrku. Of the other words, marrkungurn is a demonstrative form found only in Marrku, and tharriba for ‘trepang’ has already been discussed. (17)

=Iw (≠G) mangany $ + suppletive plural stem ra deletion of relevant segments

–

Paradigm borrowing in Marrku and its implications

49

three-way tense system (as in Garig/Ilgar and Iwaidja), plus an imperative and an irrealis (also as in Garig/Ilgar). There is likely to have been some form of perfective vs imperfective contrast in the past, an inference based on the existence of a certain number of different past forms of the same verb (e.g. malawuthi and manathawuthi for ‘they spoke’, and manathun and manathuwa for ‘they died, were sick’). Again this parallels, in a prima facie way, the past contrasts found in Garig/Ilgar and Iwaidja, but we lack the translations, revealing discourse contexts, or full paradigms that would help us make a more definitive statement. Nonetheless, we can at least say that there are no clear differences between the TAM semantics of the various languages, within the limited evidence we have. (b) the difficulty of segmenting roots (in both languages) and the frequency of suppletive forms makes roots hard to isolate as a borrowed unit; both Weinreich (1968: 35) and Heath (1978: 105–7) mention these as factors which hinder morphological borrowability but would aid borrowing of inflected stems. And they are factors which apply especially to Marrku, but also to Garig/Ilgar and Iwaidja. Considering our five Marrku verbs again, we should not be misled by the confident and arbitrary stem forms cited at the head of each paradigm. For ‘say’, we have man, an and hanyman in the present, ngany, many and hanyman in the past, and man and hanyman in the future. For ‘be’ we have ldayi in the singular and lkuyi in the plural; for ‘return’ we have arrun, rarruny, hirarruny and rarrun in the singular and rangkuny or yarngkuny in the plural, and even for a relatively regular verb like ‘talk’ we have lawuthi, thawuthi and yawuthi as alternate stem forms. Taking the same verb in Iwaidja, stems vary from ldaharrama in ‘I talk’ (nga-ldaharrama), kaharrama in ‘you talk’ (an-kaharrama) and raharrama in ‘(s)he talks’ (no segmentation possible, since underlyingly r combines the hardening prefix K- with the now-invisible initial ld). It is often said in studies of language contact (e.g. Thomason and Kaufman 1988) that formal invariance favours borrowability. However, in situations where speakers already have expectations of high morphological variability, based on the structure of both languages they speak (or, put differently, their language does not dispose them to operate with “morphemes” in their analyses of their language’s morphology), families or paradigms of inflected words may be the units they operate with, so having morphemic, invariant stems may simply not be part of the game.

8.2 Sociolinguistic The Cobourg Peninsula region, like many parts of indigenous Australia and indeed many other parts of the world (Southern New Guinea, Vaupes, Mandara

50

Nicholas Evans

Mountains of Cameroon, etc.), was traditionally characterised by widespread and intense multilingualism (cf. Evans 2011). For people whose indexical clan language was Marrku, it is quite likely that the standard repertoire included Garig/Ilgar and Iwaidja in addition to other languages of the region. In such situations, it may have been quite normal not to expect everything to have its own distinct lexical items or language-specific formatives in the emblematic language: signalling of group membership could have taken place quite adequately by using Marrku-specific words and constructions with sufficient frequency (say, 50–70% of words per utterance, on average). It is worth drawing a parallel with the semiotics of special registers in Bininj Gun-wok here: the respect register Gun-gurrng (Evans 2003b, Garde 2013) does not have distinct lexical items for every concept, but it distributes them sufficiently across the frequency curve that almost every utterance will include each one, thereby being immediately identifiable as belonging to a respectful tenor. Recent work by Höder (2014) suggests an interesting way to view this issue, though his Diasystematic Construction Grammar approach. This allows for communicative systems in which multilingualism is the norm, and in which some constructions are language-specific while others are associated with more than one language. These constructions, could, in principle, include fully inflected verb paradigms, as in the case of the ‘work’ and ‘want’ verbs discussed above.

Figure 3: Monolingual and interlingual constructional elements in a multilingual setting (Höder 2014: 45)

In the Marrku context, the verb paradigms for ‘work’, ‘want’ and ‘be from’ would be shared with Garig/Ilgar through interlingual links, as would (parts of) the prefix paradigm for the reflexive pronoun, while the paradigms for other inflected verbs would be confined to Marrku.

Paradigm borrowing in Marrku and its implications

51

9 Paradigm borrowing, subsystem integrity, and the genesis of mixed languages In the preceding section I made some suggestions regarding how both structural and sociolinguistic features of Marrku and its multilingual setting may have favoured the borrowing of morphological paradigms, both hosted and recombinant.25 We now view the issue from another angle, namely of how the data we have considered can be seen as a first step in the development of a mixed language. Seifart (2012: 475) proposes the “Principle of Morphosyntactic Subsystem Integrity” in language contact: in situations where various grammatical morphemes are borrowed, these tend to be morphosyntactically interrelated, rather than being random collections of forms or sets of forms that are best described by well-known borrowability hierarchies

He gives, as an example, the case of Resígaro (Arawakan, Colombia), which has borrowed entire paradigms of inflectional and derivational morphemes from Bora (Bora-Muinane, Northwest Amazon), including: – about 20 noun class and gender markers – six number markers – eight bound grammatical roots that are used to form e.g. numerals and demonstratives – other pro-forms Most of the borrowed morphemes he considers belong to subsystems for expressing countable units (through nominal classification), number marking, and quantity. Seifart (2012: 498) goes on to argue that his Principle of Morphosyntactic Subsystem Integrity, and the Resígaro data: may also open a new perspective on the treatment of mixed languages, which . . . defy a common characterization in terms of current models of contact-induced language change. Mixed languages may differ from cases like Resígaro in the extent to which material from different languages is mixed, but at least for some cases of mixed languages, a primary determinant seems to be the maintenance of the integrity of morphosyntactic subsystems.

25 It might be objected that on the Höder analysis just outlined, it may be more accurate to refer to some of them as ‘shared’ rather than ‘borrowed’, thereby avoiding claims of directionality. But against this the clear grounding of the morphological paradigms in principles which hold generally in Garig/Ilgar but only with the relevant verbs in Marrku makes it clear that there is, in fact, a directionality involved.

52

Nicholas Evans

In a number of mixed languages, including some that are considered to be close to the ‘prototype’ of mixed languages (Bakker 2003: 124), the sets of etymologically distinct morphological material seem to be divided precisely along the lines of tightly integrated morphosyntactic subsystems, e.g. verbal inflectional subsystems in Copper Island Aleut . . . Resígaro, under the analysis proposed here, constitutes an intermediate case, a ‘missing link’ between borrowing and language mixing that may eventually help to bridge the perceived gap between ‘normal’ contact-induced changes and mixed languages.

The Marrku data discussed above presents some further cases of wholesale borrowing of morphological subsystems – whether the ‘sub-’ in the subsystems is characterised by being just a subpart of a word class (i.e. a subset of the class of all inflected verbs), or, in the case of reflexives, by a prefixal system confined to combination with a single element. Like the Resigaro data discussed by Seifart, as well as the Turkish-influenced varieties of Romani discussed by Igla and a number of other cases, it illustrates the first step in a process by which organised inflectional series can be transferred from one language to another, preserving their forms, their meaning, and the paradigmatic relations between the forms. To get from the Resigaro, Marrku or Turkish-influenced Romani cases to a full-fledged mixed language, two further steps are needed; it is an unanswered question for further research what the relationship is between them. (a) within the linguistic system itself, the relevant morphological subsystems need to be extended more widely (e.g. generalising the prefixes in Marrku from ‘work’ and ‘want’ to other verbs) so that they become a more productive part of the system. (b) within the sociolinguistic matrix, the repertoire of languages needs to change so that it is no longer a matter of dipping into an integrated Höderian ‘diasystematic constructional pool’, but instead the words or constructions are regarded as being monolingual elements. We know that this happened with Michif, for example, as changes in racial and legislative identity of its speakers isolated them from the Canadian French and Cree speech communities which had earlier been part of their bilingual repertoire, leaving Michif speakers lacking a knowledge of French, Cree or indeed both. Of course, in another situation, such as abiding and stable multilingualism of the sort that appeared to be the traditional norm on the Cobourg Peninsula, this sociolinguistic tradition may never have occurred, leaving the relevant subsystem elements as permanently shared elements in the diasystematic constructional pool.26 26 And it is likely, in fact, that the Michif case is atypical. Apart from Michif, all other mixed languages are symbiotic mixed languages i.e. speakers are bilingual in one of more of the source languages.

Paradigm borrowing in Marrku and its implications

53

In closing, one final remark needs to be made about the relationship between paradigm borrowing and the postulation of genealogical links. In Evans (2000), the shared morphology in the person/number prefixes on reflexive pronouns in Marrku was taken as crucial evidence for the inclusion of Marrku within a postulated Iwaidjan family. If we reconsider this, now seeing it as a case of paradigm borrowing, the phylogenetic picture changes: the evidence for including Marrku in the same family as Garig, Ilgar, Iwaidja, Mawng and Amurdak dwindles to the point where, apart from some borrowed lexicon, nothing particular is shared beyond what is also shared with a raft of other non-Pama-Nyungan languages such as Gunwinyguan (e.g. 1sg prefixes in nga-, verb stems like *thuwa ‘sicken, die’). Adjudicating on when we are dealing with shared inheritance and when we are confronted with paradigm borrowing is not straightforward, but the accumulating evidence from languages like Resígaro and some Romani dialects – as well as from other languages discussed in the present volume – suggests that paradigm borrowing is more common than once was believed.

Acknowledgements My work on Marrku and the broader linguistic context of Croker Island has been supported by a number of grants, most importantly those from the Australian Research Council (‘Non-Pama-Nyungan Languages of Northern Australia’ and ARC Laureate Project ‘Wellsprings of Linguistic Diversity’) and the Volkswagen Stiftung, DoBeS Program (Yiwarruj, yinyman, radbiyi lda mali: Iwaidja and other endangered languages of the Cobourg Peninsula (Australia) in their Cultural Context.). I thank these institutions, as well as my employers (University of Melbourne and Australian National University) for their support, and volume editors Felicity Meakins and Carmel O’Shannessy and one anonymous referee for their trenchant commentaries on an earlier version, and Susan Ford for assistance with the manuscript. Without the timely efforts to record Marrku in the 1960s, by Bernhard Schebeck and Heather Hinch, none of the work reported on here would be possible. My deepest thanks go to †Joy Williams, †Charlie Wardaga, †Mick Yarmirr and Khaki Marrala, who helped me learn something of Marrku while it was still possible – and especially Joy Williams who worked long and hard with me on the transcriptions of the heritage Marrku recordings.

54

Nicholas Evans

References Bakker, Peter. 1997. A language of our own: The genesis of Michif, the mixed Cree-French language of the Canadian Métis. New York: Oxford University Press. Capell, Arthur. 1963. Marrgu field notes and word list. Unpublished. Capell, Arthur & Heather E. Hinch. 1970. Maung Grammar, Texts and Vocabulary. The Hague: Mouton. Cense, A.A. 1979. Makassaars-Nederlands Woordenboek. ‘s-Gravenhage: Martinus Nijhoff. Charola, Erika. 2002. The verb phrase structure of Gurindji Kriol. Melbourne: University of Melbourne Honours Thesis. Davies, Jennifer. 1989. Umbugarla: A sketch grammar. Melbourne: University of Melbourne Honours Thesis. Dixon, Robert M. W. 2002. Australian languages: Their origins and development. Cambridge: Cambridge University Press. Elšík, Viktor & Yaron Matras. 2006. Markedness and language change: The Romani sample (Empirical Approaches to Language Typology 32). Berlin: Mouton de Gruyter. Evans, Nicholas. 1992. Macassan loans in Top End languages. Australian Journal of Linguistics 12(1). 45–91. Evans, Nicholas. 1997. Macassan loans and linguistic stratigraphy among the Iwaidjan languages. In Patrick McConvell & Nicholas Evans (eds.), Archaeology and linguistics: Global perspectives on ancient Australia, 237–259. Melbourne: Oxford University Press. Evans, Nicholas. 1998. Iwaidja mutation and its origins. In Anna Siewierska & Jae Jung Song (eds.), Case, typology and grammar: In honour of Barry J. Blake, 115–149. Amsterdam: John Benjamins. Evans, Nicholas. 2000. Iwaidjan, a very un-Australian language family. Linguistic Typology 4(1). 91–142. Evans, Nicholas. 2001. The last speaker is dead – long live the last speaker! In Paul Newman & Martha Ratliff (eds.), Linguistic field work, 250–281. Cambridge: Cambridge University Press. Evans, Nicholas. 2003a. Introduction: Comparative Non-Pama-Nyungan and Australian historical linguistics. In Nicholas Evans (ed.), The non-Pama-Nyungan languages of northern Australia: Comparative studies of the continent’s most linguistically complex region, 3– 25. Canberra: Pacific Linguistics. Evans, Nicholas. 2003b. Bininj Gun-wok: A pan-dialectal grammar of Mayali, Kunwinjku and Kune. 2 vols. Canberra: Pacific Linguistics. Evans, Nicholas. 2009. Dying Words: Endangered languages and what they have to tell us (The Language Library). Maldon & Oxford: Wiley-Blackwell. Evans, Nicholas. 2011. A tale of many tongues: Documenting polyglot narrative in North Australian oral traditions. In Brett Baker, Ilana Mushin, Mark Harvey & Rod Gardner (eds.), Indigenous language and social identity. Papers in honour of Michael Walsh, 291–314. Canberra: Pacific Linguistics. Evans, Nicholas, Joy Williams Malwagag and Khaki Marrala. 2006. Majila Inkawart. Jabiru: Iwaidja Inyman. Gardani, Francesco. 2008. Borrowing of inflectional morphemes in language contact. Frankfurt: Peter Lang.

Paradigm borrowing in Marrku and its implications

55

Gardani, Francesco, Peter Arkadiev & Nino Amiridze (eds.) 2014. Borrowed morphology: an overview. In Gardani, Arkadiev & Amiridze (eds.), Borrowed Morphology, 1–23. Berlin: Mouton de Gruyter. Garde, Murray. 2013. Culture, interaction and person reference in an Australian language. Amsterdam: John Benjamins. Harvey, Mark. 2002. A grammar of Gaagudju. Berlin: Mouton de Gruyter. Heath, Jeffrey. 1978. Linguistic diffusion in Arnhem Land (Vol. 13). Canberra: Australian Institute of Aboriginal Studies. Höder, Steffen. 2014. Constructing diasystems: grammatical organisation in bilingual groups. In Tor A. Åfarli & and Brit Mæhlum (eds.), The sociolinguistics of grammar, 137–152. Amsterdam: John Benjamins. Igla, Birgit. 1996. Das Romani von Ajia Varvara: Deskriptive und historisch-vergleichende Darstellung eines Zigeunerdialekts. Wiesbaden: Harassowitz. Kossmann, Maarten. 2008. On the Nature of Borrowing in Cypriot Arabic. Zeitschrift für Arabische Linguistik 49. 5–24. Kossmann, Maarten. 2010. Parallel System Borrowing: Parallel morphological systems due to the borrowing of paradigms. Diachronica 27(3). 459–487. Mailhammer, Robert. 2009. Towards an aspect-based account of the verb categories of Amurdak. Australian Journal of Linguistics 29(3). 349–391. Matras, Yaron. 2002. Romani: A linguistic introduction. Cambridge: Cambridge University Press. Matras, Yaron. 2007. The borrowability of structural categories. In Y. Matras & J. Sakel (Eds.), Grammatical borrowing in cross-linguistic perspective, 31–73. Berlin: Mouton de Gruyter. McConvell, Patrick. 2008. Mixed Languages as outcomes of code-switching: Recent examples from Australia and their implications. Journal of Language Contact 2. 187–212. Meakins, Felicity. 2010. The development of asymmetrical serial verb constructions in an Australian mixed language. Linguistic Typology 14(1). 1–38. Meakins, Felicity. 2011. Borrowing contextual inflection: Evidence from northern Australia. Morphology 21(1). 57–87. Meakins, Felicity. 2012. Which Mix? – Code-switching or a mixed language – Gurindji Kriol. Journal of Pidgin and Creole Languages 27(1). 105–140. Meakins, Felicity & Carmel O’Shannessy. 2010. Ordering arguments about: Word order and discourse motivations in the development and use of the ergative marker in two Australian mixed languages. Lingua 120 (7). 1693–1713. Muysken, Peter. 2000. Bilingual speech: A typology of code-mixing. Cambridge: Cambridge University Press. O’Grady, Geoffrey N., Carl F. Voegelin & Florence M. Voegelin. 1966. Languages of the world: Indo-Pacific fascicle six. Anthropological Linguistics 8(2). O’Grady, Geoffrey N., Stephen A. Wurm & Kenneth L. Hale. 1966. Aboriginal languages of Australia (a preliminary classification). Map, drawn by Robert M. Watter, Department of Linguistics, Victoria, B.C. O’Shannessy, Carmel. 2012. The role of code-switched input to children in the origin of a new mixed language. Linguistics 50 (2). 305–340. O’Shannessy, Carmel. 2013. The role of multiple sources in the formation of an innovative auxiliary category in Light Warlpiri, a new Australian mixed language. Language 89(2). 328– 353. Schmidt, Wilhelm. 1919. Die Gliederung der australischen Sprachen. Vienna: MechitharistenBuchdruckerei.

56

Nicholas Evans

Schuchardt, Hugo. 1928. Hugo Schuchardt-Brevier. Ein Vademecum der allgemeinen Sprachwissenschaft. Zusammengestellt und eingeleitet von Leo Spitzer. Halle (Saale): Max Niemeyer Verlag. Seifart, Frank. 2012. The Principle of Morphosyntactic Subsystem Integrity in language contact: Evidence from morphological borrowing in Resígaro (Arawakan). Diachronica 29(4). 471– 504. Seifart, Frank. 2013. AfBo: A world-wide survey of affix borrowing. Leipzig: Max Planck Institute for Evolutionary Anthropology. http://afbo.info (accessed 18 March 2015). Singer, Ruth. 2011. Typologising idiomaticity: Noun-verb idioms and their relations. Linguistic Typology 15: 625–659. Singer, Ruth. 2015. The dynamics of nominal classification: productive and lexicalised uses of gender in Mawng. Berlin & New York: de Gruyter Mouton. Thomason, Sarah G. 1997. Mednyj Aleut. In Sarah G. Thomason (ed.), Contact languages: A wider perspective, 449–468. Amsterdam & Philadelphia: John Benjamins. Thomason, Sarah G. & Terence Kaufman. 1988. Language contact, creolization and genetic linguistics. Berkeley: University of California Press. Walker, Alan & David Zorc. 1981. Austronesian loanwords in Yolngu-Matha of northeastern Arnhem Land. Aboriginal History 5(2). 109–134. Walsh, Michael. 1981. Maps of Australia and Tasmania. In Wurm, Stephen A. & Shirô Hattori (eds.), Language atlas of the Pacific area, part 1: New Guinea area, Oceania, Australia. Canberra: The Australian Academy of the Humanities, in conjunction with the Japan Academy. Weinreich, Uriel. 1968 [1953]. Languages in contact: Findings and problems. Berlin: Mouton de Gruyter.

Ilana Mushin and Janet Watts

2 Identifying the grammars of Queensland ex-government Reserves: The case of Woorie Talk Abstract: The 1897 Protection of Aborigines and the Restriction of the sale of Opium Act in Queensland had a profound effect on the nature of language contact in that State throughout much of the twentieth century. In particular the Act allowed for the controlled movement of Aboriginal people from anywhere in the State to Government-run Reserves – Cherbourg, Palm Island and Woorabinda. This resulted in people from more than 40 different language backgrounds coming together, often for the first time. By the twentieth century these languages in contact would have included existing contact languages deriving from the spread of NSW Pidgin as well as Traditional Australian Languages. The Reserves also served as depots for Aboriginal Labour in rural industries such as Pastoralism, Agriculture and Mining, providing a context for more sustained contact with the varieties of English spoken in rural Queensland. Today these ex-Reserves, built from a diversity of dispossessed Indigenous Queenslanders, are well-established communities. However we can see the footprint of their contact histories in the vernacular languages that are widely spoken and acquired by children in these places. In this paper we focus on the vernacular language of Woorabinda (‘Woorie Talk’) as a case study of the ways in which ex-Reserve varieties may be understood as contact languages. Allridge (1984) began this quest by identifying features of NSW Pidgin, Roper River Kriol and Torres Strait Islander Creole in ex-Reserve varieties, working with the important, but limited, Queensland Speech Survey materials collected in the 1960s. She concluded that ex-Reserve vernaculars developed from earlier Pidgin languages that depidginised as speakers shifted to English. More recently Munro & Mushin (2016), on the basis of recent field work, argued that the sociohistorical circumstances that gave rise to Woorie Talk were comparable with Winford’s (2000) analysis of the emergence of the Barbados ‘Intermediate Creole’, Bajan. In this paper we analyse the nature of the superstrate and substrate influences that form part of the development of Woorie Talk grammar. We focus in particular on features of the tense system that emerged from the regional varieties of English spoken by Queensland settlers with whom Woorabinda residents would have had regular contact. We also examine evidence for substrate

58

Ilana Mushin and Janet Watts

inputs that came directly from Queensland traditional languages that would have been spoken in Woorabinda, rather than indirectly via the expansion NSW Pidgin into Queensland. Our analysis thus provides new insights into the contact language status of varieties that are commonly labelled as ‘Aboriginal English’.

1 Introduction In this paper we show how contemporary vernacular languages spoken in Queensland communities that were once Government-run reserves can be understood as contact languages. Reserve varieties emerged in the early-to-mid twentieth century as a consequence of Queensland government policies that resulted in new patterns of contact between Aboriginal and Torres Strait Islander groups. By the twentieth century many of these groups had already experienced a significant decline in the use of traditional languages, following the displacement and decimation of the Aboriginal population in the nineteenth century. Most groups had shifted to speaking contact varieties that developed from the spread of NSW Pidgin into Queensland. Reserve varieties are therefore a twentieth century product of contact between traditional languages, English-lexifier contact languages and English. Historically they have been called “Aboriginal English” (e.g. Alexander 1965, 1968; Readdy 1961; Eades 1983, 2014), although as Allridge (1984) and Angelo (2013) show, they display features also found in Australian creoles. To this extent they appear similar to varieties of Aboriginal English spoken in other parts of Australia (e.g. Kaldor and Malcolm 1991, Malcolm 2000). In recent work, Munro and Mushin (2016) argued that the vernacular spoken in the Central Queensland ex-reserve community of Woorabinda – here called “Woorie Talk”1 – is an example of an “intermediate creole” (Winford 2000), similar in its story of formation to the Barbadian creole Bajan. Munro and Mushin (2016) focused on the social conditions and historical factors that gave rise to Woorie Talk, providing a foundation for understanding its status as a contact language through an analysis of the history of the reserve. Regardless of its classification as a contact variety, we know that Woorie Talk developed as a distinct variety during the twentieth century and shows input 1 The name “Woorie Talk” emerged from community consultation around a language awareness poster, developed in collaboration with the Language Perspectives group of the Indigenous Schooling Support Unit of the Queensland Department of Education, Training and Employment (DETE). It is not a widely used term within the community – the variety is typically negatively identified as ‘Rubbish English’ or ‘Slang’.

The case of Woorie Talk

59

from three types of sources: nineteenth century contact varieties already in use, colonial English varieties, and a plethora of traditional languages which came together on reserves. Understanding the grammar of Woorie Talk requires us to investigate the balance of these sources in its evolution and to take into account further changes from the mid-twentieth century on, as input from traditional languages continued to decline and access to Standard Australian English increased. Our focus here is on identifying more precisely how the three kinds of input – English, nineteenth century contact varieties and local traditional languages – have contributed to the structure of Woorie Talk as it is spoken today. In doing so we aim to progress the scholarship on mainland Queensland Indigenous vernaculars, which have been commonly termed Aboriginal English (e.g. Flint 1968; Kaldor and Malcolm 1991; Eades 2014).2 The term “Aboriginal English” includes varieties identified with Aboriginal people in metropolitan and regional urban settings, as well as those which developed in reserves and missions (Eades 2014). The umbrella label of Aboriginal English thus obscures the range of contexts in which varieties have emerged in different parts of Australia, and the different combinations of languages in contact that have impacted their grammatical and lexical development. The Queensland Speech Survey, conducted in the 1960s by Elwyn Flint and his students, used the term “Aboriginal English” for the contemporary varieties spoken in Queensland reserves and missions, including Woorabinda, Palm Island, Cherbourg, Doomadgee, Yarrabah, Bloomfield River, Aurukun and Weipa. Recordings made as part of this Queensland-wide survey mostly consisted of school students and adults being asked to talk among themselves in an effort to sample informal ways of speaking English in these communities.3 A number of theses produced during this period focused on varieties spoken in the reserve communities of Woorabinda (Alexander 1968), Yarrabah (Alexander 1965), Cherbourg (Readdy 1961) and Palm Island (Dutton 1965). The map below shows the location of these reserves, and the Hope Vale mission in Northern Queensland. The map also shows the traditional Indigenous languages of the regions surrounding the reserves along with some additional languages, such as Kuku Yalanji and Gunya, that we discuss by name in the sections below.

2 The only major exception is a vernacular spoken around the tip of Cape York, that Crowley and Rigsby (1979) named “Cape York Creole” and which is clearly an expansion of Torres Strait Creole (Shnukal 1988) to the mainland. 3 This replicated data collected in non-Aboriginal communities, which focused on high schools in rural and urban locations.

60

Ilana Mushin and Janet Watts

Figure 1: Locations of Australian languages and communities mentioned in this chapter4

4 This is only a subset of the language groups which were removed to Yarrabah, Woorabinda and Cherbourg during the twentieth century.

The case of Woorie Talk

61

These four studies identified a vernacular language variety spoken in each community that differed systematically from Standard Australian English (SAE) varieties and largely focused on how grammatical (mostly phonological) features differed from Standard Australian English.5 They did not compare these varieties with each other, nor consider their origins, nor examine how features of these varieties may have emerged from language contact. In more recent general overviews of Aboriginal English varieties there is an acknowledgement of their contact origins (e.g. Malcolm and Grote 2007). For example, it is has been established that Queensland vernaculars, especially those spoken south of Cape York Peninsula, had their origin in the northwards spread of NSW pidgin, following the pathways of colonial settlement, especially via the pastoral industry in the nineteenth century (e.g. Dutton 1983; Allridge 1984). Over time in Queensland, Aboriginal speakers of these pidgin varieties were more likely to come into contact with fellow speakers of such contact languages and/or speakers of English varieties and less likely to be in sustained contact with speakers of traditional languages. By the turn of the twentieth century, very many Queensland languages south of Cape York had ceased to be fully transmitted to the next generation. According to Mühlhäusler (1979), these contact varieties that were expansions of earlier pidgin varieties (e.g. NSW/QLD Pidgin – Dutton 1983) became further influenced by English, the language of the dominant culture. Under this analysis we expect to find varieties in most Queensland communities with a relatively large footprint from English. However two questions remain. The first is which features of earlier contact varieties, such as NSW Pidgin, remain in contemporary use. The second is whether, and how, contemporary Queensland reserve varieties have diversified, through both internal and contact-driven means, to be distinguishable from other Aboriginal English varieties and indeed other contact varieties more generally. In an unpublished honours thesis, Allridge (1984) examined the transcribed data from the four aforementioned theses on reserve varieties produced in the 1960s. Her focus was on identifying features of the varieties spoken in Yarrabah, Palm Island, Woorabinda and Cherbourg that were also found in other Australian contact languages: Kriol (Sandefur 1979), Torres Strait Creole (Shnukal 1988), Cape 5 We use the Australian Curriculum definition of Standard Australian English (SAE) as “The variety of spoken and written English language in Australia used in more formal settings such as for official or public purposes, and recorded in dictionaries, style guides and grammars. While it is always dynamic and evolving, it is recognised as the ‘common language’ of Australians.” (Australian Curriculum and Reporting Authority Website, http://www.australiancurriculum.edu.au/ Glossary?a=E&t=Standard%20Australian%20English#, accessed March 11th 2015). This definition appears consistent with the use of the term in the Queensland Speech Survey materials.

62

Ilana Mushin and Janet Watts

York Creole (Crowley and Rigsby 1979), NSW Pidgin (Troy 1990) and QLD Pidgin (Dutton 1983). Her findings showed that taken collectively, these varieties all retained features that were associated with earlier contact languages such as NSW Pidgin, and contemporary creoles spoken elsewhere (e.g. the use of bin past tense found in NT Kriol but also attested in the Palm Island and Yarrabah data; retention from NSW/QLD Pidgin of a -pela/-fella suffix on some pronouns). Allridge’s work was an important first step in establishing the contact language origins of these Queensland reserve varieties. However her results did not attempt to explore which features occurred in which of the four varieties. To what extent do they share grammatical and lexical features? How can we account for similarities and differences between them? How are they similar or different to other Australian contact varieties? Are the varieties recorded in the 1960s still used in these communities? How have they changed? Addressing these questions requires us to look more closely at the vernaculars spoken in each community, updating the data collected in the 1960s with more recent recordings. We focus here exclusively on the Woorabinda variety – Woorie Talk – to explore how intersections of multiple languages and varieties have contributed to its development. Our data come from two time periods, 1960s and 2011–13. The 1960s data are from Alexander (1968). The 2011–13 data are approximately 50 hours of audiovisual recordings of unelicited interactions between adults, adults and children, and children and children, in informal and institutional settings.6 We have supplemented this corpus with samples of language that were collected for a language awareness poster produced in collaboration with the Queensland Department of Education and Training Indigenous Schooling Support Unit (Language Perspectives, 2015). In the next section we outline the history of Woorie Talk and present examples from contemporary language use. In section 3 we examine superstrate influences, illustrated by a detailed analysis of future marking. In section 4 we consider substrate inputs and investigate which features may have come directly from traditional languages spoken on the reserve, and which may have already been present in nineteenth century pidgin varieties.

2 Woorie Talk We consider Woorie Talk to be an example of a larger set of reserve varieties on the basis of common experiences of language contact across all reserves in 6 The data was collected as part of the ARC Linkage Project LP100200406. The majority of the recordings in the overall corpus are child-child interactions but we have drawn from both adult and child language for the purposes of this paper.

The case of Woorie Talk

63

Queensland. The establishment of reserves was part of Queensland Government policy under the 1897 Aborigines Protection and Restriction of the Sale of Opium Act (henceforth “The Act”) which lasted until 1971. New reserves were set up in the early twentieth century by the Queensland State Government, adding to pre-existing missions that also came under The Act. Reserves were ostensibly sites of “protection” for Aboriginal people from the degenerating influence of the wider mainstream. However in practice they served as places of confinement and punishment and as labour depots until the 1960s (e.g. Forde 1990; Blake 2001; Watson 2010). Importantly, The Act officially sanctioned the removal of Aboriginal and Torres Strait Islander people from all parts of the State to these reserves where they had to remain unless assigned to work contracts. The varieties of language that emerged in all of these reserves are therefore a consequence of not only the traumas of colonisation, and its devastating impact on Queensland’s Indigenous linguistic diversity, but also of a plethora of language groups living together in close proximity, some of whom had never been in contact before. There is evidence that by the turn of the twentieth century, groups from most of inland Queensland, from the far southwest to the Gulf of Carpentaria, were already using contact varieties that developed with the spread of the Pastoral industry into these regions from the 1840s (Dutton 1983). However there is also evidence that while traditional language transmission had been radically disrupted in the nineteenth century, many of the groups who were moved into reserves still used their traditional languages as their primary languages (Mushin, Angelo and Munro, 2016). Reserves thus became homes to multiple traditional languages, with many of their speakers removed far from their homelands, as well as varieties of already established contact languages resulting from the expansion of NSW Pidgin throughout Queensland. The strict controls over the movement of people in and out of reserves under Queensland law reinforced an isolation from larger mainstream centres of communication, which prevented sustained contact with speakers of Standard Australian English (Munro and Mushin, 2016; Mushin, Angelo and Munro, 2016). Contact with non-Aboriginal people occurred mostly during work assignments where reserve residents worked alongside not only people of BritishAustralian origin, but also people of Chinese and Melanesian origin. Aboriginal people returned to the reserves between work contracts, bringing these additional language influences with them. As Mushin, Angelo and Munro (2016) show, until the 1960s there were few opportunities for residents to learn Standard Australian English.7 7 During the 1960s changes to The Act allowed for greater mobility and also schooling outside of reserves. Furthermore reserve schools became administered by the Queensland Education Department as part of the State School system.

64

Ilana Mushin and Janet Watts

There is evidence that a range of English-based varieties was spoken even from the earliest days of the reserve. This is summarised in the following quote from a local newspaper article that was published the year the entire population of the Taroom Reserve (1908–1926) was moved 250 km north to Woorabinda to make way for an irrigation scheme (that was never built):8 They are a most heterogeneous lot. Some are well educated and speak first-class English; some still speak pidgin English. They have come from all parts of the State, some even from Mornington Island. There is one blackfellow there who, when he first came, was unable to hold converse with any of his countrymen; his language was entirely foreign to that of any other abo. in the camp. (Morning Bulletin, August 8, 1927)

Alexander (1968), in her thesis on “Woorabinda Aboriginal English,” described the language situation as diglossic with what she called “Australian English” as a High (H) variety, used in interactions with non-Aboriginal people and in nonAboriginal institutional settings, and what she called “Aboriginal English” as a Low (L) variety, used by Woorabinda residents among themselves: (. . .) the familiar L form of Aboriginal English at Woorabinda has characteristics which would enable it to be termed a pidgin in this sense. It is certainly not a full language, inasmuch as it is a familiar variant of Australian English. It is also not a full language in that it does not exploit the full range of grammatical forms of English as a whole.

We cannot tell whether reserve residents in 1927 or the 1960s considered their home variety, characterised as a pidgin in the above quotes, as a distinctive Woorabinda way of speaking. There is certainly evidence that they did in the twenty-first century. Woorabinda residents today regularly recognise that their home variety is different from Standard Australian English (as used for example as school language, and language of ‘mainstream’ Australia), but also as different from Aboriginal ways of speaking in other communities. Both adults and children are able to identify lexicon and grammar that form a part of this code. Indigenous people we interviewed in Woorabinda in 2012 who were recent arrivals to the community (e.g. because they had married in, or came for employment) also recognise Woorie Talk as different from vernaculars spoken elsewhere. There is however a long-standing stigma attached to Woorie Talk, as illustrated in Alexander’s (1968) above characterisation of the Aboriginal English vernacular as “not a full language”. Indeed the association of Woorie Talk by its own speakers, as well as by others, with deficient English continues to the 8 For the purposes of contact history, we consider Taroom and Woorabinda to be the same reserve.

The case of Woorie Talk

65

present day and has constituted a significant challenge in our own attempts to develop a grammatical description of this variety. Its low status both within Woorabinda and in the mainstream Australian community makes people less willing to be recorded when speaking, and more prone to targeting English when being recorded. Nonetheless we have been able to record samples of Woorie Talk where participants were engaged in everyday activities and interactions that would have taken place even if we had not been recording (e.g. adults supervising children at a playgroup, children interacting in a playground), which is in contrast with the conversations recorded by Alexander and Flint that were set up solely for the purposes of recording. The following four examples illustrate Woorie Talk between residents in Woorabinda, both adults and children. We have numbered and boldfaced utterances within these extracts that illustrate some of the features of Woorie Talk we take up in the following sections:9 Extract 1: Adult/Adult Edw: Ah put dat fruit (dat) out too early; (dat way/today)? (0.9) Edw: (1) I’s all dere da table. {It’s on the table} (0.7) Edw: jes do da ting. ((cutting action with hands)) Sar: ( ) c’n do all dat.= Edw: Doreen do da cheese, an dat ting dere; (2) I’s all dere la {It’s all over there} (uh fruit),

Extract 2: Mother/Teen Daughter Dau: Mum, (0.8) um Sy said (3) she goin today. {she was going today} (0.9)

9 The orthography used in these examples is based around community discussions held in 2012 and 2013 in designing a language awareness poster (Language Perspectives, 2015). It currently mixes Standard English spellings with new conventions for writing Woorie Talk words, especially function words, including pronouns. We recognise that use of English spelling conventions for Woorie Talk does obscure its differences from more standard English varieties, however we are not yet in a position to fully systematise the way Woorie Talk is written. The transcriptions of conversation follow Conversation Analytic conventions for the transcription of timing and prosody as follows: [ ] = overlapping talk; (x.x) = timed silence; ? = final rising intonation; , = level (continuing) intonation, ; = partial falling intonation; . = complete falling intonation; : = lengthened segment; (xxx) = uncertain hearing.

66

Ilana Mushin and Janet Watts

Edw: no; (4) Tasha go w’d dem. {Tasha went with them} Oh jes gid awa:y; Do:n; Ah don wanna go Rocky, (5) youfla always wanna go Rocky Show. {you(pl) always want to go to Rocky Show} Ah ca:n. (0.5) Edw: (6) Ah gat no money¿ {I don’t have any money} (0.7)

Extract 3: Primary School-aged Children/Adult Mar: (7) ah sta:b you in da eyeball?, {I’ll stab you in the eyeball} (0.4) Loc: (8) ahll stab you in da e:[ye? ] {I’ll stab you in the eyes} Unc: [if y]ou go stab someone in the eyeball [an they lose their e:yeball,] Loc: [ ah chuck you in da neck ] before you [do dat; da:; ] Unc: [guess where] you’d be goin; (0.3) Mar: where; (0.3) Unc: you won’t be stayin wid mum an dem, (.) you be goin somewhere e:lse; (0.6) Mar: (sdaa:::;) (0.6) Unc: yea:h;

Extract 4: Primary school aged children. Geo: What ah need no:w? {What do I need now?} (0.7) Geo: (9) ahmma copy[:: e:::h; ] {I’m going to copy} Nay: [ahll make you one ay?] (0.5) Mir: bro make me one; (0.2) Nay: yeah;

These extracts were selected to give a flavour of Woorie Talk. Extract 1 comes from a conversation between two middle-aged women as they supervised a playgroup. Edwina is organising with Sarah-Lee to put out the morning tea food and

The case of Woorie Talk

67

drink for the children.10 Extract 2, from the same recording as Extract 1, features one of the middle-aged women and her teenage daughter discussing whether the daughter can go to the Rockhampton Agricultural Show, a major annual event about two hours drive from Woorabinda. Extract 3 was recorded as an Elder member of the community (named ‘Uncle’ here) attempted to break up a fight between two boys. Extract 4 comes from an interaction between three primary school-aged children, boys and girls. Some of the grammatical features in these examples are associated exclusively with Woorie Talk, some of these features are found in other Indigenous vernaculars, some in non-Indigenous vernaculars, and some shared by all three types: – Location: Stative locations indicated by a deictic demonstrative (dere), juxtaposing a NP without a preposition or case form to explicitly indicate the locative relationship, as in (1) all dere da table. Destination place names are also bare NPs, as in (5) wanna go Rocky Show. Prepositions based on English forms also occur in locative adjuncts, as in (7) ah stab ya in da eyeball, as they do in Standard Australian English. – Future marking: In these examples we find future expressed with bare VPs, as in (7) ah stab you, and with variant first person pronouns which appear to have fused English auxiliaries, as in (8) Ahll stab you and (10) Ahmma copy. We discuss the full range of future marking in our corpus, and its relationship with English and nineteenth century pidgin forms, in section 3. – Personal Pronouns: Plural forms are 1st person ufla (not in these examples), 2nd person youfla, as in (5), and 3rd person dem, as in (4). The -fla suffix derives from NSW Pidgin (Baker 1993) and variants of this are found in a range of Australian contact varieties, including varieties of Aboriginal English (Koch 2000). Plural forms are invariant for all grammatical relations, as they are in Kriol (Angelo and Schulze-Berndt 2013). – Negation: Negative possession is formed using the verb “get” plus negative “no”, as in (6) Ah gat no money. Negative particles do:n and ca:n appear to derive from English negative auxiliaries (kan is also attested in Kriol – Angelo and Schultze-Berndt 2013). – Discourse marker: The final particle la is used in Woorie Talk and attested in other reserve varieties in contexts where the speaker is trying to achieve joint attention with their interlocutor on a specific object. It frequently collocates with deictic demonstratives ere and dere, and with pointing gestures, as in (2) I’s all dere la (pointing towards the table where the fruit is located) (Gourlay and Mushin 2015). – Phonology: lack of interdental fricatives, no syllable-final obstruent clusters (see Baker and Bundgaard-Nielson (this volume)). 10 The names are pseudonyms

68

Ilana Mushin and Janet Watts

These extracts do not illustrate all of the grammatical features of Woorie Talk. There are also many lexical features which are common to reserve varieties and which are often identified by community members as distinctive of Woorie Talk. For example, when Woorabinda adults and high school children were being interviewed about language variation for the Language Awareness poster (Language Perspectives, 2015), they typically responded with words and formulaic expressions, including words with traditional language origins (e.g. gundoonoos ‘children’, bulloo ‘grandfather’, womba ‘crazy’, budgeroo ‘money’), as well as words with English origins, sometimes archaic (e.g. pinnin (< to pink (hit with a bullet)) – ‘throw stones at’, porki untin (< porcupine hunting) – ‘echidna hunting’; nebid (< never) – negative exclamation). Some of these are widespread in Indigenous vernaculars across Australia (e.g. bogi ‘swim, wash’, yaduman ‘horse’, gammin ‘false, rubbish’), a feature of the spread of Pidgin from NSW, as the north and west of Australia were colonised during the nineteenth century (Baker 1993; Walsh 1991). Some of these words are also attested in our conversation corpus, showing that they remain in active use. Grammatical features such as the form and distribution of prepositions and determiners, negation and question formation, and the full system of tense/ aspect/mood marking are subject to ongoing analysis. Even in these short conversational extracts, however, we can already recognise that some features are also attested in nineteenth century pidgins (e.g. the -fla pronoun suffix); some features appear to be derived directly from English grammar (e.g. negative particle do:n and the use of gonna future); and some features are attested neither in nineteenth century Pidgin nor in English (e.g. the discourse marker la), leaving open the possibility of direct substrate input or more recent grammatical innovation. Extract 3 above also supports the observation made in Alexander (1968) that there is variation in the community between what we are calling Woorie Talk and talk that appears to target Australian English (Standard or non-standard). In this extract the Uncle appears to target a more English-like grammar than the other speakers represented in this selection of extracts. For example he uses English auxiliaries (you won’t be stayin), although the lack of inflection of the be-auxiliary (you be goin), and the use of the serial verb (go stab) shows this is not Standard Australian English. More research is required before it is possible to determine whether the use of English here is typical for this Elder (as it is with some Elders), or whether he is targeting Standard Australian English, because he is admonishing the children on school grounds, an institutional setting associated with the use of Standard Australian English. Our observations of the conversation corpus show that a significant number of older adult Woorabinda residents, like the Elder in Extract 3, use more features associated

The case of Woorie Talk

69

with non-standard or Standard Australian English than speakers in other age groups. In order to characterise the grammar of Woorie Talk we need to develop a better understanding of the sources for its grammatical features: as developments from superstrate and substrate inputs found in pre-existing contact varieties; as features brought in from twentieth century contact with colonial English; or as features adopted from traditional languages spoken in Taroom and Woorabinda. In the next section we show how such an analysis can be developed for superstrate inputs through an analysis of future marking in our corpus. In section 4 we focus on features of Woorie Talk that may reflect direct substrate inputs. Together these features help to distinguish Woorie Talk as a product of its own unique social and linguistic history.

3 Future time reference and superstrate input to Woorie Talk This section explores how we can identify superstrate influences in the development of Woorie Talk, by exploring one part of its Tense/Aspect/Mood system. Although analysis of the whole system would be of interest in determining the relative inputs into this key component of grammar, we have limited our discussion here to the expression of future time. Future marking has already received some attention in studies of Aboriginal English varieties, especially in terms of its relationship with modal meanings (Eades 1983; Harkins 1994; Malcolm 1996). Kriol future marking has been explored in Hudson (1983), Munro (2011), Sandefur (1979) and Angelo and Schultze-Berndt (this volume). We are therefore in a good position to assess how marking of future time in Woorie Talk compares with what we find in other Australian contact languages. In this section we also explore the complexity of English inputs to Woorie Talk, discussing the evidence for the influence of early contact varieties and colonial English dialects spoken in the first half of the twentieth century on the development of forms used in future time contexts. Examples of future time reference in Woorie Talk can be seen in the extracts above. Examples (8) ahll stab you in da eye and (9) ahmma copy include forms derived from English will, and gonna. Example (7) ah stab you in da eyeball?, shows future time referred to with a simple verb form, in a way that is not possible in standard English varieties, but has been attested in QLD Pidgin (Dutton 1983).

70

Ilana Mushin and Janet Watts

Constructions in standard varieties of English associated with the expression of future time include modal auxiliaries, such as can, may, must, will, shall, constructions with be going + to-infinitive, with the simple present, and with the present progressive. Other constructions with infinitival or gerund-participial complements of certain verbs, such as I intend to see her tomorrow (Huddleston and Pullum 2002: 210), or with semi-auxiliaries, e.g. be about to, be due to, be destined to (Quirk et al. 1985: 236) may be used. Previous studies of future marking in Aboriginal English in south-east Queensland (Eades 1983) and Alice Springs (Harkins 1994) describe the use of the following forms to express future time reference: might, gonna, will (and ll), wanna and present tense (Eades 1983: 299), and can, gonta, will, wanta, like to, might and should (Harkins 1994: 101). To the extent that modal auxiliaries and present tense forms are used, the system is clearly structurally like Standard English. The focus of these studies was to consider semantic differences between different forms, and show how forms which on the surface appear very similar to Standard Australian English are used to express different meanings, in line with differing cultural conceptualisations, particularly around the certainty of future events. For example, where I’ll go and see him expresses a commitment to future action in Standard Australian English, in South-East Queensland Aboriginal English the same words express a conditional prediction about the future, which Eades (1983: 298) translates as “If nothing else happens I’ll go and see him”. Other Aboriginal English varieties appear to have a smaller set of “future auxiliaries” (Butcher 2008: 633) or ways of expressing a specifically future meaning (Malcolm 1996: 157). The forms will, gonna, going to (variably preceded by a form of be), and “unmarked” or present tense future forms are attested, along with gotta or go in some areas. In QLD Pidgin (as in NSW Pidgin, Troy 1990: 94–95; Malcolm and Koscielecki 1997: 26), future time reference was expressed with a simple verb form (or less commonly a verb with an -in suffix), sometimes combined with the adverbs of time directly and by and by (Dutton 1983: 101). By and by with a sense of ‘presently/soon/shortly’ has been attested in English since the sixteenth century (Romaine 2004: 491), and was used in nineteenth-century British English in clause initial and final positions (Baker 1993: 34 in Romaine 2004: 491), as it was in NSW and QLD Pidgin, and developed into a future tense marker in other pidgins and creoles of the Pacific (See Schultze-Berndt and Angelo, this volume). In Cape York Creole (Crowley and Rigsby 1979: 191) and Torres Strait Creole (Shnukal 1988: 42), go is used as a future marker preceding the verb, for example Im go kambek “He will return” (Crowley and Rigsby 1979: 191), and pre-sentence modifiers like klosap and baimbai indicate the proximity of the future event.

The case of Woorie Talk

71

Varieties of Kriol in northern Australia express future time with the markers garra and -l, (Hudson 1983: 29; Munro 2011: 472; Sandefur 1979: 128). We therefore see systematic differences between the marking of future time in Aboriginal Englishes and in Creole varieties. How is future time expressed in Woorie Talk? Alexander (1968) identified four ways of referring to future time for Woorie Talk: will, gonna, simple forms and verb +-in/ing, as in the following examples: (10)

I’m gonna see if that’s still goin ere.

(gonna)

(11)

But she pass probly.

(simple verb form)

(12)

Where you spendin yeh Easter?

(verb + -in/-ing)

The examples of the will forms occur with be, as what Alexander (1968: 90) calls the “future form of the copula”. (13)

I’ll be goin to Mount Morgan.

(14)

Doctor West’ll be here on Thursday.

All of these constructions occur in our data, as follows.

3.1 Will, ‘ll and won’t These forms appear to be derived from English will, and its contracted and negative forms ‘ll and won’t. These forms are also found in varieties of Aboriginal English (Eades 1983: 298; Harkins 1994; Malcolm 1996: 157). The number of tokens as a proportion of all future forms in the Woorabinda corpus (24%) is much smaller than in Malcolm’s (1996) Western Australian data (39%). In our data ll is the most common of the three will forms, and, similarly to the Australian mixed languages Gurindji Kriol and Light Warlpiri (Meakins 2007: 412; O’Shannessy 2013: 343), it is not restricted to first person pronouns as it is in varieties of Kriol (Hudson 1983: 30; Munro 2011: 473). (15)

Ooh ahll take dat pen off you an make it scribble on ya face (03:45.0-120904-2-4B)

(16)

Mia; tomorrow we’ll make cake for im;

(130605_adults2)

(17)

She’ll beat you ay, she smarter dan you ay;

(02:19.0-130509-2-2)

(18)

Dey’ll eat fish

(05:53.1-121112-2-7)

72

Ilana Mushin and Janet Watts

3.2 Gonna Constructions using gonna are common in Woorie Talk, as in varieties of Aboriginal English (Malcolm 1996: 157). There is variation in whether forms based on the English be auxiliary are used. When the subject is the first person singular pronoun, the form most frequently used is ahm, as in (19) and (20). This appears to be derived from the English first person singular pronoun + auxiliary am.11 (19)

Ahm gonna draw the stadium;

(04:38.8-120904-2-8A)

(20)

Ahmonna copy dat dere;

(13:14.8-120605-1-4A1)

With second and third person subjects, gonna directly follows the subject most often, as in (21) and (22), though there are some more English-like examples where a form of copula be precedes gonna, as in (23) and (24). (21)

You gonna get growled?,

(03:51.4-120903-2-3A1)

(22)

Ma mum gonna take me to hospital?,

(06:36.5-110908-P-1)

(23)

It’s gonna blow up Maya

(01:58.3-120905-2-1A1)

(24)

I made a wye, an you’re gonna make a tee;

(01:17.5-120605-1-4A2)

3.3 Simple verb A simple verb form, as used for present time reference, can be used in future time contexts. This construction occurs most frequently with first person singular subjects, and many of the recorded utterances are offers and threats, suggesting that it might be associated with the speaker’s immediate intentions. (25)

Ah swap you a pen; Ah let ya old it til tomorrow;

(17:39.7-130508-3-4)

(26)

A: E knock ya out B: E won, ahll knock im straight out

(00:01.9-120604-2-3A)

11 In response to a question from one reviewer, ahm occurs only before gonna, verbs with –ing and before adjectives, prepositions and nouns with a present tense sense (e.g. ahm cold), which corresponds with where forms of be are used in Standard Australian English. Because ahm is found in both future and present time reference contexts it seems unlikely that –m is giving the future meaning.

The case of Woorie Talk

(27)

73

If ufla git two porkis,{echidnas} we share some wit da old people (Poster: Language Perspectives, 2015)

This use of simple verb forms differs from the future use of the present tense in Standard English, in that it can be used in a broader range of contexts, to talk about events that are not planned or regular occurrences. The simple form is attested in other Aboriginal English varieties and in NSW and QLD Pidgins. It appears to be used relatively frequently in Woorie Talk, when compared, for example, to the very low proportion of the equivalent “unmarked” tokens found in Malcolm (1996: 157).

3.4 Verb + -in/-ing Verbs with an -in or -ing suffix can be used with future as well as present time reference in Woorie Talk. There is variation in whether the form is realised as -in or -ing, but the -in form is by far the more common (37/40 tokens in our corpus).12 As with gonna, there is variation as to whether an auxiliary is used before a verb + -ing construction expressing future time. (28)

Dey goin Wednesday (show);

(08:04.4-120605-2-2A1)

(29)

After library we goin big lunch;

(03:11.0-130508-3-4)

(30)

Dats going in da garbage;

(05:12.0-121112-2-7)

The examples above outline the range of forms expressing future time reference in the corpus. In the rest of this section the relationships between the Woorie Talk system we find in our data and its possible sources are analysed in more detail. As noted earlier, we know that many of the first residents of the Woorabinda reserve were speakers of earlier contact varieties which spread through Queensland during the second half of the nineteenth century. As a result, QLD Pidgin (which developed from NSW Pidgin, see Troy 1990; Dutton 1983), provided an important input to Woorie Talk. However in Woorie Talk, the time adverbs attested in QLD Pidgin (directly and by and by) are absent, and did not grammaticalise into future time markers as they did in Kriol. Woorie Talk shows additional input from English, in constructions with will and gonna that are not regularly found 12 A systematic analysis of variation across all future marking strategies in Woorie Talk is ongoing, but is in too preliminary a phase to report here.

74

Ilana Mushin and Janet Watts

in the earlier pidgin varieties, and which appear to be modelled more directly from English, consistent with other varieties of Aboriginal English. At the same time, the simple form is still used, and can be used to express uncertain or unplanned events. This is seen clearly in the example in (11) above from Alexander (1968: 61), But she pass probly. This is unlike the future use of the present tense in varieties of English, which represents “future assumed to be fact”, and is restricted to statements about inalterable plans (Leech 2004: 65). The verb + -in form is used, as it was in QLD Pidgin, but now variably occurs with forms derived from the English auxiliary. What was the nature of the further contact with English that led to the shift away from NSW/QLD Pidgin time adverbs to a more English-like system? We would expect that the English varieties the first residents of Taroom and Woorabinda reserves had the most contact with were varieties of early Australian English. In the mid-nineteenth century, before Woorabinda was established, there would have been a more diverse range of varieties of English spoken in Queensland, as a large proportion of the non-Indigenous population had migrated from overseas. This would have included regional varieties from England, Ireland, Scotland (and to a lesser extent, Wales). It would also have included second language speakers of English, or of varieties like Chinese Pidgin English, as many labourers were from China, Germany and India (Rechner 2005: 59). By the time the Taroom reserve was established in 1911, however, only 25.9% of the population of Queensland was born outside Australia, compared to 67.3% when Queensland became a state in 1861 (Queensland Treasury 2009). Of the population born overseas, four fifths were from the United Kingdom and Ireland, with over half of these born in England, and around a third in Ireland. According to Schneider (2007: 121), the period from 1901 to 1942 was a phase of nativisation in the history of Australian English, with the stabilisation and beginning recognition of its linguistic features occurring alongside a growing sense of nationhood during Federation and World War I. Varieties of English spoken in Queensland in the early twentieth century have not been described in detail. However, it is likely that the kind of work done by Taroom and Woorabinda residents in agricultural industries, as stockmen and in other station work for men (Forde 1990: 28, l’Oste-Brown et al. 1995: 18), and in domestic roles for women (Forde 1990: 29), would have brought them into contact with speakers of the kinds of non-standard varieties associated with rural communities (Shnukal 1978) and “working-class and country men” (Pawley 2008: 362). So it is expected that non-standard Englishes would have had a greater input to Woorie Talk than Standard Australian English (the influence of non-standard varieties of English on Aboriginal Englishes is also noted by Malcolm (2008: 216) and Schneider (2007: 122)).

The case of Woorie Talk

75

As expected, forms used for future time reference in Woorie Talk bear similarities to those found in non-standard and informal Englishes, in line with what we know about non-standard Australian English of the late-nineteenth and earlytwentieth centuries. Future time in non-standard Australian and British Englishes is distinguished from Standard Australian English in the higher rate of contracted forms or “casual pronunciations”, like gonna (Anderwald 2004: 181; Pawley 2008: 364) and ll. In Britain, the contracted forms gonna, gotta and wanna were first recorded in regional dialects or non-standard varieties, at roughly the same time as the non-contracted forms (e.g. going to) greatly increased in frequency (Krug 2000: 177). In Australian English written fiction, a rise in the use of going to occurred in the nineteenth century and contracted forms of will (ll and won’t) greatly rose in frequency between the nineteenth and twentieth centuries (Collins and Yao 2014: 514). Collins (2014: 9) notes in general that trends seen in written language tend to occur more often in spoken language. The origin of gonna in non-standard varieties of English, and its likely rise in popularity in Australian English through the second half of the nineteenth century fits with the timing of the development of Woorie Talk, and the prevalence of gonna. Other aspects of future time reference in Woorie Talk, however, are not found in Standard or other non-standard varieties of Australian English. As described above, the simple form of the verb is used to express future time in unplanned contexts in a way that differs from English. This may reflect QLD Pidgin input, as simple forms of the verb were commonly used to express future time. We also see variable use of forms of be with gonna and verb + -in/-ing, with the form ahm being used most commonly for first person singular pronouns. This shows similarities to other language varieties such as Kriol, where there is a restriction of the -l form to first person singular subjects, realised as al and wil (Hudson 1983: 30), and African American English, where there is variation in auxiliary use with gon and gonna (Green 2002: 36). In summary, while there is clearly more work to be done to determine the precise semantic and pragmatic environments for the use of these different ways of marking future time in Woorie Talk, we have been able to show how these forms may have emerged in that variety, and how future marking in Woorie Talk is similar to, and different from, future marking in other Australian contact varieties.

4 Identifying substrate inputs In the previous section we presented an analysis of the superstrate origins of one feature of Woorie Talk grammar – future marking – suggesting that reserve varieties like Woorie Talk have retained some features from earlier pidgins (e.g.

76

Ilana Mushin and Janet Watts

simple verb forms) while losing others (e.g. adverbs), and accounting for the more direct impact of continuing contact with English through the twentieth century. In this section we turn to features of Woorie Talk that have emerged from substrate input from traditional languages still spoken within the reserves during the early twentieth century. In order to identify substrate impacts on Woorie Talk, we need to clearly understand the nature of contact, not only between substrate and superstrate languages, but also among the substrate languages themselves. As we did for superstrate inputs in section 3, we need to separate out the features from Indigenous languages that were part of the nineteenth century spread of NSW Pidgin from features which developed after the establishment of Queensland Government reserves in the twentieth century. As noted earlier, Queensland Government reserves, such as Woorabinda, were established to receive Indigenous people from all over the State. The original intake was dominated by language groups from the immediate surrounding areas of Taroom (in the 1910s) and Woorabinda (in the 1930s), including Gunggari, Bidjara, Gangulu, and Wadjigu. However there were also significant numbers of removals of people from languages groups as far afield as Kuku Yalanji and Gugu Yimidhirr in Far North Queensland, and Kulili and Gunya in the far Southwest (see map on page 13 for locations of these languages). Until about the late 1950s, groups from different geographic areas kept to themselves within the Woorabinda reserve, living in separate camps (Munro and Mushin 2016). Interviews we conducted in 2012 with community members who grew up in the 1930s to 1950s suggest that while heritage languages were still spoken within camps to some extent, official sanctions against their use meant that they were not part of common parlance across the reserve. Furthermore, as the twentieth century progressed, maintenance and transmission of heritage languages became more difficult, not only because of prohibitions on speaking them, but also because children resided for the most part in dormitories, away from adult family members, and because Woorabinda residents of working age were constantly on the move. Such movements might have been the result of work assignments to different parts of Queensland, but they might also have been the result of further removals to and from other reserves, often for punishment, but sometimes for family reunions. Whatever the reason, the people who came to live together in Woorabinda throughout the period of The Act (1897– 1971) had little or no control over whether they remained there, nor for how long. Three important factors relating to substrate contact emerge from this social situation. Firstly, the fact that so many Queensland languages were already in significant decline as L1s by the turn of the twentieth century means that we do not expect substrate inputs to be as significant as found in contact varieties like

The case of Woorie Talk

77

Kriol where the substrates were still spoken as first languages during the period of creole-formation (see Munro and Mushin 2016 for a comparison of Woorie Talk and Kriol in this respect). Secondly, the removal of people from all over Queensland results in at least 40 different language groups being represented in each reserve (albeit with different ratios of L1 speakers of each language in each reserve) that in turn means that features in reserve varieties might have multiple sources. Thirdly, the constant movement of people in and out of reserves, and between reserves, allows for features which may have originated in one reserve to spread between reserve varieties. We see this pattern most distinctly in the distribution of the discourse particle la, illustrated in Extract 1 above (I’s all dere la), which functions to direct joint attention to a specific object. Details of the formal and pragmatic features of la are presented in Gourlay and Mushin (2015). Among Indigenous Queenslanders, this particle is highly recognisable as a feature of reserve varieties and is attested in Woorabinda, Palm Island, Cherbourg, and Yarrabah.13 In regional towns and in the Brisbane-Ipswich area, the use of la is commonly associated by Indigenous people with “Mission Talk” (i.e. reserve varieties), and has spread into town varieties in southern Queensland and northern NSW (e.g. Moree – Hitchen 1992). La is not attested in vernaculars spoken in Cape York or north-west Queensland communities. To date la has not been attested in NSW or QLD Pidgin. It is not found in Kriol, although the Kriol form luk following a demonstrative may share some functions with la (Brett Baker pc). Meakins (2011: 57) similarly identifies a clitic =rla in Gurindji Kriol which regularly co-occurs with demonstratives, but this clitic appears to derive directly from the Gurindji focus clitic =rla, and would therefore be unrelated to the la found in reserve varieties. From this we can conclude that the particle la that is found in Woorie Talk emerged in one of the Queensland reserves and spread to other reserves with the movement of people, and that it was not retained from earlier pidgin varieties. Indeed, as outlined in Gourlay and Mushin (2015: 19), there are a number of plausible sources for la that come from the Indigenous languages removed to Queensland Government Reserves. For example, in Kuku Yalanji, a language from north of Cairns, a clitic =la and its allomorph =lu are described as examples of

13 There appears to be a lack of consensus among Queensland Indigenous people as to whether la is associated with a particular reserve variety or with Queensland varieties more generally. Some Woorabinda residents claim that it is more frequent in Woorie Talk than in other varieties.

78

Ilana Mushin and Janet Watts

attention-getting devices in discourse and may attach to all parts of speech (. . .) These clitics seem to request the audience to pay special attention and could be translated as ‘hark’ or ‘pay attention’. (Patz 2002: 115)

The attention-getting functions of the Kuku Yalanji clitic appear closely aligned with the contemporary functions of la as a particle for achieving joint attention on a specific object. The Kuku Yalanji people who were removed to reserves, including Woorabinda, were still speaking Kuku Yalanji as their first language, and indeed there are still people in Woorabinda today who know some of the language. It is possible that attention-getting clitics are relatively widespread in that region of far north Queensland that includes Kuku Yalanji, however, there are few detailed descriptions such languages. Even where some language description material is available, it rarely considers discourse phenomena. Another possible source for the particle la is the particle pala/bala, a form found in a number of Pama-Nyungan languages. The functions of this particle are various but they include demonstratives (e.g. Yanykuntjatjara – Goddard 1986), and “focus” functions (e.g. Bilinarra – Meakins and Nordlinger 2014: 397). In Margany/Gunya, a language that was spoken in southwest Queensland near what is now the town of Cunnamulla, Breen (1981: 346) describes bala as a “focus on demonstratives”, co-occurring with demonstratives to mean “this one now”. There is certainly functional overlap with la in reserve varieties, but we are not in a position to speculate whether bala was used to manage joint attention in the way that la clearly does in reserve varieties. The languages of southwest Queensland were among the earliest casualties of the colonisation and so even by the 1960s, when Gavan Breen began his survey of western Queensland languages, there were very few speakers of any languages of this region remaining. Similar conditions exist for the languages of the regions surrounding Taroom and Woorabinda, all of which bear some relation to each other. The little archival data that remains for these languages does not assist us in determining the extent to which particles like bala, which place a “focus on demonstratives”, or which serve to gain an interlocutor’s attention on an object, were features of the south western and central Queensland languages which came together at Woorabinda. If people from different parts of northern and southern Queensland were already familiar with phonologically similar forms that served to direct joint attention, then this may have reinforced its incorporation into the new language. The fact that even among the sparse descriptions of Queensland languages we can find forms which bear both phonological resemblance and functional overlap with the particle la, suggests that this particle is indeed an example of a

The case of Woorie Talk

79

substrate feature that emerged in reserve varieties independently of earlier pidgin spread during the twentieth century.14 The phonology of Woorie Talk also appears to have incorporated at least one local substrate feature in addition to features retained from earlier contact varieties. As noted earlier some of the segmental phonology of Woorie Talk mirrors features found in the earlier pidgins, including the use of dental stop consonants where English would have interdental fricatives (e.g. ting ‘thing’, dat ‘that’), and simplification of final consonant clusters (e.g. negative particle don). These features clearly derive from the lack of such fricatives and consonant clusters in traditional Indigenous languages (see Baker and Bundgaard-Nielson (this volume) for a discussion of Kriol obstruents). One feature of Woorie Talk which does not seem to be widespread across Australian contact languages is the use an intervocalic trilled [r] between nonhigh vowels where other Aboriginal contact languages might use a voiced alveolar stop (e.g. budda [bara] ‘brother’). The widespread word for horse, yarraman, which is used in Woorie Talk is spelled ‘yaduman’ by consensus, reflecting that Woorie Talk speakers consider the [r] in this environment to be an allophone of /d/. This allophony may have emerged from variation between a trill and an alveolar stop in Bidjara, as reported in Breen (1973).15 Bidjara people constituted a prevalent language group of Woorabinda. Bidjara was one of the languages recorded by Elwyn Flint in Woorabinda in the 1960s. Our interviews with elders in 2012 indicates that it was still being spoken in camps through the middle of the twentieth century, although no longer transmitted to children. In this section we have shown that reserve varieties such as Woorie Talk did adopt features from traditional Queensland languages that are not found in the 14 A number of audience members at the workshop presentation of this paper raised the particle la(h), found in contemporary varieties derived from Chinese Pidgin (e.g. Singapore English and Malay English), as a possible additional source for the la found in reserve varieties. It is true that Queensland Indigenous people, especially those working in industries under The Act, often worked alongside Chinese people, and there is a long history of contact in the nineteenth century between speakers of Chinese Pidgin English (and/or Chinese dialects) and Queensland Indigenous people. However in general there is little evidence of direct transfer of Chinese Pidgin features into Aboriginal Pidgins (Meakins 2014: 373), with some evidence that Chinese workers tended to use the Aboriginal Pidgins in contact situations. Furthermore, the functions of lah as described for contemporary Asian English varieties do not significantly overlap with the functions of la in reserve varieties (see Gourlay and Mushin 2015: 19–20 for further discussion). 15 Mary Laughren (MS) also observes some alternations between an alveolar stop and a trill in the recordings of speakers of Bidjara and the closely related language Gunggari recorded by Elwyn Flint.

80

Ilana Mushin and Janet Watts

records of NSW or QLD pidgin. These inputs are fewer than the substrate influences we see in the grammars of other Australian contact languages (e.g. Roper River Kriol – see Dickson, Ponsonnet and Nichols this volume) as we might expect given the parlous state of traditional language maintenance in early twentieth century Queensland. Aside from the two examples discussed above, most other substrate impacts appear to be lexical (e.g. gundunoo ‘child’ from gandu/kandu/andu in Gungarri, Bidjara and Margany/Gunya – Laughren (MS)), but more research is required to ascertain the extent to which such words entered Woorie Talk from the reserve in the twentieth century, and which words were already present in nineteenth century contact varieties that contributed to the emergence of Woorie Talk.

5 Conclusions Our aim in this chapter has been to examine the ways in which inputs from NSW/QLD pidgin, colonial English varieties and traditional Indigenous languages have contributed to the development of Indigenous vernaculars that emerged in Queensland Government reserves in the twentieth century. Earlier work on these varieties did acknowledge contact origins, as summarised in Allridge (1984), but did not investigate the nature of the language contact situations that gave rise to these varieties. In Australian contact linguistics one hypothesis to account for the considerable reliance on English in such varieties has been to assume that at some stage, presumably during the early part of the twentieth century, speakers of pidgin varieties (who may also have been speakers of heritage languages) began to shift towards speaking English, the dominant language of the Australian mainstream, as their target language. This is the process of “de-pidginisation” proposed in Mühlhäusler (1979), and taken up in Allridge (1984), to account for reserve varieties that have been the focus of the current chapter. To date, however, there has yet to be a detailed empirical study of the process of depidginisation as a mechanisim of language change in a community in Australia. Our discussion here provides some evidence that the development of varieties such as Woorie Talk involved both incorporation of notable amounts of English features as well as retention (including internally-motivated development) of pidgin and substrate features. To understand contemporary Indigenous vernaculars such as Woorie Talk thus requires a historical perspective that develops an account of how all of these factors have impacted the contemporary grammar.

The case of Woorie Talk

81

In comparing the data for Woorie Talk used in Alexander (1968) with the corpus collected in the 2010s we see little evidence of a shift towards English in the latter part of the twentieth century, either sudden or gradual. While the numbers of people who are proficient in English may have increased with the advent of better access to Australian mainstream centres through education, media, and the rescinding of the extreme controlling measures of The Act in 1971, there is little evidence that the Woorie Talk we have recorded in the twenty-first century is vastly different from that recorded in the 1960s. This suggests that Woorie Talk has attained a degree of stability as a community vernacular, although there was consensus among the older and younger Woorabinda residents that we interviewed that there are differences between young people’s vernacular and Woorie Talk. It is believed that the young people’s vernacular has recently been influenced by African American culture, via the mass media. In accounting for the combination of language influences on Woorie Talk we recognise that the development of Queensland reserve varieties from the early twentieth century did indeed involve to a greater extent as English as an input language than we see in some other contact varieties such as Kriol and Torres Strait Creole. This is partly due to the diversity of language groups within Woorabinda, many of which had already largely shifted from their traditional languages to speaking contact varieties prior to the establishment of the reserves. The greater influence from English, either directly or indirectly via pidgin input is thereby a consequence of the more significant decline of traditional language use that precluded these substrate languages from making a more substantial contribution. In this paper we focused on one small part of the grammar, the marking of future time, to show that the Woorie Talk system utilises structures found in NSW and QLD Pidgins, as well as early Australian English. We are finding similar blends of such systems in work being undertaken on the rest of the TAM system, prepositions and the determiner system that will be the subject of future publications. Our close examination of one variety among Queensland reserve varieties has also shown that despite the declining use of substrate languages as L1, and the diversity of remaining languages, which include languages from south and north, eastern and western Queensland, we can nonetheless identify features in Woorie talk that appear to have developed independently of the nineteenth century pidgins, and have likely sources in the particular constellation of traditional languages which came into contact within the reserves. Such developments help to preserve the distinctiveness of reserve varieties as products of the particular confluence of social and historical factors that we outlined here. We expect that close examination of other contact varieties which have come under the umbrella term “Aboriginal English” will reveal that at least

82

Ilana Mushin and Janet Watts

some of the variation we find across such varieties, within Queensland and across Australia more generally, can be accounted for in terms of local patterns of language contact. We hope that research such as we have presented in this chapter can lead to more detailed studies of individual varieties that project the diversity of Australia’s colonial history and its impact on Indigenous languages.

References Alexander, Diane Helen. 1965. Yarrabah Aboriginal English. Brisbane: University of Queensland Honours thesis. Alexander, Diane Helen. 1968. Woorabinda Aboriginal Australian English. Brisbane: University of Queensland MA thesis. Allridge, Clare. 1984. Aboriginal English as a Post-Pidgin. Canberra: The Australian National University Honours thesis. Anderwald, Lieselotte. 2004. The varieties of English spoken in the Southeast of England: morphology and syntax. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider & Clive Upton (eds.). A handbook of varieties of English: a multimedia reference tool. Vol. 2: morphology and syntax. Berlin & New York: Mouton de Gruyter. Angelo, Denise. 2013. Identification and assessment contexts of Aboriginal and Torres Strait Islander learners of Standard Australian English: challenges for the language testing community. Papers in Language Testing and Assessment 2(2). 67–102. Angelo, Denise & Eva Schultze-Berndt. 2013. Kriol structure dataset. In Suzanne Maria Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber (eds.). Atlas of pidgin and creole language structures online. Lepzig: Max Planck Institute for Evolutionary Anthropology. http://apics-online.info/contributions/25 (accessed 23 March 2015) Baker, Philip. 1993. Australian influence on Melanesian Pidgin English. Te Reo 36. 3–67. Blake, Thom. 2001. A dumping ground: A history of the Cherbourg settlement. Brisbane: University of Queensland Press. Breen, J. Gavan. 1973. Bidyara and Gungabula: grammar and vocabulary. Linguistic Communications 8. 1–227. Melbourne: Monash University. Breen, J. Gavan. 1981. Margany and Gunya. In Robert M. W. Dixon and Barry Blake (eds.) Handbook of Australian languages, vol. 2, 275–393. Canberra: The Australian National University Press. Butcher, Andrew. 2008. Linguistic aspects of Australian Aboriginal English. Clinical Linguistics & Phonetics 22(8). 625–642. Collins, Peter. 2014. Quasi-modals and modals in Australian English Fiction 1800–1999, with comparisons across British and American English. Journal of English Linguistics 42(1). 7– 30. Collins, Peter & Xinyue Yao. 2014. Grammatical Change in the Verb Phrase in Australian English: A Corpus-based Study. Australian Journal of Linguistics 34(4). 506–523. Crowley, Terry & Bruce Rigsby. 1979. Cape York Creole. In T. Shopen (ed.). Languages and their status. 153–207. Cambridge, Mass: Winthrop. Dutton, Tom. 1965. The Informal Speech of Palm Island Aboriginal Children, North Queensland. Brisbane: University of Queensland MA thesis.

The case of Woorie Talk

83

Dutton, Tom. 1983. The origin and spread of Aboriginal Pidgin English in Queensland: A preliminary account. Aboriginal History 7(1–2). 90–122. Eades, Diana. 1983. English as an Aboriginal Language in Southeast Queensland: University of Queensland PhD thesis. Eades, Diana. 2014. Aboriginal English. In Harold Koch & Rachel Nordlinger (eds.). The languages and linguistics of Australia: A comprehensive guide. 417–448. Berlin & Boston: Walter de Gruyter. Flint, Elwyn. 1968. Aboriginal English: Linguistic description as an aid to teaching. English in Australia 6. 3–21. Forde, Therese. 1990. “Confinement and control” A history of Woorabinda Aboriginal community 1927–1990. Brisbane: University of Queensland Honours thesis. Goddard, Cliff. 1986. Yanykunytjatjara Grammar. Alice Springs: Institute for Aborginal Development. Gourlay, Claire & Ilana Mushin. 2015. ‘Up dere la’: Final particle la in a Queensland Aboriginal Vernacular. Australian Journal of Linguistics 35(1). 1–26. Green, Lisa. 2002. African American English: A linguistic introduction. New York & Cambridge: Cambridge University Press. Harkins, Jean. 1994. Bridging two worlds: Aboriginal English and crosscultural understanding. Brisbane: University of Queensland Press. Hitchen, Michael Moy. 1992. Talkin up: Aboriginal English in Moree M.Litt. Armidale, NSW: University of New England M.Litt. thesis. Huddleston, Rodney & Geoffrey K. Pullum. 2002. The Cambridge grammar of the English language. Cambridge: Cambridge University Press. Hudson, Joyce. 1983. Grammatical and semantic aspects of Fitzroy Valley Kriol. Darwin: Work Papers of SIL-AAB Series A, Volume 8. Kaldor, Susan & Ian Malcolm. 1991. Aboriginal English – an overview. In Suzanne Romaine (ed.). Language in Australia. 67–83. Cambridge: Cambridge University Press. Krug, Manfred G. 2000. Emerging English modals: a corpus-based study of grammaticalization. Berlin & New York: Mouton de Gruyter. Language Perspectives. 2015. Youfla whichay? Woorie way [poster]. Cairns: Indigenous Schooling Support Unit, Education Queensland. Laughren, Mary (MS) Elwyn Flint’s 1965 recordings of Maric Languages of the Warrego Maranoa region of Queensland. Unpublished Manuscript. L’Oste-Brown, Luke Godwin, Gordon Henry, Ted Mitchell & Vera Tyson. 1995. Living under the Act: Taroom Aboriginal reserve 1911–1927. Brisbane: Dept. Environmentand Heritage, Queensland. Leech, Geoffrey. 2004. Meaning and the English verb. Harlow, England: Pearson/Longman. Malcolm, Ian. 2008. Australian creoles and Aboriginal English: morphology and syntax. In Kate Burridge & Bernd Kortmann (eds.). Varieties of English 3: The Pacific and Australia. Berlin: Mouton de Gruyter. Malcolm, Ian. 2000. Aboriginal English Research: an Overview. Asian Englishes 3(2). 9–31. Malcolm, Ian. 1996. Observations on variability in the verb phrase in Aboriginal English. Australian Journal of Linguistics 16(2). 145–165. Malcolm, Ian & Ellen Grote. 2007. Aboriginal English: Restructured variety for cultural maintenance. In Gerhard Leitner & Ian Malcolm (eds.). The Habitat of Australia’s Aboriginal Languages: Past, Present and Future. 153–179. Berlin & New York: Mouton de Gruyter.

84

Ilana Mushin and Janet Watts

Malcolm, Ian & Marek M. Koscielecki. 1997. Aboriginality and English: Report to the Australian Research Council. Mt Lawley: Centre for Applied Language Research. Meakins, Felicity. 2007. Case-marking in contact: the development and function of case morphology in Gurindji Kriol, an Australian mixed language. Melbourne: The University of Melbourne PhD thesis. Meakins, Felicity. 2011. Case-marking in contact: the development and function of case morphology in Gurindji Kriol. Amsterdam: John Benjamins. Meakins, Felicity. 2014. Language contact varieties. In Harold Koch & Rachel Nordlinger (eds.). The Languages and Linguistics of Australia: A comprehensive guide, 362–411. Berlin: Walter de Gruyter. Meakins, Felicity & Rachel Nordlinger. 2014. A Grammar of Bilinarra: An Australian Aboriginal language of the Northern Territory. Berlin: Mouton de Gruyter. Morning Bulletin. 1927. Woorabinda. Morning Bulletin, Rockhampton, Monday, August 8, 1927. p.3. http://nla.gov.au/nla.news-article54605556 (accessed 29 September 2014). Mühlhäusler, Peter. 1979. Remarks on the pidgin and creole situation in Australia. AIAS Newsletter 12. 41–52. Munro, Jennifer. 2004. Substrate language influence in Kriol: The application of transfer constraints to language contact in Northern Australia: University of New England PhD thesis. Munro, Jennifer. 2011. Roper River Aboriginal language features in Australian Kriol: Considering semantic features. In Claire Lefebvre (ed.). Creoles, their Substrates, and Language Typology. Amsterdam: John Benjamins. 461–87. Munro, Jennifer & Ilana Mushin. 2016. Rethinking Australian Indigenous English-based speech varieties: Evidence from Woorabinda. Journal of Pidgin and Creole Languages 31(1). Mushin, Ilana, Denise Angelo & Jennifer Munro. 2016. Same but different: understanding language contact in Queensland Indigenous settlements. In Jean Christophe Verstraete & Di Hafner (eds.). Land and language in the Cape York Peninsula and Gulf Country. Amsterdam: John Benjamins. 383–408. O’Shannessy, Carmel. 2013. The role of multiple sources in the formation of an innovative auxiliary category in Light Warlpiri, a new Australian mixed language. Language 89(2). 328–353. Patz, Elizabeth 2002. A grammar of the Kuku Yalanji language of North Queensland Canberra: Pacific Linguistics. Pawley, Andrew. 2008. Australian Vernacular English: some grammatical characteristics. In Kate Burridge & Bernd Kortmann (eds.). Varieties of English: The Pacific and Australasia. Berlin: Mouton de Gruyter. Queensland Treasury. 2009. Historical Tables, Society, 1859–2008 (Q150 Release) http://www. qgso.qld.gov.au/products/tables/historical-tables-society/index.php (accessed 3 September 2014). Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A comprehensive grammar of the English language. New York: Longman. Readdy, Coral Ann. 1961. South Queensland Aboriginal English. Brisbane: University of Queensland Honours thesis. Rechner, Judy Gale. 2005. Taroom Shire: pioneers, magic soil and sandstone gorges. Taroom, Qld: Taroom Shire Council. Romaine, Suzanne. 2004. English-lexicon pidgins and creoles of the Pacific. In Raymond Hickey (ed.). Legacies of colonial English: studies in transported dialects. New York: Cambridge University Press.

The case of Woorie Talk

85

Sandefur, John R. 1979. An Australian creole in the Northern Territory: a description of NgukurrBamyili dialects (part 1). In Work Papers of SIL-AAB Series B, Volume 3. Darwin: Summer Institute of Linguistics. Schneider, Edgar W. 2007. Postcolonial English: varieties around the world. Cambridge: Cambridge University Press. Shnukal, Anna. 1978. A sociolinguistic study of Australian English: Phonological and syntactic variation in Cessnock, N.S.W.: Georgetown University PhD thesis. Shnukal, Anna. 1988. Broken: an introduction to the creole language of Torres Strait. Pacific Linguistics C-107 Canberra: Research School of Pacific Linguistics, Australian National University. Troy, Jakelin. 1990. Australian Aboriginal contact with the English language in New South Wales: 1788 to 1845. Canberra: Dept. of Linguistics, Research School of Pacific Studies, Australian National University. Walsh, Michael. 1991. A nagging problem in Australian lexical history. In Tom Dutton, Malcolm Ross & Darrell Tryon (eds.). The Language Game: Papers in Memory of Donald C. Laycock. Canberra: Pacific Linguistics. 507–19. Watson, Joanne. 2010. Palm Island: Through a long lens. Canberra: Aboriginal Studies Press. Winford, Donald. 2000. “Intermediate” creoles and degrees of change in creole formation: the case of Bajan. In Ingrid Neumann-Holzchuh & Edgar W. Schneider (eds.). Degrees of restructuring in creole languages, 217–245. Amsterdam & Philadelphia: John Benjamins.

III Transfer of form: Lexical

Patrick McConvell

3 Kinship loanwords in Indigenous Australia, before and after colonization Abstract: This paper begins with a discussion of medieval kinship loanwords into English from French. As well as illustrating how these terms (uncle, aunt, cousin, nephew etc.) are drawn from outside the lineal core of the family, they also show how the takeover from earlier Anglo-Saxon terms was not immediate but went through a phase of plurality or overlay of some centuries in which there was kin-switching between the old and new systems, parallel to code-switching, a phenomenon also to be discussed in relation to Australian Indigenous languages as they borrow kinship terms. As well as these collateral terms highlighted in English, affinal (in-law) terms are also commonly loanwords in many languages. Turning to Australia, examples are given of long-distance affinal Wanderwörter (travelling words), including ramparr a word for ‘mother-in-law’ which turned into lamparr(a) ‘father-in-law’ as it travelled east and exploded across the Northern Territory in the last 150 years. In contrast, linguistic prehistory can show us examples of local borrowing of grandparent kinship terms to fill gaps as systems change, and rarer examples of more wholesale local borrowing. Moving on to the more recent era when English and Pidgin-Kriol have had impact on the situation, terms like uncle and aunt when used in Indigenous contexts tend at first to have the ‘Aboriginal’ meanings ‘mother’s brother’ and ‘father’s sister’ but with increasing kin-switching into the wider English meanings. Pre-existing features of some Aboriginal languages like Guugu Yimidhirr predispose them to move in this direction but this language and others add hybrid compounds like cousin-ngamu (mother’s brother’s daughter). ‘Cousin’ is most often cross-cousin (mother’s brother’s or father’s sister’ child) in earlier stages of Pidgin-Kriol based on traditional kin-classification, but in an area in the Northern Territory it also intriguingly means ‘mother-in-law’, because this feature was adopted from a language around the Queensland border on the advancing eastward path of Pidgin-Kriol and the cattle industry in the late nineteenth century.

1 Introduction Kinship terms tend to be inherited often from quite distant proto-languages. However there are exceptions – in English for instance a number of kinship terms were borrowed from French in the late Middle Ages, displacing earlier terms and changing the kinship system. These included uncle, aunt, and cousin.

90

Patrick McConvell

Matras (2009: 169–171) proposes a hierarchy of borrowability of kinship terms whereby collateral terms (like those just cited), and affinal terms (spouse, in-laws) are more commonly borrowed than lineal terms (parents, siblings etc.). In traditional Australian languages, change tends to follow the pattern suggested by Matras, with more affinal terms borrowed, some travelling long distances as Wanderwörter along with changes in marriage systems. In a few cases (e.g. Nyulnyulan) a larger proportion of terminology of all kinds is borrowed from neighbours. As English and Pidgin-Creole varieties entered the scene after colonization, some terms began to be adopted alongside the traditional Indigenous ones, finally replacing them in some cases. The most common of these borrowings included versions of uncle, aunt and cousin but with meanings adapted to a different system (e.g. uncle as ‘mother’s brother’ only). Other lineal terms were also apparently adopted (mama, papa, granny, brother, sister) but care has to be taken not to assume borrowing as there are similar Indigenous terms (Hendery and McConvell 2013). There is contextual variation in meaning and function of these loanwords in speech, from those closer to standard Australian English to those closer to traditional Indigenous systems.

2 Kinship loanwords in medieval Europe 2.1 Uncles, aunts and cousins A striking diffusion of kinship loanwords took place from French into Germanic languages in the late medieval period. Those entering English included uncle, aunt and cousin, but they were also borrowed into other languages, such as German, Dutch and Swedish. Icelandic was exempt from this due to its relative isolation. Why did this happen? There is surprising lack of agreement among those who have examined this. Goody (1959; 1983: 262–278) attempted to turn the question on its head and rather asked why Anglo-Saxon kinship terms survived in the consanguineal terms for close family: father, mother, son, daughter.

2.2 ‘Plurality’: nephew Goody (1983: 271) also made the interesting observation that in late Medieval and early Modern English, the different terms and systems coming by inheritance from Anglo-Saxon and by borrowing from French were actually used alongside

Kinship loanwords in Indigenous Australia, before and after colonization

91

each other for several hundred years. This type of mingling, or we might say kinswitching – parallel to the more general notion of code-switching – has been studied not so much in linguistics or anthropology as by historians of social institutions who comb literary and biographical texts for examples, e.g. Tadmor (2004). In this field the phenomenon is known as plurality because there is more than one term for what is apparently the same kin-type. The metaphor of ‘onion skin’ in an ‘onion model’ (MacFarlane 1978: 146–147; Tadmor 2004: 147) has been used to portray the different layers of terminology and system. At the same time, but separated from the above ‘plurality’ studies by a seldom breached disciplinary wall (except notably for the anthropologist Goody), there has been an upsurge in interest in anthropology in variation in kinship systems and terminologies. This has been conceptualized by Kronenfeld (2009) in terms of overlays, that is, one system is substituted for another in speech to signal a particular discourse orientation. For instance a skewing system (where kin of adjacent generations are referred to by the same term, one type of which is Omaha skewing) is used when common affiliation to a lineage and inheritance is at stake, and a more neutral base system without skewing otherwise. Similar discourse motivations for skewing in Australian Aboriginal cases were pointed out by Rumsey (1983) independently of Kronenfeld, and McConvell (2012) surveying the range of skewing in Australia, linking to the work of Kronenfeld. The analysis of plurality/overlay is not necessarily linked to the history and etymology of the forms of kinship terms involved, but it may be. As in the case of the new layer in Germanic languages, the later layer is borrowed from French. In some cases the different forms are doublets, that is the different forms are descended ultimately from the same proto-form, but by different routes, one by consistent inheritance and another by later borrowing from a different branch of the same family. Occasionally the borrowed words have remained close in form to the ancient proto-forms and their inherited cousins, and may initially pose a problem of deciding whether the word is inherited or diffused. This is at first glance the case with English nephew. This word descends from a proto-Indo-European root *népōts (Mallory and Adams 1997: 392). It is agreed by most scholars that the senses of this word in proto Indo-European included ‘grandson’ but some believe that it was already polysemous in the protolanguage, with also the sense ‘nephew’.1 Most etymologies give the proximate

1 This polysemy has been argued to be part of the general operation of Omaha skewing (extension of kinship term meaning to an adjacent generation in the paternal line) in proto-IndoEuropean (pIE), which also explains some other equivalences such as ‘grandfather /MF’ = ‘uncle/ MB’ and the reason why MB, later MB/FB, in Latin is avunculus ‘little grandfather’, the source of French oncle, English uncle etc. (Friedrich 1966). Some sources identify the original meaning of pIE *népōts as ‘grandson’ (Hettrich 1985), some ‘nephew’ (Benveniste 1969: 188–190).

92

Patrick McConvell

source of English ‘nephew’ as Old French neveu (the same spelling as Modern French), which is thought to have entered Middle English in the thirteenth century. Some sources cite the Old French form as having a ‘grandson’ meaning, as well as ‘nephew’, although it is unclear if this is true of all eras and dialects.2 However the Anglo-Saxon form nefa is similar in form, and does have the ‘grandson’ – ‘nephew’ polysemy; it is descended from proto-Germanic *nefô which has the same polysemy (Ringe 2006: 96). In the operation of plurality/ overlay, which continued until at least the eighteenth century, the other terms borrowed from French like uncle, aunt and cousin were decidedly different from the Anglo-Saxon terms. But in the case of nephew the similarity of sound-form may have led to some degree of hybridization. In terms of meaning, there is a clear ‘retention’ of the meaning ‘grandson’ alongside ‘nephew’ in the English Bible translations of the sixteenth century. On the one hand this is interpreted as a Germanic legacy, since this polysemy was present in proto-Germanic, and additional senses found also in early modern English, such as ‘cousin’, are found in Dutch neef today for instance. On the other hand, scholars have seen here the influence of the meaning of Latin nepos ‘grandson’ on the learned writing of the time, and later the influence of this meaning through the Bible translations themselves (Tadmor 2004: 148). Tadmor (2004: 148–150) argues that MacFarlane’s use of the ‘onion model’ presents an abrupt temporal divide between when two layers are ‘fused’ (polysemy) and when they separate (monosemy with another term taking over the previous second sense). In fact Tadmor cites evidence from English kinship terminology – specifically the archaic ‘grandson’ meaning of nephew coexisting with the narrow nephew meaning, versus the loss of the ‘grandson’ meaning – to show that ‘this terminological shift . . . was a very slow process and . . . for a long time plural usages could have existed simultaneously’. She hints at the kind of motivation the same people had for using the new ‘grandchildren’ rather than the archaic (in this sense) ‘nephews’ in different contexts.3 I have dwelt on this example because it illustrates the kind of issue the dynamics of kinship presents when we face a language contact context more familiar to some readers than that of Australia. Also it is good to be able to examine situations which have been more documented over hundreds of years longer than in Australia, and the often slow and complex process of different

2 A dictionary of the Norman variety around the time of the conquest of England does not give a ‘grandson’ meaning for neuz, neweu, newe ‘nephew’ (Kelham 1779: 163). 3 See Jones (1990) on the complexity of overlaps, overlays and change in Medieval German kinship, and commentary in McConvell (2013: 9) including on reflexes of pIE *nepots ‘nephew etc.’.

Kinship loanwords in Indigenous Australia, before and after colonization

93

systems exerting pressure on each other. There are obviously huge cultural differences in the situations: one interpretation of the English evidence (if this is indeed correct) that meanings were affected by similarities of terms to classical literary sources is clearly not a factor in Australia for instance. But there are basic commonalities in the ways in which kinship terminology changes and is affected by language contact. We will return to some of the issues raised here, of long periods of plurality/overlay of systems, and of etymology and discriminating between loans and inheritances when we look into the Australian Indigenous situation.

3 Borrowing of kinship terms between Australian traditional Indigenous languages 3.1 Affinal Wanderwörter Affinal terms, designating spouses and in-laws, appear to be among the most highly borrowed kinship terms in Australia, following the world-wide trend identified by Matras (2009: 169–171). Some of these are notable Wanderwörter ‘travelling words’ that are borrowed successively into languages of different groups and families across long distances (Haynie et al. 2014). In the case of two such affinal Wanderwörter in northern Australia, I have argued that changes in distribution and meaning which accompany diffusion tell a story of the diffusion of new marriage patterns and which kin controlled the betrothal of women, mothers-in-laws or fathers-in-laws (McConvell 2015). The path of travel of these two roots %ramparr and %tyamVny+ are shown on the maps in Figure 1 and 2, respectively). One is the root %ramparr,4 which began in the Worrorran family in the North Kimberley with the meaning ‘screen’ (referring to avoidance behavior) then centred on the wife’s mother, subsequently diffusing to the Nyulnyulan family where it changed meaning to wife’s father, then as it was borrowed into the Ngumpin-Yapa subgroup of Pama-Nyungan, underwent a regular sound change in that subgroup to lamparr; addition of -a again by a regular sound change in one language, Mudburra, then rapid and recent diffusion east as the

4 I use the symbol % to indicate an early and general form of the diffusing root. The asterisk * is inappropriate because this is not a proto-form nor does it stand for a set of cognates in the sense of forms descended from one proto-form.

94

Patrick McConvell

Figure 1: Map of spread of the Wanderwort ramparr > lamparr(a)

cattle industry created new networks and spread the word lamparra and the associated marriage conventions further east. The other kinship Wanderwort on Figure 2 is the term %tyamVny+. This is in fact the northern diffusion of a widespread Pama-Nyungan inherited root *tyam(p)Vny which is primarily ‘mother’s father’ in Pama-Nyungan (Walmajarri, tyamirti; Gurindji tyawityi etc.). In many languages, especially in the northern diffusion of the root into non-Pama-Nyungan, it has an extension to ‘crosscousin’ by Alternate Generation equivalence and thence to ‘spouse’ and/or ‘sibling-in-law’ by the consanguineal-affinal or ‘spouse-equation’.5 McConvell (2013b, 2015) argues that it is the affinal meaning of this term which is the ‘cutting edge’ that enabled it to diffuse over such a wide area of non-PamaLanguages. It would make sense if the diffusion accompanied a new form of marriage but the exact details have not been worked out yet.

5 For instance Ngalakgan tyaminy ‘spouse’. Extension types come from Scheffler (1978) and are reformulated and given a diachronic interpretation in McConvell (to appear).

Kinship loanwords in Indigenous Australia, before and after colonization

95

Figure 2: Tyamuny

3.2 Local borrowing – gap fillers Affinal Wanderwörter are not the only types of kinship term borrowing that can be recognized. If we reconstruct a sub-group proto-language sometimes we find that certain terms cannot be reconstructed because that term is represented by different loanwords in different languages. For instance in Ngumpin-Yapa ‘father’s father’ is such a term. Here are the terms found and their loan sources: Table 1: Father’s father in Ngumpin-Yapa

Language(s)

Form

Loan source (proximate < earlier)

Form

Direction of borrowing

Eastern Ngumpin Western Ngumpin Yapa (Warlpiri)

kaku kirlaki warringiyi

Western Mirndi < Jarragan Kija < Bunuban Arrernte

kakung kirlaki arrenge-ye

North–South North–South South–North

96

Patrick McConvell

Each branch of the Ngumpin-Yapa (Eastern Ngumpin, Western Ngumpin, Yapa) subgroup borrowed the term for ‘father’s father’ (FF) from a neighbouring language on the side of the subgroup where it was situated: east, west and south, respectively. The other grandparental terms in Ngumpin-Yapa are all inherited. They can either be traced all the way back to proto-Pama-Nyungan (like *tyamu(ny) ‘mother’s father’ (MF)) or at least to proto-Ngumpin-Yapa (*tyatya ‘mother’s mother’ (MM), *ngaputyu ‘father’s mother’ (FM)). It looks as if all the branches independently required an additional FF term at some stage and grabbed one locally. So what kind of kinship system would have existed before this happened and why did they need to change? The most likely earlier system is one in which there was an equivalence of MM(B) and FF(Z), the two parallel grandparents and siblings.6 This is like the Yolngu (North-East Arnhem Land) system where the term maari covers both these kin types. However the Yolngu also felt the need of a way to distinguish FF from MM which they do by adding an optional ‘paternal’ suffix -mu(ngu) to maari’. The loan source history for kaku also presents us with another example of an earlier term for parallel grandparent(‘s siblings) becoming differentiated into MM and FF. The term for MM(B) in Jarragan can be reconstructed as *kangke(+ gender suffix) and this is still the form in Kija, with kangkayi the address form. The reflexes in the other northern Jarragan languages are Miriwoong kakayi and Gajirrabeng kake-ng MMB; kake-ny MM with the respective gender suffixes. It appears that -ngk- has become -k- in this branch of the family, quite possibly through the operation of a retrogressive form of Nasal Coda Dissimilation (noting that both gender suffixes are nasals). In Gajirrabeng however, FF appears as kakung. The vowel change marking a shift to paternal is not so obvious to explain. The eastern neighbours, Western Mirndi, have no /e/ central vowel and it will be rendered by different vowels in different environments when items containing it are borrowed. In this case, with velar segments preceding and following, the result would be /u/. It seems then that this kakung could be a back-borrowing of a doublet from Mirndi after the semantic change to FF had taken place. This makes it likely that the original borrowing of kakeng MM/FF

6 I am reluctant to suggest that the MM(B) term in pNGY *tyatya meant MM+FF, since there is no solid case where a tyaty+ term means FF. Some tyatyi and tyatye mean MF, to the east. There are many possible cognates in eastern and southern Australia meaning ‘elder sibling’ (further discussed in relation to ‘sister’ look-alikes in section 6.2). Since the Alternate Generation equivalence of ‘elder sibling’ with ‘parallel grandparent’ (MM,FF) is relatively common the connection is plausible.

Kinship loanwords in Indigenous Australia, before and after colonization

97

as kakung was into an environment where there was a narrowing to FF because a dedicated term for MM already existed.7 This is a complex language contact interaction, but this reinforces the point made in the introductory section on English that the reality can be one of multi-staged language interaction and ‘plurality’/overlay of forms with different meanings. The Eastern Ngumpin languages have exceptionally high rates of borrowing of general vocabulary for Australia, mostly from non-Pama-Nyungan neighbours of the Jarragan and Mirndi families. Gurindji for instance has over 40% loanwords in general lexical lists of 800 and 200 words including much core vocabulary (McConvell 2009; Bowern et al. 2011). The figures for the kinship domain are only slightly less at 39%, but most of them are attributable either to being affinal or gap-filler borrowings.8 Wholesale kinship term borrowing is not encountered.

3.3 Local borrowing – massive replacement In contrast to the selective borrowing of kinship terms to fill a gap in constructing a new more complex system, while retaining the rest of the inherited terms, there are cases where a large number of terms have been borrowed, reducing the inherited component. The case here is that of the Nyulnyulan family (nonPama-Nyungan) in the West Kimberley, abutting the Ngumpin-Yapa subgroup discussed above to the west. The languages of the Nyulnyulan family have been recorded as having very low loan proportions on a standard 200 word list (Bowern et al. 2011; the two figures are minimum and maximum with different formulas): Bardi Nyulnyul Nimanburru Ngumbarl

2% 1% 0% 1%

3% 2% 1% 4%

Yawuru

6%

8%

Nyikina

3%

7%

7 This is an example of ‘semantic narrowing in borrowing’ – see McConvell and Ponsonnet (2013). 8 The calculation is based on an 18 word list of which 7 are likely loans (McConvell 1997: 218– 219). This list does not include trirelational terms, which seem to have a higher loan count.

98

Patrick McConvell

This is in stark contrast with the two Eastern Ngumpin languages in the sample: Mudburra

40%

42%

Gurindji

48%

48%

Yet some domains in all or most of the Nyulnyulan languages, including kinship, have high loan proportions, including quite old loans from Pama-Nyungan languages even in basic vocabulary. The dominant direction of borrowing is opposite to that of the Eastern Ngumpin languages discussed above (where north-south, Non-Pama-Nyungan to Pama-Nyungan is dominant). Whether or not the general lexical loan figures may need eventual readjustment, a small set of kinship terms (grandparents) in one language (Nyulnyul) will be the focus here (Table 2), to provide an example of the phenomenon of massive loan replacement which is fairly typical of the family in kinship. Kami MM and tyampa-/tyamu MF are clearly Pama-Nyungan loans and go back to proto-Pama-Nyungan in the first case and probably in the second also, although they were no doubt borrowed from a nearby sub-group located in the Kimberley/Pilbara. Kaparli is also widely distributed in western Pama-Nyungan as a term for FM, in Western Desert and Arandic, eastern Ngayardic (Nyamal, Kariyarra) and western Marrngu (Karajarri), the latter the probable proximate loan source for Nyulnyulan.9 Kalu+ also kaalu and karlu in some languages, meaning FF, have a much more restricted distribution, found only in Nyulnyulan and coastal Nyangumarta, a near Marrngu (Pama-Nyungan) neighbor to the west. The alternative of this term originating in Nyulnyulan and being borrowed into Nyangumarta certainly has better prospects than this direction for the other terms. One contrary indication is the presence of the suffix -rtu in Nyulnyulan, for instance in Nyulnyul kalurtu in Table 2. This appears to be a form of a widespread Pama-Nyungan kinship suffix (McConvell 2008: 316–318)10 and is found with several kinship terms in Coastal Nyangumarta (O’Grady n.d.) although not

9 Kaparli competes with kami as the term for the fused grandmother term in Western Desert (FM = MM) and one or the other is found in different dialects (McConvell 2015). This term may be cognate with another widespread (and possibly proto-) Pama-Nyungan term for FM: ngapari. the sound changes involved cannot be justified at this stage, but if they were, the ‘alternative’, already implausible, of regarding the source of kaparli as being in Nyulnyulan would be completely ruled out 10 Nyangumarta and other Pama-Nyungan languages in the area have vowel-harmony which explains i > u (Sharp 2004: 60–69) -rta in the Nyulnyulan form kamirta ‘MM” is probably cognate with * -rti/rtu but the vowel requires explanation.

Kinship loanwords in Indigenous Australia, before and after colonization

99

with FF which has a different widespread Pama-Nyungan kinship suffix -tyi in karlutyi (McConvell 2008: 318–321).11 If however the root karlu is truly Indigenous to Nyulnyulan, this then presents us with a curious inversion of the situation of kaku ‘FF’ in Eastern Ngumpin. Once again the FF term is the odd one out, but in this case it appears possible that this is an old Indigenous term and all the other grandparent terms were borrowed from Pama-Nyungan in the south. It is difficult to ascertain with present evidence what the original meaning was of karlu, or why the other three grandparent terms were borrowed. There is evidence in the form of the grandparent terms that shows they are quite old in Nyulnyulan and if this was borrowing into this language, it was ancient. There are two diagnostic sound changes which affect words in a regular fashion in Nyulnyul, which can tell us how old the words are in the language through a method called linguistic stratigraphy (McConvell and Smith 2003). Bowern (2004: 16fn. 15) identified two ordered changes in Nyulnyul illustrated by McConvell et al. (2014; Rangan et al. 2015) for a loanword from PamaNyungan (Marrngu) of original form larrkarti ‘boab’. 1.

Regressive vowel harmony > *larrkirti

2.

Final vowel deletion > larrkirt

Under favourable circumstances these changes are able to be calibrated to yield an absolute chronology from historical or archaeological dates. For instance the term tyimpil ‘quartzite or glass spear point’ in Nyulnyul, which derives from tyimpila originating in the Worrorran language family of North Kimberley, has change 2, final vowel deletion. On the other hand tyimpil has not undergone change 1 or it would appear as tyimpal. Kim Akerman (p.c) estimates that the Nyulnyul would have entered the trade in tyimpila in the early to mid-nineteenth century. Larrkarti, which underwent both sound changes, yielding larrkirt, must have entered the Nyulnyul language before that. The same method of linguistic stratigraphy can be used for Nyulnyul grandparent terms. Clearly FM kapirl (< %kaparli) and MM kamart (< %kami-rta) have undergone both changes which puts them into the older stratum with larrkarti > larrkirt ‘boab’ before the nineteenth century and not the later stratum of tyimpila > tyimpil ‘quartz/glass point’. Two-syllable words like tyamu > tyaam do not undergo change 1, regressive vowel harmony, just lengthening of the first vowel, so 11 Kaal in the meaning ‘father’ also occurs in Nyulnyul (McGregor 2011: 776). This would be a regular outcome of regressive vowel harmony and final vowel deletion operating on *kalu, just as tyaam is found descended from *tyamu MF, but the meaning of kaal F is puzzling. See McConvell (to appear) for a possible explanation in terms of filiocentricity.

100

Patrick McConvell

cannot assist with discriminating the stratum of the word. Because the second and third vowels of kalurtu are the same, regressive vowel harmony has no effect so again we cannot tell what stratum kalurt belongs to.12 Table 2: Nyulnyul grandparental terms KINTYPE

ELKIN

MCGREGOR

FF FM MM MF

kalod kabil kamad djam

kalud, kalurd kabirl kamard jaam

Additional senses mSC fSC13 MFB, MFZ, mDC

Reconstruction

Related K-Karajarri; M-Mangarla

*kalurtu *kaparli *kamirta *tyamu(ny)

kalurtu (K) kaparli kami tyampartu tyamu (M)

4 The meaning of English ‘uncle’ and ‘aunt’ in Aboriginal languages 4.1 Uncle and aunt adopted into an Aboriginal system The terms ‘uncle’ and ‘aunty’ were absorbed readily and quickly into the speech of Australian Aboriginal people, over much of Australia, used not only in Pidgin/ Kriol varieties but also inserted into traditional vernacular speech. Today they are so ubiquitous that the traditional terms are becoming obsolescent in some languages, even when the languages themselves are still spoken. The terms generally had and have the meaning associated with the ‘bifurcate merging’ classification common in most of the traditional languages, that is the reflex of ‘uncle’ (angkul etc) is mother’s brother and ‘aunty’ (anti) is father’s sister. Father’s brother is referred by the term for ‘father’ (including dedi etc. in Kriol) and mother’s sister is ‘mother’ (mami, mama in Pidgin/Kriol and Aboriginal English).

12 However McKelson’s version of FF in Karajarri (2005) is kalityi. If the original form of the root was kali, then one might argue for regressive vowel harmony having operated in kalurtu > kalurt. It seems more likely that ka(r)lu was the original form as O’Grady (nd) gives both karlutyi and karlityi, with the likelihood that regressive vowel assimilation is responsible for the karlityi variant. 13 McGregor (2011: 776) gives this junior reciprocal ‘grandson’ as the first meaning of kapirl, with FM second. Nekes and Worms (1953: 532) give only FM. McGregor (2011: 770) gives tyipi as FM. Since this is not apparently cited by Nekes and Worms, and retains a final vowel where established words have lost theirs, it is likely to be a new loanword (of unclear provenance). Other Nyulnyulan FM words are ng, nasal assimilation, is mentioned by her as regular sound changes in the history of Bularnu (1996: 15) 22 It is possible that the cross-cousin meaning could be derived from ‘father-in-law’ which is often equated to ‘mother’s brother’ in Dravidianate systems. Mother’s Brother’s Child could have been equated to Mother’s Brother by Omaha skewing. There is some evidence of Omaha skewing in the Warluwarric region, for instance in Wakaya (Breen 1970; see also McConvell 2012 for Omaha skewing not far away).

Kinship loanwords in Indigenous Australia, before and after colonization

107

that the Aboriginal English term cousin was used in that same area for ‘motherin-law’ – daughter-in-law’ (Breen 197023). This unusual equivalence had a disproportionate effect on the semantics of the newly arriving Pidgin-Kriol term katyin ‘cousin’ as it spread with the cattle industry to the west.

6 Mama, Papa and the core family 6.1 Versus Goldilocks From Matras’s hypothesis, and from the inverted perspective of Goody on the retention of core nuclear family terms in the Middle Ages as other kin terms colonized the periphery, one might expect to find a similar resistance in Australian Aboriginal languages to adopting core terms, both in language contact between traditional Indigenous languages and between them and the advancing English based varieties. We have seen a couple of types of cases where some consanguineal terms were adopted from other Indigenous neighbouring languages. The examples here were mainly grandparental terms, though, not parents and children. Core terms coming in from either English or traditional Indigenous languages have retained their classificatory extension to siblings. That is to say, for many Aboriginal groups, terms like mother or mum continue to be applied to mother’s sisters and mother’s parallel cousins, long after the traditional language ceased to be spoken and people had moved into urban environments. This extension is of course the other side of the coin of the more English-like use of aunty for FZ not MZ. However here there is often overlap, overlay or ‘plurality’ as the historians of English society call it: the same people may use mum to mean mother’s sister at one time, and strictly only ‘mother’ at another, and ‘aunty’ to mean both MZ and FZ at one time, and strictly only FZ at another. I introduced the term ‘kin-switching’ parallel to code-switching, for this type of

23 This is not in Bularnu but in the nearby Warluwarric language Wakaya. The traditional Indigenous language term is not mangatha or a related term, but limpirnngi, in which -ngi is a feminine suffix. The form with the masculine suffic -ngu means ‘son-in-law’. Spencer and Gillen (1904) give a somewhat different form of the stem for ‘Worgaia’ with a different suffix -ntha probably from a Mirndi language: limpari-ntha. The stem may be related to the term lamparra for ‘father-in-law/son-in-law’ which spread across from the west in the late nineteenth century. The vowel change a > i and subsequent rr > rn require further explanation.

108

Patrick McConvell

variation: this needs to be studied both historically and synchronically, from the viewpoint of social motivations, but to my knowledge has not been, in Australia.

6.2 Baby-talkers, proto-worlders and ‘tjitjas’ We need to avoid some of the types of research agenda that lead us in unproductive directions. Two of these have claimed to offer global explanations for items in the core family kin lexicon, primarily the forms mama and papa. A group of researchers whom I call ‘Baby-talkers’ (Jakobson 1960) think that these ‘nursery words’ are constantly recreated in different languages across the world by natural interactions between caregivers and children. Consequently it can be fruitless to try to seek their history by normal methods of historical linguistics, etymology or reconstruction. ‘Proto-Worlders’ (Bancel et al. 2011) are ostensibly at the opposite extreme, seeking to reconstruct such terms not within the normal bounds of historical linguistics (around 10,000 years) but back to the dawn of humankind and language, at least 100,000 years ago. What both these approaches share is a lack of concern when it comes to such terms for the painstaking task of following the trails of inheritance and of language contact over time – the kind of approach which I have followed in this chapter. As it happens both the terms mama and papa are widespread in Australia but the main exemplars have the opposite meanings from those expected by both the baby-talkers and the proto-worlders: mama usually ‘father’ and papa quite frequently ‘mother’, although other words for ‘mother’ are very common, notably ngama and variants (Hendery and McConvell 2013). All these words can be reconstructed and their paths as loanwords tracked where that is relevant. It is possible that *mama is the proto-Pama-Nyungan word for ‘father’, for instance but further work is needed to establish this solidly. The chapter cited (Hendery and McConvell 2013) also looks into the borrowing of forms like ‘mama’, ‘mum’, ‘mummy’ and ‘papa’, and also ‘daddy’ from English and English-based Pidgins-Creole varieties in Australia. This is necessary in the first instance because these elements can interfere with proper analysis of the traditional Indigenous words of similar forms. One of the findings however is that actually very few of these terms have been thoroughly borrowed into the traditional Indigenous languages. One apparent exception to this is the word for ‘sister’. Tyitya and similar terms are commonly used especially in south-eastern Australia. However while there may be some loans from Pidgin sister (often pronounced tyitya) among these, it is also true that there are many local traditional words in Victoria,

Kinship loanwords in Indigenous Australia, before and after colonization

109

New South Wales and Queensland of the form tyityV which mean ‘sister’ or ‘brother’. The fact that the meaning varies in this way in itself casts doubt on the etymology form English or Pidgin. Also there are many other forms in the region meaning ‘brother’ or ‘sister’ of the form tyatyV. These are quite likely to be related to the tyityV forms by a change of vowel, the details of which have not been fully worked out. In any case this is another cautionary tale, like that of the Baby-talkers and the Proto-Worlders, that we should not jump to conclusions that the solution to the history of the Aboriginal words lies outside our normal parameters. Their history can be and should be worked out with the tried and true methods of historical linguistics, which include careful attention to language contact.

7 Conclusions This paper began by delving into the history of French kinship terms entering English in the medieval period. But entry of a word is not the end of the story. The old words stay on for a time, sometimes a long time, performing a tango with the new ones, the two layers overlapping and overlaying and the meanings of the words changing as they interact with each other. This kind of complex dance is what has happened everywhere including in Aboriginal Australia before colonization, between the old languages, and after colonization when English and the new contact languages established themselves and spread. We cannot know as much of the detail of the earlier stages as perhaps we can in the literate culture of England hundreds of years ago, but much can be accurately inferred from the methods of historical linguistics, and for the recent period and the present we can see the ferment and change before our eyes. Patterns of usage and meanings of kinship terms are transformed, often by language contact. Creativity powers innovations such as the new compound kinship terms using the kin-switching between systems and terms to make new hybrids which keep the old and the new alive while remaking both. The detailed study of contact, overlay and spread of new kinship terms and systems have also illuminated both the history of social and marital arrangements before colonization and those which accompanied colonization, as in the spread of Pidgin-Kriol varieties along with the cattle industry in northern Australia.

110

Patrick McConvell

Abbreviations for kin M F B Z S D W H e y m f

mother father brother sister son daughter wife husband elder [elder brother or sister] younger male propositus (sometimes called ‘male speaking’) female propositus (sometimes called ‘female speaking’)

References Aklif, Gedda. 1999. Ardiyooloon Bardi Ngaanka: One Arm Point Bardi Dictionary. Halls Creek: KLRC. Alpher, Barry. 1991. Yir-Yoront lexicon: Sketch and dictionary of an Australian language. Berlin: Mouton de Gruyter. Arthur, Jay. 1996. Aboriginal English: a Cultural Study. Melbourne: Oxford University Press. Bancel, P.J., A. Matthey de l’Etang, and J. Bengston. 2011. Back to proto-Sapiens Pt. 2: the Global Kinship Terms Mama, Papa and Kaka. In D. Jones and B. Milicic eds. Kinship, Language and Prehistory: Per Hage and the Renaissance of Kinship Studies, 38–45. Salt Lake City: University of Utah Press. Benveniste, Emile. 1973. Indo-European language and society. Coral Gables: University of Miami Press. Translated by Elizabeth Palmer from Vocabulaire des Institutions IndoEuropéennes. 1969. Paris: Les Editions de Minuit. Bowern, Claire. 2004. Bardi verb morphology in historical perspective. Massachusetts: Harvard University. Unpublished PhD dissertation. Bowern, Claire, Epps, P. Grey, R. McConvell, P. et al. 2011. Does Lateral Transmission Obscure Inheritance in Hunter-Gatherer Languages? PLoS ONE (Public Library of Science) 6(9). 1–9. Breen, J. Gavan 1970. Wakaya word list. Document 0047 of the Aboriginal Studies Electronic Data Archive (ASEDA). Australian Institute of Aboriginal and Torres Strait Islander Studies: Canberra. Breen, J. Gavan. 1990. Warluwarra Wordlist. ASEDA document 0253. Canberra: AIATSIS. Breen, J. Gavan. Bularnu vocabulary with Indjilandji words added. In Breen, J.G. 1980. Bularnu phonology and grammar with semantic vocabulary. Friedrich, Paul. 1966. Proto-Indo-European Kinship. Ethnology 5. 1–36. Goody, Jack. 1959. Indo-European Society. Past and Present 16. 88–91. Reprinted in J. Goody ed. 1969. Comparative Studies in Kinship. 235–9. London: Routledge and Kegan Paul. Goody, Jack. 1983. The Development of the Family and Marriage in Europe. Cambridge: Cambridge University Press.

Kinship loanwords in Indigenous Australia, before and after colonization

111

Haynie, Hannah. Bowern, C. McConvell, P. et al. 2014. Wanderwörter in languages of the Americas and Australia. Ampersand 1. 1–18. Hendery, Rachel and Patrick McConvell. 2013. Mama and Papa in Indigenous Australia. In P. McConvell, I. Keen and R. Hendery eds. Kinship Systems: Change and Reconstruction, 217–238. Salt Lake City: University of Utah Press. Hettrich, Heinrich. 1985. Indo-European Kinship Terminology in Linguistics and Anthropology: Anthropological Linguistics 27(4). 453–480. Jakobson, Roman. 1960. Why ‘Mama’ and ‘Papa’? In Selected Writings: Phonological Studies. 536–545. The Hague: Walter de Gruyter. Jones, William. 1990. German Kinship terms: Documentation and Analysis. 750–1500. Berlin: de Gruyter. Kelham, Robert. 1779. A Dictionary of the Old French or Norman language. London: Edward Brooke. Kronenfeld, David. 2009. Fanti Kinship and the Analysis of Kinship Terminologies. Urbana: University of Illinois Press. Laughren, Mary. 1982. Warlpiri structure. In, ed. By Heath, J., Merlan, F. and Rumsey, A. eds. Languages of kinship in Australia, 72–85. Sydney: Oceania Monographs. Lissarague, Amanda. 2007. Dhanggati grammar and dictionary with Dhanggati stories. Nambucca Heads: Muurrbay Aboriginal Language and Culture Co-operative. Macfarlane, Alan. 1978. The Origins of English Individualism: The Family, Property and Social Transition. Oxford: Blackwell. Mallory, J.P. and D.Q. Adams. 1997. Encyclopedia of Indo-European Culture. London: Fitzroy Dearborn. Matras, Yaron. 2009. Language Contact. Cambridge: Cambridge University Press. McConnel, Ursula. 1950. Junior Marriage Systems: Comparative Survey. Oceania 20. 107–143. McConvell, Patrick. 1997. Long-lost relations: Pama-Nyungan and Northern kinship. In P. McConvell and N. Evans eds. Archaeology and Linguistics: Aboriginal Australia in Global Perspective, 207–236. Melbourne: Oxford University Press. McConvell, Patrick. 2007. The PNy kin suffix -li/-lu. http://wiki.pacific-credo.fr/index.php? title=AustKin_Publications. McConvell, Patrick. 2008. Grand-daddy Morphs: the importance of suffixes in reconstructing Pama-Nyungan kinship. In C. Bowern, B. Evans & L. Miceli eds. Morphology and language history: in honour of Harold Koch, 313–328. Amsterdam: Benjamins. McConvell, Patrick. 2009. Loanwords in Gurindji, a Pama-Nyungan language of Australia. chapter and database in M. Haspelmath & U. Tadmor eds. Loanwords in the World’s Languages: A Comparative Handbook, 790–822. Berlin: Mouton de Gruyter. McConvell, Patrick. 2012. Omaha Skewing in Australia: Overlays, Dynamism and Change. In T. Trautmann & P. Whiteley eds. Crow-Omaha: New Light on a Classic Problem of Kinship Analysis, 205–222. Tucson: The University of Arizona Press. McConvell, Patrick. 2013a. Introduction: Kinship Change in Anthropology and Linguistics. In P. McConvell, R. Hendery and I. Keen eds. Kinship Systems: Change and Reconstruction, 1– 18. Salt Lake City: University of Utah Press. McConvell, Patrick. 2013b. Proto-Pama-Nyungan Kinship and the AustKin project: Reconstructing Proto-terms for ‘Mother’s Father’ and their Transformations. In P. McConvell, R. Hendery and I. Keen eds. Kinship Systems: Change and Reconstruction, 192–216. Salt Lake City: University of Utah Press. McConvell, Patrick, Thomas Saunders and Stef Spronck. 2014. Linguistic prehistory of the Australian boab, In Gawne, Lauren and Jill Vaughan (eds.), Selected Papers from the

112

Patrick McConvell

44th Conference of the Australian Linguistic Society, 277–292. Melbourne: University of Melbourne. McConvell, Patrick. 2015. Long-Distance Diffusion of Affinal Kinship Terms as Evidence of Late Holocene Change in Marriage Systems in Aboriginal Australia. In. P. Toner ed. Strings of Connectedness: Essays in Honor of Ian Keen. pp. 287–315. Canberra: ANU Press. McConvell, Patrick. to appear. Enhancing The Kinship Anthropology Of Scheffler With Diachronic Linguistics And Centricity. In W. Shapiro ed. [title & publisher tba]. McConvell, Patrick and Barry Alpher. 2002. On the Omaha Trail in Australia: Tracking Skewing from East to West. Anthropological Forum 12(2). 159–176. McConvell, Patrick and Mike Smith. 2004. ‘Millers and mullers: the archaeolinguistic stratigraphy of seed-grinding in Central Australia’ In H. Andersen ed. Language contacts in prehistory: studies in stratigraphy, 177–200. Amsterdam: Benjamins. McConvell, Patrick and Maia Ponsonnet. 2013. Results and Prospects in the Study of Semantic Change: A Review of From Polysemy to Semantic Change (2008). Journal of Language Contact 6. 180–196. McGregor, William. 2011. The Nyulnyul language of Dampier Land, Western Australia. Canberra: Pacific Linguistics. McKelson, Kevin and Wangka Maya Pilbara Aboriginal Language Centre on behalf of the Karajarri people. 2005. Karajarri – English Dictionary English – Karajarri Finderlist and Topical Wordlist. South Hedland: Wangka Maya Pilbara Aboriginal Language Centre. Nekes, Hermann and Ernest Worms. 1953. Australian Languages. New edition 2006. Ed. William McGregor. Berlin: Mouton de Gruyter. O’Grady, Geoffrey. N.D. Unpublished fieldnotes. MS 321 in AIATSIS catalogue. Powell, Fiona. 2002. Transformations in Gugu Yimithirr kinship terminology. Anthropological Forum 12(2). 177–194. Rangan, Haripriya, Bell KL, Baum DA, Fowler R, McConvell P, et al. 2015. New Genetic and Linguistic Analyses Show Ancient Human Influence on Baobab Evolution and Distribution in Australia. PLoS ONE 10(4). e0119758. Reay, Marie. 1949. Native Thought in Rural New South Wales. Oceania 20. 89–118. Ringe, Donald. 2006. From Proto-Indo-European to proto-Germanic: a linguistic history of English. Oxford: Oxford University Press. Roth, W.E. (1948[1897]). Ethnological Studies Among the North-West-Central Queensland Aborigines. Brisbane: Edmund Gregory, Government Printer, William St. Rumsey, Alan. 1983. Kinship and Context among the Ngarinyin. Oceania 51. 181–192. Scheffler, Harold. 1978. Australian Kin Classification. Cambridge: Cambridge University Press. Sharp, Janet. 2004. Nyangumarta: A Language of the Pilbara Region of Western Australia. Australian National University, Canberra: Pacific Linguistics. Spencer, Baldwin and Frank Gillen. 1904. The Northern Tribes of Central Australia. London: Macmillan. Tadmor, Naomi. 2004. Family and Friends in Eighteenth-Century England: Household, Kinship and Patronage. Cambridge: Cambridge University. Troy, Jakelin F. 1994. The Sydney Language. Canberra: the author/Australian Dictionaries Project. AIATSIS. Tsunoda, T. 2003. A Provisional Warrungu Dictionary. Tokyo: Department of Asian and Pacific Linguistics, Institute of Cross-Cultural Studies. Wrigley, Matthew. 1992. Jaru Dictionary: Draft Edition, June 1992. Halls Creek: Kimberley Language Resource Centre.

David Nash

4 Placenames evidence for NSW Pidgin1,2 Abstract: NSW Pidgin is documented from fragmentary source material, and evidence is sparse for its probable nineteenth century spread through Queensland to northern Australia. I adduce another source of data bearing on NSW Pidgin’s formation and spread, that found in placenames. I argue that NSW, and northern Australia too, there are ‘Pidgin placenames’: ones that are neither purely of Indigenous origin, nor entirely imposed by the incoming colonists and settlers. The study considers a NSW Pidgin landform word widely incorporated into placenames (balga ‘hill’), the NSW Pidgin words gabun ‘big’, ngarang ‘little’, budjari and merrijig ‘good’, yarraman ‘horse’, piccaninny ‘child’, blackgin and lubra ‘Aboriginal woman’. The study also covers some words derived from English, such as blackfellow ‘Aboriginal person’ and sugarbag ‘native honey’. Unlike a lot of introduced placenames, the Pidgin placenames were likely formed by local residents, and show a mix of the formal and semantic patterns of Indigenous versus introduced placenames. The study shows how the limited corpus of NSW Pidgin can be somewhat augmented, and throws a little more light on the geographic and temporal extent of NSW Pidgin.

1 Introduction The pattern in the distribution some official Australian placenames provides a perspective on the propagation of nineteenth century Pidgin, from its entry at 1 This study relates two long held interests of Patrick McConvell's: contact languages, and how languages preserve traces of layers of (pre)history. More particularly, Patrick has shown how placenames can illuminate these domains: McConvell 2002 considered contrasts between Aboriginal and European ways of assigning placenames, and McConvell 2009: 387–390 showed how placename analysis can contribute 'evidence of previous occupation'. 2 The initial version of this paper received (jointly) the NSW Geographic Names Board’s Murray Chapman Award for 2013. An earlier version was presented to the 12th Australian Languages Workshop, 8–10 March 2013, The University of Queensland’s Moreton Bay Research Station, North Stradbroke Island. I am grateful for comments there, for anonymous reviewers’ comments, and for further discussion of the topic with Denise Angelo, Harold Koch, Jen Munro, Petter Næssan, Jeff Siegel, and Jane Simpson. I have made appreciative use of Steele’s (2005) lexical database for NSW. Figures 1 and 6 are reproduced with permission. The maps have been made with QGIS 1.8.0, and Natural Earth (free vector and raster map data from naturalearthdata.com).

114

David Nash

Sydney the capital of the first state, New South Wales (NSW). This paper investigates some limited aspects of the spread of NSW Pidgin through placenames presumably formed within it. After setting out the scope of the study, and the difficulties with the available data, I consider first words which did not spread beyond NSW, and then words which did spread through Queensland and to northern Australia. I conclude with less numerous examples found in other parts of the mainland. The dating of the spread of NSW Pidgin is not addressed, because the date of bestowal of the relevant placenames has not been collated and often proves to be unobtainable in any detail. Also the study does not cover other interlanguage placenames, such as conscious portmanteau combinations, ‘blended loans’, or coinages during language reclamation in recent years.

2 Pidgin placenames The starting point for distinguishing types of modern Australian placenames is the framework established by Hercus & Simpson (2002: 1): In Australia we have two sets of placenames, one superimposed over the other. These are the set of networks of placenames that Indigenous Australians developed to refer to places (the Indigenous placename networks), and the set of placenames that Europeans developed to refer to places (the introduced placename system). Table 1: Australian placename sets after European Contact (primary sets shaded) L1:

Indigenous Australian

European: English etc.

Lexifier Australian English etc.

The cells of Table 1 on the main diagonal represent the two main sets identified by Hercus & Simpson (2002: 1). Speakers of Australian languages typically named places according to various patterns which contrast in a number of ways with those of English and other European languages. The two systems of placenames differ in a number of ways, for example, whether the placenames form networks, how they act as mnemonics, what uses are made of the land, and what counts as a significant feature. (Hercus & Simpson 2002: 10)

Placenames evidence for NSW Pidgin

115

With respect to the third aspect, as the authors put it, In fact, it is exceedingly difficult to find Indigenous placenames that make explicit reference to habitation. That is, we do not find names of the form ‘X Camp’ (Hercus & Simpson 2002: 13)

In Australia the Indigenous placenames contrast formally with those of colonial origin. External modification with a descriptive adjective, compounding, commemoration of a distant person or other entity, or incorporation of a landform generic in the placename are patterns rarely found in Australian Indigenous placenames, while being commonplace in the placenames introduced with European settlement (Hercus & Simpson 2002). These contrasting tendencies can be brought to bear on the study of those placenames that are candidates for having arisen where an Indigenous language has been in contact with English. In contrast, the introduced placenames commonly show other strategies of coinage: commemoration (of a named person or a placename from Europe), topographic descriptors (such as Mount Remarkable), and relative location (East Arm) Hercus & Simpson (2002: 14–18). This last European-style strategy uses a modifier (such as a cardinal direction, or relative elevation or age or size) to relate two nearby distinct places. This pattern is commonplace with English placenames (especially in England, e.g. Great Haseley and Little Haseley in Oxfordshire) but is rarely found with Australian Indigenous placenames. For example, the use of an external modifier indicating size is to be contrasted with what can be called the internal diminutive, as occurs in the -dool placenames of central north NSW (Nash 2014a). The situation of Table 1 increased in complexity as soon as the English colonists adopted placenames from an Aboriginal language, populating the upper right cell of Table 1, that is, representing placenames borrowed from an Australian language into colonial English. The remaining cell (bottom left) represents imported (European) placenames some of which would have been adopted by Aboriginal people and thus borrowed into their languages.3 However, none of these four sets is the subject of the present study. Rather, I turn attention to

3 The earliest borrowings we are now aware of from outside Australia into an Australian language would have been Macassan placenames (MacKnight 1976) borrowed into Top End languages. The first borrowing from a European language would have been after the establishment of the English colony at Port Jackson in 1788, where the first borrowed placename may well have been the very name England, as witness the Sydney Language reply P. & W. Englánda ‘In England.’ recorded by Dawes (Dawes & Anonymous 2009: Notebook B, 33) (compare Berıwȧl ‘England’ (Dawes & Anonymous 2009: Notebook B, 4)) from the sense ‘a great distance off’ (Dawes & Anonymous 2009: Notebook C, 9).

116

David Nash

another possible source of placenames in nineteenth century NSW, namely the kind of language used between the NSW Aboriginal people and the incoming colonists. This variety of talk, not in itself a full language, has been termed NSW Pidgin, and ‘was the vehicle by which many words from NSW Aboriginal languages entered Australian English . . .’ (Harkins 2008: 405). I draw on the descriptions of NSW Pidgin by Troy (1993, 1990, 1994) and Amery & Mühlhäusler (1996). A vocabulary extracted from Troy (1994) by Amanda Lissarrague (Harkins 2008: 405) has been published by Wafer et al. (2008). Once NSW Pidgin arose, first around Sydney and other early European settlements, the possibilities for placename origin widened from Table 1 to those of Table 2. Table 2: Australian placename sets with Pidgin placename sets shaded light grey (primary sets dark shaded) L1:

Indigenous Australian

European: English etc.

Lexifier Australian Pidgin English etc.

This study focusses on placenames which likely arose in the common way of talking used by Aborigines and the English-speaking colonists to communicate with each other. There is little previous literature on placenames as part of pidgin (or creole) lexicons (just J. Nash 2014b and two references there cited). Pidgin placenames are likely to have been co-constructed and arose in interaction between speakers of an Australian language and speakers of the colonial English. We can expect that the co-construction would have been mediated partly by NSW Pidgin in regions where it was the lingua franca.4

2.1 Example of origin A rare glimpse of how some placenames would have arisen in the nineteenth century pastoral context is afforded by this reminiscence of Surveyor A. H. Chesterman. 4 Such a scenario has been argued for as the likely origin of the English word budgerigar (Nash 2013).

Placenames evidence for NSW Pidgin

117

As showing how the aboriginals occasionally coined words the name of an old outstation on the Lachlan can be cited. In the very early days the hutkeeper there had a wooden leg – ‘Waddy’ being the aboriginal word for ‘stick’ the blacks called the place ‘Waddy-man’, a combination of an aboriginal word and an English word. This name was subsequently corrupted into ‘Waddy-mandow’. Strangely enough, some years after I first heard of the above, a very old aboriginal of the Upper Murray (now, I believe, dead), in supplying me with about two hundred native words for simple and ordinary terms, upon being asked what a blackfellow would call a wooden leg replied ‘Waddy-mundoh’. – A. H. Chesterman’s letter, Tumut, 6 January 1900, pp. 1–2 (Royal Anthropological Society of Australia 2004: 420–5)

Clearly the expression compounded NSW Pidgin mandowi ‘foot’ (Wafer et al. 2008: 813) and the even more common waddy (wadi ‘piece of wood’ adopted into Australian English). Chesterman’s version ‘Waddy-man’ would have been a shortening of ‘Waddy-mandow’ (rather than the reverse). Both words wadi and mandowi originated in the Sydney Language, but the compound apparently arose in the colonial period.5 The origin of Waddy Mandoe Crossing (28°49′S 151°02′E) and Waddy Mundoey Creek (28°47′S 151°03′E) (just over the Queensland border, hundreds of kilometres north of the Lachlan River) presumably had something in common with Chesterman’s account: ‘Reportedly an Aboriginal term, language and dialect not recorded, given to describe a European man with a wooden leg. Information provided by H. Bracker (–) pastoralist?.’6 The people who were bestowing placenames across Australia in the nineteenth century were typically English speakers whose work took them to the areas less well-known to the colonists. Whether they were explorers, prospectors, pastoral workers, teamsters or surveyors, they engaged in various ways with local Aboriginal people and had the opportunity to more or less jointly arrive at a name for a place of mutual interest. This option appears not have been taken as often as the alternatives of adopting an Indigenous placename, or imposing a name from English, but it is nevertheless the focus of this study.

2.2 Some difficulties with the data Research on placenames bestowed in colonial Australia has to contend with scarcity and unreliability of information available about the origin of the names. 5 I use these conventions: bold for reconstituted spellings and italics for quoted spellings and for modern spellings. The cover symbol N is used for an indeterminate nasal consonant. 6 The State of Queensland (Department of Natural Resources and Mines) httpsː//www.dnrm. qld.gov.au/land/place-names/search/queensland-place-names-search

118

David Nash

This even applies to introduced names where history records some of the circumstances of bestowal (such as explorers’ commemorative names), and more so to Indigenous names adopted by the English-speaking settlers. For the latter type there is usually some uncertainty about the original reference of the placename, and about its phonetic and phonological form (usually transmitted through an ad hoc English-based spelling); and it is uncommon for any constituent morphology to have been recorded. A complicating factor is that some placenames have been assigned as a copy of a name from a distant location. A placename putatively of pidgin origin (because it incorporates a word clearly from, say, NSW Pidgin) may actually have been assigned solely by standard English speakers, and origin information is usually no longer available to help us distinguish these instances. Despite this, there is a noticeable coherence to be seen in the overall geographic patterns in the occurrence of the placenames considered below, and so, assuming copy names are in the minority, we can have confidence in spatial and other generalisations about them. Another aspect of variation in the data is more readily dealt with. As is common in a populous built environment, modern Australian placenames often share the same base name (usually in the same spelling) in the one locality. For instance, the city of Brisbane on the Brisbane River which rises in the Brisbane Range; the name has proliferated in the modern city and applies to railway stations, airports, a hospital, and so on. These placenames can be usefully grouped into ‘toponymic sets’: placenames which involve the same base name (usually in the same spelling) in the one locality (Nash 2014a: 39–40). Typically the places in such a set are differentiated by feature type, and/or by various derivations of the base name. From each toponymic set one representative has been selected for mention in this study; it has been chosen as much as possible to the earliest assignment of the placename. In the absence of definite historical information on which to base the choice, names referring to natural features are preferred to those referring to manmade features), so that, for example, for Boree Creek the creek would be chosen from the various other designations sharing the same name (Parish, Town Railway Station and School).

2.3 Scope of this study The study began by matching some distinctive words of nineteenth century NSW Pidgin (in their conventional English spelling) with placenames in the official Australian gazetteer (Geoscience Australia 2010) and in the Geographical Names Register (GNR)(Geographical Names Board 2011). English-origin words in NSW

Placenames evidence for NSW Pidgin

119

Pidgin also may well have been bestowed in a placename, but it would be more difficult, given the sparse nature of the records, to distinguish these from placenames derived directly from English.7 Even so, it is likely that many placenames involving, say, Blackfellow, arose in the context of NSW Pidgin. There are types of interlanguage placename which are outside the scope of this study. One type is placenames in Australia which, as far as we know, arose as conscious portmanteau combinations. These are few in number; examples where the two words were from the one language, or two Indigenous languages, include nineteenth century Gungaderra (Gunghalin and Gininderra) in southern NSW (now ACT, ACT Planning and Land Authority (ACTPLA) 2013), and, outside NSW, twentieth century Yankaponga (SA) and Warrabri (NT) (Hercus & Simpson 2002: 5–6). Garawarra, the name of a National Park south of Sydney, is said to be ‘a combination of Gara (Garie) and Illawarra’,8 a coinage which has been claimed since its inception.9 Where an Indigenous stem has been combined with an ending from an introduced language, names were coined such as Glenunga (SA) (Hercus & Simpson 2002: 5–6), Bundaberg (Queensland, said to be the section term Bunda combined with the German berg ‘mountain’, though possibly after Bunda was already borrowed in to the local English). Other related types beyond the scope of this study are what might be termed ‘blended loans’ (such as NT Mongrel Downs < Mangkururrpa, Blue Sunday < Puluthanti, and other examples discussed by Koch (2009: 126)), or calqued placenames. A NSW placename of this kind was Warangesda (Elphick & Elphick 2004), coined by the Rev. Gribble and his followers around 1880 when they established near Darlington Point on the Murrumbidgee a Mission with this name, which is a partial calque from Biblical Hebrew beth hesda ‘house of mercy’ with Wiradjuri warang ‘camp’ substituted for beth. Also beyond the scope of this study are modern coinages in the context of language reclamation, notably the Adelaide placenames discussed in detail by Amery & Williams (2009 [2002]).

7 One example could be the southeastern Queensland city Toowoomba if its name was indeed an Aboriginal pronunciation of ‘The Swamp’ (‘The naming of Toowoomba’, http://www. toowoombarc.qld.gov.au/facilities-and-recreation/libraries/local-history-library/7871-thenaming-of-toowoomba). 8 The Bush Club Newsletter, Summer 2012, p. 4 http://www.bushclub.org.au/down/Newsletters/ Summer_2012_Newsletter.pdf 9 J. V. T[urner], The Sydney Bush Walker Annual April 1934, p. 8 http://www.bushwalkermagazine. org/the-annuals/1934-SBW-annual-s.pdf (PDF page 13)

120

David Nash

3 NSW examples 3.1 Cabonne and Nyrang placenames Two well attested distinctive words of NSW Pidgin are gabun ‘big’ (Koch 2011: 507n26) and ngarang ‘small’, both derived from the Sydney Language (Troy 1994: 708–789). These are found in only a few NSW placenames, concentrated in the central west of the state, as can be seen in Figure 1. – Boree Cabonne, Boree Nyrang: adjacent homesteads, from the mid-nineteenth century on Boree Creek; compare the terms recorded for two similar tree species Boree ‘Big Myall’ and Boree Nyrang ‘Yarran’ (Royal Anthropological Society of Australia 2004: PDF p203), burri ‘boree tree, or weeping myall, Acacia pendula’ (Grant & Rudder 2010, Richards 1902–03: 102) – Gamboola Cabonne, Gamboola Nyrang: adjacent nineteenth century homesteads)10 – Cabonne has also been adopted, in the twentieth century, as the name of a Local Government Area and Shire, based at Molong (and containing Boree Cabonne and Gamboola Cabonne) – Molong Nyrang, previously near Molong (Molong Historical Society 2013) – Eurow Nyrang Mountain (compare Mount Eurow ~ Eurow Mountains nearby) – Nerang Cowal, a lake near the larger Lake Cowal – Nyrang Creek (33°31′S 148°37′E), and another Nyrang Creek in the Clarence Valley (29°25′S 153°14′E) – Wowong Narang Lagoon (previously Wowang Warang Lagoon, 34°34′S 145° 52′E), near Wowong Lagoon (34°33′S 145°51′E) in the Parish of Wowong, near Carrathool – Nyrangi Tuppal Creek ~ Nerangi Tuppal Creek (35°44′S 145°16′E, Native Dog Creek, previously Nyrangi Creek ~ Nerangi Creek) (compare Tuppal Creek, 35°45′S 145°30′E) Note that in these names (except for Nyrang Creek), the sense of ‘big’ and ‘small’ provides what can be called an external modification. Nyrang Creek shows a pattern (literally ‘little creek’) not otherwise found with Australian Indigenous placenames, as mentioned above in the Introduction.

10 Gamboola Nyrang had been part of Gamboola Cabonne, previously Tea-Tree Creek Station (Molong Historical Society 2013). J.P.M. Long suggested (p.c., 24 August 2005) that the Gamboola pair of names were a later imitation of the Boree ones; all the properties were at times owned by the Smith family (Mac. Smith 1976).

Placenames evidence for NSW Pidgin

121

Figure 1: Distribution of Cabonne and Nyrang etc placenames

On the other hand, the head-modifier order within the Cabonne and Nyrang placenames is the opposite of the usual English sequence, while being a normal pattern for NSW languages. Thus the placenames involving Cabonne and Nyrang have the hallmark of having been formed by speakers of NSW Pidgin, combining properties drawn from the first languages of both settlers and Aborigines. Thus, unlike placenames imposed by explorers, the Pidgin placenames were likely formed by local residents, and show a mix of the formal and semantic patterns of Indigenous versus introduced placenames. Pidgin placenames variously show head-modifier or modifier-head order, from competing directions of functional transfer (Siegel 2012). As it happens, we have a contemporary observation about the two possible orders: A nice little point that would have afforded scope for a very pretty display of oratorical abilities, was raised in the Supreme Court on Thursday. A man was indicted for setting fire to a stack of wheat at Boree Nyrang, when it appeared in evidence that the name of the place was Nyrang Boree. The Chief Justice said it was a pity for the Attorney General to tie himself up by local description; the Court had extensive jurisdiction, and it would

122

David Nash

have been sufficient to lay it in New South Wales. A gentleman having suggested to Mr Attorney General that Nyrang was the native word for little, and that it was only called Nyrang to distinguish it from another Boree, the learned gentleman contended that Boree Little and Little Boree were synonymous, and that it did not matter whether it was called one or the other. The Judge, however, was not satisfied, and said, that he might as well be called Dowling James, instead of James Dowling. Mr Attorney-General responded, by saying, that it would not matter whether he said James Dowling Chief Justice, or Chief Justice James Dowling. This view of the case seemed to be a poser to His Honor, who said the case had better go on, and as the prisoner was acquitted on the merits, the point of course fell to the ground. (‘Jargon’, Sydney Herald, 8 May 1837, page 2 http://nla.gov.au/ nla.news-article12862532)

3.2 Bulga placenames The word balga, bulga ‘hill’ is listed as NSW Pidgin by Wafer et al. (2008: 814); it was first recorded c. 1824 as bōlgăr ~ bōlgărăr ‘hill’ by the Rev. LE Threlkeld (1824: 132) in his ‘Karee’ vocabulary from just north of Sydney, and then about 1834 in the Sydney Language as balgah ‘a big hill’ and narrang balgah ‘a little hill’11; also as bulga ‘hill’ (also in bulga-gíli ‘humpback’ with gíli ‘back’) in a vocabulary mostly of the Sydney Language by Rowley in Ridley (1878: 259). Figure 2 shows the distribution of the places formed with bulga. Just as with gabun ‘big’, ngarang ‘small’, the distribution of these bulga placenames shows that balga did not spread north into Queensland or beyond.12 The instance Bulgaback in central Victoria was effectively identified as pidgin in origin by the anthropologist Howitt: Bulgurback Creek – Crung-grurk – This creek is shown on the maps between Castleburn and Cobbannah. The word Bulgurback is corrupt, being part native and part English. Bulgur, mountain back – behind – that is, behind the mountain. (Smyth 1878: 189)

Bulga still occurs in a few official placenames as a generic or feature type. The clearest example is Oaky Bulga (in the Holbrook district). Others are The Bulga, Black Bulga, and Sappa Bulga (see below), and we can surmise from the name Cooba Bulga Stream (in theUpper Hunter) there there was once an associated *Cooba Bulga (possibly involving the Hunter River Language kuparr ‘red ochre’, Lissarrague 2006: 119). An unclear case is Bolga Parish in northern Victoria (‘Bolga Native for “high hill.”’ Martin 1944: 15) with a Mount Bolga. 11 The Rev. Richard Taylor, Notebook 1833–1835, page 150, 77-166-4/2, Alexander Turnbull Library, Wellington, NZ. 12 The occurrences in Victoria, and in names such as Murrumbulga, may be irrelevant. The three outliers, all bores called Bulga Well in inland SA (30°41′S 134°21′S) and WA (27°40′S 121° 43′E, 23°01′S 115°21′E) are probably an accidental similarity.

Placenames evidence for NSW Pidgin

123

Figure 2: Placenames involving bulga

The salience of Bulga as a feature type has diminished since the nineteenth century. Bulga Mountain and Black Bulga Mountain are versions of the name of a volcanic plug in the eastern Warrumbungles in north central NSW (31°13′S 149° 41′E, Figure 3) which has also been known as The Bulga and Black Bulga. The existence of the alternate distinctive name Moolagundi13 supports the notion that Bulga was not its original name. Where placenames with bulga included another term, note that it was often a descriptive word (such as Oaky, Black). This would fit with these placenames having arisen as semi-generic or ‘intermediate’ names, along the lines of Red Gum Creek as categorised by Hercus (2002: 69). Hence it seems that bulga was set at one stage to be spread into general Australian English as a landform term, but this did not eventuate. Actually billabong is the about the only landform term from an Australian language borrowed into Australian English, and it figures in many introduced placenames; others partially adopted are cowal and gibber. Such words enter into placenames which may well have been formed within Australian English (that is, not in the context of direct contact with an Australian language) and have lost any trace of pidgin origin. 13 ‘Questions answered’, Australian Town and Country Journal 7 Nov 1906, 23 http://nla.gov.au/ nla.news-article71542640

124

David Nash

Figure 3: Looking south to The Bulga ~ Black Bulga ~ Bulga Mountain ~ Black Bulga Mountain ~ Moolagundi (photo: © 2013 David Nash)

3.2.1 Sappa Bulga The placename Sappa Bulga in central NSW is curious because of the obscurity of the word Sappa. The earliest record I have found of it is the label Sappa Bulgas on the range southwest of Dubbo, at the western extent of a hand-drawn geology map (Stutchbury 1852b), and two mentions in the related newspaper account: The northern edge of the granite trends to the westward as far as, and including the great range denominated the Sappa Bulgas, and Harvey’s Range, to the south-west. . . . Descending the mountain, the highest and principal point of the Sappas (Nambajong) to the plains, at a distance of five miles, the sandstone is again reached, . . . (Stutchbury 1852a)

(The variant Sappa Bulgas appears to be short for Sappa Bulga Ranges, and Bulga Ranges is another recorded variant.) There are exotic occurrences of the name Sappa: a letter name from the ancient Semitic language Ge’ez, a species of tree of Northern Jamaica, a former Ethiopian capital city, and a Township across Nebraska and Kansas (USA). The last of these is a colonial placename in an English-speaking area, and as such invited further investigation. However it turns out that Sappa Township is named after the nearby Sappa Creek which was so-named by 1866 (Pattison 1934: 89), using a word for ‘beaver’ from the local language (Rogers 1967: 11).

Placenames evidence for NSW Pidgin

125

Since we do not know the circumstances of the naming of Sappa Bulga, we can hardly choose any one of those possible origins. There is another closer possibility however, that the initial consonant was written as the alveolar fricative s but represented a pronunciation of the interdental (laminodental) stop dh in the local Aboriginal language. This leads us to the word which became the name of the nearby town of Dubbo, “a Wiradhuri Aboriginal word meaning ‘cap’ or ‘head covering’. It refers to the red clay cap worn by Wiradhuri women when in mourning.” (Appleton & Appleton 1992: 92). Grant & Rudder (2010) reconstitute the word as ‘dhabu a head covering, a cap, a net, a ceremonial helmet, a roof’, with the first vowel a matching the modern pronunciation of the placename Dubbo. The relevant early records include: – Dabburang ‘pipe-clay’ (Günther 1892: 78) – Dubo ‘a net cap’ (Günther 1892: 81) – Dthoo’ boo ‘a helmet (ceremonial); cover (on head)’ (Richards 1902–03: 180) from which the first vowel of the ‘cap’ word should be reconstituted as u rather than a, thus dhubu. The discrepancy in the vowel counts against dhubu as the source for Sappa, but a connection with Dabburang ‘pipe-clay’ (Günther 1892: 78) is a possibility. Nevertheless Sappa Bulga does appear to contain Bulga ‘mountain’.

3.3 Gullen placenames A few words from Australian languages denoting topographic features have been sufficiently adopted by Australian English as to recur as feature types in placenames. The officially recognised ones relevant to NSW are Billabong, Cowal (Nash 2008), and Warrambool (Nash 2011). There could be a few more generic feature terms from NSW languages discernible in recorded placenames of NSW, but which did not become more widely adopted. Besides Bulga, discussed above, a possible case is Gullen, where it can be related to NSW Pidgin kalin ‘water’ (Wafer et al. 2008: 813), in turn from the Wiradjuri and Wangaaybuwan word galiN ‘water’ (Austin 1997: 30).14 There is a cluster of placenames involving Gullen, most in the Yass district, as shown on Figure 4. Two of these were associated with a water–related meaning in the nineteenth century: 14 Note also the record Bean Collen ‘Plenty of water’ [Dubbo Police] (Royal Anthropological Society of Australia 2004: 418) (compare Wiradjuri biyang ‘many, much, all’ Grant & Rudder 2010, Biang ‘many’ Günther 1892: 73). It is unclear whether this was a placename; it could just show that the composite expression was used in NSW Pidgin.

126

David Nash

Figure 4: Gullen placenames

–

– –

Grabben Gullen (“Known to the locals as ‘Grabby’, the name Grabben Gullen comes from a Wiradjuri Aboriginal word meaning ‘small water’.” http:// strlhistorymatters.blogspot.com.au/2009/01/grabben-gullen.html) I have not been able to elucidate the Grabben element. Gullen Ridge previously The Gullen Range (34°34′S 149°27′E), said to be from ‘swamp’ (McCarthy 1963) Narrangullen Creek (35°00′S 148°42′E, southwest of Yass), possibly involving ngarang ‘small’ (as in section 3.1).

Further possible examples and discussion are in the Appendix.

3.4 Other NSW examples Two common words in NSW Pidgin were baal ‘no, negative’ (from the Sydney Language) and, later, gammon ‘nonsense; pretence; ‘humbug’ ’ (from British criminal cant) (AND), though neither is listed by Wafer et al. (2008: 812–815). The latter continues in north Australian Kriol as geman ‘pretend, fake; falsely, lyingly, supposedly’ (Lee 2014). Gammon Creek (31°49′S 150°26′E) is of unknown

Placenames evidence for NSW Pidgin

127

origin but could reflect the same word.15 The combination of the two was also a common expression, and was said to have become the placename Beargamil (33°01′S 148°17′E), though the changes involving l are not explained: Their stadium was at the Waterfall, and Bale Gammon Springs was their corroberee ground (now called Beargamil). (‘Bunbury and the Billabong. In 1860’ Western Champion (Parkes), Thursday 21 December 1916, page 22 http://nla.gov.au/nla.news-article112321812)

4 Beyond NSW NSW Pidgin arose in the period when NSW had a wider reference – the term New South Wales originally encompassed the eastern two-thirds of the Australian continent, and was reduced to its modern extent in the mid-nineteenth century. The study is extended in this section beyond the modern borders of NSW, along the spread of NSW Pidgin north through Queensland and on to the Top End (northern part of the Northern Territory) and the Kimberley region of Western Australia. The general route was summarised by Munro (2000) in her map (reproduced as Figure 5), which for Queensland focusses on an inland route. Other research proposed that inland Queensland was reached as offshoots from a main spread north along the Queensland coast (east of the Great Dividing Range) from the 1840s to the 1870s, as shown on ‘Post-contact languages of Queensland 1800 to present’ (Wurm et al. 1996: Map 8). This latter view is supported by the distribution of placenames, which can be seen on the maps in this section to predominate along the Queensland coastal belt. The section begins with budjari and merijig, two words for ‘good’ attested also in what became Victoria and southern Queensland. Then we look at the spread of yarraman ‘horse’ across Queensland. Next we consider examples which extend further to northern Australia: first, placenames incorporating lubra and black gin (two words for ‘Aboriginal woman’) and blackfellow ‘Aborigine’, and then two examples which are also extend to northern Australia but are absent in inland southern Queensland: piccaninny ‘child’, sugarbag ‘native honey’. Most of these (but not budjari, merijig, or black gin) are in use in Kriol in north Australia (as yarraman, lubra, blekbala, biginini ~ pikanini, shugabeg, respectively) (Lee 2014).

15 Alternatively, the name could be like the Gammon Ranges in SA, and associated Gammon Hill and Gammon Creek, which were possibly named in “relation to gammon of bacon – Gammon Hill has a banded appearance not unlike a cut of bacon”. http://maps.sa.gov.au/plb/

128

David Nash

Figure 5: Path of pastoral industry and NSW Pidgin (Munro 2000: 260, Map 2)

4.1 Budgery and Merrijig The two main words meaning ‘good’ in NSW Pidgin were budjari (from the Sydney Language) and merijig ‘good’ (from Wathawurrung, the language of the Geelong-Ballarat area in Victoria) (Wafer et al. 2008: 814). Placenames with Budgery, Boodjerie and similar spellings are found across NSW and southern Queensland, and one each in Victoria and far southeast SA, with two outliers in rural WA; see Figure 6. Note Budgerydickeys Springs16 based on a man’s Pidgin name, incorporating budjari as a prenominal modifier. Similarly for Budgeree in Gippsland (Victoria), “Also note there was an eastern Kulin man named ‘Budgerry Tom’ in the 1840s” (Clark & Heydon 2002). Placenames with merrijig (apparently the only spelling of merijig in placenames) turn up in three locations in Victoria, and one in 16 Also spelled Budgery Dick Springs, Sydney Morning Herald 7 Sep 1888, page 9 http://nla.gov. au/nla.news-article13695953

Placenames evidence for NSW Pidgin

129

Figure 6: Placenames with budgery etc. and merrijig. Expanded in the east to show detail

central Queensland, also shown in Figure 6. Neither word made it to north Australian Kriol. The distribution of placenames with these two words conforms to the general picture of the spread of NSW Pidgin by the mid-nineteenth century (Figure 5).

4.2 Yarraman The word yarraman ‘horse’ has been recorded in vocabularies of languages across the eastern Australian mainland, and into northeast NT, as we know from Walsh’s (1989) comprehensive survey, as summarised on his map reproduced in Figure 7. The word originated near Sydney (AND) and spread with NSW Pidgin. As we would expect, the distribution of placenames with yarraman (shown in Figure 8) conforms largely to the areas where the word has been recorded in a vocabulary. The main difference is that there are no yarraman placenames in the NT, despite the word being used in a number of languages there including Barunga and Ngukurr Kriol (Lee 2014): apparently the usage to coin placenames with yarraman ceased before the spread of NSW Pidgin beyond Queensland. By contrast, only one placename appears to involve nanto ‘horse’: Nanto Waterhole (28°25′S 136°30′E) on the west of Lake Eyre.

130

David Nash

Figure 7: Distribution of yarraman and nantu ‘horse’ (Walsh 1989: 508)

The good fit between the distribution of placenames with yarraman (Figure 8) and the vocabulary records (Figure 7) adds confidence for the applicability of placename evidence in revealing contact language history.

4.3 Blackfellow The word blackfellow ‘Aborigine’ originated in NSW Pidgin but has long been used more widely in Australian English (AND), and so placenames involving blackfellow (in various spellings) may not have been coined in the context of a pidgin. Nevertheless it is interesting note their distribution: of the 152 places shown in Figure 9, none are in Arnhem Land, and only one in WA south of the Kimberley. Blackfellow contains the nominal suffix written -fela in the Pidgin and Kriol orthography, and has been studied by Baker (1996) and Koch (2011). Parallel to

Placenames evidence for NSW Pidgin

131

Figure 8: Distribution of placenames with yarraman

the word English word black, in NSW Pidgin blackfellow has two possible readings: as a noun (denoting a person) and as an adjective (denoting the colour). However, very few other nominals in -fela occur in Australian placenames. The clear examples are Longfella Pass in the NSW Shoalhaven,17 in a couple of places in Victoria (Wildfellow and Poor Fellow mentioned in section 5.1), and a 17 Longfella Pass 35°21′S 150°15′E and Longfella Ridge 35°21′S 150°14′E; ‘Pass in use by stockman and earliest record Townsend 1848. Also access for Aborigines en-route to hinterland. Longstanding name and in use by bushwalkers c 1940’ (Geographical Names Board 2011).

132

David Nash

Figure 9: Distribution of placenames with blackfellow (abbreviated Bf in labels), black fellow, blackfella, black fella

pair on the Northern Territory coast near Mandorah: One Falla Creek and Two Fella Creek. There is a tributary of the Robinson River in the NT Gulf district is Koolfella Creek of unknown origin, also written Coolafella Creek in 1933 court documents (Finnane & Paisley 2010: 28). So, adjectives in -fela are for some reason quite rare in recorded placenames. Hence we can assume that most if not all placenames with blackfellow most likely involve the noun not the adjective; that is, a name such as Blackfellow Swamp might have been meant as ‘black swamp’ but most likely was ‘swamp associated with blackfellow(s)’. The distribution of placenames with blackfellow fits with the conclusion of Baker (1996: 535): The NSW innovations are blackfellow, whitefellow and that fellow. These spread northwards into QLD and clockwise to Victoria, South Australia and Western Australia.

except for the absence of evidence of spread from SA to WA.

Placenames evidence for NSW Pidgin

133

Figure 10: Placenames with black gin ~ blackgin, and lubra

4.4 Blackgin and Lubra The word gin denoted a woman, from dyiin ‘woman’ in the Sydney Language. It was used in the earliest NSW Pidgin in combinations white gin and black gin; only in later times has it acquired pejorative connotations. The latter combination, sometimes written as one word, occurs in placenames along the main spread of NSW Pidgin in NSW and Queensland. There are also a few occurrences in placenames in the Victoria River District of the NT (but not in the Gulf or Top End), and rural WA: see Figure 10. Note the absence of the word in central Australia, SA, Victoria, and Tasmania. A synonymous word was lubra, thought to be from the language of southeast Tasmania, and recorded in English contexts from 1830 (AND). As well as the official placenames scattered across mainland Australia, there is an unofficial Lubra Creek on Kangaroo Island (SA) of which Taylor (2002: 91–100) presents the origin stories.18 Placenames with lubra are absent from NSW and Tasmania, and while the word was used in contact between English and Aboriginal people in the nineteenth century, it also 18 Thanks to Joshua Nash for this reference.

134

David Nash

Figure 11: Placenames with piccaninny, piccaninnie, picaninny

became part of colonial English and we cannot be confident that placenames in lubra arose in the situation of a contact language. The same caveat would apply to some extent to placenames with black gin. The examples in the WA Goldfields are one focus of this uncertainty; note that southern WA generally lacks the other putative pidgin placenames considered in this study.

4.5 Picaninny ~ Piccaninny etc The word Piccaninny denotes a child, and by extension was used as a modifier to mean ‘young, small’; pikanini (Wafer et al. 2008: 814). This word is a key indicator of NSW Pidgin, and, as it came from South Seas Jargon (originally from Portuguese-based pidgin in the Atlantic, OED), it is not associated with any particular Australian language (Amery & Mühlhäusler 1996: 35). This word is the only one of the Pidgin words considered here which occurs in placenames in Tasmania (apart from Blackfellows Crossing). The placenames with Piccaninny or Picaninny are widely scattered across the mainland eastern states and the Kimberley, with one occurrence in the NT Gulf district (Piccaninny Creek, likely from Kriol); and in SA are some placenames with the spelling Piccaninnie, and a Picaninnie Creek in western NSW: see Figure 11.

Placenames evidence for NSW Pidgin

135

Figure 12: Placenames with sugarbag

4.6 Sugarbag The word sugarbag denotes wild honey, the hive, and has been extended to the bee itself (Arthur 1996: 61–62). Unlike all the examples considered above, sugarbag is simply from English, and apparently the word was not used in early NSW Pidgin; the first record in from 1830, after the spread north to Queensland. The distribution of placenames with Sugarbag avoids the southern Australia, and is along the Queensland coast, and from Cape York Peninsula to the Kimberley; see Figure 12. Of course, some of the Sugarbag placenames could be from standard Australian English, deriving from the term for the hessian bag used for bulk sugar, as one of the occurrences in NSW is known to be: Sugarbag Creek (Ku-Ring-Gai LGA): Named after ‘The Sugarbag Gang’. A group of youngsters from the area in the 1930’s. The name was derived from the sugar bags they wore on their heads, peaked in the front and hanging down their backs. This copied the headgear used by coal merchants. Geographical Names Board (2011)

Nevertheless I include sugarbag placenames in this study, if only to contrast the distribution of those placenames known to originate in NSW Pidgin.

136

David Nash

5 Sporadic examples The NSW Pidgin option suggests itself as the possible origin of various other placenames from the early contact situation in Australia, even though we otherwise lack good evidence of a pidgin or creole having been spoken in the locality. In turn this can add to the picture of the paths by which nineteenth century contact languages spread across Australia. The locations in this section are not on the paths indicated in Figure 5.

5.1 Victoria There are just a few placenames tagged as derived from Pidgin among the more than 3300 in Clark and Heydon’s (2002) Dictionary of Aboriginal placenames of Victoria. Some have been covered above under Bulga (including Mount Bolga in Victoria), Budgery, Blackfellow, Piccaninny, and Lubra. The other two are Gabo Island and Kinnabulla. Additionally, the names Wildfellow Creek (37°22′S 146°09′E) and Poor Fellow Me Creek (38°41′S 146°11′E) in southern Gippsland look to have been formed in Pidgin, although there is no extant source information.

5.1.1 Gabo Island The one origin that has been proposed for the name Gabo is that it was an Aboriginal pronunciation of the prominent feature about 10km away, Cape Howe. Pidgin for ‘Cape Howe’. According to Dent (Papers 2000: 4) “Gabo Island. Pidgin for Cape Howe, the original Aboriginal name was Werrenganno of unknown meaning” (Clark & Heydon 2002).

The nearest centre of population has been Mallacoota, and Gabo Island is visible from there and in the direction of Cape Howe (which had been named by Captain Cook when he passed by on 20 April 1770).

5.1.2 Kinnabulla Kinnabulla is a locality name and a railway station (35°54′S 142°48′E) in northwest Victoria, purportedly derived from local Aboriginal people who would ask local land holder Stephen Laver to kin-na-bulla (Kill a bullock) for them; referring to wild cattle in their district (O’Callaghan 1918: 58, per Clark & Heydon 2002).

Placenames evidence for NSW Pidgin

137

It strikes me that if the source was indeed (Pidgin) English, it is more likely to have been “skin a bullock”.

5.2 South Australia South Australia and Western Australia south of the Kimberley have been regarded as outside the corridors whereby NSW Pidgin flowed out of early NSW, and this is generally confirmed by most of the above placename distributions. There are a couple of exceptions, however, as can be seen on the maps for piccaninny, lubra and blackfellow: these have a sizeable fraction of their occurrences well inside SA. This could be the related to the wider (and unstigmatised) use of these three words in Australian English in the mid nineteenth century. Note also that the few instances of lubra and black gin in southern WA are clustered in the Eastern Goldfields (around Kalgoorlie), which were populated from eastern states in the 1890s gold rush.

5.2.1 Coober Pedy Næssan (2010) covers in detail the evidence about the origin of the placename Coober Pedy, an opal mining locality in central SA dating from 1915. The earliest newspaper mention with an explanation of the name is: Coober Pedy, the name given to this ‘dugout’ village, was chosen because in aboriginal lingo it signifies ‘white man living in a hole’. (COOBER PEDY! The Register (Adelaide) 29 Nov 1921, page 5 http://nla.gov.au/nla.news-article63348165)

Considering that the original miners were in contact with the local Aboriginal people, who spoke the Kukata language, Næssan (2010: 229) concludes the name is a Kukata lexical ‘loan blend’ composed of Parnkalla-originating kupa (as ‘whitefella’) and Kukata (Western Desert) piti (‘quarry’)

The evidence is open to a slightly different interpretation, whereby the name kupa piti was coined jointly by some Kukata and some of the opal miners. While it may be inferred that there was a pidgin in use in this contact situation, little evidence of it has survived: six documents for the Far North West region according to Wurm et al. (1996: Map 10).

138

David Nash

6 Conclusion As well as the recognised two sets of placenames in Australia, there is another less common kind. These are neither wholly Indigenous, nor wholly introduced, but likely arose in the contact language situation in nineteenth century NSW. We have investigated some of the likely candidates, and conclude that NSW Pidgin was the likely context for the formation of the dozens of placenames involving Cabonne, Nyrang, Bulga, Piccaninny, Yarraman, and so on. The study shows how the limited corpus of NSW Pidgin can be augmented. The various geographic ranges of these placenames conforms more or less to the historical spread as studied by Wurm et al. (1996) and Munro (2000), and add evidence of spatial and temporal detail. The correlations between the distribution of these placenames and the historically known spread paths of NSW Pidgin supports the relevance of placename evidence. In particular, there is some indication that a pidgin was in use also in Victoria and SA, and the WA Goldfields.

Appendix: Other gullen placenames Further to the examples in subsection 3.3, there are other placenames involving gullen in the Wiradjuri area. – Boorungullen Chain of Ponds, also Deadmas Creek (34°06′S 149°01′E, north of Yass) – Derringullen Creek (34°44′S 148°53′E, west of Yass), possibly from a Wiradjuri stem meaning ‘red’, cf. Dirrundirrung ‘red, evening red, applied to Europeans on account of their rosy cheeks’ Günther (1839: 45)19 While we can no longer ascertain whether ‘red water’ is the correct gloss for the Derringullen name, the applicability of this gloss can be ascribed to the soil erosion instigated by the nineteenth century pastoral expansion, as with many other colonial placenames of water features in southern Australia involving Red or Yellow, including the nearby Yellow Creek (34°46′S 148°58′E). – Craigengullen on the lower Lachlan River, in Gelam Parish (34°22′S 144°11′E) (Hanson 1889: 252)20 19 Apparently garbled when published as Diren-direng ‘red’, Dironbirong ‘the red streams of clouds in the evening; adj., red, said of white men’ (Günther 1892: 80–81). 20 Also mentioned in ‘News & Notes’, The Courier (Brisbane), 23 January 1864, page 3 http:// nla.gov.au/nla.news-article3168164 and as “Darlot H., Craigengullen (Cregingelong) £130.” in ‘List of Runs and Rents for the Year 1871’, Australian Town and Country Journal, 14 January 1871, page 10 http://nla.gov.au/nla.news-article70464215.

Placenames evidence for NSW Pidgin

139

None of these placenames have any recorded analysis, but galiN ‘water’ is likely to be involved. An alternative candidate for the gullen element is a Wiradjuri plural suffix -galang (Günther 1892: 57; Grant & Rudder 2010),21 or the stem ngalan ‘a light’ (Günther 1892: 90) (noting the n preceding the element gullen in all the examples). While the circumstances of the assignment of these names is not recorded,22 the gullen placenames names are all in the southern part of Wiradjuri territory, and four or five are clustered in the Yass district. This, together with the fact that the gullen element is spelled the same way (while galiN ‘water’ received various other spellings elsewhere), suggests there was a local convention favouring these names. The proposed structure of these placenames is similar to a set in SA of compound placenames where the second element is the generic word for ‘water’, kawi, kapi were often reduced to awi (Hercus & Potezny 1999). Further north in Wiradjuri territory there are other placenames where however the galiN ‘water’ occurs first: – Cullen Bullen ‘lyre birds. Also claimed to be the place of many waters’ (Reed 1970: 33). Appleton & Appleton (1992: 81) state “[1st u as in ‘pup’; 2nd u as in ‘bull’]”. The Bullen element may have been erroneously glossed from bulen-bulen ‘lyre bird’ (“a nineteenth century form ‘bulln-bulln’”) in Woiwuru, a language of Victoria (Hercus 1969 per Blake 1981: 89). – Cullengoen, near Curban on the Castlereagh River downstream from Gilgandra, presumably a version of placenames Cullengoengoen ‘red water’ Cullen-goin ‘blood and water’ (Royal Anthropological Society of Australia 2004: 40,153), involving Wiradjuri guwaN ‘blood’. On the meaning compare Derringullen above. – Cullenbone (32°29½′E 149°30½′E, in the Gulgong area) ‘meeting of waters’ (McCarthy 1963), apparently involving the Ngiyampaa Comitative suffix -buwan

21 The plural -galang is evident in Wambangalong (32°28′S 148°32′E) ‘grey kangaroo in numbers’ (McCarthy 1963) based on wambuwuny ‘grey kangaroo’ (Grant & Rudder 2010), and perhaps in the name Cadiangullong Creek (33°23′S 149°00′E) near Cadia Creek (33°26′S 149°0′’E, a tributary of the Belubula River). 22 The Senior Constable at Yass sent ‘the attached lists of places and there [sic] meaning given by Aborigines. The lists consist of the whole of the information that could be collected by the Police in the Yass Sub District. There are a great many places in the District, which have been named by Aborigines, but the oldest blacks are unable give any reason why the names were given.’ (Letter to Superintendent of Police, Goulburn, 2 October 1899) (Royal Anthropological Society of Australia 2004: PDF page 413)

140 –

David Nash

two forms of the same word: Cullengoral ‘the meeting place of waters in the Gulgong district’ (Reed 1970: 33); and Cullingral ‘a deep water hole in the Merriwa district’23 (Reed 1970: 33)

That a generic word such as ‘water’ is the base of these placenames could be an indication that they are not ancient, but formed during the early period of European settlement. However the suffixal elements are not recorded as NSW Pidgin, and these combinations would appear to have been formed within the Wiradjuri (or Ngiyampaa) language.

Abbreviations AND

Ransom (1988)

OED

Oxford English Dictionary

References ACT Planning and Land Authority (ACTPLA). 2013. Search for street and suburb names. http:// www.actpla.act.gov.au/tools_resources/place_search. Amery, Rob & Peter Mühlhäusler. 1996. Pidgin English in New South Wales. In Stephen Adolphe Wurm, Peter Mühlhäusler & Darrell T Tryon (eds.), Atlas of languages of intercultural communication in the Pacific, Asia, and the Americas, 3. Chap. 5, 33–51. Berlin, New York: Walter de Gruyter. http://books.google.com.au/books?id=oCx0D0iE2QoC. Amery, Rob & Georgina Yambo Williams. 2009 [2002]. Reclaiming through renaming: the reinstatement of Kaurna toponyms in adelaide and the adelaide plains. In Luise Hercus, Flavia Hodges & Jane Simpson (eds.), The land is a map: Placenames of indigenous origin in Australia, chap. 18, 255–276. Canberra: ANU E Press. http://epress.anu.edu.au?p=29191. Appleton, Richard & Barbara Appleton. 1992. The Cambridge dictionary of Australian places. Cambridge, UK; New York: Cambridge University Press. Maps on end papers. Arthur, Jay Mary. 1996. Aboriginal English: a cultural study. Melbourne: Oxford University Press. Austin, Peter. 1997. Proto Central New South Wales phonology. In Darrell Tryon & Michael Walsh (eds.), Boundary rider: Essays in honour of Geoffrey O’Grady, vol. C-136 Pacific Linguistics, 21–49. Canberra: Research School of Pacific and Asian Studies, Australian National University. Baker, Philip. 1996. Productive fellow. In Stephen Adolphe Wurm, Peter Mühlhäusler & Darrell T Tryon (eds.), Atlas of languages of intercultural communication in the Pacific, Asia, and the Americas, vol. 2, chap. 48, 533–536. Berlin, New York: Walter de Gruyter. http:// books.google.com.au/books?id=oCx0D0iE2QoC. 23 ‘The first runs were taken up in the late 1820s. An early pastoralist was Charles Blaxland (son of explorer Gregory Blaxland) who established Cullingral station to the immediate south and south-west of town.’ (‘Merriwa’, Sydney Morning Herald, 8 February 2004 http://www.smh. com.au/news/New-South-Wales/Merriwa/2005/02/17/1108500197645.html)

Placenames evidence for NSW Pidgin

141

Blake, Barry J. 1981. Australian Aboriginal languages. a general introduction. London: Angus & Robertson 2nd edn. http://books.google.com.au/books? id=q9WCAAAAIAAJ. Clark, Ian D & Toby Heydon. 2002. Dictionary of Aboriginal placenames of Victoria. Northcote, Vic: Victorian Aboriginal Corporation for Languages. httpsː//web.archive.org/web/ 20120324042016/http://vaclang.ozhosting.com/. Title from container: Database of Aboriginal placenames of Victoria. Dawes, William & Anonymous. 2009. The notebooks of William Dawes on the Aboriginal languages of Sydney. School of Oriental and African Studies (SOAS), London. http://www. williamdawes.org. Elphick, Beverley Gulambali & Don J Elphick. 2004. The camp of mercy: an historical and biographical record of the Warangesda Aboriginal Mission/station, Darlington Point, New South Wales. Canberra: Gulambali Aboriginal Research [rev. ed.] edn. Finnane, Mark & Fiona Paisley. 2010. Police violence and the limits of law on a late colonial frontier: The Borroloola case in 1930s Australia. Law and History Review 28(1). 141–171. http://www98.griffith.edu.au/dspace/bitstream/handle/10072/36549/66390_1.pdf? sequence=1. Geographical Names Board. 2011. Geographical names register (GNR) of NSW. http://www.gnb. nsw.gov.au/place_naming/placename_search. Geoscience Australia. 2010. Gazetteer of Australia 2010. http://www.ga.gov.au/place-names/. Grant, Stan & John Rudder. 2010. A new Wiradjuri dictionary. O’Connor, A.C.T. Restoration House. “English to Wiradjuri, Wiradjuri to English, categories of things and reference tables”. Günther, James William. 1839. Vocabulary of the Aboriginal dialect called Wirradhurri spoken in the Wellington district, etc., etc. of New Holland. MS6; microfilmed in 2000 by Academic Microforms, Caithness, Scotland; held at AIATSIS Library call number MF 266. http://www. therai.org.uk/archives-and-manuscripts/manuscript-contents/0006-guenther-james-ms-6/. Günther, James William. 1892. Vocabulary of the Aboriginal dialect called Wirradhuri spoken in the Wellington district, etc., etc. of New Holland chap. Appendix D, 56–120. Sydney: Charles Potter, Govt Printer. http://www.newcastle.edu.au/Resources/Divisions/Academic/ Library/Cultural%20Collections/pdf/09alpendixd.pdf. Hanson, William. 1889. The pastoral possessions of New South Wales. Sydney: Gibbs, Shallard & Company. http://ia700308.us.archive.org/18/items/pastoralpossessi00hans/pastoralpossessi00hans.pdf. Harkins, Jean. 2008. NSW contact languages. In James William Wafer & Amanda Lissarrague (eds.), A handbook of Aboriginal languages of New South Wales and the Australian Capital Territory, chap. 15, 402–412. Nambucca Heads, NSW: Muurrbay Aboriginal Language and Culture Co-operative. Hercus, Luise. 2002. Is it really a placename? In Luise Hercus & Jane Simpson (eds.), The land is a map: Placenames of indigenous origin in Australia, chap. 5, 63–72. Canberra: Pandanus Books in association with Pacific Linguistics. http://epress.anu.edu.au? p=29191. Hercus, Luise & Jane Simpson. 2002. Indigenous placenames: an introduction. In Luise Hercus & Jane Simpson (eds.), The land is a map: Placenames of indigenous origin in Australia, chap. 1, 1–23. Canberra: Pandanus Books in association with Pacific Linguistics. http:// epress.anu.edu.au?p=29191. Hercus, Luise A. & Vlad Potezny. 1999. Finch versus Finch-Water: a study of Aboriginal placenames in South Australia. Records of the South Australian Museum 2(31). 165–180.

142

David Nash

Koch, Harold. 2009. The methodology of reconstructing Indigenous placenames Australian Capital Territory and south-eastern New South Wales. In Aboriginal placenames: Naming and re-naming the Australian landscape, chap. 5, 115–171. Canberra: ANU E Press. http:// press.anu.edu.au?p=17331. Koch, Harold. 2011. Substrate influences on New South Wales Pidgin. The origin of -im and – fela. In Claire Lefebvre (ed.), Creoles, their substrates, and language typology, vol. 95 Typological Studies in Language, 489–512. John Benjamins Publishing. http://books.google. com.au/books? id=3cpxAAAAQBAJ&oi=fnd&pg=PA489. Lee, Jason. 2014. Kriol–English interactive dictionary. http://ausil.org/Dictionary/Kriol/indexen.htm. Lissarrague, Amanda. 2006. A salvage grammar and wordlist of the language from the Hunter River and Lake Macquarie. Nambucca Heads, NSW: Muurrbay Aboriginal Language and Culture Cooperative. Mac. Smith, Bertha. 1976. Smith, John (1811–1895), vol. 6 Australian Dictionary of Biography, Melbourne: MUP. http://adb.anu.edu.au/biography/smith-john-4607/text7579. MacKnight, C.C. 1976. The voyage to Marege’: Macassan trepangers in northern Australia. Melbourne: Melbourne University Press. Martin, A. E. (Archibald Edward). 1944. Place names in Victoria and Tasmania. Sydney: N.S.W. Bookstall Co. http://handle.slv.vic.gov.au/10381/182144. McCarthy, Frederick D. 1963. New South Wales Aboriginal place names and euphonious words, with their meanings. Sydney: The Australian Museum, V.C.N. Blight, Govt. Printer 4th edn. McConvell, Patrick. 2009. ‘Where the spear sticks up’: the variety of locatives in place names in the Victoria River District. In Harold Koch and Luise A. Hercus (ed.), Aboriginal placenames: Naming and re-naming the Australian landscape. 359–402. Canberra: ANU E Press. McConvell, Patrick. 2002. Changing places: European and Aboriginal styles. In Luise Hercus, Flavia Hodges, and Jane Simpson (eds.), The land is a map: Placenames of Indigenous origin in Australia. 50–61. Canberra: Pandanus Books. Molong Historical Society. 2013. List of property and locality names in the Molong district New South Wales. http://www.users.on.net/quincejam/mhsres.html. Munro, Jennifer M. 2000. Kriol on the move: A case of language spread and shift in northern Australia. In Jeff Siegel (ed.), Processes of language contact. Studies from Australia and the South Pacific Collection: Champs linguistiques, chap. 9, 245–270. Saint-Laurent, Quebec: Les Éditions Fides. http://books.google.com.au/books?id=9_LsfbKqUW8C. Næssan, Petter. 2010. The etymology of Coober Pedy, South Australia. Aboriginal History 34. 217– 233. http://search.informit.com.au/documentSummary;dn=661136067797473;res=IELHSS. Nash, David. 2008. Examining the name element / feature type cowal. Presented to ANPS/ CGNA workshop, Wollongong. Nash, David. 2011. What’s a Warrambool? Posted on Endangered Languages & Cultures blog, 19 June 2011. http://www.paradisec.org.au/blog/2011/06/what’s-a-warrambool/. Nash, David. 2013. The smuggled budgie: case study of an Australian loanblend. In Robert Mailhammer (ed.), Lexical and structural etymology: Beyond word histories, vol. 11 Studies in Language Change, 293–312. Berlin: De Gruyter Mouton. http://www.degruyter.com/ view/product/180256. Nash, David. 2014 a. The suffix -dool in placenames of central north NSW. In Ian Clark, Luise Hercus & Laura Kostanski (eds.), Indigenous and minority place names – Australian and international perspectives Aboriginal History Monographs, chap. 3, 39–55. Canberra: ANU Press & Aboriginal History Inc. http://press.anu.edu.au?p=286811.

Placenames evidence for NSW Pidgin

143

Nash, Joshua. 2014b. Norf’k placenames and creole toponymy. Journal of Pidgin and Creole Languages 29(1). 135–142. 10.1075/jpcl.29.1.04nas. O’Callaghan, Thomas. 1918. Names of Victorian railway stations, with their origins and meanings, together with similar information relative to the capital cities of Adelaide, Sydney, Brisbane and a few of the border stations of New South Wales and South Australia. Melbourne: Government Printer. Pattison, John J. 1934. With the US Army along the Oregon Trail, 1863–66. Nebraska History 15. 78–93. http://www.nebraskahistory.org/publish/publicat/history/full-text/NH1934Army_ Trail.pdf. Ransom, WS. 1988. The Australian national dictionary: a dictionary of Australianisms on historical principles. Melbourne: Oxford University Press. http://australiannationaldictionary. com.au/. Reed, Alexander Wyclif. 1970. Aboriginal place names and their meanings. Sydney: AH & AW Reed. Richards, C[harles]. 1902–03. Wirra’athooree. Wirrai’yarrai’. Wirrach’aree’. Wirra’jerree’ (or, Aboriginal dialects). Science of Man 5. 5:81–83 (G); 6:98–102 (G,B); 7:114–119 (B,G,W); 8:133–138 (M,N,Ny); 9:146–149 (Ng); 10:165–168 (Ng,J,Dth); 11:180–183 (Dth,Y); 12:198– 201 (Y,Gw,I,E). Ridley, William. 1878. Language of the Aborigines of Georges River, Cowpasture and Appin. The Journal of the Anthropological Institute of Great Britain and Ireland 7. 258–262. Rogers, Jean McKee. 1967. History of Harlan County 1870–1967. http://www.rootsweb.ancestry. com/ neharlan/countyhistory.pdf. Royal Anthropological Society of Australia, (RASA). 2004. Royal Anthropological Society of Australia manuscripts dated 1900. PDF file ‘anthropological society of aus roll 1’. CDROM, Geographic Names Board of NSW. http://acms.sl.nsw.gov.au/item/itemdetailpaged. aspx?itemid=421736. Siegel, Jeff. 2012. Two types of functional transfer in language contact. Journal of Language Contact 5(2). 187–215. 10.1163/187740912X639247. http://booksandjournals.brillonline. com/content/journals/10.1163/187740912x639247. Smyth, Robert Brough. 1878. The Aborigines of Victoria: With notes relating to the habits of the natives of other parts of Australia and Tasmania compiled from various sources for the Government of Victoria. Melbourne: J. Ferres. http://www.aiatsis.gov.au/collections/ exhibitions/collectors/smyth.html. Steele, Jeremy M. 2005. The Aboriginal language of Sydney: a partial reconstruction of the indigenous language of Sydney based on the notebooks of William Dawes of 1790–91, informed by other records of the Sydney and surrounding languages to c.1905. North Ryde Macquarie University Macquarie University ResearchOnline. http://hdl.handle.net/ 1959.14/738. Stutchbury, S. 1852a. Camp near Cobbe[? ]a, on the Talbragar River, July 1, 1852. The Richmond River Express and Tweed Advertiser 2. http://nla.gov.au/nla.news-article31734072. Stutchbury, S. 1852b. [Geological map of area south of Dubbo to Lachlan River]. State Library of New South Wales, Call No. Ca 85/35. http://acms.sl.nsw.gov.au/album/albumView.aspx? itemID=977102&acmsid=0. Taylor, Rebe. 2002. Unearthed: The Aboriginal Tasmanians of Kangaroo Island. Adelaide: Wakefield Press. http://books.google.com.au/books?id=sO01v0hIf18C. Threlkeld, Lancelot. 1824. Specimens of the language of the Aborigines of New South Wales to the northward of Sydney. Threlkeld Papers, 1815–1862, ML A382 or CY Reel 820, frames

144

David Nash

129–138, State Library of New South Wales, Sydney. http://acms.sl.nsw.gov.au/item/ itemDetailPaged.aspx?itemID=1037188. Troy, Jakelin. 1990. Australian Aboriginal contact with the English language in New South Wales: 1788 to 1845 (Pacific Linguistics B-103). Canberra: The Australian National University. Troy, Jakelin. 1993. Language contact in early colonial New South Wales 1788 to 1791. Canberra: Aboriginal Studies Press. Troy, Jakelin Fleur. 1994. Melaleuka: a history and description of New South Wales pidgin. Canberra The Australian National University Ph.D. Appendix 21: New South Wales Pidgin Wordlist, pages 708–789. Wafer, James William, Amanda Lissarrague & Jean Harkins. 2008. A handbook of Aboriginal languages of New South Wales and the Australian Capital Territory. Nambucca Heads, NSW: Muurrbay Aboriginal Language and Culture Co-operative. Walsh, Michael. 1989. A nagging problem in Australian lexical history. In T.E. Dutton, M.D. Ross & D.T. Tryon (eds.), The language game: papers in memory of Donald C. Laycock (Pacific Linguistics C-110), 507–519. Canberra: Research School of Pacific Studies, The Australian National University. Wurm, Stephen Adolphe, Peter Mühlhäusler & Darrell T Tryon. 1996. Atlas of languages of intercultural communication in the Pacific, Asia, and the Americas, vol. 3. Berlin, New York: Walter de Gruyter. http://books.google.com.au/books?id=oCx0D0iE2QoC.

Greg Dickson

5 Rethinking the substrates of Roper River Kriol: the case of Marra Abstract: The Roper River variety of Kriol is an endogenous creole that has stabilised within the last century while in direct contact with speakers of 6–10 local Indigenous languages. Kriol is well placed to inform creolisation theory because its relationship to substrate languages can be explored more directly and with greater confidence than is possible in the case of many exogenous creoles. Munro (2004, 2011) applied Siegel’s (1999, 2008) Transfer Constraints approach to Roper Kriol and found the theory correctly predicted the transfer of most features examined. Despite this, it is argued here that the survey of potential substrates could be more nuanced and provide a clearer view of the development of Kriol. Drawing on historical research and recent fieldwork with young Kriol speakers, this paper argues that the Marran language family, and in particular, Marra, has more greatly influenced the lexicon of Roper River Kriol than other local languages. Evidence is drawn from four domains: (i) historical evidence, (ii) kin terminology, (iii) non-English based verbs and (iv) ethnobiological nomenclature. The prevalence of Kriol verbs also found in Marra is of particular interest given that substrate derived verbs are typically thought to occur less in creoles than nominals. The argument that Marra wields greater influence than other local language on Roper River Kriol has implications for creolisation theories such as Transfer Constraints approach, suggesting that weighting potential substrate languages variably may improve their ability to predict the features of the resulting Kriol.

1 Introduction Kriol is the name given to the English-lexified creole language – or more precisely, several closely related varieties of Kriol – spoken across a large area of northern and north-western Australia by an estimated 20,000 speakers, almost all Indigenous. Scholarly research on Kriol began in the 1970s, the most prominent publication at this time being a brief grammar (Sandefur 1979) based on the Kriol varieties spoken in the communities of Barunga (formerly Bamyili) and Ngukurr. Subsequent major works include the compilation of a dictionary by Summer Institute of Linguistics (SIL) (SIL-AAIB 1996; also Lee 2004), a sociolinguistic

146

Greg Dickson

description of Kriol literacy practices (Rhydwen 1996), the completion of the Holi Baibul in 2007 and a description of referring expressions (Nicholls 2009). In addition, specific aspects of pidginisation, creolisation and substrate influences have been examined. Hudson (1983) provides detailed observations on the variety of Kriol spoken in Fitzroy Valley, Western Australia and in particular, the grammatical and semantic relationship between Kriol and one of the most prominent substrates of that variety, Walmajarri. The Fitzroy Valley variety of Kriol is, however, geographically distant and not contiguous with Roper Kriol and will not be examined further in this paper. Harris (1986) provides a thorough socio-historical and socio-linguistic account of creolisation in the Roper River Region. His research was carried out at a time when Bickerton’s (1981) Bioprogram Hypothesis was at its most influential and hence did not consider substrate languages as significant to creolisation processes. Munro’s earlier work (e.g. 1995) discusses the spread of Kriol and questions the idea that Kriol emerged independently at various locations. Harris and Munro both find (though with some points of difference) that the creolisation of the Roper Kriol variety centres on the former Roper River Mission, established in 1908. Munro’s most recent work (2004, 2011) considers substrate influences on the Roper Kriol variety, applying the Transfer Constraints approach developed by Siegel (1999, 2008) to test the influence of various substrates on linguistic features of Roper Kriol. This paper focuses on the Roper variety of Kriol and aims to contribute to our understanding of Kriol and its origins, building upon previous work of scholars such as Harris and Munro. I consider primary and secondary sociohistorical material to revisit previous characterisations of the linguistic ecology of the region during and post-creolisation, revealing that Marra speakers had a more enduring period of contact with emerging Kriol users than had been previously acknowledged. I build upon this information and survey new data from recent fieldwork to explore the lexicon of Kriol, paying particular attention to Kriol as spoken by young people who do not speak a traditional language. I assess the lexical influence of the original languages of the region, showing that Marra and closely related languages have made disproportionately greater contributions to Kriol than other languages. With this examination, I provide a more nuanced picture of variable levels of influence among the region’s original languages, contributing to a better understanding of contemporary Roper Kriol and how it came to be. This has implications for creolisation theories such as the Transfer Constraints approach, suggesting that a more nuanced appraisal of potential substrates could result in more accurate predictions of the structure of creoles. A particular feature of this paper is that the data on contemporary Kriol is informed primarily by young Kriol speakers, aged between 20 and 35 who typi-

Rethinking the substrates of Roper River Kriol: the case of Marra

147

cally have little knowledge of traditional languages. In using data from young Kriol speakers this study differs from much of the previous research on Kriol which has been predominantly informed by senior Kriol speakers who either spoke a traditional language or had significant contact with speakers of traditional languages (Sandefur 1979; Munro 2004; Nicholls 2009). The fact that young Kriol speakers have little knowledge of and often little contact with traditional languages means they can provide clearer evidence of substrate influences, as there is less scope for bilingualism in a traditional language to influence their language use. As a result, findings regarding lexical influences of various substrate languages given below are more robust than if they were drawn from Kriol speakers with greater knowledge of particular substrates.

2 Substrates of Roper Kriol – evaluating previous characterisations In the contemporary language ecology of the Roper River Region, Roper Kriol is the first language of most of its several thousand residents. The largest community in the area is Ngukurr, established in 1968 as a secular incarnation of the Roper River Mission which had been established 60 years prior. Ngukurr has a population of over 1,000, virtually all first language Kriol speakers. A century ago, the original, traditional languages of the area were widely spoken but are now either no longer spoken or in various states of severe endangerment, known only to elderly people and some adults. The youngest generations of Kriol speakers in Ngukurr are the first for whom traditional languages do not feature significantly in their language ecology. Despite the decline of traditional languages, there is no evidence of decreolisation occurring although language change is inevitably evident. The Roper River Region in which Roper Kriol developed is traditionally a linguistically complex region (see e.g. Heath 1978a; Harvey 2012). Languages on and immediately north of the Roper River such as Ngandi, Nunggubuyu and Ngalakgan belong to the Gunwinyguan language family. Even further north lie Yolngu languages like Ritharrŋu and Wägilak which are Pama-Nyungan languages. South and at the mouth of the Roper River are Alawa, Marra and Warndarrang which have been grouped as members of the Marran family (see e.g. Wurm 1971). The status of Warndarrang as a Marran language is not clear-cut however. It has been described as a fringe member of either the Marran family (Heath 1978a) or the Gunwinyguan family (Harvey 2012). For the purposes of this study, Warndarrang is considered to be part of the Marran family, along with Marra

148

Greg Dickson

and Alawa, following classifications such as those by Wurm (1971) and Heath (1978a). During the 60-year existence of the Roper River Mission (1908–1968), people belonging to all these language groups took up permanent residence there. Community members at Ngukurr still identify strongly with the languages of their heritage, regardless of level of fluency. It is clear that the population is diverse in terms of linguistic affiliations, with all the original language groups represented by a significant proportion of the population. This linguisticallycomplex demography is frequently described by community members and researchers alike. A typical example found in research is from a detailed socioeconomic study of the late 1990s: The extent of the mission catchment area . . . is reflected in the seven main language groups currently represented in the settlement – Marra, Ngandi, Alawa, Nunggubuyu, Ritharrngu, Wandarang [sic] and Ngalakan [sic]. (Taylor et al. 2000: 15)

Yet common descriptions like the above are a potential point of confusion in that they risk conflating demographic descriptions with descriptions of linguistic ecology. Social groupings in the Roper Region are known to be tied to or barely distinct from linguistic groupings and territorial affiliations. Essentially, land, language and identity are inextricably linked (see Merlan 1981). It is common for Roper Kriol speakers to say that ‘my language is X [e.g. Ngandi]’, but this is a socio-territorial identification and does not in any way indicate proficiency in that language. Conflating demographic descriptions with descriptions of linguistic ecology are potentially more serious when applied to discussions of substrate influence where a clear understanding of local language ecologies is desirable. Earlier Kriol scholars have not always clearly distinguished demography from language ecology. See for example Sandefur’s summary of the “languages represented at Ngukurr:” “Today there are nine major traditional languages represented at Ngukurr (Mara, Wandarang, Alawa, Manggarai, Ngandi, Ngalakan, Nunggubuyu, Rembarrnga and Ritharrngu)” (Sandefur 1985a: 208, Sandefur’s spellings). This description purports to describe linguistic ecology but is actually a description of socio-territorial identifications of Kriol speakers. Harris’ sociohistorical account (1986) of creolisation in the Roper River Region avoids this problem in that little attention is paid to influences of local languages. The lack of attention paid to substrate influence was largely due to the impact of Bickerton’s Bioprogram Hypothesis at the time (see e.g. Bickerton 1984) which foregrounded the influence of Universal Grammar in creolisation and rendered substrate languages inconsequential.

Rethinking the substrates of Roper River Kriol: the case of Marra

149

In contrast, Sandefur did not closely subscribe to the Bioprogram Hypothesis and some of his later work is critical of Bickerton’s theories (e.g. 1985b). In discussing substrate influences, Sandefur observed that it “is most readily observable in the Kriol lexicon” and that “one of the most significant factors contributing to dialect differences in Kriol is the traditional Aboriginal language environment” (Sandefur 1985b: 210). He briefly provides further detail: For example, manuga ‘money’ (from ‘stone’) was borrowed from one of the languages around Ngukurr. It is commonly used at Ngukurr, and known by Kriol speakers in the communities immediately surrounding Ngukurr, but it is virtually unknown by Kriol speakers elsewhere. Some language borrowed words, however, have become regionalised. Gajinga ‘damn it’ (originally a reference to the genitals) is also from a local Ngukurr traditional language, but it is now used by Kriol speakers throughout the Roper River and Bamyili areas. (Sandefur 1985b: 210–211)

Here, Sandefur (1985b) identifies the following: – Traditional languages make a salient contribution to Kriol lexicon(s) – This in turn contributes to geographic variation across Kriol dialects – Substrate lexemes may be localised or may undergo diffusion into neighbouring varieties – Semantic shift and/or pragmatic differences may occur when lexemes are borrowed or transfer into Kriol Sandefur does not go on to investigate these aspects in detail. For example, in the quote above, he offers a localised Kriol word, manuga ‘money’, correctly stating that it is “borrowed from one of the languages around Ngukurr” but does not report its etymology. Manuga is in fact a Marra word,1 an example that foreshadows findings presented below arguing that Marra is Kriol’s most prominent substrate. Since these early studies of Kriol, Bickerton’s theories have been challenged and it is now widely accepted that local languages have influenced Kriol. More recent studies acknowledge and describe substrate languages and identify typological differences, for example: The substrate languages of Kriol can be divided into two families based on typological similarities and genetic descent; Ngalakgan, Nunggubuyu and Ngandi are Gunwinyguan languages. Marra and Warndarrang are Marran languages; Alawa is typologically similar to the Marra languages. There is disagreement as to which family the Mangarrayi language belongs. (Nicholls 2009: 9) 1 In other languages of Ngukurr manuga ‘stone/money’ occurs as gudaru (Alawa), gu-jundu (Ngandi), gu-birn (Ngalakgan), ṉuga (Nunggubuyu), ṉoka’ (Ritharrŋu) and in Warndarrang, Heath (1980a) gives three forms, including manuga (the other two being ligarr and marligarr).

150

Greg Dickson

Nonetheless it is still uncommon for research on Kriol to carefully consider the following questions individually: a. What was the linguistic ecology of the Roper River Mission and surrounding region when creole was emerging? b. What was the linguistic ecology of Ngukurr in the post-creolisation era? c. What contributions have individual local languages made to Kriol and are the contributions equal or variable? Munro (2004, 2011) systematically surveyed substrate influences of Roper Kriol using the Transfer Constraints approach (Siegel 1999, 2008). The Transfer Constraints approach predicts the structural features of creoles based on features transferring from substrate languages given appropriate conditions, such as salience, congruence and frequency of substrate features. In using the Transfer Constraints approach, Munro’s analysis of substrate influence focuses on grammatical functions rather than considering lexical borrowings. In setting the foundation for her analysis, Munro, like other researchers, identifies a group of substrates relevant to the development of Kriol: The substrate languages of Roper Kriol are those of the Indigenous language groups that maintain custodial relationship to their land, also known as country, in the Roper River region: Alawa, Marra, Ngalakgan, Warndarrang, Mangarrayi, Ngandi and Nunggubuyu. (Munro 2004: 4)

Munro also makes some attempt to assess differing degrees of influence among various substrates and ultimately eliminates some languages from her application of the Transfer Constraints approach: Neither Warndarrang nor Ngandi are included in the comparative analyses . . . primarily because Warndarrang shares many typological features with Marra, and Ngandi is closely related to Nunggubuyu (see e.g. Heath 1978a). The low numbers of Warndarrang and Ngandi speakers also implies that these groups had minimal impact on language contact. (Munro 2004: 9)

The above quote shows Munro combining typological factors with ecological factors to discount Warndarrang and Ngandi as significantly influential. Additionally, no mention is made of two other prominent languages of the region, Ritharrŋu and Wägilak. I argue that the appraisal of possible substrate influences of Kriol can be developed further, by re-examining typological and ecological factors separately. By doing so, I suggest that theories such at the Transfer Constraints approach can be better served by a more nuanced appraisal of the linguistic ecology in which creoles develop.

Rethinking the substrates of Roper River Kriol: the case of Marra

151

Considering typological factors, Munro excludes Ngandi from her application of Siegel’s Transfer Constraints approach “primarily because . . . Ngandi is closely related to Nunggubuyu” (Munro 2004: 9) but this may overlook the complex relationship between the two languages. While Heath does indeed state that “Ngandi and Nunggubuyu are closely related” (Heath 1978a: 4) he also goes on to say that “it would be misleading to exaggerate their genetic proximity” (Heath 1978a: 5). He points out examples such as verbal suffixes which, although obviously cognate, carry out different grammatical functions and that “a great many high-frequency stems (nouns, verbs, etc.) are in fact not cognate” (Heath 1978a: 5). Warndarrang was similarly excluded from Munro’s survey on typological grounds as it “shares many typological features with Marra” (Munro 2004: 9). Since Munro’s work however, Harvey (2012) has argued that the relationship between Marra and Warndarrang is predominantly lexical and attributable to borrowing and claims that there is insufficient evidence to place them in the same family. Heath placed them in the same family but did note that they “are rather divergent from each other” (1978a: 7) pointing out that it is difficult to reconstruct much of the morphology of Proto-Warndarrang-Marra-Alawa. Munro also considered demography and linguistic ecology in determining likely substrate influence. Munro is correct in saying that the languages excluded from her study, Ngandi and Warndarrang, had low numbers of speakers compared to other substrates. Warndarrang is the weakest of the original languages of the region and hasn’t been fully spoken since the 1970s.2 Ngandi is noticeably more viable than Warndarrang, although also weak. Yet Ngandi is in a similar situation to Ngalakgan, which Munro did include in the pool of influential substrates, suggesting inconsistency in selecting one over the other. Both Ngandi and Ngalakgan were spoken by small numbers of people until the last fully fluent speakers passed away in the 2000s. A 2001 language census found both languages in comparable states of endangerment: Table 1: Active and passive fluency in Ngalakgan and Ngandi in 2001, self-reported (Lee and Dickson 2003: 48–50)

Speak & Understand Understand but speak a little Total number of people surveyed who affiliated with the language

Ngalakgan

Ngandi

15 17 233

9 28 309

2 Note that two common substrate Kriol verbs, moi ‘threaten’ and maj ‘curse’ (i.e. to proclaim as sacred) are attested in existing documentation as occurring exclusively in Warndarrang. Given that Warndarrang has had some unique lexical impact on Kriol, then perhaps it is worth considering Warndarrang more broadly as having substrate influence.

152

Greg Dickson

Another issue with existing characterisations of influences of Roper River Kriol is the common omission of Ritharrŋu/Wägilak3 (Heath 1980c) which is the most widely spoken traditional language in Ngukurr today (Lee and Dickson 2003) and has been for at least half a century. In the 1970s, Heath noted that: Compared to many other languages in the Northern Territory, Ritharrŋu is still quite viable. It is spoken by reasonably substantial groups at Ngukurr (Roper River) and Lake Evella, as well as a number of outstations. Children seem to be learning the language well in most cases, although at Ngukurr many of them now speak English (in creole form) among themselves. (Heath 1980c: 3)

The likely reason that Ritharrŋu and Wägilak are not usually considered to have influenced Kriol is that speakers of these languages did not arrive in Ngukurr in significant numbers until the 1940s (Harris 1986: 231), probably after creolisation had occurred. Note though that Munro (2004) and also Siegel (2008: 226), unlike Harris, do not claim that creolisation processes were complete by this period so there is a case to be made for Ritharrŋu and Wägilak to be considered as having potential substrate influence, at least in terms of Munro’s study. At the very least, it seems reasonable to consider Ritharrŋu and Wägilak as adstrate languages that may have contributed to contemporary Kriol given the 70-year history in which those languages have been a significant part of the language ecology of the Roper River Mission/Ngukurr. Aside from the above issues, Munro’s study makes the strongest attempt to date to differentiate levels of influence among substrate languages of Kriol. Munro summarises the ultimate choice to use only Ngalakgan, Nunggubuyu, Marra and Alawa in her application of the Transfer Constraints approach as follows: Warndarrang and Ngandi having the least number of speakers, now and presumably in the past, are regarded as the least influential of the substrates. The remaining four languages were highly represented in the wider Roper River Region. Features from these languages that were transferred to individual speaker varieties of the expanding NT Pidgin would, therefore, be most likely retained in the creole due to their high frequency. (Munro 2004: 78)

Despite the issues already discussed, this reduced group represents an evidencebased attempt to differentiate levels of influence among various substrate languages. However a further potential issue arises in that in applying the Transfer

3 Heath’s grammatical description names the language as Ritharrŋu, but mentions that “this is, strictly, a name for one of the matha (clan) groups” (Heath 1980c: 2). Wägilak is also a common name used in reference to an equivalent or near-equivalent language spoken by the members of the Dhuwa moiety. Strictly speaking, Ritharrŋu is spoken by members of the Yirritja moiety.

Rethinking the substrates of Roper River Kriol: the case of Marra

153

Constraints approach, Munro doesn’t allow for different languages to have variable degrees of influence. The local languages are selected in a dichotomous or binary manner, where they are either wholly excluded from having any possible influence, or included and afforded the same degree of potential influence as the other languages selected. In real terms, actual levels of influence would be more nuanced due to varying typological features and complex linguistic ecologies. Evidence presented below demonstrates that Marra and Marran languages have greater influence – at least lexically – upon Kriol than other local languages. It is possible that applying languages in a binary manner to the question of substrate influence could result in skewed analyses of creolisation processes.

3 The case of Marra as the most prominent substrate language The correlation between Marran languages and Kriol has been noted previously. In early work on Alawa, Sharpe (1972) noted an obvious relationship between Alawa and Kriol (at the time an officially unnamed language, often referred to as Pidgin English): In Alawa, tense-aspect-mood and case are indicated by suffixation of auxiliary stems and substantives respectively; in PE [Pidgin English] they are indicated by preposed words. However the contrasts distinguished are found to be in nearly all respects identical. In surface structure the languages are very different; in deep structure and semantically they are almost identical. . . (Sharpe 1972: 9)

Another publication noting Marra’s unique lexical contribution to Kriol is a compilation of stories in Kriol and English authored by Kriol-speaking students and graduates of Deakin University: Blekbala Stori (Deakin University 2004). In the introductory material, Cherry Daniels (of Ngukurr) and Rhonda Bunbury (Deakin University) note a lexical point of interest found in the collection of texts: . . . there is no Kriol or English equivalent to the Mara [sic] word galagala, meaning tree platform for the dead, in the story ‘Holigel’. Other Mara words in this story include migamiga for leech . . . and wanguluwan meaning a poor person, without wealth or parents (Deakin University 2004: 15).

Graber (1987) offered a concise analysis on the Kriol particle na which has more functions in Kriol, particularly as a discourse marker, than its English etymon now. Graber also notes the correlation between na and similar particles in

154

Greg Dickson

traditional languages. One of the languages he identifies is Marra, in which the particle mingi has similar functions to the Kriol na. Data presented below builds on this preliminary evidence to make the case that Marran languages and Marra in particular wield greater influence on Kriol than other local languages. The evidence provided relates to four areas: (1) historical evidence; (2) kinterms; (3) non-English based verbs occurring in Kriol; and (4) ethnobiological nomenclature (specifically relating to traditional medicine and lizard taxa). Much of this data is primary data collected during fieldwork in Ngukurr and the surrounding region between 2010–2013, which in turn was built upon six years of applied linguistic work in the region.

3.1 The case of Marra – historical evidence Harris (1986) provided the first detailed sociohistorical account of the introduction of an English-based pidgin into the Roper River Region in the late 1800s and its subsequent creolisation, which he attributes to the establishment of the Roper River Mission in 1908. He characterised the demographics of the population of the new mission as such: By . . . 1909, over two hundred people had gathered at the mission. They were the remnants of the Mara, Wardarang, Alawa, Ngalakgan and Ngandi tribes, the southernmost members of the Rembarrnga and Nunggubuyu tribes and some of the easternmost members of the Mangarayi tribe. (Harris 1986: 235, original spellings)

The implication here is that if the “remnants” of the Marra arrived at the mission in 1909, none were left on their own territories and they were no longer living lives deeply connected to the pre-contact period. This implication is false. Primary oral evidence gathered during fieldwork combined with other historical evidence demonstrates a period of forty years (or more) where Marra people continued to live on country with little regular contact with English and nonIndigenous people. A Marra elder who I worked closely with on language documentation until her passing in 2013 was brought to the mission in the 1940s when she was “might be. . . around about twelve years old” (Collins 1998: 25). Prior to that, she had subsisted ‘on country,’ with relatives. She described to me her lack of knowledge of English upon arriving at the mission sometime in the 1940s: Nana Ingglish gana gu-ngarl-ngamiyi, marluy. Ganiyi Marra gana ngarl-ngarl-ngamindini. Gana guwarda-guwarda-ngalgujinji wala wul-wayarra wul-agagurr gana ngarl-walajunyirlana. Ingglish gana ngarl-walamindininya. Ga gana nginarra, gu-nga-yalya nana Ingglish, marluy. Guwarda-guwarda-ngalgujinji wala walaya wulagagurr, gana niwanji gaya la skul. . . junggu Marra gana ngarl-ngamindini. Marra, daway ngini.

Rethinking the substrates of Roper River Kriol: the case of Marra

155

[I didn’t speak any English at all.4 I was still speaking Marra. I would listen to them, the other children, talking to each other. They were speaking English. And me, I didn’t understand English at all. I was listening and listening to those children, who were there at school. . . only Marra I was speaking. Marra, my language.] (20110114MARRAfrNGUgd01a. wav_01:17:16)

Similarly, another elder involved in recent documentation of the Marra language reported leaving her traditional lands and lifestyle after her third child was born sometime around 1948. Her oral history provided in Marra (excerpted below) elaborates on earlier evidence she had given in a land claim hearing5 in 1980 (see Numamurdirdi 1980: 242–245): Mingi nga-galuni nana nanggaya balwayi. Manjayu. Nana gayi: Abaju. Roy Hammer. Niwanjanji:::: warrnggu nana nanggaya dud-ngayaganyi wuninggi. . . nanggaya. . . bla Malangaya dedi, nana Jagwilyim. Gayarra gana niwanjanji warrnggu, lujim-ngalguma bla Abaju-mob. Niwanjanji gayarra:::: warrnggu nga-niwirlini win.garra-yurr na. Win.garrayurr nga-niwirlini, warrajarri Ropa. Mingi ngalgu-galuni wala walaya. Wul-ngina: Abaju, Manjayu. . . ngani gayi. . . Jagwilyim. Ngalgu-galuni mingi. [I had (my) eldest (child) at that time. Manjayu. The other one: Abaju. Roy Hammer. We lived there (Limmen River) a long time until the next one came along, that one. . . Malangaya’s father. Jack William. We stayed there until I lost (the father) of Abaju and company. We stayed there for a long time until we came here. We came here. To higher land: the Roper River Mission. I had those (children) then. My (children): Abaju, Manjayu, and the other one. . . Jack William. I had them at that time.] (TN_20100714MARRAgroupNUMgd02a_00:03:57)

Secondary evidence gathered for a separate Marra land claim (Olney 2002) further confirms the enduring presence of Marra people on Marra country well into the twentieth century, contrary to what Harris had implied: . . . the Marra retained contact with their country during the 1920s, 1930s and 1940s. For instance, Smith [Langford-Smith] a missionary at the Roper Mission from the late 1920s to the mid 1930s wrote that: “The natives at the Limmen River are still in their native state”. (Olney 2002: 66, citing Langford-Smith 1936: 255)

4 The speaker’s reference to English here likely refers to any English-based varieties she encountered, including Standard English, Pidgin English or an English-based creole. 5 The Aboriginal Land Rights (Northern Territory) Act (1976) allowed Aboriginal people to legally claim title over their traditional territories. Land claim hearings involved the presentation of anthropological and sociohistorical evidence to support claimants in their efforts to demonstrate their traditional affiliation with and occupation of the claimed land.

156

Greg Dickson

These patterns extended beyond World War II when government patrol officers were employed to traverse the Northern Territory and report on Aboriginal populations. The Maria Island and Limmen Bight Land Claim book reports that: In late 1944 patrol officer Harney noted Marra people as inhabiting the Limmen River and coastal areas to the south. . . . In 1951 patrol officer Sweeney recorded 25 people from the Limmen River who visited the Roper Mission demonstrating that Aboriginal people were living at least part of the time near the claim area. (Olney 2002: 66)

This evidence indicates that for at least 35 years after the Roper River Mission was established in 1908, first language Marra speakers continued to live on their own land, presumably with little influence from pidgin, creole or English. Beyond that initial 35-year period, there is further evidence that Marra speakers continued living permanently or semi-permanently on their country up until at least the 1950s. Those who resided on Marra country would have continued to use the language communicatively. This suggests that while creolisation was taking place at the Roper River Mission in the first half of the twentieth century, pidgin and emerging creole speakers were coming into contact with Marra speakers who knew little or no English for up to four decades. This would have created language contact situations over an extended period that provided opportunity for the Marra language to contribute to the emerging creole. This historical evidence was not noted by Harris (1986), nor in the historical background provided by Munro (2004). Actual linguistic data from this period is more difficult to uncover. For most of the mission’s existence (1908–1968), Pidgin English and the emerging creole was actively discouraged by missionaries, as was the use of traditional languages like Marra.6 With languages other than English held in low regard, there is little written documentation of anything but English from this period. However, a key piece of evidence comes from a 1918 report by one of the missionaries, written for church members in urban Southern Australia. An interesting aspect of Joynt’s short report is that it contains an early rare example of lexical influence from traditional languages(s). In discussing the diet of Aboriginal people, Joynt lists: Yams, lily (chow-chow), lily (yalbourn), lily (guniyah), small nuts, black and green plums, three kinds of black currants, white and red currants, wild orange, wild banana, creeper (passion fruit), cucumbers, wild melon, wild potato, wild rice (two kinds – swamp and sandridge), wild cocoanuts (nuts of a palm tree), pandanus seeds, nut from nutwood tree, wild fig (guninyarra). (Joynt 1918: 4) 6 There are some instances of acknowledgement or tolerance of languages other than Standard English throughout the mission’s history, but overall traditional languages and the emerging creole were denigrated and discouraged.

Rethinking the substrates of Roper River Kriol: the case of Marra

157

This list includes four lexemes (bolded) from local languages and their inclusion suggests that they must have been frequently used to warrant their presence in a document written for non-Indigenous audiences with negligible interest in Aboriginal languages. These four lexemes provide clues as to which traditional languages were prevalent or had influence in the mission’s early years. Table 2 shows the etymology of these lexemes: Table 2: Local Aboriginal language food names occurring in Joynt (1918) Joynt (1918)

Contemporary Kriol

Chow-chow

Jojo

Lilystalk (stalk of Nymphaea violacea)

Jawjaw: Marra, Alawa, Warndarrang, Ngalakgan, Ngandi

Yalbourn

Yarlbun

Seedpod of lily

Yarlbun: Marra, Alawa, Warndarrang

Guniyah

Garnaya

Bulb of lily

Garnaya: Marra, Warndarrang

Guninyarra

not commonly used

“Tree with large reddish figs, Ficus racemosa” (Heath 1981: 457)

Gurninyarra: Marra, Alawa, Nunggubuyu7

Referent

Other language attestations

Note that the only common language for all four lexemes is Marra, which suggests that Marra speakers and their language had some prominence in the mission even in its first decade when Pidgin English was possibly a lingua franca and undergoing creolisation. Almost one hundred years later, the prominence of Marra in the etymologies of non-English based Kriol words remains, as outlined below.

3.2 The case of Marra – evidence from kinterms Kinship systems in Australian Aboriginal languages and cultures are extremely prominent and complex, reflected linguistically in large classes of highly specific kinterms and in more distinctive ways such as kinterms permeating the syntax of some languages (e.g. Evans 2003; Garde 2013). In discourse and person reference in both traditional languages and Kriol, kinterms are used far more frequently than personal names, to the point where it is possible to have a close personal relationship with someone and be fully aware of their kinship networks and 7 Heath gives two names for this tree in Nunggubuyu (1982). The other one, yibunung, is the only one that occurs in the volume of ethnographic texts (Larrangana in Heath 1980b: 480), which suggests it may be the preferred Nunggubuyu term and that gurninyarra is possibly a borrowing from Marra/Alawa.

158

Greg Dickson

relationships without even being sure of their legal English name. Kinship and kin terminology involves not just consanguineal and affinal kin, but also extends to virtually the entire population with the inclusion of classificatory kin. As Blythe (2010: 465) summarises, “a very Aboriginal way of speaking [is one] in which kinship is placed front and centre”. Given the importance of kinship and kin terminology in Aboriginal cultures, contemporary Kriol data in this domain is particularly relevant to this study. An additional reason for examining kinterms is that root forms are generally salient and unbounded, with comparatively less morphology than is found in other parts of speech. This potentially increases their borrowability and is thus a useful domain in which to observe substrate lexical influence. Recent fieldwork with young Kriol speakers determined a set of thirty-two kinterms in current use among Kriol speakers in Ngukurr (Dickson 2015). This is a larger and richer set than had been previously documented (see e.g. Nicholls 2009: 64) and, as shown below, examining the forms suggests that Marra has had a disproportionate impact in relation to other local languages. Of the thirty-two kinterms in current use in Roper Kriol, ten are derived from Indigenous languages.8 These non-English based kinterms in Roper Kriol, with rudimentary glosses, are: Table 3: Kriol kinterms derived from languages other than English Kinterm

Gloss

baba barn.ga gudi gabarani lambarra muluri abuji amuri gagu abija

sibling cross-cousin (mother’s brothers’ children/father’s sisters’ children) reciprocal term used between father/son and aunt/nephew reciprocal term used between uncle/nephew and mother/son father/son-in-law mother-in-law’s brother father’s mother + siblings father’s father + siblings mother’s mother + siblings mother’s father + siblings

Note that these kinterms are in fact all reciprocal, so while gagu refers to ego’s mother’s mother (and siblings) it also refers to a man’s sister’s daughter’s children and to a woman’s daughter’s children. The above glosses are also 8 Regarding the twenty-two English-derived kinterms: seven are derived from non-kinterms and fifteen are based on English kinterms. Note though that English-based kinterms do not refer to the same kin categories as their etymons. For example, anti ‘aunty’ refers to father’s sisters, not mother’s sisters and isn’t limited to consanguineal (blood) kin but extends to a large number of classificatory kin.

Rethinking the substrates of Roper River Kriol: the case of Marra

159

misleading because these kinterms refer to a large number of classificatory kin in addition to consanguineal and affinal kin. Three of the above kinterms – gudi, gabarani and muluri – were not listed in the Kriol Dikshenri, suggesting that nonEnglish lexical influences have to date been under-described. Looking at the precise etymologies of the terms given above, three are attributed to Aboriginal languages from beyond the local region and geographically outside the pool of languages that could be considered as substrates: barn.ga is found in Wambaya (Nordlinger 1998), a language found a few hundred kilometers south of the Roper Region, and has cognate forms in other distant languages. The explanation for the presence of this term and some Englishbased terms (e.g. banji ‘sibling-in-law’) is that they featured in NT Pidgin English which was introduced in the area in the late 1800s along with the expansion of the pastoral industry (Harris 1986). Gudi and gabarani were previously undocumented and appear to be recent additions, added to the lexicon after creolisation had taken place. Evidence that the terms are recent borrowings includes their high frequency among younger speakers and low frequency among senior speakers. These two terms originate in languages from Central Arnhem Land (outside the Roper River Region) such as Dalabon and Bininj Gunwok (see Dickson 2015 for more on these kinterms). This results in seven remaining kinterms that are attested in local languages. All seven occur in Marra, Alawa or both: Table 4: Kriol kinterms and equivalents in Marra and Alawa Kriol term

Gloss

Marra form

Alawa form

baba

sibling

baba (elder sibling only)

baba (elder sibling only)

lambarra

father-in-law

lambarra

-lambarrgarra

muluri

mother-in-law’s brother

muluri

?

abuji

father’s mother

mimi

gardigardi, jabjab

amuri

father’s father

murimuri

ganggu

gagu

mother’s mother

gugu

gugu

abija

mother’s father

bijaja

gardigardi, jabjab

Table 4 shows that three Kriol kinterms – lambarra, baba and muluri – occur in Marra in the exact form.9 Another three Kriol forms are related to Marra terms – 9 Cognates of baba occur as a sibling kinterm across much of Australia, including the Roper Region, however the exact form baba is not typical. Lambarra is also attested in other Australian languages, however Marra is the only language in the immediate region in which that exact form is documented.

160

Greg Dickson

gugu~gagu, murimuri~amuri, bijaja~abija – where both the phonological forms and the kinship categories they refer to are similar in both languages. Expanding the above to all languages of the Roper River region shows that several kinterms are attested in other languages but not with the same frequency: Table 5: Occurrence of Kriol kinterms in traditional languages of the Roper Region Language

occurs in exact form

occurs with varied form

doesn’t occur

unknown10

Marra Alawa Warndarrang Nunggubuyu Ngandi Ngalakgan Ritharrŋu

3 1 0 1 0 0 0

3 2 3 1 4 2? 1?

1 3 2 5 3 5 5–6

— 1 2 — — — —

The only non-Marran language that shares a majority of the non-English Kriol kinterms is Ngandi, however none occur in the exact form and some occur in arguably dissimilar phonological forms such as wawa (cf. baba ‘sibling’) and gokgok (cf. gagu ‘mother’s mother’. Heath’s documentation of Ngandi (1978b) found a fifth related form ngamuri (cf. amuri ‘father’s father’) but argues that it “appears to have originated in Warndarrang (my/our father’s father), passed into creole and thence into Ngandi (and other languages)” (Heath 1978b: 40). Table 5 supports the suggestion that Marra has made the largest lexical contribution to the domain of kinship of any Aboriginal language. Marra’s influential role is further evidenced when we consider the common kinterm muluri ‘brother of mother-in-law.’ The only Aboriginal language this form occurs in is Marra. Additionally, the ‘auxiliary’ kinterm, jiwa, only occurs in Marra (see Heath 1981: 477) and Alawa (see Sharpe 2001: 54). Jiwa is usually glossed as ‘widow’ by Kriol speakers, but more specifically it refers to a widow of a recently-deceased person and a group of close sibling-in-laws (of the deceased) who are culturally obliged to go into hiding during a mourning period. The wide use of both these terms among Kriol speakers in Ngukurr, in addition to other data presented above, suggests Marra has a unique degree of lexical impact.

10 Kin category or kinterm not documented in key source.

Rethinking the substrates of Roper River Kriol: the case of Marra

161

3.3 The case of Marra – evidence from non-English based Kriol verbs Recent research has documented around eighty non-English based verbs in use in Kriol, sixty of which were known to all or most adults in Ngukurr (see Dickson 2015). This significant lexical contribution from substrate languages in the category of verbs is perhaps somewhat unexpected in creoles. Typically, substrate-derived lexemes in creoles are characterised as most prevalent in nominal word classes, such as Nicholls’ suggestion that “words commonly derived from the substrate languages in heavy Kriol include body part terms, place names, animal and plant names” (2009: 23). This corresponds with borrowing hierarchies that class nouns and adjectives as more easily borrowed than verbs. It is suggested that morphological and syntagmatic factors constrain the borrowability of verbs, especially when the languages in contact are typologically dissimilar (Winford 2003: 51–52). Counter-examples exist showing that nouns are not always more easily borrowed than verbs. Wohlgemuth (2009: 245–246) cites the example of German where the need to assign borrowed nouns a gender acts as a barrier to borrowing whereas verbs are accommodated more easily into the language. In the case of substrate-derived verbs transferring into Kriol during creolisation or being borrowed after creolisation, it is clear from the data provided below that Marra-derived verbs are quite common in Kriol. Unlike kinterms and other nominals which often occur in bare forms or with little morphology, the borrowability of verbs into Kriol is potentially influenced by the complex verbal morphology found among the languages of the Roper Region. The boundedness of verbal material in Marra coupled with the typological incongruence between Kriol and Marra should act as barriers to lexical borrowing (see e.g. Thomason and Kaufman (1988: 72–73). Yet, the structure of the verb complex in Marran languages, which typically include semantically salient uninflected coverbs, is actually well suited for the lexical transfer of coverbs. This was also found to be the case with Gurindji verbs occurring in high rates in Gurindji Kriol (Meakins and O’Shannessy 2012). Complex verb structures in Marra are briefly described below, including a brief comparison with verbs of Gunwinyguan languages to the north, demonstrating how coverbs in Marran languages are good candidates for adoption into Kriol. Marran languages feature complex verbs with uninflected coverbs that carry the semantic weight of the verb. The auxiliary verb that is suffixed to the coverb compulsorily carries pronominal information fused to a semantically-bleached verb inflected for various TAM categories. This is illustrated in the following example showing the use of gubarl ‘scavenge’ as a Marra coverb:

162 (1)

Greg Dickson

Wajinja-yinja nirrirri wardabirr ngaba 3SG.eat(PRS)-REDUP small goanna and gubarl-gubarl-arlindu scavenge-REDUP-3SG.go(PRS) It (dingo) eats small lizards and it scavenges (Marra GD060301MAR.NGU_03fr_00:00:17)

Coverbs are semantically salient in Marra and also phonologically ‘robust’, making them a candidate for borrowing (see also Meakins and O’Shannessy 2012). This phonological ‘robustness’ is due to (a) coverbs being the first element in the verb complex, (b) not undergoing any phonological processes or inflection, and (c) typically causing phonological changes to pronominal prefixes which are the first element in the auxiliary verb that follows the coverb. Examples (2) and (3) demonstrate a fortition process, where the final /d/ of the coverb ngambud ‘be submerged’ causes the bilabial glide occurring in the pronominal prefix, shown in its bare form in (2), to harden to /g/: (2)

Mingi wanga now 3SG.go(PST)(PUNCT) S/he just went. (Marra MT_20110113MARRAfrmtNGUgd02_00:01:05)

(3)

Gaya bugi gana ngambud-ganga there EMPH REL be_submerged-3SG.go(PST)(PUNCT) It went under (submerged) right there. (Marra 20110901MARRAmtNGUgd01_00:06:46)

In other environments, coverbs cause lenition in the pronominal prefix that follows. Example (4) contains the full form of the pronominal prefix niwi- (1PLEXCL), whereas in (5), the lateral-final coverb gal ‘grow’ results in the deletion of /n/: (4)

Ganagu warri-niwiyurrayi gana NEG return-1PLEXCL.go(PST)(POT) REL We never went back to the east. (Marra 20110114MARRAfrNGUgd01_01:19:09)

nguwirri east(ALL)

(5)

Ganagu gal-iwiyurrayi dijei NEG grow-1PLEXCL.go(PST)(POT) this_way We didn’t grow up here. (Marra FN_20100714MARRAgroupNUMgd02_00:01:24)

Rethinking the substrates of Roper River Kriol: the case of Marra

163

The combination of phonological robustness and semantic salience allows coverbs to be good candidates for transfer in language contact situations. This is exemplified in the following Kriol example containing the verb gubarl ‘scavenge’ which was already seen in (1) as a Marra coverb. In examples (1) and (6), the form and semantics of gubarl are paralleled, describing the habits of local fauna in both instances: (6)

Im laigi gubarl daga from rabishdamp eniweya 3SG likes.TR scavenge food from rubbish.dump anywhere It (crow) likes to scavenge food from the rubbish dump (or) anywhere. (Kriol DR_20110629KRIOLdrkmcdNGUgd01a_00:36:55)

As mentioned above, similar coverb structures found in Gurindji (an unrelated language) are demonstrated to be a factor in the high degree of Gurindji-derived verbs transferring to Gurindji Kriol, making up a third of the all verbs (Meakins and O’Shannessy 2012).11 Complex verb structures like that found in Marra and other Marran languages are not found in Gunwinyguan languages, including those of the Roper Region such as Ngalakgan, Ngandi and Nunggubuyu.12 Verbs in these languages are agglutinative and polysynthetic and often incorporate nouns into the verb complex. Pronominal prefixes are always the first constituent in the verb complex and they lack a semantically-salient uninflected verbal component that would

11 Note however, that Meakins and O’Shannessy (2012) propose that a “crucial factor” (237) in the borrowability of coverbs is their degree of boundedness. That is, whether coverbs in the original languages occur in loose-nexus or close-nexus verbal structures, defined by how easily a coverb can be separated from the inflecting verb. They suggest that the low number of Warlpiri-derived verbs in Light Warlpiri is attributable to Warlpiri verbs being ‘tight-nexus verbs’, identifying aspects such as Warlpiri complex verbs forming a single phonological phrase. While Gurindji complex verbs are considered ‘loose-nexus’ and it is suggested that this is a factor in the high degree of borrowing of Gurindji verbs into Gurindji Kriol. The Marra/Kriol data does not support this theory as Marra verbs would be considered tight-nexus, akin to the Warlpiri verbs, where coverbs only occur bound to an inflecting verb and form a single phonological phrase. Yet Marra has contributed over 30 verbs to Kriol – a greater number than that of Warlpiri verbs found in the mixed language Light Warlpiri. 12 Munro (2004) and subsequently Siegel (2008: 227) incorrectly suggest that Roper River substrate languages are all agglutinative, however Marran languages and Nunggubuyu have complex morphophonemic processes and fused auxiliary verbs forms (also often suppletive) that are not agglutinative, as shown in example (1)–(5).

164

Greg Dickson

be easily borrowed. Examples (7), (8) and (9) show typical verb structures in the three Gunwinyguan languages of the Roper Region: (7)

barra-ja-ngu-tjjini 3PL>A-now-eat-PRS They eat it. (Ngandi – Heath 1978b: 224)

(8)

nga-ngu-yarrga-gambana 1SG-NEUT-flipper-cook(PRS) I’m cooking the flipper. (Nunggubuyu – Baker, Horrack et al. 2010: 66)

(9)

burru-mirlarr-miny 3NSG-be_born-PST.PUNCT They were born. (Ngalakgan – Merlan 1983: 180)

New data from recent fieldwork revealed a larger set of substrate-derived verbs in common use in Roper Kriol than had been previously documented. Fifty of the seventy-eight verbs described in Dickson (2015) were not previously documented in Kriol. Of the full set of seventy-eight verbs, sixty were deemed to be in common use and thirty-four of those sixty were not listed in the Kriol Dikshenri (Lee 2004). This establishes existing documentation of Roper Kriol as underestimating the amount of substrate-derived lexemes present in the language which has potential ramifications for any subsequent analysis of substrate influences. The set of sixty widely-known non-English based verbs is presented in Table 6, broadly grouped according to etymology: As Table 6 shows, Marra and the other Marran languages, Alawa and Warndarrang, contribute significantly more verbs to the Kriol lexicon than other local languages. Over half (33 of 60) of the widely-known non-English based verbs occurring in Kriol are attested in Marra and a further six are attested in other Marran languages (but not Marra). In the majority of cases, these Kriol verbs exist in Marran languages as uninflecting coverbs, as shown with the example of gubarl ‘scavenge’ in (1) and (6) above. The etymologies of the set of sixty common verbs is represented in Figure 1.

Rethinking the substrates of Roper River Kriol: the case of Marra

165

Table 6: Summary of etymologies of sixty common non-English based Kriol verbs Etymology

Verb ‘gloss’

Marra only (N=11)

barlai ‘be late, miss stg’ gulaj ‘nod’ many ‘walk quickly’ ngar ‘have an erection’

gardaj ‘grab’ mangala ‘jump on bandwagon’ nyal ‘help fight, take sides’

gubarl ‘scavenge’ manjal ‘be weak’ ngaja ‘ask for stg’ waranga ‘be lost, confused’

Marra and other languages (N=22)

bal ‘pound’ didi ‘delouse’ dirrwu ‘dive in’ gululu ‘thunder’ jarlu ‘lead by hand’ munyurrim ‘refine’ nyip ‘retreat, be scared’ yarryarr ‘scatter’

bardap ‘jump in fright’ dinggal ‘limp’ gilgil ‘crawl’ gumbu ‘urinate’ mangumangu ‘elope’ ngarra ‘peep’ warl ‘desire’

bilk ‘be weak’ dirr ‘fart’ gudid ‘carry, convey’ jalk ‘stab’ mirnim ‘flicker’ ngayab ‘be quiet’ yalala ‘be satisfied’

Alawa only (N=4)

birrij ‘dodge’ nyangarrim ‘be selfish’

dilbak ‘tip over’

jal ‘copy’

Warndarrang only (N=2)

maj ‘curse’

moi ‘threaten’

Other neighbouring languages13 (N=7)

bagai ‘be relaxed’ burdudup ‘piggyback’ nyurr ‘grumble’

baku ‘vomit’ ngangga ‘be burrowed’

birr ‘doubt’ nyang ‘chew’

gula ‘argue’

guna ‘defecate’

gai ‘praise’ kayai ‘be horny’ ngum ‘hit on back’ wurruwurru ‘rock baby’

jawak ‘vent anger publicly’ mal ‘preen’ nyirrk ‘obsess’

Nonbogi ‘swim’ neighbouring Australian languages14 (N=3) Unknown (N=11)

gabai ‘beckon’ juljul ‘be randy’ muny ‘curse’ nyumunyumu ‘thrust’

Figure 1 shows that the etymology of eleven of the sixty most common verbs could not be ascertained. This means that two-thirds (33 of 49) of common Kriol verbs with known etymologies also occur in Marra. However, twenty-two of the 13 The etymologies of verbs from neighbouring non-Marran languages (as best as could be determined) are: bagai – Yolŋu Matha, baku – Gurindji; birr – Ngandi, Ritharrŋu/Wägilak; burdudup – Nunggubuyu, Yolŋu Matha; ngangga – Ritharrŋu/Wägilak, Nunggubuyu; nyang – Yolŋu Matha; and nyurr – various, including Ngalakgan and Yolŋu Matha. 14 These verbs are derived from verbs commonly found in Pama-Nyungan languages or derived from New South Wales languages. Their presence in Kriol is attributed to transfer from NT Pidgin that developed from pidgins used in New South Wales and Queensland.

166

Greg Dickson

Figure 1: Language of origin of 60 commonest non-English based verbs in Roper Kriol

thirty-three Marra-Kriol shared verbs occur in other languages as well, so their presence in Kriol cannot be attributed solely to the influence of Marra. The high number of verbs with multilingual etymologies it is not unexpected given the high level of cognates between Marra and the other Marran languages, Alawa (17% cognates) and Warndarrang (54%) (see Harvey 2012). Indeed, most of the twenty-two verbs derived from Marra and other languages are derived only from languages within the Marran language family. Despite this, there remain eleven verbs which appear to have transferred directly from only Marra to Kriol, significantly more influence than is attested from any other Indigenous language. It is clear from the evidence presented in this section that Marra has contributed more verbs than any other local language to the lexicon of Kriol and that collectively Marran languages have made greater contributions than any other language family. This was not previously been documented. The degree to which this is attributable to social factors (indicated by the historical evidence given in §3.1) or typological factors (i.e. the feature of uninflecting coverbs in Marra and related languages) is unclear. There are two main hypotheses to explain the disproportionate prevalence of Marra verbs in Kriol: (1) social and

Rethinking the substrates of Roper River Kriol: the case of Marra

167

historical factors that saw an extended period of contact between Marra speakers and emerging creole speakers or (2) the transferability and borrowability of coverbs due to the their phonological robustness and semantic salience within Marra’s complex verb structures. Of course, these hypotheses are not mutually exclusive.

3.4 The case of Marra – evidence from ethnobiology The final domain in which I offer evidence of Marra’s disproportionately greater lexical impact upon Kriol is ethnobiology, or more specifically, the subdomains of traditional medicine (locally referred to as ‘bush medicine’) and lizards. In 2013, I interviewed fourteen young first language speakers of Kriol in Ngukurr (aged 22–35), surveying their knowledge in these domains.15 Each completed free-listing tasks that tested how many types of bush medicine and lizards they could recall. Collectively, they described 21 types of bush medicine and 10 lizard taxa. Individually, the average number of taxa each listed was 6 types of bush medicine and 4.4 lizard taxa. Listing 21 bush medicine types provided 78 tokens of reference to bush medicine taxa. Reference was achieved by using English-derived names, Aboriginal language-derived names and, in a few instances, descriptions or deixis (i.e. pointing). Of those 78 tokens, 45 were instances of Kriol speakers providing a name derived from local languages. In 40 of those 45 instances, the name used was attested in Marra but no names were exclusive to Marra. A key part of the study was determining which taxa are the most salient among Kriol speakers. Regarding bush medicine, nine taxa were listed by four or more of the study participants. Two of those were referred to only with English-derived names or descriptions. The other seven most salient bush medicines were identified by at least some of the interviewees by a name derived from local languages. In all cases where a ‘salient’ medicine was listed with a local language derived name, that form occurred in Marra. Table 7 summarises the etymologies of the names used to refer to the nine most salient bush medicines:

15 The fourteen study participants were of mixed gender, education, linguistic heritage and family backgrounds. Only two of the cohort reported speaking an Aboriginal language fluently (but not as a L1).

168

Greg Dickson

Table 7: Etymology of non-English based names of most salient bush medicines known to young Kriol speakers

Taxon

Salience measure16

Tokens

With local names

Etymology (of local name)

Dumbuyumubu Mayarranja Ngalangga Guyiya/dogbul Warlan Smeligras Plamtri Gulban/titri Barnarr/mabultri

0.7026 0.2969 0.2847 0.2149 0.1969 0.1852 0.1110 0.0691 0.0564

14 8 9 5 7 7 6 4 4

14 4 6 4 7 0 0 2 2

All local languages Marra/Alawa/Warndarrang All local languages Marra/Alawa/Warndarrang All local languages n/a n/a Marra/Alawa All local languages

Table 7 shows that when young Kriol speakers refer to bush medicines using a name from a local language, it is either (a) a name occurring in all local languages or (b) a name occurring only in Marran languages. There were no instances of any of the more salient medicines being referred to with an Indigenous name derived exclusively from a non-Marran language. The examination of taxonomic knowledge of lizards resulted in the 14 participants identifying 10 types of lizards across 58 tokens. Most of the tokens featured English-based lizard names or descriptions given in Kriol. 14 tokens (24%) were names derived from local languages and were used to refer to four different taxa. 12 of those 14 tokens were names occurring in Marra. The remaining two non-Marra names were Wägilak names, given by the one participant who had fluency in Wägilak. The most noteworthy aspect of the non-English nomenclature used to refer to lizards was the third most salient lizard, a type of small goanna (Varanus acanthurus), which was listed by 9 of the 14 participants. All nine used the name dabulun to refer to this lizard and this name is attested only in Marra. The only other lizard with which multiple participants used the same nonEnglish based name to refer to it was Tata lizard (Diporiphora spp); eight participants listed it and two used the name garn.gulugulu, a name occurring in Marra, Alawa, Nunggubuyu and possibly Warndarrang. The evidence drawn from this ethnobiological research is consistent with the findings of other domains: that Marran languages and Marra in particular have a disproportionately greater impact on the lexicon of Kriol than other languages of the region. 16 The salience measure utilises both number of tokens and position in free-lists, taking into account that taxa listed first or earlier are more salient to interviewees than those listed later.

Rethinking the substrates of Roper River Kriol: the case of Marra

169

4 Conclusion The above sections provide evidence from four domains – historical, kin terminology, verbs and ethnobiology – that Marra has had more to offer lexically to Roper Kriol than any other local languages. Looking at each domain in isolation does not necessarily provide a compelling argument, but the findings are consistent enough across each domain that they do appear to show that the lexicon of Kriol has been disproportionately influenced by Marra and, to a lesser extent, other languages of the Marran family (Alawa and Warndarrang). It is less apparent that other languages such as the Gunwinyguan languages to the immediate north and west – Nunggubuyu, Ngandi and Ngalakgan – have made significant contributions to the Kriol lexicon. The data used to arrive at this finding is new data gathered on recent fieldwork, drawn primarily from young Kriol speakers with little knowledge of traditional Aboriginal languages. It demonstrates that previous research and documentation of Kriol had under-described the occurrence and frequency of non-English based lexemes in Kriol. In particular, the prevalence of verbs derived from local languages (especially Marra and other Marran languages) is somewhat unexpected given the common notion that verbs are weaker candidates than nouns for borrowing and lexical transfer. This new information allows us to more accurately consider substrate influences on Roper Kriol and re-assess tools used to predict substrate influence such as the Transfer Constraints approach. The Transfer Constraints approach (Siegel 1999, 2008) predicts substrate features that will and will not occur in a creole, regulated by two main principles: availability constraints and the reinforcement principle. Availability constraints determine which features would be used by individuals in a contact language situation, based on aspects such as whether the feature is perceptually salient and whether there is a congruent feature in the target language. Once a substrate feature becomes a variable in the pool of features of an emerging creole, the reinforcement principle determines whether it will be retained. The reinforcement principle considers how typologically common a feature is across substrate languages, thereby determining whether speakers would reinforce the use of the feature and lead it to become a feature of the creole. Munro (2004, 2011) applied the Transfer Constraints approach to Roper Kriol and used the reinforcement principle of frequency to develop predictions of which substrate features would be retained in the creole. To determine the frequency of features, Munro (2004) assessed a reduced selection of languages – Ngalakgan, Nunggubuyu, Alawa and Marra – and omitted others (Ngandi, Warndarrang) citing typological similarities and negligible roles in linguistic

170

Greg Dickson

ecology. The local language(s) spoken most widely today, Ritharrŋu/Wägilak, were also omitted assumedly because speakers of those languages were latecomers to the creolising communities. The languages selected were then surveyed in a way that suggests they have equal influence; that they constituted a level playing field. The data presented in this chapter brings the picture of substrate influences on Roper Kriol into higher resolution by (a) examining primary and secondary historical information pertaining to Marra people and (b) examining the lexicon of Kriol for contributions from local languages. I have shown that the contact period between Marra speakers living on country and residents of the Roper River Mission using an emerging creole was more significant than had been previously thought. I also argue that the lexicon of Kriol shows that Marra and, to a lesser extent, other Marran languages (Alawa and Warndarrang) have contributed more to the lexicon of Kriol than other local languages. This builds the case that Marra is a more dominant or influential substrate language than other languages regularly named in the pool of substrates of Roper Kriol. These findings have the potential to inform future considerations of substrate influences on not only Roper Kriol but also on creolisation models such as the Transfer Constraints approach. It is possible that a more nuanced consideration of the linguistic ecology and sociohistorical context of creolisation periods could – as with the case of Marra – determine variable levels of influence. Another way to determine variable levels of influence is to examine key domains within the lexicon for substrate-derived items, as has been done in this study. In the present study it has been particularly useful to make extensive use of data from young Kriol speakers with little knowledge of substrate languages to accurately assess the lexical contributions of those substrates. It has also been useful to look beyond nominal word classes where substrate-derived lexemes are most salient or thought to be more commonly found. It should be noted though that lexical influence does not necessarily correspond with grammatical influence. For example, Hawai’i Creole includes a number of Japanese derived lexemes but little grammatical influence is evident. This is attributable to Japanese people comprising a large proportion of the population only after creolisation (Siegel 2000). Despite the Hawai’ian example, evidence from the lexicon is still a potential indicator of broader influences. Given a sufficiently nuanced examination of potentially variable substrate influences, the Transfer Constraints approach could be tailored to account for variable levels of influence among substrate languages by weighting them differently, rather than apply a subset of them in a dichotomous or binary manner.

Rethinking the substrates of Roper River Kriol: the case of Marra

171

Acknowledgements The first manifestation of this paper was an essay for a course on Pidgins and Creoles delivered by Marlyse Baptista at the 2011 Linguistic Institute, University of Colorado at Boulder. It was then presented at the Language Contact Workshop (ANU, Canberra) in 2014 run by Carmel O’Shannessy and Felicity Meakins. These three people have been key to its development, so thank you. Thanks also to residents of Ngukurr and Minyerri who have taught me much about their lives and languages. In particular: Jason Farrell, Dwayne Rogers, Betty Roberts, Freda Roberts†, John Joshua, Maureen Thompson†, Cherry Daniels, Arnold George, Glen Blitner, Norma Joshua†, Kamahl Murrungun, Anthony Daniels, Aaron Joshua, Daniel Wilfred, Dorianne Roberts, Amelia Huddleston, Angelina Joshua, Tom E Lewis, Naomi Wilfred, Cleo Wilfred† and many more. Thank you too for professional support and wisdom from Nicholas Evans, Sophie Nicholls, Salome Harris, Inge Kral, Murray Garde, Harold Koch, Johanna Rendle-Short, Sarah Cutfield, Maïa Ponsonnet, Christian Döhler, Julia Colleen Miller, David Nash, Jane Simpson, Eugenie Collyer, Fleur Parry, Brighde Collins and Pascale Dettwiller among others.

Abbreviations A ALL EMPH EXCL NEG NEUT NSG PL POT PRS

noun-class marker allative emphasis exclusive negative neuter non-singular plural potential present

PST PUNCT REDUP REL SG TR

1 2 3 –

past punctual reduplicate relative singular transitive first person second person third person morpheme break

References Baker, Brett, Kate Horrack, Rachel Nordlinger and Louisa Sadler. 2010. Putting it all together: Agreement, Incorporation, Coordination and External Possession in Wubuy (Australia). LFG10. 64–84. Bickerton, Derek. 1981. Roots of language. Ann Arbor: Karoma.

172

Greg Dickson

Bickerton, Derek. 1984. The language bioprogram hypothesis. Behavioral and Brain Sciences 7(2). 173–222. Blythe, Joe. 2010. Self-Association in Murriny Patha Talk-in-Interaction. Australian Journal of Linguistics. 30(4). 447–469. Collins, Emma. 1998. Freda’s Story. Aboriginal and Islander Health Worker Journal. 22(3). 24– 27. Deakin University. 2004. Blekbala Stori. Geelong: Deakin University (Faculty of Arts). Dickson, Greg. 2015. Marra and Kriol – the loss and maintenance of knowledge across a language shift boundary. Canberra: Australian National University. PhD Thesis. Evans, Nicholas. 2003. Context, Culture, and Structuration in the Languages of Australia. Annual Review of Anthropology. 32. 13–40. Garde, Murray. 2013. Culture, Interaction and Person Reference in an Australian Language: An Ethnography of Bininj Gunwok Communication. Amsterdam/Philadelphia: John Benjamins Publishing. Graber, Philip. 1987. The Kriol particle na. SIL Working papers in Language and Linguistics. 21. 1–21. Harris, John. 1986. Northern Territory Pidgins and the Origins of Kriol. Canberra: Pacific Linguistics. Harvey, Mark. 2012. Warndarrang and Marra: A Diffusional or Genetic Relationship? Australian Journal of Linguistics. 32(3). 327–360. Heath, Jeffrey. 1978a. Linguistic diffusion in Arnhem Land. Canberra: Australian Institute Of Aboriginal Studies. Heath, Jeffrey. 1978b. Ngandi grammar, texts, and dictionary. Canberra: Australian Institute of Aboriginal Studies. Heath, Jeffrey. 1980a. Basic materials in Warndarang: grammar, texts and dictionary. Canberra: Pacific Linguistics. Heath, Jeffrey. 1980b. Nunggubuyu myths and ethnographic texts. Canberra: Australian Institute Of Aboriginal Studies. Heath, Jeffrey. 1980c. Basic materials in Ritharngu: grammar, texts and dictionary. Canberra: Pacific Linguistics. Heath, Jeffrey. 1981. Basic Materials in Mara: Grammar, Texts and Dictionary. Canberra: Pacific Linguistics. Heath, Jeffrey. 1982. Nunggubuyu dictionary. Canberra: Australian Institute Of Aboriginal Studies. Hudson, Joyce. 1983. Grammatical and semantic aspects of Fitzroy Valley Kriol. Darwin: Summer Institute of Linguistics, Australian Aborigines Branch. Joynt, R. D. 1918. Ten years’ work at the Roper River Mission station, Northern Territory, Australia. Melbourne: Church Missionary Society. Langford-Smith, Keith. 1936. Sky Pilot’s Last Flight. Sydney: Angus & Robertson. Lee, Jason (ed.). 2004. Kriol-Ingglish Dikshenri. http://www1.aiatsis.gov.au/ASEDA/docs/0739Kriol/index.1.html. (2 October, 2014). Lee, Jason and Greg Dickson. 2003. State of Indigenous Languages of the Katherine Region. Katherine: Diwurruwurru-Jaru Aboriginal Corporation. Meakins, Felicity and Carmel O’Shannessy. 2012. Typological constraints on verb integration in two Australian mixed languages. Journal of Language Contact (5). 216–246. Merlan, Francesca. 1981. Land, Language and Social Identity in Aboriginal Australia. Mankind 13(2). 133–148. Merlan, Francesca. 1983. Ngalakan grammar, texts, and vocabulary. Canberra: Pacific Linguistics.

Rethinking the substrates of Roper River Kriol: the case of Marra

173

Munro, Jennifer. 1995. Kriol on the move: An investigation to the spread of a creole language in Northern Australia. Armidale, NSW: University of New England. Honours dissertation. Munro, Jennifer. 2004. Substrate Language Influence in Kriol: The Application of Transfer Constraints to Language Contact in Northern Australia. Armidale, NSW: University of New England. PhD dissertation. Munro, Jennifer. 2011. Roper River Aboriginal language features in Australian Kriol: Considering semantic categories. In Claire Lefebvre (ed.), Creoles, Their substrates and language typology, 461–487. Amsterdam; Philadelphia: John Benjamins. Nicholls, Sophie. 2009. Referring Expressions and Referential Practice in Roper Kriol (Northern Territory, Australia). Armidale, NSW: University of New England. PhD dissertation. Nordlinger, Rachel. 1998. A Grammar of Wambaya, Northern Territory (Australia). Canberra: Pacific Linguistics. Numamurdirdi, Topsy. 1980. Evidence. Limmen Bight Land Claim. Transcript of Evidence. Canberra: Australian Government Printer. Olney, Justice. 2002. Maria Island and Limmen Bight River Land Claim (Claim No 71) and part of Maria Island Region Land Claim (Claim No 198). Darwin: Office of the Aboriginal Land Commissioner. Rhydwen, Mari. 1996. Writing on the backs of blacks. St. Lucia, Qld: University of Queensland Press. Sandefur, John. 1979. An Australian creole in the Northern Territory: A description of NgukurrBamyili dialects (Part 1). Darwin: Summer Institute of Linguistics. Sandefur, John. 1985a. Aspects of the Socio-political History of Ngukurr (Roper River) and its effect on Language Change. Aboriginal History. 9(1). 205–219. Sandefur, John. 1985b. Dynamics of an Australian Creole System. Papers in Pidgin and Creole Linguistics. 4. 195–214. Siegel, Jeff. 1999. Transfer Constraints and substrate influence in Melanesian Pidgin. Journal of Pidgin and Creole Languages. 14(1). 1–44. Siegel, Jeff. 2000. Substrate influence in Hawai’i Creole English. Language in Society. 29(2): 197–236. Siegel, Jeff. 2008. The emergence of pidgin and creole languages. New York: Oxford University Press. SIL-AAIB. 1996. Kriol Dictionary: With Kriol to English and English to Kriol: First Draft. Darwin: SIL-AAIB. Sharpe, Margaret. 1972. Alawa phonology and grammar. Canberra: AIATSIS. Sharpe, Margaret. 2001. Alawa Nanggaya Nindanya Yalanu rugalarra: Alawa-Kriol-English Dictionary longer edition. Adelaide: Caitlin Press. Taylor, John, John Bern & Kate A. Senior. 2000. Ngukurr at the millenium: a baseline profile for social impact planning in south-east Arnhem Land. Canberra: Centre for Aboriginal Economic Policy Research, Australian National University. Thomason, Sarah, and Terrence Kaufman. 1988. Language contact, creolization, and genetic linguistics. Berkeley; University of California Press. Winford, Donald. 2003. An introduction to contact linguistics. Oxford: Blackwell. Wohlgemuth, Jan. 2009. A typology of verbal borrowings. Berlin: Mouton de Gruyter. Wurm, Stephen. 1971. Classification of Australian Languages, including Tasmanian. In Thomas Sebeok (ed.) Current Trends in Linguistics (Volume 8): Linguistics in Oceania. The Hague, Paris: Mouton. 721–778.

IV Transfer of form: Phonological

Rikke Bundgaard-Nielsen and Brett Baker

6 Fact or furphy? The continuum in Kriol Abstract: The Australian language Kriol is spoken by at least 20,000 people across northern Australia. Creoles (including Kriol, see Sandefur [1979, 1984, 1986] following DeCamp [1971]) are often described as existing on a continuum, ranging from ‘basilectal’ (substrate-like) to ‘acrolectal’ (superstrate-like) speech. Indeed, Sandefur used a continuum model to explain high levels of variation in the pronunciation of Kriol lexical items in the 1970s. The source of the creole continuum in other contact situations is argued to be a sociolinguistic situation where creole speakers are in continuous contact with speakers of the superstrate languages, which over time leads to great variation in the output of each speaker. We present findings from a series of three experiments examining Kriol phonology and language behaviour by native speakers: 1) an acoustic study of Kriol obstruents by different generations of speakers, 2) a perception study of (English and Kriol-like) obstruent discrimination by first language Kriol speakers, and 3) an acoustic study of English obstruent production by first language Kriol speakers. These studies converge on the same conclusion that, with the exception of the oldest generation, native Kriol speakers produce and perceive Kriol obstruents in a manner consistent with a single, invariable, phonological system, and that they transfer Kriol phonology and phonetics into the perception and production of English obstruents in a manner consistent with classic first language to second language transfer. Indeed, Kriol phonology cannot be described using a continuum model: it is internally consistent, shares traits with both sub- and superstrate languages, and is identical to neither. We argue that most previously observed inter- and intra-speaker ‘variation’ is an artifact of a diverse population of Kriol speakers, including many L2 speakers of Kriol exhibiting cross-linguistic transfer from their (substrate) native languages into their second language (Kriol).

1 Introduction Prevailing views of Northern Australian Kriol suggest that speakers’ language behaviour exists on a ‘creole continuum’, a situation in which speakers have command of a range of ways of expressing themselves, some closer to standard

178

Rikke Bundgaard-Nielsen and Brett Baker

English, others closer to the original pidgin, with many somewhere in between these two extremes (e.g. Schultze-Berndt, Meakins, and Angelo 2013). Other creole speech communities, such as those in Jamaica (DeCamp 1971) and Guyana (Bickerton 1973; Rickford 1987) have been described as existing in such a situation. In this chapter, we question the appropriateness of the ‘creole continuum’ model for Roper Kriol, the variety spoken around south-eastern Arnhem Land (Northern Territory), in particular with respect to phonology, which is where the majority of evidence for such a continuum has been put forward (e.g. Sandefur 1979, 1986; compare with Sharpe 1975; see also Fraser 1977 on the closely related dialect of Fitzroy Crossing, WA). We argue that the variability previously observed in the 1970s can be attributed to the effects of second language learning, and inter-generational transmission of these effects. We base our arguments on interviews and lexical surveys with literate speakers from three distinct generations, and a series of speech production and perception experiments summarized here: an acoustic study of the voicing characteristics of Roper Kriol obstruents by three first (L1) and a single older second language (L2) speakers, a study of obstruent perception by L1 Kriol speakers, and an acoustic study of the voicing characteristics of English by four Kriol speakers with different language learning histories. We discuss the importance of researching individual language histories in assessing the nature of intra- and inter-speaker variation, and thus the course of development of the language in the speech community. We conclude that Roper Kriol, for L1 speakers, is a language with little phonological or lexical variability beyond that normally observed in languages. In addition, we find no evidence for a Kriol continuum ranging from a pidgin-like phonological system to a Standard Australian English (SAE) – or Aboriginal English-like phonological system. Rather, SAE has been learnt as an L2 by several generations of Kriol speakers and maintains its status as a completely separate L2 code in the Kriol-speaking community (consult Sandefur 1986: 44). There is no evidence of inter-generational transmission of English, by Kriol speakers. In the following sections, we give a brief history of the formation of Roper Kriol (Section 2.1), the claims that have previously been made of Roper Kriol phonology (Section 2.2) and especially its characterization as a ‘continuum’ (Section 2.3). We then provide a brief outline of an alternative framework for the understanding of Roper Kriol and for the previously observed variability in Kriol phonology (Section 2.4). Thirdly, we summarize the results from the first stages of a three-year project designed to elucidate the phonological structure of Kriol. Our discussion is based on the intuitions of literate native speakers (Section 3) and three experiments:

Fact or furphy? The continuum in Kriol

179

(1) An acoustic study of Kriol obstruents in elicited and spontaneous speech (based partly on Baker, Bundgaard-Nielsen, and Graetzer 2014), from both L1 and L2 speakers (Section 4); (2) A study of Kriol perception of English and Kriol-like stop voicing contrasts (Bundgaard-Nielsen and Baker 2014) (Section 5); (3) An acoustic study of English obstruent production by Kriol speakers with different language learning histories (Baker and Bundgaard-Nielsen 2014) (Section 6). We demonstrate that these three studies triangulate on a consistent view of Kriol phonology as a typical (rather than unusual) language system with characteristics from the substrate languages and the lexifier English, and that such multi-faceted approaches are essential to understanding a complex sociolinguistic situation such as the one we find in the current communities of Ngukurr and Numbulwar.

2 The creole continuum, and its application to Northern Australian Kriol 2.1 Roper Kriol Roper Kriol is a major variety of the largest Indigenous language in Australia. It is spoken in western, central and southeastern Arnhem Land (northern Australia) in a number of Aboriginal communities including Ngukurr, Beswick, Barunga and Numbulwar, which is the community where the majority of the research for this paper was undertaken. Ngukurr and Numbulwar are situated in southeastern Arnhem Land; Ngukurr is a larger settlement on the Roper River, while Numbulwar is a further three hour drive north, on the mouth of the Rose River. The 2005 National Indigenous Languages Survey (AIATSIS 2005) estimates around 20,000 first language speakers of Kriol across northern Australia.1 Roper Kriol is an English-lexified creole language, meaning that its vocabulary is largely derived from English, but it is not mutually intelligible with SAE (Sandefur 1979, 1986) and differs from SAE grammatically in significant ways (Munro 2011; Sandefur 1979, 1986). Roper Kriol developed from a reduced form of English (a ‘pidgin’), becoming a creole when it became the L1 of a significant proportion of the speech community, though it is not clear when, how or why this shift took place. Early 1 The true number may be much higher, although it is very difficult to know because of the tendency of Aboriginal people to under-report their use of Kriol in favour of heritage languages.

180

Rikke Bundgaard-Nielsen and Brett Baker

work, by Harris (1986; perhaps under the influence of Bickerton 1975), claims that Kriol emerged ‘abruptly’ at Ngukurr (then Roper River Mission) soon after its establishment in 1908. Sandefur (1986), similarly, claims that Kriol had been the first language of four generations of Ngukurr residents by the 1970s. In contrast, Munro (2000) argues that Kriol emerged gradually, not just in the Mission but in the surrounding pastoral properties, firstly as an ‘expanded pidgin’ (an L2; Siegel 2008), and that children then began acquiring this expanded pidgin as a first language between the 1940s–1960s, likely partly due to the dramatic demographic shifts caused by the Second World War. The reasons for language shift in the area, however, remain obscure. A similar time-frame appears applicable to the other major dialect – Fitzroy Valley Kriol – which is spoken in the Kimberley region of Western Australia (Fraser 1977; Hudson 1983). Fraser (1977) divides speakers into two groups based on age; those under around 30 (in 1974) speak a creole as their L1, while older speakers use the contact variety as an L2. Based on interviews with several generations of Roper Kriol speakers, we suggest that Kriol was spoken in Numbulwar amongst some children (at least) from the early 1950s (see also Sandefur and Harris 1986), when the mission was founded at Numbulwar (then called Rose River Mission), but that Kriol was not the first language of all children at that time, and that many adults likely were still speaking Kriol as an L2 (Baker, Bundgaard-Nielsen, and Graetzer 2014). Before 1950, there was no settlement at Numbulwar and many of the later residents were living around Ngukurr, at least some of the time, where all of them appear to have been exposed to Kriol.

2.2 Characteristics of Kriol phonology Previous work on Kriol, primarily from the 1970s and 1980s, proposes that Kriol speakers have a phonological system which can be modeled in terms of a continuum which ranges from a basilectal system at one end to an acrolectal system at the other, broadly speaking. The basilectal system is described as being similar to that of the substrate languages; the acrolectal system is closer to English, while not entirely identical to it (Sandefur 1979: 29).2 In general, the basilectal system is claimed to have a single series of obstruents, non-contrastive for voice, no fricatives or affricates, a simple 3 or 5place vowel system, and to not allow onset clusters. The precise characteristics

2 This has also been explained as a ‘sociolectal continuum’, with continuum endpoints labeled (following the Indigenous terminology) ‘heavy’ and ‘light’, respectively. ‘Heavy’ Kriol is basilectal, while ‘Light’ Kriol is more mesolectal (Sandefur 1986: 50).

181

Fact or furphy? The continuum in Kriol

of the basilectal phonological system, of course, depends on which substrate languages are taken to be the basis for Kriol, but all of the likely substrate languages lack fricatives and affricates, and many of them also lack a contrast between two series of stops. A maximal inventory is presented in Table 1. The acrolectal end of the continuum is typically described as more English-like, or similar to Aboriginal English (Sandefur 1979; compare with Fraser [1977]). According to Sandefur (1984: 2), “these two end subsystems are connected by a series of implicationally related ‘leveling’ patterns. The end result is that a given word may be pronounced several ways [. . .] Every speaker utilizes several variants”. Table 1: Phoneme inventories in the Roper substrate languages Labial Fortis Lenis Nasals Laterals Tap Approximants

p b m

w

Laminodental

ApicoAlveolar

Apicopostalveolar

Laminopostalveolar

Velar

Glottal

t̪ d̪ n̪ l̪

t d n l ɾ

ʈ ɖ ɳ ɭ

c ɟ ɲ

k g ŋ

ʔ

ɻ

j

Notes: the lamino-dental series is limited to Nunggubuyu/Wubuy, Ngandi, Enindhilyakwa, and Ritharrngu. The fortis/lenis contrast, and contrastive glottal stop, are limited to Ngalakgan, Ngandi and Ritharrngu. Note that none of these languages has contrastive fricatives or affricates

One of the primary arguments for describing Kriol as consisting of two separate, and partially mutually exclusive, phonological systems arises from the observation that Kriol is characterized by a continuous switching back and forth between the two subsystems in the following way: The majority of Kriol words [have] several alternate pronunciations (e.g. jineg, jinek, sinek, sineik, sneik ‘snake’) [. . .] Except for the extreme heavy and light variations of some words, most Kriol speakers control virtually all pronunciations in their active everyday speech. No Kriol speaker speaks with a consistently light pronunciation. [. . .] With few exceptions, every stream of Kriol speech will contain some words with heavy pronunciations and some with light pronunciations. Within the same conversation and even within the same sentence, it is not uncommon for Kriol speakers to use more than one of the pronunciation alternatives (Sandefur 1986: 50).

Indeed, this variation is the primary reason that a continuum model (be it a postcreole continuum or a sociolectal continuum) has been invoked. The acrolectal end of the continuum is said to be the result of ‘English influence’ and evidence of ‘decreolization’:

182

Rikke Bundgaard-Nielsen and Brett Baker

As Creole [i.e. Kriol] remained in contact with English, it was continually being influenced by English. Aborigines began to perceive and reproduce sounds that occurred in English but not in Aboriginal languages (Sandefur 1979: 42).

A typical example of such supposed English influence, and decreolisation, can be found in observations of consonant clusters impossible in the substrate languages (e.g. current sneik [IPA] for English ‘snake’ versus reported earlier jinek [ɟinek]). Vowel diphthongs and monophthongs not present in the substrate languages are also given this analysis, along with the presence of obstruents (especially fricatives and affricates) not found in the substrate languages. The reported variants jinek and sneik illustrate all three of these influences. Later, Sandefur (1986: 49) claims that “virtually all [variation in Kriol] is systematic and explicable, and that Kriol should be considered to be a ‘dynamic continuum system’, but not a post-creole continuum nor an SLA continuum”, while also claiming that “the so-called Kriol decreolization continuum is in fact a Kriol-to-English interlanguage continuum” (1986: 65). Yet elsewhere (Sandefur and Harris 1986: 185), it is claimed that “Kriol is a continuum in the sense that there are a number of subsystems within it which are linked together by gradation rather than being discrete”. The contradictory nature of these and other statements presents a very difficult case to untangle, and likely also reflect changes in the author’s viewpoints over time. Finally, the use of a continuum model to explain the characteristics of Kriol may also be influenced by the fact that Kriol was perceived, by Kriol speakers themselves, as a ‘single linguistic system’: In [Kriol speakers’] view, pidgin and creole are not discrete varieties, the one spoken as a second language in contrast to the other which is spoken as a first language. Indeed, they are overlapping and interacting sections of the continuum of one language (Sandefur 1986: 50).

The characteristics of Kriol presented in the quotes above are common to the descriptions of Roper Kriol not just by Sandefur (1979, 1986), but also Sharpe (1975), and in the case of Fitzroy Valley Kriol by Fraser (1977) and Hudson (1983: 22). In the present chapter, we concentrate on the Roper Kriol system.

2.3 The sociolinguistic situation of other creole continua The application of a post-creole continuum model to Roper Kriol in an attempt to explain observed variance deserves careful consideration. In particular, attention must be paid to the systematic differences in the sociolinguistic setting and time-depth of the supposed decreolisation of Roper Kriol compared to other creoles, particularly in the Caribbean.

Fact or furphy? The continuum in Kriol

183

As outlined in Section 2.1 above, Roper Kriol is considerably younger than the Caribbean creoles, such as Jamaican Creole English (DeCamp 1971) and Guyanese Creole (Bickerton 1973; Rickford 1987), which have been among the primary sources in the theoretical literature on creole formation and change, and which have a history stretching back to the seventeenth century. Likewise, as discussed in the present section, Kriol does not appear to exist in a sociolinguistic situation where speakers have the necessary ‘continuous and extended’ access to the lexifier as is the case for many other creoles, nor the social stratification found in other creole-speaking societies and said to be a determining factor in the variation (DeCamp 1971). But even here, it is not the case that all Caribbean creoles exist in such a situation; we discuss some of these cases below. Note that, if the time-frame proposed by Harris (1986) for the development of Kriol is accepted, this would require both creolization and de-creolization within 100 years in the case of Roper Kriol; that is, three or four generations of speakers. Primarily on the basis of analyses of Caribbean creoles, the creole continuum is typically defined as “a cline of lexical, phonological, and grammatical features ranging from those closest to a standard form of the creole’s lexifier language (the acrolect) to those furthest from the lexifier language, and therefore most ‘creole-like’ (the basilect)” (Siegel 2008: 235). This definition is very close to the way in which Sandefur (in most places) describes Kriol. This notion of a creole continuum is usually attributed to DeCamp (1971), who used it to describe the dialect continuum in Jamaica between the broadest form of the local creole and the local version of Standard English. He says of Jamaica that: There is no sharp cleavage between creole and standard. Rather there is a linguistic continuum, a continuous spectrum of speech varieties ranging from the ‘bush talk’ or ‘broken language’ [. . .] to the educated standard [. . .] Each Jamaican speaker commands a span of this continuum, the breadth of the span depending on the breadth of his social contacts (DeCamp 1971: 350).

Since DeCamp, the continuum has been applied to other creole situations as well, in particular Guyana (Bickerton 1973, 1975; Rickford 1987). Importantly for DeCamp (1971), the continuum is seen as the result of decreolization – the gradual modification of the creole in the direction of the lexifier (Siegel 2008: 236) (for alternative views on the sources of variation in creole speech communities, see for instance Alleyne [1971], Escure [1982], Winford [1997]). Hence, the creole continuum is a ‘post-creole’ continuum, in societies where contact has been maintained between the creole and the lexifier language, for a period of time, presumably, sufficient to enable the creole to begin approaching the standard formally. This was typically the situation in societies where social barriers

184

Rikke Bundgaard-Nielsen and Brett Baker

had become more permeable, enabling movement of creole speakers upwards in the social scale. Unlike the situation in Jamaica, however, where continuous and extended contact with the standard form of the lexifier has affected the form of the creole over time, we find no parallel changes in the structure of Kriol due to continuous influence from English (although this research is not reported on here). Instead, it appears that SAE has remained a second language and a completely separate L2 code for adult L1 speakers of Kriol, primarily acquired through formal schooling (consult also Sandefur 1986: 44). The degree to which second dialects or standard forms of a language can become parts of a ‘continuum’ for speakers is a controversial issue, and would certainly deserve more study than we can afford it here; but see e.g. Escure (1982, 1997); Siegel (1997) for discussion, both of whom argue against the continuum model for creoles spoken in Belize and Melanesia, respectively. The sociolinguistic situation in Melanesia in particular is very similar in many respects to that in Kriol-speaking communities of northern Australia, because Melanesian creole varieties exist in parallel with English and their substrate languages, but examining the specifics would take us beyond the scope of this chapter. We argue here that previously observed variation in Kriol may not be the result of ‘decreolization’ as predicted by a post-creole continuum model, but rather the presence of significant numbers of Kriol speakers with other first languages in the context in which Kriol was first described in the 1970s.

2.4 Characteristics of Kriol speakers in previous work Section 2.3 suggests that Roper Kriol is unlikely to exist in a classical ‘postcreole’ sociolinguistic situation, with continuous and extended access to the lexifier resulting in decreolization over time. The consequence of this is, of course, that the continuum model may not be the best model to account for the characteristics of Kriol (or for any observed ‘variation’ in Kriol phonology). Rather, it is highly likely that the Kriol speakers who provided this ‘highly variable’ Roper Kriol output may not necessarily have been L1 speakers of Kriol, making a continuum account for the variation observed by Sandefur even less plausible.3 In an observation, Sandefur (1979: 18) notes that: 3 Sandefur bases his work on discussions and interviews with speakers of Kriol with highly variable language learning backgrounds. One of his main informants (CD, aged approximately 70) is well-known to the second author. She affirms that as a child, she was mainly spoken to, by her caregivers, in a number of traditional languages, rather than Kriol. She also claims to have spoken back to her caregivers in traditional languages, most of the time, placing this speaker firmly in the Gen1 category, discussed below.

Fact or furphy? The continuum in Kriol

185

Creole is not the mother tongue of everyone in the Roper River area. Many people, though mostly older people, speak their traditional Aboriginal language as their mother tongue. A third generation of Creole mother tongue speakers, however, is currently emerging.

Similarly, in his (1986) study, Sandefur writes: “Single-range Kriol speakers, [. . .] who speak Kriol as their mother-tongue, are for the most part younger than the mid-thirties” (1986: 45).4 This is the speaker group invited to participate in the present research. We refer to these speakers as ‘third generation’ (Gen3) speakers. We suggest, based on these observations from the 1970s by Sandefur, that at that time, the productions of Kriol in the community were highly variable precisely because not all Kriol speakers were L1 speakers. It is clear from these and other quotes by Sandefur that he did not carefully distinguish between the language acquisition histories of Kriol speakers when reporting their behaviour, despite being explicit about these differences in language behaviour between L1 and L2 speakers of Kriol in the community: There are, however, some Kriol speakers who tend to have consistently heavy pronunciation in Kriol. These are mostly mother tongue speakers of a traditional language who speak Kriol as a second language and who speak no (Aboriginal) English. (Sandefur 1986: 50).

In the research that follows, we clearly distinguish between the Kriol produced by L1 speakers of the language who acquired it from parents who were also L1 speakers, from that produced by speakers with other language acquisition histories. We show that these differences in acquisition history are systematically related to differences in linguistic behaviour. Specifically, we distinguish the Kriol spoken by residents of Ngukurr and Numbulwar according to their position in a temporal (generational) scale of acquisition and minimally between first generation, second generation, and third generation groups. We define these groups as follows: Generation 1 (‘Gen1’) speakers acquired Kriol as an L2, having already acquired a traditional (i.e. autochthonous) language as an L1. For most of our participants, this language was Wubuy/Nunggubuyu (Heath 1984) or one of the neighbouring languages such as Djambarrpuyngu from NE Arnhem Land (Wilkinson 1991) or Enindhilyakwa from Groote Eylandt (van Egmond 2012). They may have acquired Kriol in adulthood, or as older children. Gen1 speakers acquired their L1 from caregivers who were also L1 speakers of one or more Indigenous languages and typically (late) L2 speakers of some variety of Kriol 4 He differs here from statements elsewhere in the same publication where he claims (1986: 20), for instance, that Kriol (in the 1970s) was the “mother tongue of four generations” of speakers, and (1986: 43) “in the Roper River region, creolization took place at the turn of the century” (i.e. in the early 1900s).

186

Rikke Bundgaard-Nielsen and Brett Baker

or pidgin also. Gen1 speakers are typically older than 60, that is, born before the 1950s (when the Rose River mission was established in what is now Numbulwar). These speakers typically spent part of their childhood at Ngukurr or other Kriolspeaking settlements such as Mainoru or Mountain Valley where they were exposed to Kriol speakers. The degree to which they use Kriol on a daily basis, and their level of competence, varies a great deal, depending on their social networks and their residence patterns. Most would be dominant in a traditional language and use Kriol only at times with younger family. Generation 2 (‘Gen2’) speakers are those who acquired Kriol as their L1, usually alongside a traditional language, as a second L1, from caregivers who were primarily L1 speakers of a traditional language and L2 speakers of Kriol. In other domains of linguistics, these speakers would typically be referred to as simultaneous, or early sequential, bilinguals. They are likely to be dominant in Kriol. Many of these speakers are balanced bilinguals, equally comfortable in either Kriol or their traditional language. They also differ in subtle but important respects from ‘Gen3’ speakers in likely perpetuating characteristics of their parents’ substrate L1-influenced L2 Kriol phonological systems in their Kriol productions. In Numbulwar, community members born roughly between 1950 and 1970 fall into this category. Finally, we classify as Generation 3 (Gen3+) those speakers who acquired Kriol as an L1 from caregivers who were also L1 speakers of Kriol. The parent/ caregiver generation, for true Gen3+ speakers, may in turn belong either to the Gen2 or Gen3+ group. If the proposed language acquisition account of Kriol presented here is correct, this group is likely to display the same degree of consistency in their language behaviour (segmental perception and production) as speakers of any other (non-contact) language. Most people in Numbulwar born after 1977 fall into this category. We pay careful attention to the individual language learning histories of Kriol speakers because we believe that this will provide a more detailed and accurate understanding of the characteristics of Kriol, of the reported variation in Kriol, and, perhaps, a greater understanding of the relationship between Kriol and English. We also believe that such an approach will be useful in identifying the specific challenges speakers of Kriol face when using English as an L2. Paying careful attention to the language learning history of all speaker/listener participants also safeguards against conflating the performance of L2 and L1 users of Kriol in terms of both the speech perception and speech production studies reported below. Indeed, the results of the third experiment reported on here (see Section 6) amply illustrate the effect of L1 interference on a speaker’s L2 productions – and serve as a great (if inverse) example of the kind of variation that exists in the Kriol speaking community, due to the idiosyncratic language learning histories of the speakers rather than inherent variation in Kriol

Fact or furphy? The continuum in Kriol

187

phonology. We believe that it is likely that a substantial amount of the previously observed ‘variation in Kriol’ is an artefact of overlooked L2 acquisition phenomena; see section 4.3 for a dramatic illustration of this kind of Kriol from a Gen1 speaker. We have one final point to address with respect to the likelihood of a continuum model accounting for previously observed variation in Roper Kriol. While it appears unlikely that Roper Kriol exists in a post-creole continuum in a DeCamp-ian sense, and that variation in Roper Kriol should be accounted for by such a continuum rather than a language shift and second language acquisition phenomenon, the creole continuum itself also gives rise to the fundamental theoretical problem that the “two ‘ends’ of the supposed continuum [do] not match up with each other” (McEntegart and Le Page 1982: 10); in this case the two ‘ends’ consist of conflicting phonological systems, according to Sandefur (1979, 1984). Such a continuum model requires that a given listener is able to perceive and produce the phones at the acrolectal end, correctly perceiving and producing the distinction between, for instance, /p/ and /b/ while losing this ability at the basilectal end where no such distinction is reportedly made. Further complicating this task is the fact that any interpretation of others’ speech will be perpetually bogged down by the need for continuous assessment of which range of the continuum is being used – for each phrase and potentially even word – by the interlocutor. This leads to an additional problem for language acquisition: how can learners identify the targets for word-learning, and how is it possible to develop native phonological contrasts of a kind that can be ‘turned off’ and back ‘on’ again as speakers move up and down the continuum between basilectal and acrolectal phonological systems? Indeed, normal segmental perception is so highly automatic and largely outside conscious control (Diehl, et al. 2004) that it is impossible to imagine a speaker who would be (under this proposal) able to assign phones to a system that contrasted for voice in one part of their phonological system but to a single undifferentiated series in another, depending on their judgements (moment by moment, potentially) of the range of the system that was being used. We point this out only with respect to stop voicing, without going into the additional problems caused by the lack of affricates or fricatives proposed for the basilect, or the differences in vowel systems, phonotactics and many other differences. The results of the perception study presented in Section 5 suggest that Kriol speakers cannot, as the continuum model would have it, ‘switch on’ English phonological categories that are not supported by Kriol (like /s/ vs. /z/) in a sociolinguistic context where this would be appropriate, nor can they ‘turn off’ their expectation (and dependency) on a native Kriol phonetic realisation of shared phonological categories (in the case of /p/ vs /b/).

188

Rikke Bundgaard-Nielsen and Brett Baker

3 The Kriol lexicon In order to examine the obstruent inventory of Roper Kriol and assess the degree to which speakers use and accept a wide range of pronunciations of Kriol words, as previously reported by earlier researchers, we conducted an extensive lexical survey with literate speakers of Kriol (see Baker et al 2014 for full report). In doing so, we depart from the traditional approach of identifying minimal pairs to establish the phonological inventory of a given language. There is no wellestablished writing tradition in Kriol, and we cannot rely on minimal pairs defined by loaned English spelling for cognate words (such as beg for English ‘bag’ and bek for English ‘back’, suggesting a voicing-based final velar stop contrast; SIL 1996). If we were to rely on the loaned spelling patterns in an attempt to ascertain whether Kriol indeed does have a voicing-based velar stop contrast, we face four alternative interpretations of the facts: (1) True positive: Kriol does indeed have such a contrast, if the participant produces a /k g/ distinction; (2) True negative: Kriol does not have such a contrast, if the participant fails to differentiate final /k g/; (3) False positive: Kriol does not have such a contrast but the tested pair is coincidentally (at least sometimes) produced with VOT values that fall in the ‘correct’ VOT range according to the linguist’s L1 (English) expectations (see our discussion of the Gen1 Kriol speaker VOT values in Section 4.3) (4) False negative: Kriol does have such a contrast, but the phonetic realisation of the contrast fails to match the language specific expectations of the linguist (in the case where, for instance, the phonemic boundary of Kriol /k/ and /g/ falls in a different part of the VOT spectrum than does English, leading the linguist to experience ‘variable’ production). In most cases, determining which of the four possible interpretations is the correct one is difficult, though substantial work with multiple informants may alleviate the problems. However, the matter is further complicated by the fact that Kriol may indeed have the contrast in question, but not in the supposed minimal pair as the phonological relationship between Kriol words and their English cognates is idiosyncratic and the minimal pair in question may in fact be a pair of homophones with a loaned English difference in spelling (Baker, Bundgaard-Nielsen, and Graetzer 2014), see Table 3, below. The two main informants for our lexical survey were third generation Kriol speakers (38 and 42 years of age respectively). We also interviewed older and younger speakers, with a particular focus on the question of variation in the

189

Fact or furphy? The continuum in Kriol

pronunciation of Kriol words. These speakers all identified themselves as native speakers of Kriol, although our oldest participant reported not having learnt Kriol until middle childhood. All speakers, with the exception of the oldest Gen1 informant (see section 4.3), were very firm in their identification of the ‘correct’ form of the items discussed. Indeed, they rejected the proposal that variant forms exist. When offered variant forms (such as /brog/ for English frog: Sandefur 1986), all L1 speakers suggested that ‘only old people speak like that’, or ‘perhaps people from Beswick’. Our survey thus allowed us to identify the canonical pronunciation of a large number of Kriol words as well as systematic correspondences between English and Roper Kriol obstruents, and we were thus able to establish a tentative Kriol obstruent phoneme inventory (see Table 2: for full discussion see Baker, Bundgaard-Nielsen, and Graetzer 2014). Table 2: The obstruent inventory of Roper Kriol

Stops

Labial

Dental

Alveolar

Retroflex

pb

t ̪ d̪

td

(ʈ?) ɖ

Affricates Fricatives

Alveo-palatal

Velar

Glottal

kg tʃ dʒ

f

s

ʃ

h

The survey also resulted in the selection of Kriol words (see Appendix 1 and 2 for these wordlists) eliciting all of the obstruents that have correspondents in the English source words, for use in the acoustic recordings reported below. We selected approximately 100 words for the first study (Speaker 1 in the acoustic studies presented below), and a subset of about 40 words for use in the second acoustic study (Speakers 2 and 3 below). We did not systematically select words to elicit the two proposed retroflex stops. It is important to note that the phonological relationship between the English source words and their Kriol ‘cognates’ is not always consistent. Rather, the underlying form of each Kriol lexeme needs to be specified individually, and cannot be assumed to be a direct reflection of the corresponding English word (see Table 3). We suggest that this lack of systematic correspondence between the Kriol lexemes and their English cognates may also have contributed (in addition to L2 transfer phenomena) to the perception – typically by English-speaking researchers and linguist-missionaries – that Kriol lacks a systematic voicing distinction.

190

Rikke Bundgaard-Nielsen and Brett Baker

Table 3: Examples of correspondences between Kriol and English5 Segmental agreement between Kriol and the English source word a. b. c. d.

tisim krasim frog tjetj

/ti:sim/ /kɹasim/ /fɹog/ /tʃetʃ/

< < <
|z|) –267.44 267.31 4.02

< 2e-16 *** < 2e-16 *** 5.86e-05 ***

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1 Correlation of Fixed Effects: (Intr) time time –0.998 worderVA –0.030 0.006

249

Entrenchment of Light Warlpiri morphology

Appendix 3 Table 9: Transitive verbs appearing in the Light Warlpiri texts in 2005 and 2010 LW

gloss

LW

gloss

LW

gloss

bait-im bast-im do-im faind-im hab-im jeinj-im* kam-at-im lik-im look-im old-im shut-im straik-im tow-im

bite bust do find have change bring out lick look hold throw strike tow

blok-im* brouk-im draib-im fiks-im help-im jeis-im kat-im lift-im meik-im pantirn-im pud-im spark-im teik-im

block broke drive fix help chase cut lift make pierce put cause to spark take

bamp-im* dig-im* iid-im ged-im id-im kil-im lock-im mub-im pouk-im sii-im* stiil-im tip-im* taj-im*

bump dig eat get hit hit lock move poke see steal hit touch

* denotes a verb in the 2005 texts but not the 2010 texts. All verbs in the 2010 texts also occurred in 2005.

Appendix 4 glmer analysis of number of ergative case marker tokens in children’s Warlpiri texts, 2005 and 2010, AV and VA word orders; data in Table 5 dat.glmerwrlp = glmer(erg ~ worder + (1|speaker) + (1|story), family = binomial, data = dat) summary(dat.glmerwrlp) Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) [‘glmerMod’] Family: binomial (logit) Formula: erg ~ worder + (1 | speaker) + (1 | story) Data: dat AIC BIC 265.5 280.2

logLik –128.8

deviance 257.5

Scaled residuals: Min 1Q Median –3.1784 0.1775 0.3178

3Q 0.4407

df.resid 289 Max 1.3169

250

Carmel O’Shannessy

Random effects: Groups Name Variance speaker (Intercept) story (Intercept)

Std.Dev. 1.51024 0.07691

1.2289 0.2773

Number of obs: 293, groups: speaker, 26; story, 3 Fixed effects: Estimate (Intercept) 1.3887 worderVA 1.4210 —

Std. Error 0.3655 0.5659

z value 3.799 2.511

Pr(>|z|) 0.000145 *** 0.012045 *

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1

Appendix 5 glmer analysis of dative case allomorphs, in Light Warlpiri vs Warlpiri, 2005 vs 2010 dat.glmer = glmer(casevowela ~ language + data_time + (1|story) + (1|speaker), family = binomial, data = dat) summary(dat.glmer) Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) [‘glmerMod’] Family: binomial (logit) Formula: casevowela ~ language + data_time + (1 | story) Data: dat AIC BIC 218.2 231.4

logLik –105.1

Scaled residuals: Min 1Q –2.1680 –0.4933 Random effects: Groups Name story (Intercept)

deviance 210.2

Median –0.3599

df.resid 197

3Q 0.7425

Variance 0.1294

Max 2.6525

Std.Dev. 0.3597

Number of obs: 201, groups: story, 3

Entrenchment of Light Warlpiri morphology

Fixed effects: (Intercept) languageWrlp data_time —

Estimate 164.86433 2.54644 –0.08292

Std. Error 30.62935 0.36596 0.01525

z value 5.383 6.958 –5.437

Pr(>|z|) 7.34e-08 *** 3.45e-12 *** 5.42e-08 ***

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1 Correlation of Fixed Effects: (Intr) lnggWr languagWrlp –0.020 data_time –1.000 0.013

251

V Transfer of function, structure, distribution and semantics

Denise Angelo and Eva Schultze-Berndt

8 Beware bambai – lest it be apprehensive1 Abstract: The aim of this paper is to draw attention to the use of the Kriol particle bambai as an apprehensive, i.e. a main clause modal marker indicating that an event will potentially occur but is undesirable, with associated pragmatics of warning or threat. This use is considered an extension from the temporal/ sequential function of this particle that is widespread in creole languages of the Pacific, including Kriol. We consider two potential motivations for this functional extension: substrate influence and independent grammaticalisation. The first is plausible insofar as dedicated apprehensive markers are a common trait in Australian languages, including in those that are currently in contact with Kriol and/or have previously been considered potential substrate languages, such as in the account of creolisation for Roper Kriol by Munro (2004). Apprehensive markers are also found in languages which could have influenced the precursor pidgins in New South Wales and Queensland as they expanded northwards towards the Northern Territory. In fact we show that the apprehensive function of bambai is more widely distributed in pidgin and creole languages of Australia and the Pacific than previously assumed, which could even point, potentially, to an earlier development in Australia. The possibility of independent grammaticalisation cannot, however, be excluded, since parallel developments of temporal markers to apprehensive markers are attested in a number of geographically distant and unrelated languages. The plausible link between the two functions is the semanticisation of an invited inference from ‘event about to occur’ to ‘event to be avoided’. We conclude that these two motivations are not necessarily mutually exclusive, but that both are consistent with the idea of instantaneous grammaticalisation through substrate influence in creole genesis. This paper thus contributes to our understanding of potential diachronic sources of the cross-linguistic category of apprehensive as well as to an analysis of this function in the Kriol modal system and its origins, and a more nuanced picture of regional variation in Kriol temporal and modal expressions. 1 As young, green linguists new to the Northern Territory in the beginning of the 1990s, we were both in awe of Patrick’s encyclopaedic knowledge about aspects of Aboriginal languages and cultures. We benefited from Patrick’s many, generous and encouraging conversations as we passed through Darwin. Patrick also convened the Top End Linguistics Circle (TELC) during the 1990s and made it a welcoming venue for all of us who passed through. We hope this paper will be considered by Patrick as a fitting tribute to his long-standing interest in the mechanisms and outcomes of language contact.

256

Denise Angelo and Eva Schultze-Berndt

1 Introduction In this paper we will present evidence for a semantic extension of the temporal marker bambai (‘soon’, ‘later’, ‘then’) to the function of an apprehensive modal expressing undesirable possibility, in at least some of the varieties of Kriol spoken across northern Australia. An example is (1), where bambai occurs in the spontaneous Kriol translation of a Ngarinyman clause featuring a dedicated apprehensive marker; note that both the Ngarinyman and the Kriol clauses have main clause status. Ngarinyman (Ngumpin-Yapa, Pama-Nyungan; Northern Territory) (line 1) and Kriol (line 2): (1)

Ngaja=ngali bayalan guliyan garraga. APPR=1DU.INCL bite:PRS dangerous frill.necked.lizard It might bite you and me, the dangerous frill-necked lizard. Bambai hi bait-im mi. APPR 3SG.S bite-TR 1SG.O It might bite me. (Westside/Timber Creek; ER, Ngarinyman field notes ESB, 1994)

Kriol is an English-lexified creole language that exists as a chain of dialects ranging from the Kimberleys in the west to the Gulf of Carpentaria in the east. Local varieties of Kriol are recognised by speakers and are described under a variety of geolocational epithets by linguists, such as Roper River/Ngukurr Kriol, Bamyili/Barunga Kriol, Katherine Kriol, Daly River Kriol (or Ngan’giwatyfella), Westside Kriol, Fitzroy Valley Kriol, Kimberley Kriol, Barkly Kriol or Wumpurrarni English (Meakins 2014: 379–380; Schultze-Berndt, Meakins and Angelo 2013). Our data for this paper come mainly from Westside varieties of Kriol (around Timber Creek, west of the Victoria River) or from Eastern varieties of Kriol recorded in Binjari (an Aboriginal community just to the west of Katherine), further east in Barunga (an Aboriginal community and former government reserve called Bamyili in central Arnhem Land), and nearby Beswick (also known as Wugularr), Jilkminggan (formerly Djembere, an Aboriginal community at the headwaters of the Roper River) and Ngukurr (formerly the Roper River Mission) (see Map). Where not otherwise indicated, examples come from spoken language recorded by one of the authors. The Kriol data consist of audio-recorded spontaneous spoken discourse and some written or planned spoken sources such as school texts (predominantly from the former bilingual Kriol-English program at Barunga), health video voice overs undertaken by a Kriol interpreter from Ngukurr and a

Beware bambai – lest it be apprehensive

257

Kriol bible translation,2 Holi Baibul (Bible Society in Australia 2007). We will also draw on published accounts of other Kriol varieties as well as first-hand or published data from traditional Australian languages including speakers’ direct translations from their traditional language into Kriol. It should be kept in mind that our data cover all age groups in non-Westside Kriol data, but on the Westside it is based largely on recordings with older speakers who would have acquired one or more traditional languages before, or alongside, Kriol. Excluding the Kriol bible translation, our examples of bambai amount to 75 in total, of which 35 exhibit the apprehensive function. Our findings are in contrast to those of Munro (2004: 135; 2011: 475) who finds no evidence for the apprehensive category in the Roper Kriol variety, although she does identify it as one of the “semantic features” shared by all local substrate languages present at the Roper River Mission site in their tense/ mood systems, and therefore as a likely candidate for transfer into Roper Kriol. In addressing the question of whether the apprehensive use of bambai in Kriol is a case of substrate influence, or alternatively the result of independent grammaticalisation processes, we suggest it is most likely a combination of the two. The paper is structured as follows: In Section 2, we introduce the apprehensive as a cross-linguistic category and its various manifestations in traditional Australian languages, in particular those which could be considered substrate languages for speakers of Kriol or of its pastoral pidgin predecessors. Section 3 provides an overview of the functions attested for bambai (and related forms) in historic pidgins and present day creoles of Australia and the Pacific and shows that such reflexes of English by and by are used in apprehensive contexts in at least some other Australian and Pacific pidgins and creoles apart from Kriol. In Section 4, we discuss in more detail the distribution and functions (temporal and apprehensive) of bambai in Kriol, taking into account potential regional variation. In Section 5 we propose an explicit account of the semantic and pragmatic link between the temporal and apprehensive function, taking into account parallel developments attested in other languages and the role of substrate influence. A conclusion is presented in Section 6. 2 The variety of Kriol utilised in the Holi Baibul is considered comprehensible across the Kriolspeaking area, but it predominately reflects the speech of Kriol speakers of varieties to the east of Katherine, with most translators hailing from Ngukurr and Minyerri, but also with considerable input from Barunga and Beswick. However, the Holi Baibul does also include translation work undertaken by a speaker from the eastern Kimberley and, in addition, various editing, comprehensibility and style considerations have been implemented to optimise its readability and appropriateness as much as possible for a “pan-regional” Kriol readership. Nevertheless, most Kriol speakers themselves would identify the variety in the Holi Baibul as aligning most closely to Roper River dialects spoken at Ngukurr and Minyerri (Margaret Mickan p.c.).

258

Denise Angelo and Eva Schultze-Berndt

2 Apprehensive modality While apprehensive markers are a widespread category cross-linguistically, they have not received much attention in the typological and semantic literature on modality and clause-linkage. As a general characterisation, an apprehensive marker conveys the possibility of a state of affairs that is possible, but undesirable and best avoided, often in conjunction with a sentence specifying the action necessary (or to be avoided) to prevent this state of affairs. The category is amply attested in languages of the Amazon (Vuillermet 2013), Austronesian languages (Lichtenberk 1995) and Australian languages (e.g. Dixon 1980: 380–381; Verstraete 2005: 256–265). Other terms that have been employed to label this category include timitive (Palmer 2001 [1994]: 22), admonitive (e.g. Nordlinger 1998, Meakins and Nordlinger 2013), evitative (e.g. Heath 1984: 346; Merlan 1982: 147; Munro 2004, 2011), ‘lest’ marker (e.g. Austin 1981: 225; Blake 1979: 68–75) and ‘for FEAR’ (Donaldson 1980: 285–6). Lichtenberk (1995), the only general discussion of the apprehensive category (based on a survey of Oceanic languages), recognises a number of subtypes of apprehensive markers, distinguished by their syntactic status, as shown in Figure 1 below (cf. also Vuillermet 2013). i. apprehensional-epistemic function LEST clause can function as an independent clause and it expresses both possibility and apprehension. ii. pre-cautionary function LEST clause appears in a complex sentence and encodes apprehension causing situation, while main clause contains the precaution. Both situations are prominently encoded. Two sub-types are identified, on the grounds of causality: a. negative purpose (causal) If no precaution is taken, then the apprehension causing outcome will occur (if not X, then Y), or precautionary situation averts the apprehension causing one (X so that not Y). b. in case (non-causal) Precaution is taken should an apprehension causing situation arise (X in case Y). iii. fear function LEST clause is embedded as a complement of a predicate of fearing and encodes an undesirable situation.

Figure 1: Apprehensive subtypes identified by Lichtenberk (1995)

Markers of apprehensional-epistemic modality (subtype i) are compatible with main clauses. They can be characterised as mixed modals with both epistemic and attitudinal components, since they involve the elements of a speaker’s

Beware bambai – lest it be apprehensive

259

“degree of certainty about the factual status of a proposition” plus “his or her attitude concerning the desirability of the situation encoded in the clause” (Lichtenberk 1995: 293). In other words, apprehensives can be characterised as bouletic or teleological modals, relating to a participant’s desires or goals, in addition to their epistemic value. For example, the speaker in (1), by employing the markers ngaja (in Ngarinyman) and bambai (in Kriol), signals that the event of a frill-necked lizard biting them is not only in the realm of possibility, but also undesirable from her (and the addressee’s) perspective. Adverbial subordinators or ‘lest’ markers (Lichtenberk’s pre-cautionary function, subtype ii in Figure 1 above) indicate that the subject of the main clause is apprehensive of the event encoded by a subordinate clause. Such ‘lest’ markers can be restricted to negative purposive function (‘so that . . . not’) or may have a more general ‘in case’ interpretation which does not depend on the control of the subject over the event that is to be avoided (as in Take your umbrella in case it rains / so that it does not rain). Finally, apprehensive markers can be complementisers restricted in their distribution to main verbs of fearing (subtype iii in Figure 1 above). English lest is an example of a morpheme with both adverbial subordinating and complementiser functions (López-Couso 2007). A semantically related category is the aversive inflection (also sometimes called avertive) on nominals to indicate an entity to be avoided (2). Yidiny (Yidinic, Pama-Nyungan; Queensland): (2)

Yingu this.ABS

waguu-ja man-ABS

garrba-ng hide-PRS

bama-yida. people-AVERS

This man is hiding for fear of the [strange] people (i.e. so he will not be seen). (Dixon 1980: 299; orthography adapted) In this paper, we will only be concerned with apprehensive modal markers since neither Kriol bambai nor the forms in the traditional languages to be considered here appear embedded under predicates of fearing, nor are they restricted to subordinating function. As we will see, the clauses containing these markers do exhibit strong links of pragmatic coherence with their immediate context, which often specifies the precaution to be taken in the form of a directive. Indeed, a cross-linguistic semantic link between the prohibitive and apprehensive modality is discussed in Pakendorf and Schalley (2007). This link is made in many of the language-specific characterisations of the category found in reference grammars of Australian languages, such as “This expresses the undesirability of an event, and the need to avert it. There is usually an implied injunction that the hearer be careful” (Evans 1995: 264–265).

260

Denise Angelo and Eva Schultze-Berndt

The apprehensive is widespread in Australian languages; however, the manner in which it is realised varies widely (see Dixon 1980: 459, who contrasts the variability of the apprehensive with the purposive “which has similar form in languages from every part of the continent”). Illustrative examples of specialized apprehensive markers in several Australian languages – and of some typical contexts – are presented in (1) above and (3) to (5) below. In Ngarinyman (1) the apprehensive marker is a particle. In Ngalakgan (3), it is a verbal prefix attached to a verb inflected for present tense. Nunggubuyu has both an apprehensive (“evitative”) inflectional suffix and a clitic -magi (analysed as a postposition and glossed as ‘lest’ by Heath (1984)), which can be used in combination with the suffix, as shown in (4). The inflectional suffix may also carry the modal meaning on its own, but on the other hand, the apprehensive function can be conveyed by -magi alone in combination with a non-past inflectional form (Heath 1984: 346). In Yukulta (5), the only marker is a suffix (or clitic) ‑marra. In all of the examples below, the first clause specifies a precautionary measure that should serve to avoid the outcome described by the apprehensive clause. Ngalakgan (Gunwinyguan, non-Pama-Nyungan; Northern Territory): (3)

Wanyba NEG

rerre-ngini-gah camp-POSS-ALL

Ø-rabon-jih 3SG-come-FUT.NEG

ngu-meleh-bun. 1SG/3SG-EVIT-hit:PRS

He’d better not come to my camp (or/lest) I strike him. (Merlan 1983: 97; orthography adapted) Nunggubuyu (Gunwinyguan, non-Pama-Nyungan; Northern Territory): (4)

“Ngirri-dabali-ny 2PL/1SG-remove:BEN-FUT

ngawaa-ngaa-ngun-magi!” 1SG/NC-bring.down-EVIT-lest

(Emu said) “Bring him out for me! Or else I will bring down the sky!” (Heath 1980: 45; orthography adapted) Yukulta (Tangkic; Queensland): (5)

Yararamatya-lati, whisper.IND-3PL.PRS

kunawuna-ntha child-DAT

tyirrmany-marra. wake.IND-LEST

They are whispering so as not to wake the child. (Keen 1983: 247; orthography adapted) Although apprehensives are a common trait of Australian languages, they are not always encoded by grammatical markers dedicated to this function. In

Beware bambai – lest it be apprehensive

261

some languages, a general irrealis category covers a broader range of meanings which may include the apprehensive function (see Verstraete 2005 for an overview of modal categories in non-Pama-Nyungan languages). This is the case, for example, in Wardaman, where the same irrealis prefix can express a judgement of (un)desirability (a deontic or bouletic category) when combined with a present tense suffix, and apprehensive modality (a judgment of undesirable possibility) when used without a suffix (Merlan 1994: 224). Similarly, in neighbouring Jaminjung, one of two irrealis modal categories covers the functions of negative ability, prohibitive (both in combination with negation), and apprehensive; the last two functions are illustrated in (6) (Schultze-Berndt and Caudal in prep). Jaminjung (Western Mirndi, non-Pama-Nyungan; Northern Territory): (6)

Gurrany NEG

bad step

yanj-inama, IRR.2SG>3SG-kick/step

majani maybe

lurr pierce

ya-niny-gijja. IRR-3SG>2SG-poke

Don’t step on it, (otherwise) it might poke you. (Elicitation; IP; ES97_A01_01.317) In addition to this array of inflectional possibilities, in some languages the apprehensive is also expressed – or differentiated from closely associated meanings – through the use of an invariant temporal marker, which is interesting given the semantic range of bambai that we will be examining later. We will discuss relevant examples in the context of the semantic link between temporal succession and apprehensive marking in Section 5. To summarise, main clause markers belonging to the cross-linguistic category of apprehensive modality – encoding an assessment of a state of affairs as potential and undesirable, and often carrying an illocutionary force of warning – are prominently represented in Australian languages, albeit in a variety of formal manifestations and semantic ranges. Importantly, they are attested in northern languages that are currently in contact with and/or were potential substrate languages for Kriol (as noted by Munro 2004: 135; 2011: 475). We will return to the question of the potential role of the substrate languages in the development of an apprehensive function of bambai in Section 5. First, however, we will provide an overview of the reflexes of English by and by in Australian and Pacific pidgins and creoles (Section 3) and of the functions of Kriol bambai (Section 4).

262

Denise Angelo and Eva Schultze-Berndt

3 Reflexes of by and by in English-lexified Creoles This section describes historical and contemporaneous uses of by and by reflexes in English-lexified pidgins and creoles, including those in and around the present day Kriol-speaking area, but also further afield. This material provides a background for comparison with the present uses of bambai in Kriol varieties (in section 4) and presents factors to be taken into account when positing likely mechanisms by which bambai has taken on an apprehensive function in Kriol (in section 5). A reflex of English by and by3 is classified as a “world feature” of Englishlexified creole languages by Clark (1979: 19–20), along with items such as along (comitative), been, got ‘have’, him ‘transitive suffix’, piccaninny ‘child’, plenty, savvy ‘know’, and where (relativizer). It is therefore to be found with the meaning of ‘later’, ‘eventually’ etc in historical sources as diverse as Mark Twain’s renditions of Gullah from the southern United States (e.g. Twain 1874: 592), to observations of coastal slave trade areas of West Africa (Mathews 1788: 166), and to records of Chinese Pidgin English in south-east Asia and the Pacific, including Australia (e.g. Siegel 2009: 316). Baker (1993: 34) notes that by and by is “extensively attested in pidgins and creoles around the world, including the Pacific” and indicates that it occurs with typical temporal meanings in his historical data from the Pacific. In clause-initial/final position, Baker (1993: 18) finds that by and by is attested in 22 Pacific locations prior to 1900, with the earliest token from Chinese Pidgin English in 1807, followed in order by Hawaii, New Zealand, New South Wales, California, the Marquesas, Fiji, Loyalty Islands, Tahiti, Queensland, Vanuatu, Solomon Islands, the Carolines, Rotuma, the German colony of New Guinea, New Caledonia, Papua New Guinea, Torres Strait, Kiribati, Samoa and Cook Islands.4 Other researchers have also found by and by with similar temporal meanings across the Pacific, in a nautical jargon associated with early maritime industries (e.g. whaling, sealing, sandalwood, bêche-de-mer)

3 For clarity, the form by and by will be used throughout the historical discussion, although readers should not assume this reflects the existence of a single or standard pronunciation, form or meaning. 4 Due to ongoing linguistic debates about the development of by and by as a future marker in Melanesian creoles, Baker also presents preverbal examples (feature 40) of which he finds only four in his entire historical corpus, and which could all be interpreted with temporal meaning (1993: 40–41).

Beware bambai – lest it be apprehensive

263

(Romaine 2004: 460–462), and have considered it to be a Pan-Pacific Pidgin feature (Keesing 1988: 32–33; Siegel 2011: 534). Most present day English-lexified creole languages spoken throughout the south Pacific basin have a reflex of by and by in a variety of related temporal meanings and functions, including in Australian Kriol (Harris 1986: 244; Sandefur 1984: 84), in Yumplatok, also known as Broken and Torres Strait Creole (Shnukal 1988: 113), and its dialect, Cape York Creole (Crowley and Rigsby 1979: 192), and in Yarrie Lingo (Yeatman et al. 2009). Expressions derived from by and by are also found in nearby pidgin and creole languages in Melanesia and the Pacific, including Bislama (Meyerhoff 2013: 226), Solomon Islands Pidjin (Keesing 1988: 184), Tok Pisin (Smith and Siegel 2013: 218) and Hawai’i Creole (Siegel 2011: 545; Velupillai 2013: 255). In contact varieties on the Australian mainland, by and by is represented with a temporal meaning in historical records from the convict settlement of NSW established by Britain in 1788. Troy (1994: 711) reconstructs the form baimbai from items such as bye and bye, bime bye etc in various historical written records, and determines that it belongs to the core vocabulary of New South Wales (NSW) Pidgin because it is attested across several sources from different regions. Troy (1994: 434, 711) describes it as an adverb encoding future tense (but glosses it as ‘eventually’). Further to the south, by and by is attested on Flinders Island in Bass Strait in 1837 and in the colony of South Australia in 1842 (Simpson 1996: 177, 186), with the usual temporal meaning. To the north, Dutton (1983) describes a bifurcated pathway of NSW Pidgin, into present day Queensland. NSW Pidgin was first transported to an area near the present site of Brisbane via a coastal route from around the 1820s, but by the 1840s, the pastoral industry had also made its way into southern Queensland via an inland route, bringing (a variety of) NSW Pidgin with it. In the historical Queensland sources located by Dutton (1983: 102–103), by and by has a similar temporal function to what Troy identified in NSW.5 On closer inspection, however, the temporal function in such historical sources is not restricted to future time reference; rather, the marker is also used to indicate temporal sequence (‘later, after that’) in clauses with past

5 by and by is represented as an English lexical entry, with a temporal meaning ‘soon’, in finder vocabulary lists in a number of modern day (salvage) grammars of Queensland Aboriginal languages (e.g. Breen 1981: 226; Williams 1980: 197). This is a likely indication that by and by has a degree of currency for speakers of those languages and/or Aboriginal people in those areas as a temporal expression. It is also evidence that by and by is not (strongly) associated with an apprehensive function by these speakers and/or linguists.

264

Denise Angelo and Eva Schultze-Berndt

time reference, much like in contemporary Kriol (see Section 4.1). Examples (7) and (8) are representative. Queensland Pidgin: (7)

My father said to him [Dalaipi], “You make the rain come and fill the holes again, Dalaipi.” He answered, “Byamby me makeim come.” [. . .] He answered, “I’ll make it come soon.” (south-east Queensland; mid-nineteenth century; Petrie 1904: 185, quoted in Dutton 1983: 111)

NSW Pidgin: (8)

His account was “Baal dat pfellar batter (eat). He bin gry, gry, gry all day and eat ‘im grass. Debildebil sit down long a dat pfellar, mine tink it. By-um-by put ‘im down gun, pick-um up grass, eat ‘im. Blackfellar come up behind – huh!”– illustrating by a blow of a fist on the back of his own head, and then by a quiver of his limbs, the death agony of poor Cunningham. [. . .] “That man hadn’t eaten (i.e. proper food). He kept on crying and crying and he ate grass. I think a ‘devil’ was in that man. Then he put his rifle down, picked some grass and ate it. An Aboriginal came up behind – huh!” [. . .] (Bogan River district, north-western NSW; reportedly 1870’s; White (1904), quoted in Nash: http://www.anu.edu.au/linguistics/nash/ aust/pidgin.html)

From NSW and southern Queensland, the English-based pidgin spread via the pastoral industry into northern Queensland, then westwards around the Gulf of Carpentaria, across the Northern Territory (NT) and into the Kimberley. In the early phases of non-Indigenous occupation, Queensland and the NT had a significant Chinese population in contact with Aboriginal groups through their collective involvement with mining and the pastoral industries. Records indicate that Chinese speakers of the NT pidgin, likewise, used by and by (Harris 1986: 177–178). Reflexes of by and by in temporal function are attested in records of Chinese Pidgin English from outside Australia too (Li and Matthews 2013). In any case, the records of early NT Pidgin, the precursor of contemporary Kriol varieties, present a similar picture to those for the earlier pidgins in Queensland and NSW. For this variety, too, by and by is described as a future tense marker (Harris 1986: 244), but, as shown in the narrated recount in example (9), it also has sequential function in clauses with past-time reference.

Beware bambai – lest it be apprehensive

265

NT Pidgin: (9)

This blackfeller, him away; him play about him all right, bye and bye him go cranky longa head. That fire, that mud, that star catch him. Next morning him feel all right. By and bye him say “Mee too much cold, want um big feller fire” [. . .] Then by and bye him get bad; no more eat um tucker and him die. This Aboriginal man, he was away (from his home/family); he was playing around but then he went mad in the head. The fire, the ‘mud’ and the star got him. The next morning he felt alright. Then he said ‘I’m very cold, I want a big fire’ [. . .] Then later he got bad; he didn’t eat and he died. (Northern Territory; Gee 1926: 35–36, quoted in Harris 1986: 332)

Melanesian Pidgins were influenced early on by NSW Pidgin via maritime activities, with Sydney the major port in the Southern Pacific (Baker 1993; Troy 1994; Romaine 2004; Meakins 2014). In addition, through later extensive use of Pacific Islander (Kanaka) labour on Queensland plantations, Queensland Pidgin influenced the development of Melanesian Pidgin and its descendants, Bislama in Vanuatu, Solomon Islands Pijin in Solomon Islands, Tok Pisin in Papua New Guinea and Yumplatok/Torres Strait Creole in the Torres Strait and Cape York Creole on the tip of Cape York (Crowley and Rigsby 1979; Shnukal 1988; Siegel 2011: 534). As a result, all of these creole languages use a form of by and by as a clause-initial or clause-final temporal marker, and by and by is attested in nineteenth century sources for all of them except for Yumplatok and Cape York Creole, for which historical data are not yet available. In three creole languages descendent from Melanesian Pidgins, Bislama, Solomon Islands Pijin and Tok Pisin (but not Yumplatok), by and by has followed a trajectory of development into a future/irrealis marker. In Tok Pisin, for instance, baimbai is now more common in reduced forms bai~b, and may appear in a preverb position (Siegel 2011: 544). The development of etymons of by and by from clause-initial to preverbal marker and concomitant grammaticalisation to a future marker is described in detail for Tok Pisin by Sankoff and Laberge (1973) and by Romaine (1995, 1998), who also captures considerable variability in the use of these forms. In addition to this fairly consistent picture of two related temporal functions (temporal sequence and future marking), descriptions of Australian and Pacific creole and pidgin languages offer tantalising but inconsistent glimpses of additional functions of by and by. In Tok Pisin, Romaine (1995: 410) notes that, in Pacific Pidgin data from the nineteenth century, half the tokens encode remote future, while others encode results “or later sequences of events [as in (8) and

266

Denise Angelo and Eva Schultze-Berndt

(9) above], some of which can be understood as warnings and threats, as they still do in modern Tok Pisin [as in (10)].” Tok Pisin (Papua New Guinea): (10)

Bai later

i 3SG/SRP

baimbai later (APPR)

go go

em 3SG

antap on.top i 3SG/SRP

tru really blek, black

tumach, too.much olsem like

blek. black

If it goes too far on top, it gets black, like black. (Indagen (rural setting), Morobe Province, 1987 fieldwork; Romaine 1995: 410) Further references to apprehensive uses of baimbai~bai~b have not, however, been located at this stage in published accounts of Tok Pisin (Mihailic 1971; Sankoff and Laberge 1973; Verhaar 1995; Smith 2008; Smith and Siegel 2013). In Yumplatok from the Torres Strait, Shnukal (1988: 113) treats bambai as an adverb with a purely future and sequential meaning ‘later (on), someday, eventually, afterwards’. Nevertheless, (11) and the translation provided by Shnukal clearly illustrate an apprehensive meaning, not a temporal one. Yumplatok (aka Torres Strait Creole, Broken): (11)

Ai 1SG

go MOD

go go

deya, there

bambai later (APPR)

ai 1SG

I’d better go there, or else I’ll miss the plane.

mes-e miss-TR

da DEF

plein. plane

(Shnukal 1988: 113)

Since orels (from English ‘or else’) is also attested in Yumplatok (Shnukal 1988: 77–78), there seems to be a partial functional overlap with bambai. In Cape York Creole, the related contact language spoken on the nearby northern tip of Cape York Peninsula, the form baimbai is described by Crowley and Rigsby (1979: 192) as one of a number of “pre-sentence modifiers” serving as an indicator of distant future (contrasting with klosap for the near future). Baimbai is not recorded by Crowley and Rigsby (1979: 205) in apprehensive function in this variety, whereas they find a different temporal marker, bipo (< ‘before’), with this role. By and by is also attested further afield in the Pacific in present day Hawai’i Creole where Siegel (2000: 201) has documented its use as a temporal since 1791, even earlier than Baker found (1993: 18, cited above). The history of Hawai’i Creole has some links with the maritime and plantation activities described above for Australia and Melanesia, but it also involved significantly different multi-

Beware bambai – lest it be apprehensive

267

cultural populations and historical processes, including the development and sustained use of a Hawai’ian-lexified pidgin (Roberts 2013) as well as the importation of large populations of foreign labourers (Siegel 2000). Siegel (2011: 545) describes by and by in both Hawai’i Creole and Kriol as adverbs meaning not only ‘later’ but also ‘otherwise’, a gloss capturing an apprehensive function. Siegel does not, however, regard such expressions as modals, since he claims that reflexes of by and by do not exist as a modal marker in Kriol and Hawai’i Creole (2011: 544). For Hawai’i Creole, Velupillai (2013: 255) analyses bumbye solely with a temporal function as a remote future tense marker which contrasts with the more general future tense expressed by go(i)ng/gonna. However, in (12) the expression of future possibility has the connotation of a warning, the pragmatics of an apprehensive clause. In addition to a temporal function, Sakoda and Siegel (2008: 536) explicitly assign modern bambai (with alternate spellings bumbye, by ‘m by) the function of “negative conditional” as an “adverbial connector”, as in (13). Note that (13) does not support the analysis of bambai as exclusively a marker of remote future. Hawai’i Creole: (12) Evri taim yu smok dis sigaret goin kil yu. Kam ober hea! every time 2SG smoke DEM cigarette FUT kill 2SG come over here Bambai des stik pok yu des stik kil yu. REM.FUT 3PL sticks poke 2SG 3PL sticks kill 2SG Every time you smoke those cigarettes (they) will kill you. Come here! Eventually those sticks will ruin you, those sticks will kill you. (Velupillai 2013: 255; orthography adapted) (13) Yu beta tek yo ambrela. Bambai yu get wet. 2SG better take 2SG umbrella later (APPR) 2SG get wet You’d better take your umbrella. Otherwise you’ll get wet. (Lum 1998: 225, cited in Sakoda and Siegel 2008: 536, example 87a) A number of on-line dictionary sources further attest apprehensive usages of bambai~bumbye. One beginners’ guide to Hawai’i Creole states: Bambai can also mean ‘otherwise’. If your friend is pigging out on junk food and you’re worried for their health, you could say, “Ho, no eat choke candy yah, bambai you gon get puka teeth” which means, “Wow, don’t eat too much candy, otherwise you’re going to get cavities.” (Riel 2013)

268

Denise Angelo and Eva Schultze-Berndt

In contrast to the sporadic appearances of apprehensives among the descriptions of the temporal function(s) of by and by and its derivatives in Pacific contact varieties, in descriptions of Norf’k from Norfolk Island the local reflex of by and by, bembeya, is only glossed as ‘lest’ and is listed as heading adverbial clauses, but only to express “purpose and result” with no temporal usage (Mühlhäusler 2010: 356–357, 2013: 238). Norf’k (Norfolk Island): (14) Nau dunt yu tuu gu medl en madl now don’t 2PL two go meddle and muddle en mitimiti orn d’-wieh t’-cherch bembeya yorlye kechet. and kiss on DEF-way to-church APPR 2PL get.in.trouble Now don’t you two go meddling and muddling and mitimiti (kiss and cuddle repeatedly) on the way to church in case you get into trouble. (Facebook discussion; Nebauer-Borg n.d.) There are some caveats that should be kept in mind concerning the data and discussion presented in this section. With regard to the historical data, historical records can be imperfect: The number of preserved historical records is fairly low and these are documented primarily by English speakers, often linguistically naïve, who might have preferentially heard and rendered English temporal meanings of by and by. With regard to grammatical descriptions of present day contact languages, the apprehensive function has not received the consistent attention that has been accorded to it in, say, descriptions of traditional Aboriginal languages. In summary, however, we can see that by and by and its reflexes are widespread items synchronically and diachronically in English-lexified contact varieties of the Pacific, where they are mostly recorded in temporal function associated with future time and sometimes with indicating a temporal sequence. There is also evidence – albeit of a somewhat sporadic and inconsistent nature – that reflexes of by and by have been harnessed to express apprehensive meanings in some of the creole languages of this region, despite assertions that it remains a purely temporal marker in these very languages. In addition to the range of functions just described, by and by has been involved in one grammaticalisation pathway in the south-western Pacific region, namely becoming a marker of future/irrealis in three Melanesian creoles/pidgins, but this is not directly associated with the apprehensive. In the next section, we will examine which functions are fulfilled by bambai in Kriol.

Beware bambai – lest it be apprehensive

269

4 Distribution and function of bambai in Kriol In this section we will examine the functions of bambai in varieties of presentday Kriol in northern Australia. In our Kriol data, bambai appears clauseinitially or clause-finally and has two broad functions, temporal and apprehensive. A clause-final position of bambai with temporal meaning is illustrated in (15); while (16) provides an example of bambai in clause-initial position with apprehensive meaning. (15) Ai show yu thad lilgel bambai, Denise, yu luk la im. 1SG show 2SG DEM girl later NAME 2SG look LOC/ALL 3SG I’ll show you (i.e. introduce you to) that girl later, Denise, you’ll see her. (Eastern/Beswick; spoken conversation [MS: P23–24]) (16) Ai gan lad-i yu insaid. Bambai yu dagat mi. 1SG CAN.NEG let-TR 2SG inside APPR 2SG eat 1SG [Dog: Little Pig, let me come in or I’ll destroy your house.] [First Pig:] I won’t let you in. You might eat me. (Eastern/Jilkminggan; written cartoon) We will illustrate the temporal function (4.1) of bambai before discussing in more depth the apprehensive function in different varieties (4.2).

4.1 Temporal function In its temporal function, in all varieties that we investigated, bambai can be used with future time reference, in two closely related functions: i. indefinite future, referring to an event at an unspecified time after speech time, often translated into English as ‘(sometime) later’ or ‘soon’ (as in (15)); ii. sequencing of an event with respect to a prior event, and often to be interpreted as a consequence of it (but with no implication of undesirability of that consequence), translatable as ‘then’ or ‘so that’ (as in (17)); Most frequently, the clause containing bambai has no additional overt temporal marking (i.e. is in present tense); thus the marker on its own conveys future time reference, as in examples (15) and (17).

270

Denise Angelo and Eva Schultze-Berndt

(17) “D D

en and

jeya there

M M

en and

B, B

luk look

from from

dat DEM

“Bambai (so).then

ai I

bud-a put-VTR

yu-mob 2-PL tri” tree

breig-a break-TR

thed-mob DEM-PL

im 3SG

laitha. thus

la LOC/ALL

tok talk

this DEM

kenggurru kangaroo

bushis branch

ba PURP

kug-um.” cook-TR

“D and M and B, break off those branches, from that tree there, look,” he would say, “So then I can put them with this kangaroo to cook it.” (ie. for the purpose of flavouring and keeping grit off). (Context: The speaker is illustrating how her husband might elicit assistance when preparing to cook a kangaroo in a ground oven). (Eastern/Binjari (ex Hodgson River); Staged reported speech, [MH: Tr17/9]) Overt future (garra~gada and variants) or irrealis markers (mait) in combination with bambai are quite frequently found in written texts such as the Kriol Bible and may reflect a certain degree of influence from Standard Australian English. Examples of bambai from spoken texts with future/irrealis marking are very rare in our data. Example (18), from a Kriol voice over for an information video about hospitals, is illustrative. The first version (a), featuring both bambai and the future marker gada, is the first written draft of a translation from English into Kriol, the first stage in preparing to record the voice over; while the second version (b) is a transcript of the actual spoken Kriol voice over, with no tense marking co-occurring with bambai. (18) (a) Wal, Robert im redi bla gu bla im opareishin na. well Robert 3SG ready for go for 3SG operation now Bambai im gada gud-bala soon 3SG FUT good-NMZR en im gada gu-bek then la im kantri. and 3SG FUT go-back then LOC/ALL 3SG country Well, Robert is ready to go for his operation. Soon he will be well and he will go back home. (collectively workshopped written (draft) translation from English; workshop notes) (b) Wal, tudei na, det Robert im gu bla im opareishin. well today now DEM Robert 3SG go for 3SG operation Bambai im nomo sik-bala soon 3SG NEG sick-NMZR en im lau bla gu-bek la im komyuniti. and 3SG allowed for go-back LOC/ALL 3SG community

Beware bambai – lest it be apprehensive

271

Well, Robert is going in for his operation today. Soon he’ll be well enough to return home. (English voice over) (Eastern/Ngukurr; planned spoken language; AH. [GTH41(51)]) In the eastern varieties – but not, according to our corpus data, in the Westside variety – bambai is also found with the overt past tense marker. In the past tense, bambai is only attested in clause-initial position. Two related functions can be distinguished. i. sequencing with respect to a prior event, sometimes to be interpreted as a consequence of it, translatable as ‘then’, as in (19). ii. resuming or refocussing on the main thread of a narrative as a discourse particle, often accompanied by additional discourse markers such as wal ‘well’ and na (used to mark a shift in topic or temporal frame) in (20) below. This function, more frequently found in written data, is labelled “resumptive” here and could be conveyed by ‘so anyway, . . .’, ‘now, . . .’, ‘well, . . .’ in English. When used resumptively, bambai (along with any associated discourse markers) is often in a separate intonation unit from the following clause, which in written texts is commonly rendered with a comma. (19) Mela bin ol mub deya naaa, sidan deya naaa, 1PLEX PST all move there now sit/stay there now bambai elikopta bin kam deya na, thad-lot deya na gada kemra. then helicopter PST come there now DEM-PL there now with camera (Context: During a flood people, including the speaker, moved to a dry place on a road). We all moved there and stayed there (for quite a while), then a helicopter came there, those people with a camera. (Eastern/Beswick; spoken narrative; [MS. P17]) (20) Im-bin brabli kleba-wan langa ola naja-lot enimul. 3SG-PST very wily-NMZR LOC/ALL PL other-PL animal Tharran na im det sneik. Wal bambai na, det sneik DEM:NOM FOC 3SG DEM snake well then FOC DEM snake bin go langa det wuman, en im-bin tok [. . .] PST go LOC/ALL DEM woman and 3SG-PST say It (i.e. one particularly sly animal) was the wiliest of all the animals. That animal was the snake. Well, the snake went to the woman and it said [. . .] (Kriol Bible Translation; Jenasis 3.1: http://aboriginalbibles.org.au/Kriol/Conc/root.htm)

272

Denise Angelo and Eva Schultze-Berndt

The degree of regional differentiation in the manifestations of temporal and modal marking more generally remains to be investigated in greater depth. Regarding bambai, the following tentative generalisations are offered here and summarised in Table 1. In Westside Kriol, bambai is restricted to final position in temporal function (which is moreover attested less frequently than the apprehensive function in the data) and it always has the sense of ‘(some time) later’ with future time reference. In contrast, all temporal functions of bambai – future, sequential (with both future and past reference), and past discourse/ resumptive marking – are attested in eastern dialects of Kriol, including the Holi Baibul, a text associated primarily with eastern Kriol varieties. In these eastern varieties, bambai appears clause-initially in all of these temporal functions. In the sense of ‘(sometime) later’, it may also occur clause finally (as in 15 above). Table 1: Bambai in temporal function: Range of meaning and clause position according to dialect dialect area

indefinite future clause position

Westside Kriol

✓

FINAL

eastern Kriol varieties including Holi Baibul

✓

FINAL & INITIAL

sequential clause position

✓

INITIAL

resumptive clause position

✓

INITIAL (EXTERNAL)

4.2 Apprehensive function As an apprehensive, bambai appears in a number of discourse contexts where a speaker is alerting the hearer to the potential occurrence of an undesirable event, and, possibly, to ways of avoiding this consequence. All instances of spoken bambai analysed here occur in conversation or within a narrative. All apprehensive uses of bambai are clause-initial in all Kriol varieties examined. Even though apprehensive clauses have future time reference (event time > speech time) or posterior time reference (event time > reference time, as in (24) and (27)), no additional future/modal marker is present, except for one case of co-occurrence with a future/irrealis modal, and one with an ability modal (22). Likewise, even when a narrative is in the past, overtly marked by auxiliary bin in the main clause, the bambai clause is in present tense (not overtly marked). One way bambai is used in its apprehensive function is in admonitory speech acts, which can fall anywhere along a continuum from suggesting (sensible) precautions through to giving direct warnings or even making threats. In admonitory

Beware bambai – lest it be apprehensive

273

usages, bambai often (but not necessarily) combines with a directive or prohibitive clause encoding some evasive action which could be undertaken to avoid the unpleasant outcome. Close English equivalents include ‘(do /desist p), or else q’ or ‘watch out, maybe q’, as in (21) to (23). This is probably the most frequent usage of bambai in spontaneous spoken language. (21) Ey! Bambai yundubala breik-im thet motika, liv-im. Hey! APPR 2DU break-TR DEM vehicle leave-TR Bambai Dadi graul la yu. APPR Dad scold LOC/ALL 2SG “Hey! (Watch out or else) you two might damage that car, leave it (alone)! (You’d better watch out or) Dad will tell you off!” (Westside/Timber Creek; DB; ES99_V05_02) (22) En yu mait tok: “Nomo pud-a thet beibi feis-dan laithet and 2SG IRR say PROH put-TR DEM baby face-down thus la pila. Bambai im gin safakeit.” LOC/ALL pillow APPR 3SG can suffocate And you might say: “Don’t put that baby face down like that on the pillow. Otherwise it might suffocate.” (Eastern/Binjari (ex Hodgson River); staged reported speech. MH. Tr13) As (21) and (22) show, the admonitory use is attested both in Westside Kriol and eastern varieties of Kriol. It is also amply attested in the Holi Baibul (23) (out of 965 tokens of bambai, 26 are clearly apprehensive), and is also found in contact varieties spoken in the Gulf of Carpentaria (see further below). (23)

Wal well

yu 2SG

bambai APPR

alb-um help-TR main 1SG:POSS

mi 1SG

na now

braja brother

Isau Esau

kil-im kill-TR

mi 1SG

ded. dead

Well you help me, lest my brother Esau kill me. (Bible Society in Australia 2007, Holi Baibul, Jenasis 32.11) In another use of bambai, the clause beginning with bambai encodes an undesirable potential outcome for a protagonist of the discourse (which may of course be the speaker, as in (16) rather than the addressee (or speaker and addressee); consequently, the admonitory illocutionary force is absent. Still,

274

Denise Angelo and Eva Schultze-Berndt

more often than not, the clause featuring bambai in this use is accompanied by a description of a precautionary measure which averts, or could avert, the predicted bad consequence. The cause and effect relationship between the preventative situation and the (lack of) undesirable consequence may require pragmatic reconstruction. For example, in (24), which has past time reference, the averred rising of the hill prevented the group of people potentially drowning on it. In the negative conditional in (25), it is only when a saltwater crocodile avoids taking on a buffalo single-handedly that a predictable negative outcome (from the point of view of the crocodile) is averted. Habitual burying of dead animals which are not eaten prevents an unpleasant smell in (26). In (27), the light does not directly prevent the children crawling into the flood waters, rather it allows the adults to supervise them more closely. English translation equivalents are often ‘otherwise’, ‘in case’ or ‘so that not q’. All these interpretations are however consistent with an analysis of the bambai clause as a main clause. The coherence between this clause and (usually) the preceding clause can be established at speech act level (‘do/avoid p, [I’m telling you because otherwise] it may be that q’ or (less frequently) at content level (‘p, [otherwise] q’, as in (24) and (26) (see e.g. Sweetser 1990, Stukker and Sanders 2012 for coherence at different levels). The function of bambai, in our analysis, is to flag proposition q as possible and undesirable, not to indicate the semantic link to a preceding context. (24)

Bat but

thad DEM

bambai APPR

hil hill alabat 3PL

du also

bin PST

oldei HABIT

gow-ap, go-up

lift-im-a, rise-TR-up

laithet, thus

draund. drown

But that hill also kept going up, rising, like this, otherwise they all would have drowned. (Eastside/Binjari (ex. Minyerri); Spoken narrative; [BR.Tr2.297]) (25)

Bat but

nomo NEG

bambai APPR en and

wan-bala, one-NMZR

jarran DEM

deig-im take-TR

bafalo buffalo

langa LOC/ALL

im 3SG

dreig-im drag-TR

im 3SG

atsaid out

plein. plane

But not (if there’s just) one (crocodile), otherwise that buffalo drags it out and takes it onto dry ground (and kills it). [Context: In this school text about animal behaviour – here, crocodiles and buffalos – the preceding scenario is many crocodiles encountering one buffalo, which favours the crocodiles, unlike this scenario]. (Eastside/Bamyili; Written text; Brennan 1983: 9)

Beware bambai – lest it be apprehensive

275

Jaminjung (Western Mirndi, non-Pama-Nyungan; Northern Territory) (line 1–2) and Kriol (line 3): (26) IP:

Nawij bag burr-angga-m, mirrbba burr-arra-m. neck break 3PL>3SG-get/handle-PRS buried 3PL>3sg-put-PRS

ESB: Mirrbba? buried IP:

Hm, bambai im thingk. hm APPR 3SG stink

[Context: about birds hunted by children] “They break its neck (by strangling it), (then) they bury it.” (Jaminjung) “Bury?” “Hm, otherwise it would stink.” (Westside/Kununurra; spoken narrative/ dialogue with linguist; IP, ES97_A01_01.079-083) (27)

Wi 1PLEX bambai APPR

bin PST dei 3PL

gada6 MOD

totj-im-bat torch-TR-CONT

ola PL

kid, children

yuno PART

krol-in. crawl-in

We had to keep shining the torch on the kids, you know, in case they crawled in.’ (Eastern/Beswick; spoken narrative; [MS.P13]) Table 2 summarises the distributional possibilities of bambai in its temporal functions and the apprehensive function, according to a coarse-grained division into Westside Kriol or eastern regional varieties. It should be noted that overall our data sets of bambai (in any function) are fairly small (see Section 1), which could be the result of many factors, mainly the scarcity of real conversational and interactional speech events. Moreover, apprehensive meaning can also be rendered with a more general modal (mait ‘might’), or indeed conveyed entirely pragmatically; thus, bambai is also not an obligatory feature of apprehensives. It is also possible that those Kriol expressions which are partially analogous to bambai, such as afta ‘after’, abdajad ‘afterwards’ (3SG-eat=3PL plant.food “You might eat up all the food on them.” Thet that

min, mean

bambai APPR

yu 2SG

finish-im finish-TR

taka food

fo for

MAI my

femili. family

“That means, you might finish the food (that was) for MY family.” (Jaminjung and Westside Kriol; staged conversation; IP, ES96_A08_02.073) Such “matchings” between substrate languages onto lexifier items form the cornerstone of the mechanism by which substrate influences enter creole languages. Such influences are widely recognised today, although there is some disagreement over the degree of substrate influence that should be assumed and also regarding the processes by which this occurs. Crucially for our discussion of bambai, direct relexification, as postulated by Lefebvre (1998, 2001), is not a possible explanation for the structure of many creole languages across the Pacific region, as they do not exhibit the same inflectional morphology as their Austronesian or Australian substrate languages. Here, substrate influence can still be recognised, but in more subtle ways. An approach that grapples with the structural differences between substrates and creole languages is the Transfer Constraints Approach developed by Siegel (1999, 2008, 2011; Siegel, Sandeman and Corne 2000). This approach predicts that a “match” – or transfer – of a substrate meaning to a lexifier form (akin to relexification) will only occur when a perceptually salient form exists in the lexifier – preferably a free form rather than a bound morpheme – which is recognised by speakers involved in the creolisation as having a function or meaning similar to a morpheme in the substrate (their L1). This is demonstrably the case for some bilingual Kriol speakers, who equate the free form bambai with the apprehensive structures in Ngarinyman (1) and Jaminjung (30). An additional prediction is that the transfer will only occur if the two forms are in a congruent syntactic position, which is not the case between the various morphosyntactic expressions of the apprehensive in Ngarinyman and Jaminjung and the many other possible substrate languages of Kriol when compared to the clause initial position of the apprehensive function of bambai in Kriol. A second pillar of the Transfer Constraint approach is the reinforcement principle. This explicitly takes into account that a “transferred” lexifier item will only become conventionalised, i.e. retained as a feature of a stabilised creole

280

Denise Angelo and Eva Schultze-Berndt

language, if it occurs with reasonable frequency across individuals during this process of stabilisation and levelling. The high frequency of a transferred feature in the developing creole language can be due to its high frequency in a single substrate language, or – importantly for our purposes – the existence of a similar feature in more than one of the substrate languages involved (Siegel 1999: 27– 29). In other words, a shared category across several substrate languages increases the likelihood of speakers finding a means to express this meaning in a pidgin or creole language and recognising it when others express this meaning. Certainly the apprehensive is a common feature of substrate languages in the Kriol-speaking area, and indeed in Australia more generally, and thus likely – according to this model – to exert influence on developing contact languages. Munro (2004) investigates the Transfer Constraints approach with regard to Roper Kriol and realises the need to broaden the application of the Transfer Constraint model in some respects. For instance, Munro (2004: 119–120) expands the model to encompass the transfer of semantic categories (as opposed to purely structurally defined characteristics) and acknowledges that these ought not be so strictly reliant on structural congruence. Munro (2004: 198–199) also responds to the discussion of Koch (2000) which highlights the likely role of early substrate languages in NSW in influencing the NSW and Queensland pidgins, and she finds that this is a relevant factor when analysing Roper Kriol. As a result of her examination of the substrate languages (taken to be Alawa, Marra, Ngalakgan and Nunggubuyu on the basis of historical demographic data about the Roper River Mission camps), she finds that the (augmented) Transfer Constraint model successfully predicts a number of features of Kriol. According to Munro (2004: 128–130), the category of apprehensive (Munro uses the term ‘evitative’) – which all four putative substrate languages for Roper Kriol possess – is a likely candidate for transfer into the development of Northern Territory Pidgin at Roper River and thence to Kriol. However, Munro does not find an apprehensive category in Kriol, perhaps because bambai does not actually have this function in her Roper Kriol data, or else because its apprehensive/evitative function has not been recognised (a factor in analyses of other creole languages, see section 3). She proposes that the absence of the expected apprehensive/evitative in Kriol is because the perceptually salient and available forms in English that could plausibly be reanalyzed as apprehensives – lest and or else – occur in structurally different positions from the (inflectional) apprehensives in the substrate languages. The reanalysis of a temporal into an apprehensive was not considered an expected pathway. There is also an issue with the plausibility of lest and or else as salient features in the communicative repertoire of Aboriginal people learning and using these pidgins at the time, or indeed

Beware bambai – lest it be apprehensive

281

in the speech of non-Aboriginal station workers.7 In any case, neither of these candidates was taken into (Roper) Kriol to express the apprehensive function, but, like both of them, bambai does not exhibit structural congruence with the substrate apprehensives. The presence of bambai as an apprehensive in Kriol, including eastern varieties such as Roper Kriol, shows that a differently positioned structural item can act as a host for transferred substrate meaning, and this host item can already be in place in the precursor pidgin (and therefore not immediately sourced from the lexifier). This meaning transfer need not have occurred in the Roper River region. The occurrence of apprehensive bambai outside the core Kriol speaking region (see Section 4), suggests that earlier substrate languages, in Queensland or even New South Wales, are plausible candidates for the transfer of an apprehensive function, although, to date, the available historical records yield evidence only for the temporal, but not for the apprehensive function of bambai (see Section 3). The apprehensive as a grammatical category is, however, found in a number of potential substrate languages in the path of the northwards expansion of the pastoral industry out of New South Wales (as mapped by Meakins 2014: 363), including Ngiyambaa (Donaldson 1980: 285–286), Yuwaalaraay (Giacon 2014: 468), Margany and Gunya (Breen 1981: 317–318, 328, 340), Pitta Pitta dialects (Blake and Breen 1971: 112–114), Yalarnnga (Breen and Blake 2007: 43), Kalkatungu (Blake 1979: 68–75), as well as Yukulta in the Gulf of Carpentaria region (Keen 1983; see example (5)). At any stage of the expansion of NSW Pidgin, the criterion of availability as a salient feature in the input is met: bambai in its temporal function clearly existed as an early pidgin feature in Australia, evidenced in a number of localities (see Section 3). Thus, nothing would prevent an association of an inflectional apprehensive with a free marker except a lack of semantic identifiability. Since the time and location of the apprehensive function entering the pidgin or creole has to remain speculation for the time being, it is the semantic link between temporal succession and apprehensive function that we will focus on in the following subsection. In this context, we also address the possibility of a grammaticalisation of bambai to apprehensive function independent of substrate language influence.

5.2 The grammaticalisation account A grammaticalisation account of the apprehensive function of bambai would take as its starting point the assumption that originally, bambai was transferred 7 Note that López-Couso (2007: 14), in historical corpora of English totalling a staggering 3.4 million words, found merely 322 attestations of lest.

282

Denise Angelo and Eva Schultze-Berndt

to NSW Pidgin in a function close to that of its English source, that of temporal succession, and underwent a grammaticalisation process, resulting in an extension of the marker to apprehensive function. Just as in the case of substrate influence, this independent development could have occurred at any stage during the expansion of the pidgin or after stabilisation of the creole, but we lack historical evidence to pinpoint this more precisely. Two arguments support a grammaticalisation account. First, a grammaticalisation process affecting erstwhile temporal bambai is actually attested for Melanesian Creoles such as Tok Pisin (Romaine 1995; see Section 3), albeit resulting in a more general future/modal marker rather than an apprehensive marker. Second, the grammaticalisation (or extended use) of markers of temporal succession to apprehensive markers has occurred in a number of unrelated languages, although this particular semantic change has not received much attention in the literature.8 Without claim to exhaustive coverage, we briefly discuss examples from Germanic languages, Pidgin Hawaiian, and Northern Australian languages. In colloquial German, a connective with a meaning of ‘afterwards, later’, nachher, has developed a strongly conventionalized apprehensive use when in clause-initial position and unstressed; both functions are illustrated in (31). For example, a web search for combinations of nachher with the verb gewinnen ‘win’ – expressing an event which in most cases would be associated with positive value – only yielded examples where the outcome of winning was evaluated as negative for some specific reason. German: (31)

(a)

Nachher later

(b)

Ich 1SG am PREP

gehen go:1PL.PRS

glaube, think:1SG.PRS Gewinn-spiel win-game

wir we ich 1SG

noch PART

essen. eat:INF

nehme take:1SG.PRS

lieber rather

nicht NEG

teil VPRT

8 This is not to claim that temporal markers are the only potential sources of apprehensive meaning. While the grammaticalisation of apprehensives remains an under-investigated area, other attested sources include verbs of seeing (Lichtenberk 1995: 303), prohibitives (Pakendorf and Schalley 2007), and connectives such as ‘or else; otherwise; if not’ which explicitly indicate the likely occurrence of an event in the case certain precautions are not taken. Lichtenberk (1995) claims that apprehensives more generally arise as complements of expressions of fear, but in light of the semantic links discussed in this section, such an account would seem too restrictive.

Beware bambai – lest it be apprehensive

Nachher Later/APPR und and

kann can

gewinne win:1SG.PRS das DEF

CL CL

ich 1SG

Finale finals

283

noch PART nicht NEG

zu at

Hause home

schauen. watch

I think I’d rather not take part in this lottery. I might win [a trip] and would not be able to watch the CL (Champions League [in football]) finals at home.’ (http://www.hifi-forum.de/viewthread-144-7707-4.html) A closely parallel phenomenon in Dutch, involving the temporal adverb straks ‘very soon, immediately’ is described in detail by Boogaart (2009), although with no reference to the cross-linguistic category of apprehensive. In an unrelated language, Pidgin Hawaiian, a connective meaning ‘later’, likewise, can have an apprehensive function, as illustrated in (32). Pidgin Hawaiian: (32) (a) Mahope hele aku maua me Kauiaa. later go DIR 1DU with Kauiaa Later Kauiaa and I left (the house). (b) Wau olelo iaia noho malie, mahope huhu kela kahunapule. 1SG speak 3SG stay quiet later (APPR) angry DET priest (That woman Auroria returned to the church with some girls [. . .].) I told her to be quiet, or else the priest would get angry.’ (Pidgin Hawaiian; naturalistic written data; Roberts 2013, APICS online, ex. 71–79) For our deliberations about bambai, we consider it significant that an extension of a marker of temporal succession to apprehensive function can be found in some Northern Australian languages which could be considered substrates for Kriol. In Marra, considered by Dickson (this volume) as perhaps the most important substrate language for Roper Kriol, the apprehensive (in positive polarity) is mostly marked, not by a dedicated inflection (or inflectional series), but by a combination of inflectional features (Heath 1981: 228, 230–247). The category is not distinguished formally in any way in the negative, where the sequential adverbial wuninggi ‘further, more, additionally’ with the future negative inflection conveys this sense (33a); reinforcement by wuninggi is also very common – although not obligatory – in positive apprehensive (evitative) clauses (33b) (Heath 1981: 187).

284

Denise Angelo and Eva Schultze-Berndt

Marra (Marran, non-Pama-Nyungan; Northern Territory): (33)

(a)

Wurla go:IMP

rnarriya-yurr, DEM.M.SG-ALL

wuninggi further

ngula NEG

rninggu-way 3SG>2SG-give.FUT

Go to him, or else he won’t give it to you! (b)

ngananggu-wa, 2SG>1SG-give.IMP

wuninggi further

rang-ningg-anjiyi hit/kill-1SG>2SG-AUX.EVIT

Give it to me, or else I will hit/kill you!’ (Heath 1981: 187; adpated orthography) It seems quite likely, therefore, that wuninggi, while retaining its temporal function, is already in the process of becoming re-analysed as an essential element for conveying the apprehensive/evitative function. Heath (1981: 187) explicitly comments on the parallel with “local creole English, where baymbay [i.e. by and by] means later, in a little while but is also common in ‘lest’ clauses.” In another language, Mangarrayi, situated at the headwaters of the Roper River, a dedicated particle barlaga ‘lest’9 which is transparently related to the reduplicative adverb barlarlaga ‘right now, today’, marks the apprehensive (evitative anticipatory) function, coupled with an array of non-past verbal marking (Merlan 1982: 146–7). The process of co-opting a temporal to apprehensive function seems to be even further advanced here than in Marra. Mangarrayi (Marran/Gunwinyguan, non-Pama-Nyungan; Northern Territory): (34)

(a)

Barrgji hard

rnama 2SG.hold.IMP

barlaga APPR

nya-way-(y)i-n! 2SG-fall-MPASS-PRS

Hold on tight lest you fall!’ (or: Hold on tight, or else you might fall.) (b)

Garlaji quickly

nganghma 2SG.ask:IMP

barlaga APPR

Ø-yag. 3SG-go:NPST

Ask him quick before he goes. (or: Ask him quickly or else he might go.) The semantic link between the temporal and the apprehensive function can be described by invoking the Invited Inferences model of semantic change (e.g. 9 Merlan (1982: 147) assigns two glosses and related functions, ‘lest’ and ‘before’, to this marker; the latter is applied e.g. to example (36b). We suggest that both can be reduced to the apprehensive function, especially considering that the related adverb does not convey a sense of anteriority; we have adapted glosses and translations accordingly.

Beware bambai – lest it be apprehensive

285

Geis and Zwicky 1971; Traugott and Dasher [2001] 2004: 34–40; Traugott 2004: 552–553). This model is based on the observation that a certain form, in particular contexts, may come with certain inferences, over and beyond their encoded meaning, which can be exploited (“invited”) by the speaker. Such inferences can become semanticised, i.e. become part of the encoded meaning, via a stage of conventionalisation of the inference. Such conventionalized Generalised Invited Inferences arise in contexts where the original and the new meaning are both present in the most likely interpretation and therefore ambiguity is unlikely to arise (Traugott 2004: 552–553). Temporal markers form a particularly well-described source of semantic changes based on such invited inferences, e.g. from temporal precedence to causation (English since), or from temporal simultaneity to condition (German wenn ‘when’). In the case of an extension of a marker of temporal succession into an apprehensive function, the invited inference consists of the pragmatic enrichment that an event that may be imminent (temporal) is undesirable and to be avoided (apprehensive). This additional semantic component – the undesirability of an event which is potentially about to occur – is described as “negative rhetorical direction” by Boogaart (2009). As Merlan (1982: 147) notes, a temporal expression expressing imminence, such as Mangarrayi barlaga < ‘now’ or Dutch straks ‘soon’, seems intuitively suitable to the expression of such an additional component. The conventionalisation of the implicature of undesirability may come about through frequent use of a clausal sequence in which the first clause has the illocutionary force of a directive, and the second is introduced by the temporal marker (see Section 4.2). For example, in (21), the directive ‘leave it alone’ is both preceded and followed by clauses spelling out two potential events, breaking the car, and being scolded by Dad. By the principle of relevance, the most plausible link between these clauses is to construe both potential events as undesirable, since this provides the motivation for the speaker’s directive, which in turn is construed as the instruction of how to best avoid the potential event. Moreover, a causal link has to be construed by the addressee between the two apprehensive clauses which depict breaking the car and being (subsequently) scolded. Of course directives can also be employed to induce the hearer to bring about a desirable consequence, and temporal markers can be employed in this case too, e.g. Catch some fish for us, then we can eat. Thus the semantic change to apprehensive function requires for a particular temporal marker (German nachher, Dutch straks, Kriol bambai) to be employed primarily in a context of a directive to avoid an obviously undesirable event. A diagnostic of semanticisation of the apprehensive function of the temporal marker, then, will be its potential to occur outside the directive context

286

Denise Angelo and Eva Schultze-Berndt

and even without explicit mention of the precautionary measure, and, furthermore, with an unambiguous interpretation of a negative consequence even if the context does not strongly suggest such an interpretation (such as in the German example in (31) above). In our Kriol data, the precautionary measure is frequently left unspecified or at most indirectly specified, but in most cases, the event described in the clause introduced by bambai is unambiguously negative. A less clear case is the following example (35). At first sight, learning one’s language should not be considered an undesirable consequence; however the context is that of the speaker working on a cattle station as a young girl where indeed she was discouraged from learning her language, since the station managers, who kept her in their household as a servant, considered this strengthening of her family ties an undesirable event. Westside Kriol (Timber Creek): (35)

I 3SG

neva NEG.PST

langa LOC/ALL Im 3SG

big big

reken think

let-im let-TR gel girl

mindubala 1DU

tokin talking

“Bambai APPR

yu 2SG

go go

na. SEQ len-im learn-TR

yu 2SG

langgwij”. language

She didn’t let us two go to the big girls to talk. She thought “You might learn your language”. (Spoken narrative; DB 2006, recorded and transcribed by Jenny Denton and Colleen Moerkerken) Incipient semanticisation, in this case restriction of the marker to apprehensive function, will also go hand in hand with diversification of distribution in cases where the original function is also retained. Such a diversification seems to have taken place in Westside Kriol. In our data of older Kriol speakers in and around Timber Creek and Kununurra, bambai in clause-initial position in untensed clauses always has an apprehensive function, while the temporal function is retained only for postverbal bambai. Boogaart’s (2009) discussion and examples of the “anticipated negative event” function of Dutch straks suggest a similar positional diversification. To summarise the preceding discussion, an independent development of a temporal marker, bambai, going through an independent process of semantic extension or temporal marker undergoing to a marker of undesirable possibility is plausible and is attested elsewhere. On the other hand, it seems absurd to discard

Beware bambai – lest it be apprehensive

287

the potential influence of the apprehensive function found encoded – albeit in a large variety of formal manifestations and paradigmatic positions – in the modal systems of the substrate languages and in fact as a wider areal feature, especially given evidence that individual speakers equate it with bambai in their traditional languages. As suggested by the Transfer Constraint model discussed in Section 5.1, such a widespread and stable semantic feature would be expected to transfer to a formal expression that is semantically related (or construable as related) in a pidgin or nascent creole language. Thus, the two accounts need not be considered mutually exclusive. The apprehensive function of bambai can potentially be accounted for by the concept of instantaneous grammaticalisation (Bruyn 1996), i.e. a “shortcut to grammaticalisation through direct calquing” (Keesing 1991: 331). In this account, the existing model in a substrate language motivates speakers to find means of encoding the same category, not by direct transfer but by recruiting, and extending to a grammatical function, a lexical form from the lexifier, following universally attested grammaticalisation paths (see also Plag 2002; Heine and Kuteva 2005). Of course, in the case of bambai, as we have noted elsewhere, it is the pidgin that has provided the “form”, not the lexifier per se. Intriguingly, the availability of temporal markers exhibiting an extension towards apprehensive function in at least two immediate substrate languages of Roper Kriol (wuninggi ‘later’ in Marra; barlaga related to barlarlaga ‘right now, today’ in Mangarrayi) could have facilitated the encoding of an apprehensive function by bambai in Roper Kriol. Of course, great care has to be taken in interpreting such material in terms of causality and directionality. It is, for instance, also possible that speakers of Marra and Mangarrayi – who at the time of linguistic documentation all had proficiency in Roper Kriol or its pidgin predecessor (Merlan 1982: xii), and who therefore used bambai in apprehensive function as Heath (1981: 181) observed – reinforced the propensity of these temporals to co-occur in apprehensive structures in Marra and Mangarrayi, thus raising them into greater prominence in these languages. Declining speakership in the 1970s (when Heath and Merlan undertook fieldwork) and associated phenomena, such as overregularisation, might have promoted or accelerated the influence of Roper Kriol structures on these traditional languages. Similarly, Hawai’ian Pidgin mahope ‘later’ (example (32)) might have developed temporal and apprehensive meanings which then transferred to the convenient candidate bambai in Hawai’ian Pidgin English, or it is also possible that the directionality of influence was reversed, and Hawai’ian Pidgin English speakers transferred temporal and apprehensive meanings of bambai to the closest available candidate in Hawai’ian Pidgin, mahope.

288

Denise Angelo and Eva Schultze-Berndt

6 Conclusions Evidence has been presented in this paper that bambai expresses apprehensive modality in varieties of Kriol. It is unsurprising that this function of bambai has not been identified previously, because, as already noted in Section 2, the category of apprehensive has not received much attention cross-linguistically. What attention bambai has drawn in Pacific creolistic writings largely identifies it just as a temporal connective and one of the core/common features of regional contact varieties or concerns its development as an irrealis/future marker in some Melanesian pidgins/creoles, which may well have obscured its apprehensive tendencies elsewhere. What is more, other structures can convey apprehensive meaning, such as more general modals (e.g. Kriol mait ‘might’); as a consequence, dedicated apprehensive forms may not necessarily occur in language data and this may disguise the need to seek them more rigorously. Furthermore, Kriol bambai has retained its full form and has not developed a reduced form as in the trajectory charted for reflexes of by and by in future/irrealis function in some Melanesian pidgins/creoles, thus rendering the extension to a modal function in Kriol less obvious. In Kriol, moreover, bambai is positioned (often well) outside the verb phrase, and so is not considered a natural candidate for expressing modality (although other epistemic modals like maitbi (< might be) also occur outside the verb phrase). Such formal and structural factors may have prevented a description of the apprehensive function of bambai in earlier analyses. It is clear that bambai has a long history of being dispersed via jargons and pidgins in its meaning of ‘(sometime) later’, which is common worldwide. Admittedly this temporal meaning might be purely due to what English speakers (and primary historical recorders) were able to note, as it is the meaning close to recorded English usages. In any case, bambai is now clearly used in apprehensive function in some modern day English-lexified contact languages of Australia and the Pacific, including several Kriol varieties (but perhaps not in the Kimberley), some Queensland Gulf varieties, Torres Strait Creole (but not Cape York Creole), Hawai’i Creole and Norf’k. In the creoles that use bambai apprehensively, bambai has not grammaticalised into a future marker. This observation may prove to be robust, and a mutually exclusive distribution may eventually be demonstrated which could indicate a split grammaticalisation pathway for bambai, with one branch of semantic extension becoming a future marker, the other, apprehensive. Outside of the Pacific, bambai does not seem to have developed as an apprehensive, despite its wide distribution as a temporal connective, so this might lend weight to the possibility that the apprehensive

Beware bambai – lest it be apprehensive

289

innovation for bambai spread to recipient speech communities in the region, rather than independently grammaticalising in each place, but this is by no means certain: We have seen, for instance, that other neighbouring languages have utilised (unrelated) temporal markers for apprehensive functions (e.g. German, Dutch in Section 5.2). We also showed that the functions of bambai have diverged in eastern and western varieties of Kriol: the western varieties show positional diversification between the temporal (clause-final) and the modal function (clause-initial), and do not use the temporal function in combination with tense marking and in a sequential reading (Section 4). With regard to how bambai acquired its apprehensive meaning, in support of substrate influence, we have examples of individual speakers directly equating Kriol bambai with the apprehensive structures in Australian Aboriginal languages in codeswitching utterances (Section 5.1). On a larger scale, however, we cannot know for sure where bambai took on apprehensive meaning: The northwards path of the pastoral pidgin which carried bambai through NSW, Queensland and into the NT passed through the hands (or rather mouths) of speakers of a great number of Australian Aboriginal languages, many with apprehensive structures of various kinds (Sections 2 and 5.1). We tentatively propose that evidence is mounting that bambai was gaining its additional apprehensive function before it was carried into the NT, both because of its distribution on the eastern “fringes” of Kriol, as well as its mutually exclusive distribution with the future marker bambai in Melanesia. Although substrate influence is likely to have played an important role in providing pools of speakers familiar with expressing an apprehensive category, harnessing bambai as an apprehensive is not predicted by current substratist models due to both meaning and structural mismatches with the substrate languages. The development of a temporal bambai into a marker of apprehensive modality therefore also depends on grammaticalisation processes. To this end, we have established a semantic link between temporal and apprehensive meanings, explained how the implicature of undesirability may have become conventionalised and provided examples of just such developments cross-linguistically (Section 5.2). Since the historical record is lacking, the precise grammaticalisation processes involved in how bambai accrued apprehensive function may not be proven definitively. On this basis, we have tentatively proposed that “instantaneous grammaticalisation” may be the best approach: It includes a role for substrate influence (motivating speakers to find a means of encoding the same category as expressed in their language(s)), and for grammaticalisation (recruiting a lexical form from the lexifier not by direct transfer but by employing it in a grammatical function following cross-linguistically attested grammaticalisation pathways).

290

Denise Angelo and Eva Schultze-Berndt

This discussion of bambai as an apprehensive highlights some of the conceptual issues of contact linguistics. Both substrate influence and grammaticalisation are likely to have played a part in the bambai story, the former process providing a meaning without a likely home, the latter providing the home, by means of a cross-linguistically attested and cognitively plausible pathway. It is likely, due to the very nature of language contact, that this involved recursive processes, with one feeding the other and vice versa, against a backdrop of broader social processes of population movements and new communication needs.

Acknowledgements This paper was presented at the Language Contact Workshop (6–7 March 2014) run by Felicity Meakins and Carmel O’Shannessy. The authors wish to thank the editors and one anonymous reviewer for valuable comments on the first draft of this paper. The first author would particularly like to acknowledge Barbara Raymond, Maureen Hodgson and Miliwanga Sandy who over the years provided marvelous narratives in Kriol (from which tiny excerpts appear in this paper), and who have given invaluable insights into this rich and vibrant language. Some Eastern Kriol examples are drawn from Kriol fieldwork undertaken with the assistance of the Australian Institute for Aboriginal and Torres Strait Islander Studies (AIATSIS) (grant G97/6054). Thanks are also due to the Living Archive of Aboriginal Languages (LAAL) which has included in its collection Kriol texts from the Barunga Literature Production Centre which provided materials for the Barunga Community Education Centre which operated the only official Kriol-English school bilingual program in the Northern Territory over 3 decades. The second author would like to thank the TAMEAL (Tense, Aspect, Modality and Evidentiality in Australian Aboriginal Languages) team, in particular Patrick Caudal, Martina Faller and Marie-Eve Ritz, for many enlightening discussions about tense and modality, and gratefully acknowledges financial support received from the Max Planck Society and from the Volkswagen Foundation (DoBeS Grant 82957) for the documentation of Jaminjung and Ngarinyman, and from the 7th Framework Programme of the European Commission for the TAMEAL project (Grant Agreement PIRSES-GA-2008-230818-TAMEAL). Our greatest debt is to all the speakers – of Kriol and of other Australian languages – that we have had the privilege to work with over the years.

Beware bambai – lest it be apprehensive

291

Abbreviations ABS ALL APPR AVERS BEN CONT DAT DEF DEM DET DIR DU EVIT EX FOC FUT HAB IMP INCL IND INF INT IRR LOC

absolutive allative apprehensive aversive benefactive applicative continous aspect dative definite article demonstrative determiner directional dual evitative exclusive focus future habitual imperative inclusive indicative infinitive interjection irrealis locative

M MOD MPASS NC NEG NMZR NPST PART PL POSS PRS PST PURP REM SEQ SG SRP TR VPART

1 2 3 – =

masculine modal mediopassive noun class negation nominaliser nonpast discourse particle plural possessor present past purposive remote sequential singular subject-referencing pronoun transitive marker verb particle first person second person third person morpheme break clitic break

References Austin, Peter. 1981. A grammar of Diyari, South Australia. Cambridge: Cambridge University Press. Baker, Philip. 1993. Australian influence on Melanesian Pidgin English. Te Reo 36. 3–67. Bible Society in Australia. 2007. Holi Baibul (Online Kriol Bible and Concordance). http:// aboriginalbibles.org.au/Kriol/Conc/treeMaster.htm (accessed 18 February 2015). Blake, Barry. 1979. A Kalkatungu grammar (Pacific Linguistics Series B, no. 57). Canberra: Research School of Pacific Studies. The Australian National University. Blake, Barry & J. G. Breen. 1971. The Pitta Pitta dialects (Linguistic Communications 4). Melbourne: Monash University. Boogaart, Ronny. 2009. Een retorische straks-constructie. In Ronny Boogaart, Josien Lalleman, Marijke Mooijaart & Marijke van der Wal (eds.), Woorden wisselen. Voor Ariane van Santen

292

Denise Angelo and Eva Schultze-Berndt

bij haar afscheid van de Leidse universiteit (SNL-reeks 20), 167 183. Leiden: Stichting Neerlandistiek Leiden. Breen, J. G. 1981. Margany and Gunya. In Robert M. W. Dixon & Barry Blake (eds.), Handbook of Australian languages (Vol. 2), 274–393. Canberra: The Australian National University Press. Breen, Gavan & Barry Blake. 2007. The grammar of Yalarnnga. A language of western Queensland. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The Australian National University. Brennan, Queenie. 1983. Deinja langa bush. Barunga, NT: Bamyili Press. http://laal.cdu.edu. au/record/cdu:36141/info/ (accessed 18 February 2015). Bruyn, Adrienne. 1996. On identifying instances of grammaticalization in Creole languages. In Philip Baker & Anand Syea (eds.), Changing meanings, changing functions. Papers relating to grammaticalization in contact languages, 29–46. London: University of Westminster Press. Clark, Ross. 1979. In Search of Beach-la-mar: Towards a History of Pacific Pidgin English. Te Reo 22–23. 3–64. Crowley, Terry & Bruce Rigsby. 1979. Cape York Creole. In Timothy Shopen (ed.), Languages and their status, 152–207. Philadelphia: University of Pennsylvania Press. Dixon, Robert M. W. 1980. The Languages of Australia. Cambridge: Cambridge University Press. Donaldson, Tamsin. 1980. Ngiyambaa. The language of the Wangaaybuwan. Cambridge: Cambridge University Press. Dutton, Tom. 1983. The origin and spread of Aboriginal Pidgin English in Queensland: A preliminary account. Aboriginal History 7(1). 90–122. Evans, Nicholas. 1995. A grammar of Kayardild with historical-comparative notes on Tangkic. Berlin: Mouton de Gruyter. Gee, Lionel. 1926. Bush tracks and gold fields. Adelaide: F.W. Preece and Sons. Geis, Michael & Arnold Zwicky. 1971. On invited inferences. Linguistic Inquiry 2. 561–566. Giacon, John. 2014. A grammar of Yuwaalaraay and Gamilaraay: a description of two New South Wales languages based on 160 years of records. Canberra: Australian National University. PhD dissertation. Harris, John. 1986. Northern Territory pidgins and the origin of Kriol. Canberra: Pacific Linguistics. Heath, Jeffrey. 1978. Ngandi grammar, texts and dictionary. Canberra: Australian Institute of Aboriginal Studies. Heath, Jeffrey. 1980. Nunggubuyu myths and ethnographic texts. Canberra: Australian Institute of Aboriginal Studies. Heath, Jeffrey. 1981. Basic materials in Mara: Grammar, texts and dictionary (Pacific Linguistics Series B no. 60). Canberra: Department of Linguistics, Research School of Pacific Studies, Australian National University. Heath, Jeffrey. 1984. Functional grammar of Nunggubuyu. Canberra: Australian Institute of Aboriginal Studies. Heine, Bernd & Tania Kuteva. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Hudson, Joyce. 1983. Grammatical and semantic aspects of Fitzroy Valley Kriol (Work papers of SIL-AAB, Series A, Volume 8). Darwin: Summer Institute of Linguistics. Keen, Sandra. 1983. Yukulta. In Robert M. W. Dixon & Barry Blake (eds.), Handbook of Australian Languages (Vol. 3), 191–305. Canberra: The Australian National University Press. Keesing, Roger. 1988. Melanesian Pidgin and the Oceanic substrate. Stanford: Stanford University Press.

Beware bambai – lest it be apprehensive

293

Keesing, Roger. 1991. Substrates, calquing, and grammaticalization in Melanesian pidgin. In Elizabeth Traugott & Bernd Heine (eds.), Approaches to grammaticalization (Vol. 1), 315– 342. Amsterdam: John Benjamins. Koch, Harold. 2000. The role of Australian Aboriginal languages in the formation of Australian Pidgin grammar: Transitive verbs and adjectives. In Jeff Siegel (ed.), Processes of language contact: Studies from Australia and the South Pacific, 13–46. Montreal: Fides. Language Perspectives. 2015. Mornington Island Poster. Cairns: Far North Queensland Indigenous Schooling Support Unit, Education Queensland. Language Perspectives. in press. Kowanyama Poster. Cairns: Far North Queensland Indigenous Schooling Support Unit, Education Queensland. Lee, Jason (ed.). 2004. Kriol-Ingglish Diksheneri. Kriol English Dictionary. Darwin: Summer Institute of Linguistics (SIL). www1.aiatsis.gov.au/ASEDA/docs/0739Kriol/index.1.html (accessed 18 February 2015). Lefebvre, Claire. 1998. Creole Genesis and the Acquisition of Grammar: The Case of Haitian Creole. New York: Cambridge. Lefebvre, Claire. 2001. The Interplay of relexification and levelling in Creole genesis and development. Linguistics 39(2). 371–408. Li, Michelle & Stephen Matthews. 2013. Chinese Pidgin English structure dataset. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber (eds.), Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. http://apics-online.info/contributions/20 (accessed 18 February 2015). Lichtenberk, František. 1995. Apprehensional epistemics. In Joan Bybee & Suzanne Fleischman (eds.), Modality in grammar and discourse, 293–328. Amsterdam: John Benjamins. López-Couso, María José. 2007. Adverbial connectives within and beyond adverbial subordination. The history of lest. In Ursula Lenker and Anneli Meurman-Solin (eds.), Connectives in the History of English, 11–29. Amsterdam: Benjamins. Lum, Darrell. 1998. Orphan Annie: Colouring in the eyes. In E. Chock, J. Harstad, D. Lum & B. Teeter (eds.), Bamboo Ridge: Growing up local. An anthology of poetry and prose from Hawai’i, 222–231. Honolulu: University of Hawai’i Press. Mathews, John. 1788. A Voyage to the River Sierra-Leone on the Coast of Africa; containing the account of the trade and productions of the country, and of the civil and religious customs and manners of the people; in a series of letters to a friend in England. London: B. White and Son & J Sewell. Meakins, Felicity. 2014. Language contact varieties. In Harold Koch & Rachel Nordlinger (eds.), The Languages and Linguistics of Australia: a comprehensive guide (The World of Linguistics, Volume 3), 365–416. Berlin: Mouton de Gruyter. Meakins, Felicity & Rachel Nordlinger. 2013. A Grammar of Bilinarra, an Australian Aboriginal language of the Victoria River District (NT). Berlin: Mouton de Gruyter. Merlan, Francesca. 1982. Mangarayi (Lingua Descriptive Series Vol 4). Amsterdam: North Holland Publishing Company. Merlan, Francesca. 1983. Ngalakan grammar, texts and vocabulary (Pacific Linguistics, Series B; no. 89). Canberra: Department of Linguistics, Research School of Pacific Studies, Australian National University. Merlan, Francesca. 1994. A Grammar of Wardaman. A language of the Northern Territory of Australia. Berlin: Mouton de Gruyter.

294

Denise Angelo and Eva Schultze-Berndt

Meyerhoff, Miriam. 2013. Bislama. In S. Michaelis, P. Maurer, M. Haspelmath, & M. Huber (eds.), The survey of pidgin and creole languages (Vol. I: English-based and Dutch based languages), 223–231. Oxford: Oxford University Press. Mihailic, F. 1971. The Jacaranda dictionary and grammar of Melanesian Pidgin. Milton, Qld: The Jacaranda Press. Mühlhäusler, Peter. 2010. Norfolk Island and Pitcairn varieties. In Daniel Schreier, Peter Trudgill, Edgar W. Schneider and Jeffrey P. Williams (eds.), The Lesser-Known Varieties of English: An Introduction, 348–364. Cambridge: Cambridge University Press. Mühlhäusler, Peter. 2013. Norf’k. In Michaelis, Susanne Maria & Maurer, Philippe & Haspelmath, Martin & Huber, Magnus (eds.), The survey of pidgin and creole languages (Vol. I: Englishbased and Dutch-based languages), 232–240. Oxford: Oxford University Press. Munro, Jennifer. 2004. Substrate language influence in Kriol: The application of transfer constraints to language contact in northern Australia. PhD dissertation, Armidale: University of New England. Munro, Jennifer. 2011. Roper River Aboriginal language features in Australian Kriol. Considering semantic categories. In Claire Lefebvre (ed.), Creoles, Their Substrates, and Language Typology, 461–488. Amsterdam: John Benjamins. Nebauer-Borg, Rachel. (n.d.). Reduplication in the Norf’k language (two) [Website]. www.facebook. com/permalink.php?id=1427777337438830&story_fbid=1497744690442094 (accessed 18 February 2015). Nordlinger, Rachel. 1998. A grammar of Wambaya, Northern Territory (Australia). Canberra: Pacific Linguistics. Pakendorf, Brigitte & Ewa Schalley. 2007. From possibility to prohibition: A rare grammaticalization pathway. Linguistic Typology 11. 515–540. Palmer, Frank. 2001 [1994]. Mood and modality (Second edition). Cambridge: Cambridge University Press. Petrie, Constance. 1904. Tom Petrie’s reminiscences of early Queensland, dating from 1837. Brisbane: Watson, Ferguson & Co. http://www.archive.org/details/cu31924063745495 (accessed 18 February 2015). Plag, Ingo. 2002. On the role of grammaticalization in creolization. In Glenn Gilbert (ed.), Pidgin and Creole Linguistics in the 21st century. New York: Lang. Riel, J. 2013. Beginner’s guide to Hawai’i Creole [audio] [Website]. http://matadornetwork.com/ abroad/beginners-guide-to-hawaii-creole-audio/ (accessed 18 February 2015). Roberts, Sarah. 2013. Pidgin Hawaiian structure dataset. In Susanne Michaelis, Philippe Maurer, Martin Haspelmath & Magnus Huber (eds.), Atlas of Pidgin and Creole Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. http:// apics-online.info/contributions/71 (accessed 18 February 2015). Romaine, Suzanne. 1995. The grammaticalization of irrealis in Tok Pisin. In Joan Bybee & Suzanne Fleischman (eds.), Modality in Grammar and Discourse, 1–39. Amsterdam: John Benjamins. Romaine, Suzanne. 1998. The use of bai in young children’s creolized Tok Pisin. In J. Verhaar (ed.), Melanesian Pidgin and Tok Pisin. Proceedings of the first international conference of pidgins and creoles in Melanesia, 189–203. Amsterdam: John Benjamins. Romaine, Suzanne. 2004. English input to the English-lexicon pidgins and creoles of the Pacific. In R. Hickey (ed.), Legacies of colonial English. Studies in transported dialects, 456–499. Cambridge: Cambridge University Press.

Beware bambai – lest it be apprehensive

295

Sakoda, Kent & Jeff Siegel. 2008. Hawai’i Creole – morphology and syntax. In Kate Burridge & Bernd Kortmann (eds.), Varieties of English (Vol. 3. The Pacific and Australasia), 514–545. Berlin: de Gruyter. Sandefur, John. 1984. Papers on Kriol: The writing system and a resource guide. Darwin: Summer Institute of Linguistics. Sankoff, Gillian & Suzanne Laberge. 1973. On the acquisition of native speakers by a language. Kivung 6,1. 32–47. Schultze-Berndt, Eva & Patrick Caudal. in prep. A semantic analysis of the modal system of Jaminjung. Ms, University of Manchester. Schultze-Berndt, Eva, Felicity Meakins & Denise Angelo. 2013. Kriol. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber (eds.), The survey of pidgin and creole languages (Vol. I: English-based and Dutch-based languages), 241–251. Oxford: Oxford University Press. Shnukal, Anna. 1988. Broken: an introduction to the creole language of Torres Strait (Pacific Linguistics C–107). Canberra: Research School of Pacific Linguistics, Australian National University. Siegel, Jeff. 1999. Transfer constraints in and substrate influence in Melanesian Pidgin. Journal of Pidgin and Creole Languages 14. 1–44. Siegel, Jeff. 2000. Substrate influence in Hawai’i Creole English. Language in Society 29. 197– 236. Siegel, Jeff. 2008. The emergence of Pidgin and Creole languages. Oxford: Oxford University Press. Siegel, Jeff. 2009. Chinese Pidgin English in South-eastern Australia. The note book of Jong Ah Siug. Journal of Pidgin and Creole Languages 24, 2. 306–337. Siegel, Jeff. 2011. Substrate reinforcement and the retention of Pan-Pacific Pidgin features in modern contact varieties. In Claire Lefebvre (ed.), Creoles, Their Substrates, and Language Typology, 531–556. Amsterdam: John Benjamins. Siegel, Jeff, Barbara Sandeman & Chris Corne. 2000. Predicting substrate influence. Tense modality-aspect marking in Tayo. In Jeff Siegel (ed.), Processes of Language Contact: Studies from Australia and the South Pacific, 75–100. Montreal: Fides. Simpson, John & Edmund Weiner (eds.) 1989. Oxford English Dictionary (2nd Edition) [Online]. Oxford & New York: Clarendon Press. www.oed.com (accessed 18 February 2015). Simpson, Jane. 1996. Early language contact varieties in South Australia. Australian Journal of Linguistics 16, 2. 169–207. Smith, Geoff. 2008. Tok Pisin morphology and syntax. In Kate Burridge & Bernd Kortmann (eds.), Varieties of English (Vol. 3. The Pacific and Australasia), 488–513. Berlin: de Gruyter. Smith, Geoff & Jeff Siegel. 2013. Tok Pisin. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber (eds.), The survey of pidgin and creole languages (Vol. I: English-based and Dutch-based languages), 214–222. Oxford: Oxford University Press. Stukker, Ninke & Ted Sanders. 2012. Subjectivity and prototype structure in causal connectives: A cross-linguistic perspective. Journal of Pragmatics 44,2: 169–190. Sweetser, Eve. 1990. From etymology to pragmatics: Metaphorical and cultural aspects of semantic structure. Cambridge: Cambridge University Press. Traugott, Elizabeth. 2004. Historical Pragmatics. In Lawrence R. Horn & Gregory Ward (eds.) The Handbook of Pragmatics, 538–561. Malden: Blackwell. Traugott, Elizabeth & Richard Dasher. 2004 [2001]. Regularity in semantic change. Cambridge: Cambridge University Press.

296

Denise Angelo and Eva Schultze-Berndt

Troy, Jakelin. 1994. Melaleuka: A history and description of New South Wales Pidgin. Canberra: Australian National University. PhD dissertation. Twain, Mark. 1874. A true story, repeated word for word as I heard it. The Atlantic Monthly 34 (Nov), 591–594. http://docsouth.unc.edu/neh/twain/twain.html (accessed 18 February 2015). Velupillai, Viveka. 2013. Hawai’i Creole. In Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber (eds.), The survey of pidgin and creole languages (Vol. I: English-based and Dutch-based languages), 252–261. Oxford: Oxford University Press. Verhaar, John W. M. 1995. Towards a reference grammar of Tok Pisin: An experiment in corpus linguistics (Oceanic Linguistics Special Publication 26). Honolulu: University of Hawaii Press. Verstraete, Jean-Christophe. 2005. The semantics and pragmatics of composite mood marking: The non-Pama-Nyungan languages of northern Australia. Linguistic Typology 9. 223–268. Vuillermet, Marine. 2013. The grammar of fear in Ese’eja. Paper presented at the Syntax Circle, University of California, Berkeley, Department of Linguistics (15 Nov). White, Charles. 1904. The Story of the Blacks: The Aborigines of Australia by Charles White. [Serial]. Windsor & Richmond Gazette. http://gutenberg.net.au/ebooks13/1300091h.html (accessed 18 February 2015). Williams, C. J. 1980. A grammar of Yuwaalaraay (Pacific Linguistics Series B, no. 74). Canberra: Department of Linguistics, Research School of Pacific Studies, The Australian National University. Yeatman, Allan, Attina Atkinson, Avril Sands, Bernadine Yeatman, Beverley Murgha, Carmel Canendo, David Mundraby, Esme Murgha, Evette Mundraby, Frank Canendo, Jasmine Canendo, Kay Sands, Kerryn Bulmer, Malambie Underwood, Maree Fourmile, Mina Andrews, Paul Neal, Samantha Miller, Sharna Ambrym, Silvia Barlow, Tina Yeatman & Vincent Murgha. 2009. At da Crick [Poster]. Cairns, Queensland: Far North Queensland Indigenous Schooling Support Unit, Department of Education, Training and the Arts.

Maïa Ponsonnet

9 Reflexive, reciprocal and emphatic functions in Barunga Kriol 1 Abstract: This chapter describes the reflexive, reciprocal and adverbial emphatic markers mijelp, gija and miself in Barunga Kriol, the variety of the Australian Kriol spoken in Beswick/Wugularr (Top End, Northern Territory, Australia). These markers are interesting because their distribution has evolved in recent years, resulting in further and neater distinctions. Firstly, a typologically rare distinction between two types of reciprocals has emerged, where transitive verbs and “semi-transitive” verbs receive distinct reciprocal marking. This distinction could result from contact with other Kriol varieties, and represents an interesting pattern of contact-induced change, where no actual form or function is borrowed from the source language. Secondly, the reflexive and emphatic markers, which were originally quasi-identical, have evolved to become two (or more) welldifferentiated items. Based on the analysis of these markers, this chapter examines the ways in which a creole can develop new categories, and questions the principles underlying these developments. Contact with neighbouring varieties of Kriol, as well as late substrate reinforcement, appear to have played a role in these innovations. In addition, this case study indicates that Kriol varieties can be influenced not only by their immediate substrates, but also by other Australian languages within a broader contact area, via contact between varieties.

1 Introduction This chapter describes the reflexive, reciprocal and emphatic markers mijelp, gija and miself in Barunga Kriol, the variety of the Australian Kriol spoken in Beswick/ Wugularr, near Barunga/Bamiyili (Top End, Northern Territory, Australia). The distribution of these markers has evolved in recent years, resulting in further and neater distinctions. Both Sandefur (1979) and Munro (2004) describe mijelp as a strictly reflexive marker, and gija as a reciprocal marker, with no overlap

1 I very pleased to dedicate this article to Patrick McConvell, a dear friend and occasional co-author.

298

Maïa Ponsonnet

between the two. Neither Sandefur (1979) nor Munro (2004) distinguish the reflexive from the emphatic marker miself. In comparison to these descriptions, data collected in Beswick in 2014 offers a different picture. Firstly, mijelp is used as a reflexive marker but also a reciprocal marker under certain conditions. Gija remains a reciprocal marker, but its distribution is restricted, being complementary with the distribution of the reciprocal mijelp. The respective distribution of mijelp and gija delineate a typologically rare distinction between several types of reciprocals, where strictly transitive verbs and “semi-transitive” verbs (i.e. verbs that admit an oblique object) receive different reciprocal marking. Secondly, the reflexive mijelp and the emphatic miself, which were originally quasi-identical, have evolved to become two (or more) well-differentiated items. Overall, the Barunga Kriol system of reflexive, reciprocal and emphatic markers has evolved towards further differentiation. This chapter examines the ways in which a creole can develop new categories, and questions the principles of these developments. Among plausible factors of change are contact with neighbouring Kriol varieties, as well as substrate reinforcement (Siegel 1998, 2008). Taken together, these factors suggest that in the post-colonial era, a creole can evolve towards features that are well represented among its immediate substrates, but also – via contact with other creole varieties – among other languages within a broader, pan-regional area of contact. Contact has an interesting effect in the case discussed here. I hypothesise that contact with other varieties of Kriol may have indirectly triggered the abovementioned innovative and typologically unusual treatment of semi-transitive verbs with respect to reciprocals. This scenario represents a case of contactinduced innovation that has rarely (if ever) been reported in the literature, and is therefore of interest for the typology of contact-induced language change (Heine and Kuteva 2005: 124–141). Section 2 below presents Kriol and the varieties I am concerned with here, as well as the speakers I worked with and the data they provided. Sections 3 to 5 respectively describe the reflexive/reciprocal marker mijelp, the reciprocal gija, and the expression of reciprocity for intransitive verbs. Section 6 describes the exclusive adverb miself, which fulfils the exclusive adverbial emphatic function in today’s Barunga Kriol. After these synchronic descriptions, Section 7 adopts a historical perspective. In 7.1 I explore the interwoven evolution of the reflexive/ reciprocal marker mijelp and of the reciprocal gija, and discuss scenarios and factors for this evolution. In 7.2 I discuss the historical development of the exclusive adverb miself.

Reflexive, reciprocal and emphatic functions in Barunga Kriol

299

2 Languages, speakers and data 2.1 Kriol and its varieties Kriol is an English-based creole spoken by up to 30,000 (mostly Indigenous) people (Lee and Obata 2010) in the central north of Australia. Schultze-Berndt, Meakins, and Angelo (2013) provide an overview of the grammar of Kriol, across varieties. Kriol developed in the Northern Territory in the first half of the twentieth century, on the basis of Northern Territory Pidgin (Koch 2000a). Kriol studies begun in the 1960s, and over the years linguists have distinguished a number of varieties. Towards the east of the Kriol area (i.e. “eastern varieties”, which I am concerned with in this chapter), linguists differentiate between Roper and Barunga Kriol (Harris 1986; Sandefur 1979, 1986; Rhydwen 1995; Ponsonnet 2011, 2012; Dickson 2015). Western varieties include the Fitzroy Valley variety (Hudson 1985) and the variety spoken west of Katherine around Timber Creek, which is commonly referred to as Westside Kriol. This chapter is based mostly on data collected in Beswick, a community 110km to the east of the town of Katherine, near the community of Barunga. The variety labelled “Barunga Kriol” is also spoken in Beswick – as expected given the geographical and social proximity of these communities. According to Sandefur (1986: 21), Kriol was adopted in the Barunga region towards the end of the first half of the twentieth century, i.e. a few decades later than in the Roper River region. Harris (1986) and Munro (2000, 2004) have argued that Kriol emerged in the Roper River region and diffused towards the west. However, as pointed out by Meakins (2014: 377–379), this has not been demonstrated, and parallel genesis is also a likely possibility. Irrespective of genesis scenarios (which are not in focus in this chapter) there is evidence for influences of Roper Kriol upon the variety spoken in Beswick and Barunga. Some words with non-English etyma used by Beswick speakers are also used in the Roper River region, and are known to have etyma in Australian languages around the Roper region (see Dickson, this volume and Dickson, 2015). Some of these words are rare in Barunga Kriol, and Beswick speakers assert that they come from Roper River.2 Whether these influences are past influences resulting from diffusion or more recent influences resulting from contact is a

2 Given the permeability of the boundaries between Kriol and substrate languages with respect to lexical features, it is not always possible to ascertain the origin of the words in question. Therefore, the generalised conclusion that they come from the Roper region may also reflect folk ideology. Nevertheless, there are good indications that some words used by Beswick speakers have no etymon in either English or local substrates, but have one in a Roper substrate.

300

Maïa Ponsonnet

question for further research. Borrowing via contact is plausible, since there are social interactions between the Roper River region and Beswick. Some Beswick speakers have family in the Roper River region. Roper speakers sometimes visit Beswick (and reciprocally), and speakers of all varieties can meet in Katherine, the local service town. The data made available by Greg Dickson and Salome Harris (pers. com. Sep 2014) suggest that synchronically, the Roper and Barunga varieties are alike with respect to the features described in this chapter. Thus, the present description of Barunga features may be valid for Roper Kriol as well. Given the linguistic influences from the Roper River region to Barunga/Beswick, it is possible that these innovations started in the Roper River region and spread to the Barunga region. With respect to these particular innovations, the reverse direction of influence is also plausible. I will leave questions of diffusion aside, because there is not enough published data to tease apart Roper from Barunga Kriol with respect to the features discussed in this chapter. Firstly, there is no precise description of reflexives, reciprocal and emphatic functions in contemporary Roper Kriol (Dickson 2015; Munro 2004; Nicholls 2009). Secondly, it is not possible to know whether Roper and Barunga Kriol differed in the past with respect to reflexive, reciprocal and emphatic functions. The oldest and most detailed description of the eastern varieties, provided by Sandefur (1979), merges Roper and Barunga Kriol. Munro (2004) also offers a description of reflexive and reciprocals for Roper Kriol, and her analysis matches Sandefur’s. Another source of historical information is the Summer Institute of Linguistics’s translation of the Bible (1991, 2007), where the Roper and Barunga varieties cannot be teased apart either.

2.2 Barunga Kriol, its speakers and their data Beswick (also known as Wugularr) is a small Aboriginal community (of around 500 inhabitants) located 110km to the east of the service town of Katherine (of around 10,000 inhabitants) via sealed roads. Most people who live at Beswick are Indigenous and Kriol is the main language of daily interaction. English is used with non-Indigenous residents of Beswick at the local school, clinic, supermarket and other service-providing units. The substrate languages3 of Barunga Kriol – mostly Bininj Gun-wok (Mayali on Figure 1), Dalabon, Jawoyn and Rembarrnga – are known by a significant number of older speakers. Some 3 Here “substrate” is used with the sense “local language in use prior to the emergence of creole”. Whether these languages had some influence upon the creole is independent of my labelling them “substrates”.

Reflexive, reciprocal and emphatic functions in Barunga Kriol

301

middle-aged speakers can also speak a traditional language, albeit often not fluently, and some younger ones – e.g. in their mid-twenties – have passive knowledge and can say a few words. Even though many people have some knowledge of traditional languages, actual use remains occasional (usually by elders). Most people at Beswick go to Katherine regularly, mostly to access health and social services, as well as supermarkets. They also have strong ties in neighbouring communities, which they often visit: for instance Barunga (formerly known as Bamiyili), 30km west on the road to Katherine, and Manyallaluk, further west along the same road, as well as Weemol and Bulman, 250km to the east (i.e. away from the town of Katherine) via a dirt road. The linguistic analyses presented here are based on a ≈15-hour corpus of audio and video recordings, including narratives, comments on stimuli,4 as well as semantic and grammatical elicitation (some focused on the features discussed in this chapter). Most of the data I rely on here was collected in 2014 in Beswick, but I also recorded Kriol from a few speakers who live in Weemol, Bulman and Barunga. The age of the speakers ranged from 9 to 80 years old, with very diverse linguistic backgrounds. Table 1 provides some information about the 17 speakers who contributed data for this study. All these speakers were female, except for one young boy. Observations of male speakers suggests that gender differences are irrelevant with respect to reflexive, reciprocal and emphatic functions. As will be discussed in Section 7, the linguistic items presented in the chapter are evidently undergoing a relatively fast evolution. As a result, my data contains a lot of variation – both intra- and inter-speaker. Two of the speakers I did extensive elicitation with, however, were remarkably consistent with respect to reflexive, reciprocal and emphatic features. One of them even volunteered firm metalinguistic judgments, all of them confirmed by the data within her speech and within several other speakers’ speech. Given their linguistic and biographic background, these two speakers can be considered relatively “prototypical” speakers of the type of Kriol spoken in Beswick. †Lily Bennett, born in 1951, lived in Barunga and Weemol, before moving to Beswick in 2001. She lived in Beswick from 2001 until her death in 2014. Ingrid Ashley was born in the mid-1980s and lived in Weemol until 2011, with regular extended visits to her family in Beswick. She has been based in Beswick full time since 2011, expresses her desire to integrate and deploys efforts to achieve integration. The analyses I arrived at based on a significant amount of spontaneous and elicited data from 4 Still pictures, small videos designed for elicitation and the Australian movie Rabbit-Proof Fence (Noyce 2002). Recordings of speakers commenting on movies contain less Kriol speech, as participants only speak intermittently while the movie is being screened.

302

Maïa Ponsonnet

Table 1: Participants in the study5 Initials

DOB

Recent Residence

Languages and Literacy

*† LB

1951

Beswick

English, Dalabon. Literate.

Extensive: narratives, stimuli-based, elicitation.

*IA

c. 1986

Beswick, Weemol

English. Some literacy skills.

Extensive: narratives, stimuli-based, elicitation

#QB

c. 1950

Barunga, Oenpelli, Katherine

Mayali, English, Dalabon. Literate. Kriol not her most common daily language (self report).

Extensive: narratives, stimuli-based, pure elicitation.

*MJ

1975

Barunga, Oenpelli, Katherine

Mayali, English, Dalabon. Literate. Kriol not her most common daily language (self report).

Extensive: narratives, stimuli-based.

#JJA

1963

Weemol, Beswick

English, Dalabon (many borrowings into Kriol).

Narratives, stimuli-based, pure elicitation.

*AA

c. 1984

Beswick, Weemol

English.

Narratives, stimuli-based.

?TM

c. 1985

Beswick, Weemol.

English. Literate.

Stimuli-based.

#MT

c. 1940

Weemol

Dalabon, Mayali, English.

Narratives, stimuli-based.

*ABM

c. 1994

Beswick

English. Probably literate.

Some stimuli-based and some conversation.

*KBM

c. 1992

Beswick

English. Probably literate.

Some stimuli-based.

*JP

c. 1983

Weemol, Manyallaluk

English. May be literate.

Some stimuli-based.

#JBr

c. 1960

Weemol

English, Dalabon. Literate.

Some lexical elicitation and narratives

?PA

2001

Beswick, Weemol

English.

Some narratives and some stimuli-based.

?ND

c. 1945

Beswick

Dalabon, English.

Participation to stimulibased sessions.

?JBi (male)

2005

Beswick, Weemol

English.

Participation to stimulibased sessions.

#DC

?c. 1935

Bulman

Rembarrnga, Dalabon, English.

Some short narratives.

?MJo

?c. 1945

Beswick

English. Probably some Dalabon.

Occasional participation to stimuli sessions.

Data Collected

5 “*”: at least one of the innovative features discussed below (innovative distribution of the reciprocals mijelp and gija [Sections 3 and 4], segmental distinction between reflexive/reciprocal mijelp and exclusive adverb miself [Section 6]) is attested; speakers flagged with “*” are “prototypical” Beswick speakers. “#”: it is attested that at least one of the innovative features discussed below has not been adopted. “?”: the data does not show whether the innovative features discussed below have been adopted. † indicates that the speaker is now deceased.

Reflexive, reciprocal and emphatic functions in Barunga Kriol

303

Lily Bennett and Ingrid Ashley is consistent with the data provided in smaller amounts by other “presumably prototypical” Beswick speakers (for instance ABM and KBM, two young women around 20 years old, who have spent all or most of their lives in Beswick; as well as other speakers marked with “*”). Therefore, I relied on Lily Bennett and Ingrid Ashley’s data to set a “standard” description of the reflexive, reciprocal and emphatic functions in Barunga Kriol, and assessed divergence against this standard. Throughout the chapter, I refer to these two speakers and the other standard Beswick speakers as “prototypical speakers”. The descriptions based on these prototypical speakers presents variation compared with speakers who live or have spent more time in Weemol, 250km to the east, as well as with older speakers from the Barunga region who have a different linguistic background. These differences suggest that the innovations discussed in the chapter have not (yet) spread from the Barunga/Beswick region towards the more remote Kriol-speaking community of Weemol.

3 The reflexive/reciprocal mijelp Like other varieties of Kriol, Barunga Kriol has a marker which merges reflexive and reciprocal functions: the post-verbal clitic mijelp. Mijelp comes after the aspectual verb suffixes if there are any (i.e. after -bat, as shown in (1)), which is the same slot as object pronouns. While the English etymon, myself, agrees with the person and number of its antecedent, mijelp is invariable. (1)

AP haid-im-bat=miyel feis. prop.noun hide-TR-CONT=REFL/RECP face AP is hiding his face [hiding himself face]. (20140328d_000_ABM 135 [Stim])

The fully articulated, standard pronunciation is [miʝɛlp] or [miʝɛlb] but there are variations, which must be considered in order to distinguish the reflexive/ reciprocal mijelp from the exclusive adverb miself (see Section 6). With mijelp, the second consonant is most commonly [ ʝ], but also realised as an affricate ͡ It can be realised as [j] or this consonant can be elided, especially by [ɉʝ]. younger speakers ([mijɛl], [mi-ɛl] – not a diphthong but two successive vowels). Older speakers sometimes pronounce this second consonant as [s] (sometimes clearly laminal), but prototypical speakers always have a palatal consonant if there is a consonant. The final stop varies in quality, is rarely released and more

304

Maïa Ponsonnet

often not pronounced at all ([miʝɛl]). The [l] can be dropped as well ([miʝe]). By contrast with the exclusive adverb miself, in mijelp the final consonant is never realised as a fricative. Throughout the chapter, I use standard spellings in the prose, but aim at reflecting the actual pronunciations of the items under consideration in the examples. Example (2) illustrates the reflexive use of mijelp in the singular (where the reciprocal reading is not available); example (3) illustrates reflexive use in the plural (where ambiguity is possible in theory, but is usually resolved by context); and (4) illustrates reciprocal use (which is obviously not available with singular participants). (2) Ai got la main rum ai dres-im-ap=mijel 1SG.S get LOC 1SG.POSS room 1SG.S dress-TR-UP=REFL/RECP ai-l sprei=mijel komb-ep=mijel 1SG.S-FUT spray=REFL/RECP comb-UP = REFL/RECP ta-im-ap=mijel ai kam-at na. tie.up-TR-UP = REFL/RECP 1SG.S come/go-OUT EMPH I get into my room, I dress (myself) up, I’ll (hair-)spray myself, comb myself, tie up my hair [tie up myself], and I go out then. (20140326b_001_IA 77 [ContEl]) (3) Bat yu luk im sabi dei bin aid-im=mijelp deya. but 2SG.S look 3SG.S know 3PL.S PST hide-TR=REFL/RECP DIST.DEM But you see, he knows that they’re hiding themselves there. [Three children had been walking along together, and one is now parting from the others, who keep going by themselves, without the third one.] (20140411a_000_LB 068 [RPF]) (4) Dei bin lib-um=mijal na, 3PL.S PST leave-TR=REFL/RECP now oni tubala miself gone. only 3DU.S EXCL gone They have left each other now, only these two keep going, by themselves (without the other one). (20140328a_003_LB_ND 018 [RPF]) Evans, Levinson et al. (2011: 20) state that 34% of the world’s languages have a construction that merges reflexives and reciprocals (see also Heine and Miyashita 2008: 171). Dixon (1980: 433) states that this is the default situation in Australian

Reflexive, reciprocal and emphatic functions in Barunga Kriol

305

languages. The Barunga Kriol reflexive/reciprocal marker is thus cross-linguistically unsurprising, apparently covering a fairly typical range of nuances in reflexive and reciprocal actions (König and Gast 2008a: 23–26; Dalrymple et al. 1998; Evans, Levinson, et al. 2011: 8–9). The examples above illustrate grooming events (2), which are typically expressed as reflexives across languages (Heine and Miyashita 2008: 212; Kemmer 1993: 16); actions involving the whole self instead of a second participant (in (3): hide something/hide oneself). Mijelp covers other events typical of reflexives or middle voice (see Kemmer 1993: 16– 20), such (among others) body movement (5), emotions (6). (5)

Im-in 3SG.S - PST

salki sulky

Tan-im=mijelh6 turn-TR=REFL/RECP

igin again bla DAT

sidan. . . sit/be sidan sit/be

en and

tok-tok-bat. talk:RED-CONT

He was sulky too, he was sitting. . . He turned himself around to sit and talk (together). (20140328b_000_AA 092 [Stim]) (6)

En and

dat ART

men man

im 3SG.S

fil-im=miyal feel-TR=REFL/RECP

sad. sad

And the man was feeling sad. (20140328c_003_JBi_PA 091 (JBi) [Stim]) Corresponding to what Kemmer labels “emotion middle” (Kemmer 1993: 18) and “cognition middle” (Kemmer 1993: 19), some Barunga Kriol intransitive verbs denoting emotions or cognitive states also occur with mijelp: sori mijelp ‘feel sorry’ in (7), mejin mijelp ‘imagine’ in (8). These expressions have lexicalized and do not convey any reflexive meaning. Therefore, such constructions are not productive but occur only with a small set of verbs. They will not be considered in the analyses that follow. (7) Yu sori=mijelb, longwei dat pesen im go-wei-go-wei. ART person 3SG.S go-AWAY:RED 2SG.S feel.sorry=REFL/RECP away You feel sad, when someone is a long way away. (20140408b_005_JJA_MJo 76 (JJA) [El]) (8) Tubala bin mejin-ing=mijel tu deya. 3DU.S PST imagine-CONT=REFL/RECP too DIST.DEM They (the two) of them were thinking (having ideas, having fantasies) too there. (20140407b_000_IA 222 [El]) 6 Here the final plosive is a glottal stop (“h” in Kriol).

306

Maïa Ponsonnet

Mijelp also describes a range of reciprocal events, whether simultaneous reciprocity (as in (4) above) or some more complicated combinations, for instance successive reciprocity in (9), and some other reciprocal combinations (König and Gast 2008a: 23–26; Dalrymple et al. 1998; Majid et al. 2011; Evans, Levinson, et al. 2011: 8–9). (9)

Bat but

wei

en and

wen when

bedam before

RLTVZR

im 3SG.S

dei 3PL.S

gib-it, give-TR

gib-it-bat give-TR-CONT tubala 3DU.S

ebrithing, everything

gib-it=mijel give-TR=REFL/RECP

ebrithing. everything

But when they keep giving all sorts of things to each other, and when he gives [to this person], these two give everything to each other [give things to each other all the time]. (20140326b_002_IA 026 [El]) Thus, mijelp covers a broad semantic spectrum within both the reflexive and the reciprocal domains. On the other hand, the distribution of mijelp is restricted because it does not occur with all verbs, but only with verbs that admit a direct object, i.e. transitive and ditransitive verbs. Intransitive verbs do not enter in reciprocal constructions with mijelp (although they occur with some intransitive verbs, in the lexicalized expressions illustrated by (7) and (8) above). The behaviour of verbs other than transitive and ditransitive in reciprocal constructions is discussed in Sections 4 and 5. Further examples of verbs entering mijelp constructions (reflexive and reciprocal), all transitive or ditransitive, are provided in Table 2 and Table 3. Table 2: Some verbs found in mijelp reflexive constructions Standard transitive construction

Construction with reflexive mijelp

ardim im ‘hurt him/her’

ardim mijelp ‘hurt oneself’

filim im ‘feel him/her’

filim mijelp ‘feel’ (e.g. feel drunk, tired. . .)

haidim im ‘hide him/her’

haidim mijelp ‘hide oneself’

kaberimap im ‘cover him/her’

kaberimap mijelp (+ body-part) ‘cover oneself (one’s body-part)’ e.g. kaberim mijelp feis ‘cover one’s face’

kilim im ‘hurt/kill him/her’

kilim mijelp ‘kill/hit oneself’

meikim im + nominal ‘make him/her’ + nominal

meikim mijelp + nominal ‘make oneself’ + nominal e.g. meikim mijelp sik ‘make oneself sick’

shoum im ‘show it’ (ditransitive)

shoum mijelp ‘show oneself’

Reflexive, reciprocal and emphatic functions in Barunga Kriol

307

Table 3: Some verbs found in mijelp reciprocal constructions Standard transitive construction

Construction with reciprocal mijelp

alpum im ‘help him/her’ dijim im ‘tease/harass him/her’ gibit im ‘give to him/her’ (ditransitive) graul im ‘scold him/her’ kilim im ‘hit/kill him/her’ libum im ‘leave him/her’ luk im ‘see him/her’ olim im ‘hug him/her’ upahupa im ‘kiss him/her’

alpum mijelp ‘help each other’ dijim mijelp ‘tease/harass each other’ gibit mijelp ‘give to each other’ graul mijelp ‘argue with each other’ kilim mijelp ‘hit/kill each other’ libum mijelp ‘leave each other’ luk mijelp ‘see each other’ olim mijelp ‘hug each other’ upahupa mijelp ‘kiss each other’

4 The reciprocal gija In addition to the reflexive/reciprocal mijelp, Barunga Kriol uses two other markers to encode reciprocity: the post-verbal and post-adjectival particle gija and the nominal enclitic gija (