Language in Australia

LANGUAGE IN AUSTRALIA provides a comprehensive account of the present linguistic situation in Australia, primarily from a sociolinguistic perspective. The focus is mainly on description, and the chapters aim to provide a comprehensive overview and summary of what is known about Australia's languages as well as a guide to current areas of research interest. The volume is divided into five parts. The first four contain chapters dealing with Australia's indigenous and non-indigenous languages, and the fifth is devoted to public policy and social issues related to the languages of Australia. There are also sketches of each of the major language types in Australia. These include Aboriginal languages, pidgins and Creoles, Aboriginal English, and community languages such as Greek, Italian and Serbo-Croat. No other book offers such a broad survey of the language situation in Australia. Linguists as well as non-specialists will find in this volume, a companion to Language in the USA and Language in the British Isles, a guide and reference source to the linguistic heritage of Australia.

Language in Australia

Language in Australia

edited by SUZANNE ROMAINE Merton College, Oxford

The right of the University of Cambridge to print and sell all manner of books was granted by Henry VIII in 1534. The University has printed and published continuously since 1584.

CAMBRIDGE UNIVERSITY PRESS Cambridge New York Port Chester Melbourne Sydney

PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE The Pitt Building, Trumpington Street, Cambridge, United Kingdom CAMBRIDGE UNIVERSITY PRESS The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarcon 13,28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http ://wwwcambridge.org © Cambridge University Press 1991 This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1991 First paperback edition 2004 A catalogue recordfor this book is available from the British Library Library of Congress cataloguing in publication data Language in Australia, edited by Suzanne Romaine. 000 p. 00 cm. Includes index. ISBN 0 521 32786 5 hardback 1. Australia - Languages. 2. Socio-linguistics -Australia. I. Romaine, Suzanne. 1951P381.A78L35 1991 409.94-dc20 90-1844 CIP ISBN 0 521 32786 5 hardback ISBN 0 521 33983 9 paperback

Contents

List of illustrations List of tables Notes on contributors Preface A cknowledgemen ts Introduction

SUZANNE ROMAINE

Part I Aboriginal and Islander Languages 1

Overview of indigenous languages of Australia

page ix xi xiii xvii xix 1 25 27

MICHAEL WALSH

2

Language in Aboriginal Australia: social dialects in a geographic idiom PETER S U T T O N 3 Aboriginal English - an overview IAN G, MALCOLM

49 67

AND SUSAN KALDOR

4

Communicative strategies in Aboriginal English

84

DIANA EADES

5

Language and communication in Aboriginal land claim hearings

6

7

8

Warlpiri in the 80s: an overview of research into language variation and child language EDITH L. BAVIN AND TIM SHOPEN A sketch of Kalaw Kawaw Ya KEVIN FORD AND DANA OBER

Understanding language shift: a step towards language maintenance P A T R I C K MCCONVELL

Part II Pidgins and Creoles 9

Overview of the pidgin and Creole languages of Australia

94

HAROLD KOCH

PETER MUHLHAUSLER

104

118

143

157 159

viii

Contents

10 Queensland Kanaka English PETER M U H L H A U S L E R 11 Torres Strait creole ANNA S H N U K A L 12 Kriol - the creation of a new language

174 180 195

JOHN W. HARRIS

13 A sketch of the structure of Kriol

J O H N SANDEFUR

204

Part III Transplanted languages other than English

213

14 Overview of 'immigrant7 or community

215

languages

MICHAEL CLYNE

15 Dutch in Australia: perceptions of and attitudes towards transference and other language contact phenomena

228

ANNE PAUWELS

16 German and Dutch in Australia: structures and use

241

MICHAEL CLYNE

17 Modern Greek in Australia ANASTASIOS TAMIS 18 Language variety among Italians: anglicisation, attrition and attitudes

CAMILLA BETTONI

249 263

19 First generation Serbo-Croatian speakers in Queensland: language maintenance and language shift

270

JACQUES DOUCET

Part IV Varieties of Australian English

285

20

287

A survey of regional usage in the lexicon of Australian English

21 22

P A U L I N E BRYANT

Finding a place in Sydney: migrants and language change BARBARA M. H O R V A T H Gender differences in Australian English

304 318

ANNE PAUWELS

Part V Public policy and social issues

327

23 National language policy and planning: migrant

329

languages

24 25

ULDIS O Z O L I N S

Social class differences in the lexicon D A V I D Plain English: some sociolinguistic revelations

CORSON

349 363

ROBERT D. EAGLESON

References Index

373 401

Illustrations

Figures 1.1 1.2 1.3 1.4 1.5 2.1 8.1 8.2 13.1 17.1 17.2 21.1 21.2 21.3 21.4 24.1 24.2

24.3

Ergative and accusative marking in relation to the page 39 animacy hierarchy Competing case systems in Yidiny 39 Example of an Australian phonological inventory 41 Basic phonological inventory of a hypothetical 41 Australian language Gunaarlpelyu Expanded inventory of Gunaarlpelyu 41 Preliminary statement of genetic relationships, Wik 60 Group Model of language choice in bilingualism 151 Social arenas relevant to codeswitching at Daguragu 153 Phonemes of Kriol 205 Number of overseas-born Greek-Australians in Victoria 252 Evolution of modern Greek in Australia 260 The linguistic structure of the sociolects in the periphery 309 The linguistic structure of the sociolects in the core 310 Distribution of the social characteristics across the core 311 sociolects The sociolinguistic relationship between migrants and 316 their hosts: an intergenerational model Line graph illustrating group means recorded on the 355 measure by age and social group ( N = 129) Line graph illustrating group percentage means recorded 357 on the G-L instrument by age and social group. Descriptive task-oral language ( N = 129) Line graph illustrating group percentage means recorded 357 on the G-L instrument by age and social group. Explanatory task-oral language ( N = 129)

x

List of illustrations

24.4

24.5

Line graph illustrating group percentage means recorded on the G-L instrument by age and social group. Descriptive task-written language (N = 80) Line graph illustrating percentage means recorded on the G-L instrument by age and social group. Explanatory task-written language (N = 80)

358

Approximate location of languages referred to Major classifications of Australian languages The Wik region Clan countries Language associations by clan estate, Wik Region The Western Torres Strait, showing major dialect boundaries Torres Strait showing inhabited islands Regional distribution of Cheerio-Frankfurt etc. South-East regional distribution of Dink-Dinky-Donkey Slide

28 34 54 58 63

358

Maps 1.1 1.2 2.1 2.2 2.3 7.1 11.1 20.1 20.2 20.3

119 181 294 296 300

Tables

6.1 6.2

6.3 10.1 11.1 14.1

14.2 15.1 15.2 16.1 16.2 17.1 17.2 17.3 17.4 17.5 19.1 19.2

The traditional Warlpiri system for Bound Subject page 107 Pronouns The traditional Warlpiri system for Bound Subject 108 Pronouns, analysed into morphemes for person and number The forms most frequently used by young Warlpiri 109 speakers for Bound Subject Pronouns Number of male and female Melanesians in Queensland 176 Phonological inventory of TSC 186 Home use of languages other than English: total 217 numbers for first 14 community languages and Aboriginal languages 218 Percentage language shift 237 Contingency between the A-test and the Attitudinal test 237 Contingency between the B-test and the Attitudinal test Language shift in the first and second generations: 243 German and Dutch Percentage of syntactic transfers - proximity-motivated 247 or SVO generalisation Number of regular users of Modern Greek in Australia 250 by state Medium of communication normally used by second 252 generation claimants Number of students attending Modern Greek classes in 253 Australia Parental evaluation of the linguistic skills of their older 254 and younger children Attitudes towards language maintenance, by generation 255 Media variables 274 Written items 274

xii

List of tables

19.3 Printed items 19.4 Listening practices 19.5 Domains of language use 19.6 Dominance configuration 21.1 The phonetic variants of selected AE vowels 21.2 Percent distribution of vowel variants in the peripheral sociolects 21.3 Percent distribution of Anglo-Celtic speakers over the three varieties of AE 21.4 Percent distribution of Greek and Italian speakers over the three varieties of AE 21.5 Percent distribution of all ethnic groups across the four core sociolects 22.1 Gender and major differentiators of AE varieties

276 276 278 279 307 309 312 312 313 321

Notes on contributors

is Senior Lecturer in Linguistics at La Trobe University. Since 1982 she has been doing research on Warlpiri, in particular, children's acquisition of the language.

EDITH BAVIN

CAMILLA BETTONI is Senior Lecturer in Linguistics in the Department of Italian at the University of Sydney. Her book, Italian in North Queensland (1981), is based on her extensive research on Italian in Australia. PAULINE BRYANT is a PhD student in Linguistics at the Australian National University. In addition to her research on regional variation, she has investigated intonation in Australian English.

CLYNEis Professor of Linguistics at Monash University. His books include Australia Talks (1976), Multilingual Australia (1982) and Australia - Meeting Place of Languages (1985).

MICHAEL

is in the Education Faculty at Massey University in New Zealand. His books include The Lexical Bar (1985), Oral Language Across the Curriculum, Language Policy Across the Curriculum and he is editor of Language and Education: An International Journal

DAVID CORSON

JACQUES DOUCET is Adviser in Education in the Department of Defence in France and has done research on Serbo-Croatian in Queensland. DIANA EADES is lecturing in Linguistics at the University of New England

and currently carrying out research on the use of Aboriginal English in legal interviews. is Associate Professor of Modern English Language at the University of Sydney. He is the author of Aboriginal English and the Child (with Susan Kaldor and Ian G. Malcolm, 1982) and Plain English and the Law. He is Special Adviser on Plain English to the Australian Government. ROBERT D. EAGLESON

xiv

Notes on contributors

KEVIN FORD is Principal Lecturer at the School of Australian Linguistics in Batchelor and is currently working on a dictionary of Kalaw Kawaw Ya/Kalaw Lagaw Ya. J O H N W . HARRIS is

now Director of the Zadock Institute for Christianity and Society in Canberra after having spent many years in Aboriginal communities in the Northern Territory. His most recent book is One Blood: Aborigines Encounter 200 Years of Christianity. BARBARAM.HORVATHis Senior Lecturer in Linguistics at the University of Sydney. She is the author of Variation in Australian English: The Sociolects of Sydney (198 5). SUSAN KALDOR has recently retired from her position as Associate Professor in Linguistics at the University of Western Australia. Her report, Languages for Western Australia, is a language policy document currently being implemented by the Western Australian Ministry of Education. HAROLD KOCH is Senior Lecturer in Linguistics at the Australian National University. His major research project is a descriptive and comparative study of the Kaytej language of Central Australia.

is Head of the Department of Language Studies in the Western Australian College of Advanced Education and collaborated with Susan Kaldor in a survey of Aboriginal children's English in Western Australia.

I A N G . MALCOLM

PATRICK McCONVELL is Lecturer in Anthropology at University College of the Northern Territories.

is Lecturer in General Linguistics at the University of Oxford and currently working at Bond University. He is the author of Pidgin and Creole linguistics (1986) and co-edited with Stephen Wurm the Handbook of Tok Pisin (1985). PETER MUHLHAUSLER

works for the Department of Aboriginal Affairs in Townsville and is a speaker of Kalaw Kawaw Ya.

DANAOBER

ULDIS OZOLI NS is a Lecturer in the Department

of Language and Culture Studies at Victoria College in Melbourne. His research interests include language policy and education with particular reference to migrant languages in Australia.

PAUWELsis Director of the National Centre for Cross-Cultural Communication and Community Languages in the Professions at Monash University. She is the author of Immigrant Dialects and Language Maintenance in Australia (1986) and the editor of Women and Language in Australian and New Zealand Society (1987). ANNE

Notes on contributors

xv

S U Z A N N E ROMAINE is Merton Professor of English Language at the University of Oxford. Her books include The Language of Children and Adolescents (1984), Pidgin and Creole Languages (1988) and Bilingualism (1989).

is a translator with the Australian Aborigines and Islanders Branch of the Summer Institute of Linguistics, assigned since 1973 to the Kriol Bible translation project. He has published a language course on Kriol (with Joy Sandefur, 1981), a grammar of the language (1979) along with papers describing various aspects of the language.

J O H N SANDEFUR

ANNA SHNUKALis Lecturer in Linguistics at the University of Queensland. Her description of Torres Strait Creole {Broken: An Introduction to the Creole Language of the Torres Strait) was published in 1989. PETER S U T T O N is Head of the Division of Anthropology at the South Australian Museum, in Adelaide. He has carried out extensive fieldwork in Aboriginal communities and has recently edited Dreamings: The Art of Aboriginal Australia (1988). TIMOTHY SHOPEN

is Senior Lecturer in Linguistics at the Australian

National University. ANASTASiosTAMisis Lecturer in Greek Studies at the Universities of La Trobe and Melbourne. His most recent books are Greeks in Australia (1988), The Immigration and Settlement of Greek Macedonians in Australia (1989) and Glossologia (1989). is Lecturer in Linguistics at the University of Sydney. He is currently working on a general guide to Aboriginal languages (with Paul Black) and a reference grammar of Murrinh-Patha.

MICHAEL W A L S H

Preface

The impetus for this book came from the publication of Language in the USA (Ferguson and Heath 1981) and Language in the British Isles (Trudgill 1984). This volume is a companion and complement to these two. The purpose of Language in Australia is to provide a comprehensive account of the present linguistic situation in Australia, primarily from a sociolinguistic perspective. There are at present no other books which offer such a broad survey of the language situation in Australia, although there are now works which cover selected aspects of it, for example, Clyne (1976), a sample of studies on Australian English, migrant and Aboriginal languages, Clyne (1982) on the position of community languages, Clyne (1985) on language contact, Blair and Collins (1989) on varieties of Australian English, and the surveys of Aboriginal and Torres Strait Islander languages by Dixon and Blake (1979), Dixon (1980), Blake (1981), and Yallop (1982). These and other publications are, however, important indications of the recent considerable interest in the languages of Australia. This volume is divided into five parts. The first four contain chapters dealing with Australia's indigenous and non-indigenous languages and the fifth is devoted to public policy and social issues related to the languages of Australia. There are also sketches of each of the major language types in Australia. The chapters are mainly descriptive. They aim to provide a comprehensive overview and summary of what is known about Australia's languages as well as a guide to current areas of research interest. Throughout, the contributors pay special attention to issues arising from the sociohistorical situation in which Australia's languages and language varieties coexist. As in the other major English-speaking countries, English is a relative newcomer to a linguistic scene which was already very diverse and heterogeneous. Present-day Australians speak a wide range of languages making Australia, as Clyne (1985) suggests, a meeting place of languages. While most Australians are English monolinguals, as are the majority of Britons and Americans, the continent's original inhabitants

xviii

Preface

were largely multilingual and many still are today. The languages spoken by the descendants of the original inhabitants of Australia include the following: surviving Aboriginal languages, a Creole, for example, Kriol or Cape York Creole, Aboriginal English, and a variety of Australian English. There are also many other bilingual Australians who have settled more recently and still maintain knowledge of a language other than English, for example, Greek, Italian, German, Dutch, and Serbo-Croatian, to name only a few of the new 'community' languages dealt with in part III of this volume. The editors of both Language in the USA and Language in the British Isles lament the fact that many people are relatively uninformed about the language situation in their own countries. While it may seem tedious to repeat this complaint here, it is no less true of Australia than of Britain and the United States. This volume is dedicated to the spirit of linguistic and cultural diversity and the hope that non-specialists will also find in this volume an appreciation of the rich linguistic heritage of Australia.

Acknowledgements

My biggest debt is to Bruce Rigsby for his help in the planning and undertaking of this volume. Without his support and willingness to respond to my queries, I would not havefinishedit as easily. I also received much useful advice and assistance from Michael Clyne during the preparation of part III on transplanted languages other than English. I have also had a great deal of secretarial assistance from the English Faculty Office of the University of Oxford during the preparation of this book, which has made my task as editor smoother than might normally have been possible. I am very grateful to Caroline Johnston, who assisted me in the early stages, and to Sarah Barker and Jackie Wall for their help with the final stages. Sarah's willingness in particular to undertake the typing of seemingly endless revisions was responsible for seeing the project to completion sooner rather than later. I would also like to thank Penny Carter, Marion Smith and Judith Ayling at Cambridge University Press for their sustained interest in this project.

Introduction SUZANNE ROMAINE

Australia is a geographically isolated and largely English-speaking continent surrounded primarily by non-English-speaking neighbours (apart from New Zealand). The story of its languages presents an intriguing case study for sociolinguists. While most Australians today are English monolinguals, the English language is a recent arrival. The majority of the country's original inhabitants were largely multilingual and many still are. The chapters in this book give a comprehensive overview and summary of what is known about the sociolinguistic situation of Australia's major language varieties. In my introduction I aim to provide a sociohistorical background to the evolution of the major varieties of language now found on the Australian continent. One of the most interesting developments I attempt to trace is how a new ideology of pluralism arose in the 1970s in response to social and political changes. This was in direct opposition to the earlier 'White Australia' policy, which projected an image of an ideal Australia which was monocultural, monolingual and monoracial. I also document how Australian attitudes towards language and linguistic diversity have deep historical precedents in the cultural ideology of western Europe and are paralleled in the major Anglophone nations, particularly Britain and the United States. Oppressive policies towards linguistic minorities were practised by the British for a long time 'at home' and transplanted to new colonies elsewhere. Australia has, however, recently taken steps to ensure language maintenance and to foster the development of language skills. I offer a comparative perspective on language in the United States, Britain and Australia, and consider the question of whether Australia constitutes a speech community. Finally, I make some remarks about future developments.

2

SUZANNE ROMAINE

Languages in sociohistorical perspective

Upon its arrival in Australia as a transplanted language in the eighteenth century, English quickly assumed the status of dominant language. Of considerable interest in the Australian language community is the changing ecology of a situation in which different languages have come to coexist and influence each other. One major factor here is, of course, the continued spread of English worldwide and theriseof distinctively Australian English varieties. This has been significant in the decline in use of some native Aboriginal languages, and has resulted in the addition of new Englishbased varieties (i.e. pidgin, Creole and Aboriginal English) to the communicative repertoires of many rural and urban Aboriginal communities (see Sandefur, Harris, Shnukal, and Miihlhausler, chapter 10, this volume). For a time the Australian linguistic repertoire also included Melanesian and Polynesian languages, when some 60,000 Pacific Islanders were brought in to work on the plantations of Queensland. While they seem to have had little effect on the indigenous languages of Australia, the linguistic legacy of these people survived until recently in Queensland Kanaka English. This and other early non-European linguistic contacts in Australia resulting in new varieties of English are discussed by Miihlhausler, chapter 9, this volume). New contact situations have arisen through immigration and new language communities, such as Greek and Italian, have been established. As in both the United States and Britain, there are both 'new' and 'old' immigrant communities. Australia's Dutch connection, for instance, goes back to 1606, when William Jansz and his crew explored the waters around Cape York Peninsula. In the seventeenth century Australia was known as New Holland. This was more than 100 years before James Cook's Endeavour dropped anchor in Botany Bay. While Dutch exploration continued into the late eighteenth century, the Dutch never established a permanent settlement. The first Dutch to settle there were convicts transported from England (see Duyker 1987). Most Dutch immigration has been a postwar phenomenon of the 1950s and 1960s and today for various reasons, language maintenance is the weakest among this group (see Pauwels, chapter 15, and Clyne, chapter 16, this volume). Immigration has been a consistent theme in Australian history since 1788. Historians generally recognise three major immigrant waves of 1825-60, the 1880s and 1910-30 (see, e.g. Jupp 1966). Convict settlement was planned from London and entry to Australia was controlled to a degree never true for the Americas. Australia attracted few free migrants until the 1820s. From the 1830s anyone who was British (and also white, after 1901) could enter Australia. This remained true until the 1970s. The foundations of the 'Old Australian' population were laid between the 1820s

Introduction

3

and early 1860s. A leaflet distributed in the 1880s by the Agent-General for Queensland in order to attract English immigrants described Queensland as 'no foreign land, but only England over the water' (cited in The Canberra Times, 6 August 1988). The practice of free British entry was coupled with that of virtually free travel for approved immigrants from Britain and Ireland. This gave Australia discretionary choice of immigrants and thus served to maintain the ethnic (and sex) balance. Non-Europeans were rejected regardless of their culture or education. There was also a tendency to see non-British immigrants as unskilled and potential labourers or domestics, though by the 1950s it was apparent that more immigrants from southern Europe (excluded from the United States since the early 1920s) would tip the balance in their favour. A phobia about misegenation and a desire to protect the 'white race' were involved in the debate that resulted in The Federal Immigration Restriction Act of 1901. The term 'White Australia' was used in this context and gave its name unofficially to the policy which continued into the 1960s and for which Australia became notorious. Its underlying ethos was based on the assumption that immigrants would assimilate; non-European races were undesirable because they looked different and therefore could not assimilate. Even the Aboriginal population had been divided into those of mixed descent who were expected to assimilate and 'breed out' their black traits, and tribal Aborigines who were segregated and expected to die out anyway. 'White Australia' was used to fuel national consciousness. Its aim was to produce a homogeneous English-speaking Anglo-Saxon culture. By 1945 after the three immigrant waves of 1825-60, the 1880s and 1910-30, the basic character of the Australian population was established. In 1947 Australians were 99 per cent white and 90 per cent of British origin. Virtually all spoke English, as indicated in the 1933 Census, which was the first attempt in Australian history to obtain some statistics on language use. Thus, it appeared that the social engineering of the past 100 years had succeeded. Australia had become one of a handful of homogeneous societies. The next great wave of immigration between 1947 and 1972, the largest in all, was nevertheless unique in its acceptance of non-British settlers. The generation which introduced the postwar immigration policy had been brought up with the deeply rooted belief that the 'White Australia' policy was not only necessary, but highly desirable. It was seen as a noble ideal and a rationalisation for protecting labour conditions. White Australia was regarded as essential to nation building and was based on the assumption that the Australian 'race' consisted of peoples of European origin and was not to be mongrelised. Thus, the melting-pot metaphor of assimilation

4

SUZANNE ROMAINE

implied not a biological mixture of populations (or 'races'), but unity of institutions, society and culture. After the war, however, the metaphor of'populate or perish' was used to justify large-scale immigration. At a time of economic and industrial expansion more labour than could be drawn from the traditional pool of British migrants was required. Thus, the 'White Australia' policy was ended. The ideology behind 'White Australia' persisted, however, in attitudes towards language. One of the strongest expectations of these postwar immigrants was that they should speak English in public and various organisations passed resolutions designed to enforce this by law if necessary. Wartime restrictions on the untranslated use of foreign languages in broadcasts or newspapers remained in force until 1956 (see Clyne, chapter 14, and also Ozolins, this volume). Many of the attitudes are spelled out in a leaflet issued to new arrivals in 1948 by the Australian Government. It advised immigrants on how to get on with Australians and stressed assimilation. The end result was to occur on the 'day when fellow Australians stop looking at you because your manners or speech are different, you will know you have been accepted as one of the community'. It further warned that Australians are not used to hearing foreign languages. They are inclined to stare at persons whose speech is different. Some may laugh at you or make fun of your accent. Do not let this worry or annoy you. Also try to avoid using your hands when speaking because if you do this you will be conspicuous. Australian men never wear hair nets. They regard men who do as effeminate.. . . Learn the habits and customs of the Australians and you will quickly feel at home in your new homeland. The Canberra Times, 7 August 1988

Even as late as 1967 Bill Sneddon, then Minister for Immigration could state publicly: We must have a single culture - if immigration implied multi-cultural activities within Australian society, then it was not the type Australia wanted. I am quite determined we should have a monoculture, with everyone living in the same way, understanding each other, and sharing the same aspirations. We don't want pluralism. (Cited in Bullivant 1984: 44) Only ten years later though, these statements were to appear blatantly discriminatory. The egalitarian myth of Australia as a 'classless society' was used to justify the strong assimilationist ideology which persisted into the 1960s. The attitude was sink or swim. It was regarded as contrary to egalitarian views to provide special privileges or assistance of any kind. Few facilities

Introduction

5

to assist the integration of migrants in the socioeconomic and education system were provided until the 1970s (see Ozolins, this volume). Bullivant (1984: 53-4) notes that the education system was unwilling to assist teachers in their attempts to cope with the problems of immigrant children. This enabled information that ran counter to conformist official policies and ideologies to be suppressed and excluded from the public agenda. The control of information about the education system's inability to cope with increasing numbers of migrant children even extended to the failure to collect adequate information about the distribution of migrant pupils, their knowledge of English and their performance and psychological difficulties. This was justified by the Director General of Education in New South Wales in terms of the egalitarianism myth. He said: 'We deliberately refrain from collecting any statistics in regard to school pupils from overseas. Once they are enrolled in school, they are, from our point of view, Australian children' (Commonwealth Department of Immigration 1963: 21). In the 1970s governments of both political parties in Australia moved swiftly to replace outmoded attitudes. This included 'White Australia', which was officially ended with the passing of the Racial Discrimination Act in 1975. The new Minister of Immigration under the Whitlam Government scrapped the Department of Immigration's filing system which had been based on racial classification. The preference for the British ended, and in the early 1980s the Liberals also extended the principle of refugee admission to include Asians after 1975 in the wake of the fall of Saigon. In less than ten years Australia had adopted an immigration policy no longer based on national or racial origin. This shift in policy represented the most fundamental change since 1947 (and in some respects, even the 1830s). Australia no longer sought to be 'England over the water'. It no longer expected assimilation. Welfare work shifted from the mainstream Anglo-Australian organisations to volunteers and professionals from the various ethnic communities. The very recognition that there were ethnic communities in Australia marked a major turning point away from the search for a national identity based on cultural homogeneity. Australia had become in some respects more like Canada and the United States. Free citizens had the right to choose their language and a new society could be built on people from a variety of origins. The 1970s was the decade of multiculturalism. The official ideology is spelled out in Grassby (1973). The 1976 Census, which was the first to elicit data on the use of specific languages, reported that 12 per cent of the Australian population over the age offiveclaimed to use a language other than English. Among the languages spoken were some 75-100 migrant languages. The incidence of bilingualism varies from state to state as well as across ethnic groups (see Clyne, chapter 14, this volume).

6

SUZANNE ROMAINE

Multiculturalism, language attitudes and language policy in an Aboriginal perspective It is against this sociohistorical background that recent claims made in the 1980s to the effect that Australia had become the most multicultural country in the world have to be understood, and more importantly, that the place of the Aboriginal population in the newly formed multicultural Australian speech community needs to be evaluated. White Australia celebrated its Bicentennial in 1988, yet even in 1967 voting rights were not given by all states to Aboriginal Australians. The Aborigines were also not counted in the Australian census until 1971. Before then, the Bureau of Census and Statistics adopted a very narrow interpretation of Section 127 of the Federal Constitution, which stated: I n reckoning the numbers of the people of the Commonwealth . . . aboriginal natives shall not be counted' (Yarwood and Knowling 1982: 258-9). The very early dealings between the British settlers and the Aboriginal population were at worst violent, and at best, assimilative in nature. Following the first settlement by whites the Aboriginal population was reduced from an estimated total of 250,000 in 1788 to a low point of 66,099 in 1933 (Smith 1980: 2-3, 10-55). Hughes (1988: 120) describes what happened to the Aboriginal population of Tasmania, which was all but exterminated within less than 75 years of white settlement, as the only true genocide in English colonial history (see, however, my remarks on internal colonialism below). The way to survive was to become like the whites, as is made clear in an address made by the Governor of South Australia in 1835: 'Black man, we wish to make you happy. But you cannot be happy unless you imitate white men. Build huts, wear clothes and be useful. . . love God . . . love white men . . . learn to speak English' (Broome 1982: 27). After the Second World War assimilation was adopted as the official policy towards Aborigines. This followed decades of attempted segregation in special reserves of land not required by European settlers. Here they 'were to be restrained for both their own protection and the racial purity of the broader community' (Stevens 1970: 371). The new policy of assimilation, as spelled out in 1961 at the Native Welfare Conference, was grounded in the hope of achieving a homogeneous society. Thus, Rowley (1972: 399): The policy of assimilation means that all Aborigines and part-Aborigines are expected eventually to attain the same manner of living as other Australians and to live as members of a single Australian community enjoying the same rights and privileges, accepting the same responsibilities, observing the same customs and influenced by the same beliefs, as other Australians. The Labor Government under Whitlam, elected in 1972, had also promised liberalised attitudes towards Aborigines. One of its first tasks was

Introduction 1 to establish an advisory group on the teaching of Australian languages in Aboriginal communities in the Northern Territory. As in the United States, however, the early programmes of bilingual education were assimilationist and taught Aboriginal languages only as an aid to the acquisition of English. Decisions regarding the teaching of Australian languages were made entirely or largely by outsiders. Bourke's (1980) evaluation of bilingual schools in the Northern Territory seriously questioned the rationale behind these bilingual programmes. He observed (1980: 72) that Pre-schools are being run on European lines and the programmes appear to be based on deficit theory. Some teachers even said that the children were deprived of the experiences necessary to develop their intellect. Taking these little children away from their families where they have the security to develop their language and social mores, to place them in schools where everything is different, including values and acceptable behavior patterns, is confusing. The child's intellectual and social development may even be curtailed. In the Northern Territory today the 17 bilingual programmes in operation all remain transitional (Fesl 1988: 154). Although Queensland is the state with the largest population of Aboriginal and Torres Strait Islanders, only two bilingual programmes have been implemented. Aboriginal awareness of the need to take steps to maintain traditional languages began to be articulated in the late 1970s (see Fesl 1988: chapter 11). Among the important trends are the establishment of independent Aboriginal schools, for example, at Yipirinya in Central Australia, the formation of a national body to campaign for Australian language rights, and the establishment of Aboriginal radio and television. In 1985 the House of Representatives Select Committee on Aboriginal Education recommended government support for independent schools. It has become increasingly possible for Aboriginal people to argue their own case through agencies such as the Institute for Aboriginal Development, the National Aboriginal Education Committee, The Australian Institute for Aboriginal Studies, The Aboriginal Languages Association, and community controlled language centres, for example, the Kimberley Language Centre in Broome. A National Aboriginal and Islander Broadcasting Association was also established (see Walsh, this volume). Although Aborigines had remained separate from immigrant groups represented by the Federation of Ethnic Communities Council of Australia, it is obvious that both these groups had common linguistic and other goals to be pursued. When the Senate Standing Committee was charged in 1982 with the development and implementation of a coordinated language plan for Australia, they made 16 positive recommendations for Australian languages. Among the most important was that Aboriginal people be guaranteed the major role in decision making relating to language issues.

8

SUZANNE ROMAINE

However, little or no action was taken to implement these recommendations. The Lo Bianco report (1987) on the National Language Policy reaffirmed the importance of Aboriginal involvement, but without consultation with the Aboriginal community, it proposed a three-year National Aboriginal Languages Project to be managed within the Commonwealth Education portfolio.

Language and public policy in Australia

Throughout most of Australia's history language policy at government level was ad hoc and unofficial. It was only after the 1960s that it became official and coherent. The movement to set up a national language policy is so far unprecedented in the major Anglophone countries like Britain and the US. Thus, issues concerning language planning, public policy and education are at the moment of considerable interest and significance. Ferguson and Heath (1981) note the paradox in prescriptive attitudes towards language in the English-speaking world and the general opposition to setting up a language academy or some other regulatory body. Most nations spell out in their constitutions which languages have official status for particular purposes. No government of the major Anglophone nations has ever felt the need to declare English as its official language because English has served effectively as a de facto rather than de jure official language. Australia is the first of the major English-speaking countries to formulate an explicit language policy. Australia has also taken a number of important issues related to language use on board, for example, plain English (see Eagleson, this volume). While this movement has taken hold in other English-speaking countries such as Britain (where the Plain English Movement celebrated its tenth anniversary in July 1989) and the United States, and elsewhere, for example, Finland, Sweden, France, the Netherlands, it is possible that Australia is in the vanguard of new developments. Eagleson, who is the Federal government's adviser on Plain English and adviser to the Victoria Law Reform Commission, has been instrumental in this enterprise (see chapter 25, this volume). The Australian Government has now introduced a plain language policy and Victoria launched an extensive investigation into the language of legislation. In October 1988 the Federal Government issued its Style Manual, the 'bible' for the public service, which contained a chapter by Anne Pauwels on non-sexist language (see also chapter 22, this volume) especially commissioned from the Office of the Status of Women. It describes four major aspects of sexism in English and advises writers to be careful in their portrayal of men and women. Corson's chapter discusses the implications

Introduction 9 of inequality in another area of language usage, namely, social class differences in the ability to use and comprehend specialist vocabulary needed for mastery of certain school subjects. The Minister for Employment, Education and Training, John Dawkins, also announced in June 1989 the Federal Government's plans for the formation of the Languages Institute of Australia (LI A) to see through the full implementation of the National Policy on Languages. Over one million dollars has been allocated to the institute in thefirstyear. The LI A will have a central secretariate based in Melbourne and bases initially in tertiary institutions in Brisbane, Melbourne and Sydney, but would eventually have teaching and curriculum centres in all major Australian cities. The LIA's aims are to improve the skills of school and tertiary students as well as the wider community. The activities of the institute will be wide ranging, and include, for example three research and development programmes. One of these will be a language testing unit through the University of Melbourne and the Brisbane College of Advanced Education to assess levels of language skills. This will be used for testing standards of English among migrants seeking a particular occupation. Monash University (also in Melbourne) will be the site of a study of language and society to assess language problems of groups such as migrants and the deaf. At the University of Sydney there will be a second language acquisition programme. In addition, Queensland University in Brisbane will have a national data base of language skill information and will look at computer assisted teaching methods. The Australian Advisory Council on Languages and Multicultural Education (AACLAME) has been asked by the Minister to coordinate a review of language in higher education. This review will now become a project of the new LIA. The Bicentennial Year, 1988, also saw the publication of the Australian National Dictionary, the result of ten-years labour by Ransom. Its appearance nearly a century after the first Australian dictionary (Morris 1898) constitutes a landmark in Australian lexicography. Until recently Australian lexis had been represented only in British and American International Dictionaries. The first Australian dictionary to present a comprehensive word list in which all the pronunciations, spellings and definitions are taken from Australian English usage is The Macquarie Dictionary (Delbridge et al 1981). A Dictionary of Australian Colloquialisms by Wilkes (1978) covers the more colloquial element in Australian English which was not extensively treated in The Macquarie Dictionary. As Delbridge (1990: 69) points out, Morris revealed a colonial attitude when he declared that 'Australian English meant all new words and new uses of words that have been added to the English language by reason of the fact that those who speak English have taken up their abode in Australia, Tasmania, and New Zealand'. In contrasting this venture with that of

10

SUZANNE ROMAINE

Noah Webster's in the United States, Delbridge (1990: 69) quotes the following from Webster's Dissertations on the English Language (1789): As an independent nation our honour requires us to have a system of our own, in language as well as government. Great Britain, whose children we are, and whose language we speak, should no longer be our standard. For the taste of her writers is already corrupted, and her language is on the decline. But if it were not so, she is at too great a distance to be our model and to instruct us in the principles of our language. This linguistic declaration of independence is unparalleled in Australia until the appearance of Baker's (1945) The Australian Language, whose title confidently asserted the autonomy of Australian English in the same way that Mencken's (1919) The American Language had attempted to do for American English. Baker (1945: 11) wrote: we need some better starting point than Murray's Dictionary. We have to work out the problem from the viewpoint of Australia, not from the viewpoint of England and of the judgements she passed upon our language because she did not know it as well as we do. The 1940s also saw the initiation of Mitchell's studies of the Australian English accent in sociohistorical perspective. While Mitchell (1946) declared that there was nothing 'wrong' with Australian speech, his comparison of the Australian accent with that of educated southern British English was for some an unpleasant reminder of the extent to which Australian English deviated from RP (received pronunciation), as described by Jones and other English phoneticians. The Australian Broadcasting Corporation, created in 1932, subsequently recommended Jones' (1917) norms. However, in 1941 its chairman revealed that only two out of 450 applicants for the position of announcer could be selected. Most of those recognised as suitable were Englishmen. Due to Mitchell's influence so-called 'educated Australian speech' (which Mitchell later termed 'cultivated') was subsequently adopted as the style for national broadcasting. This variety of Australian speech, while distinctively Australian, was still close to RP, and quite different from the variety which Mitchell termed 'Broad Australian'. A cultivated accent is no longer essential for the ABC. Since 1983 it has required only 'acceptable styles of educated speech' (see Leitner 1984), and now all questions concerning pronunciation, style and usage are referred to an Australian Dictionary, not a British one. The Australian National Dictionary confines itself to the meanings which words have in the Australian context, for example, boring refers to drilling for water. Australian English is a variety of English in which government servants are public rather than civil, where Moscow is a pawn shop and you can have one moral at a time rather than a whole set of them. All in all Ransom and his compilers list 10,000 distinctive contributions to the

Introduction

11

English language. Many of these distinctive elements of vocabulary reflect the contribution of Aboriginal languages, for example, dyin - 'woman' (from the first language of the Daruk people in the north and west of Sydney) has given Australian English the word gin. There is also a rich mining terminology and diminutives. When the revised edition of The MacquarieDictionary appeared in 1985, it included 1,000 headwords and phrases which had not appeared in the earlier edition. Some of these new items have a purely Australian origin, for example, koori, the Aboriginal word by which some Aborigines now prefer to be known, and gub/gubba, the Aboriginal derogatory term for a white person. Others reflect the presence of other cultures in Australia, for example, doner kebab. Most, however, are words of British stock, for example, privatisation. American words have been present in Australian English since 1850 and even more have been taken in recently, for example, to rip off, and some replace former British ones, for example, French fries instead of chips Pelbridge 1990: 74). Other distinctive Australian usages are being investigated by Bryant (this volume), and in 1985 the first computer corpus of Australian English began to be compiled. This will allow some interesting comparisons to be made between Australian, British and American usages as the latter two varieties already have corpora, that is, The Brown Corpus of American English and LOB (London/Oslo Bergen) corpus of British English. In the same year as The Australian National Dictionary appeared the Australian Bicentennial edition of the Good News Bible (1988) was published. The American Bible Society had gone to some trouble in its original Good News Bible (1976) to avoid regional peculiarities in English usage. Although this was to a large degree successful, a need was felt for a British English version, which came out in the same year. Standard Australian English is not sufficiently different from either British or American standard usage to require significant changes, but nevertheless there is about one alteration per chapter. In some cases decisions were made simply to retain an American or British item according to Australian usage, for example, American rooster and store (where the British English version has cock and shop), and British English afterwards and burnt (where the American English version has afterward and burned). Among the entirely new items of vocabulary, grammar and spelling are the following: British and American field, which is little used in Australia, becomes variously land, paddock or pastures, shorn is substituted for sheared and measurements are given in metric units. The British edition also uses metric units but seems less at ease with them (cf. American ten pounds, British five kilogrammes and Australian five kilos, see Tulloch 1989). The upsurge of interest in, rather than embarrassment at, Australian English is another indication of the new status of Australian English and a

12

SUZANNE ROMAINE

decided move away from what Phillips (1950) called the 'cultural cringe'. This is reflected in the publication of serious scholarly studies of social and regional varieties such as Horvath (1985) as well as in popular media. Horvath's chapter in this volume shows how norms have changed. Formerly, the only acceptable standard was an external prestige form, that is, RP (received pronunciation) or 'cultivated Australian'. Now it has become much more acceptable to speak with an Australian accent, as long as it is not too broad. As far as the popular press is concerned, the results of a recent contest held by the Sunday Mail (22 January 1989), for example, reveal that 'there are more Ozerisms [a term coined by the newspaper for Australian sayings, metaphors and similes] to the square metre in Queensland than mangoes in the mud after a summer storm'.

Australia in the English-speaking world: a comparative perspective With the publication of this volume we are now in a position to make comparisons across three of the major English-speaking countries of the world, the United States, the United Kingdom and Australia. Though it is tempting to cite the many parallels, particularly between the United States and Australia, it is important to recognise that each of the three countries has unique patterns of language use, despite the social (but not historical) priority of English. Ironically, the outcome of the American Revolution (as it is known in the United States, or the War of Independence in Britain) of 1775-83 was instrumental in the convict settlement of Australia. While the war was in progress convicted criminals awaited transportation, but when the war was lost, the British Government had to look elsewhere to get rid of its convicts. In July 1786 it was decided to establish a penal colony at Botany Bay. In addition to English, the United States has a number of indigenous languages spoken by the original inhabitants of the North American continent and many new languages brought by successive and continuing waves of immigration. Britain also shows a similar profile of old and new linguistic minorities. While American English is remarkably uniform compared to British English, Ferguson and Heath (1981: xxv) observe that the language situation in Australia is probably the closest parallel to that of the USA: English dominant, Aboriginal languages, and the languages of varied immigrant groups. But Australia has no counterpart to Spanish, the roles of colonial and immigrant languages are much less salient there than in the USA, and the nature of the social dialect variation in Australian English is quite different from that in American English. Much of this is true. Australia has no counterpart to Spanish, which is second in importance to English in the United States not only by virtue of

Introduction 13

number of speakers, but also because of the special political relationship between Puerto Rico and Mexico and the United States. While the roles of colonial languages and immigrant languages are undoubtedly different, it would be unfair to say they are much less salient. Firstly, with regard to colonial languages, Ferguson and Heath have in mind remnants of colonial Spanish and French. While Australia does not have direct counterparts to these (even though there are old Australian communities), colonialism produced a range of pidgin and Creole languages, some of which exist today and play important roles in Aboriginal speech repertoires. The United States and Britain have Creoles too, though in Britain they are mainly recent phenomena associated with postwar immigration from the Caribbean. It is interesting that Spanish, which is the largest minority language in the United States, has not assumed a major role in Australia, where Italian speakers comprise the largest non-English speaking ethnic group. The fact that Spanish has been comparatively better maintained in the United States than Italian is at least partly due to the fact that Spanish is an important international language with a great deal more homogeneity than Italian, which is virtually confined to its European homeland, where usage is extremely diverse and still in competition with regional dialects (see Bettoni, this volume). It is tempting to dwell on the similarities, linguistic and otherwise, between Australia, Britain and the United States. However, there are also many differences. The United States revolted against England and Australia did not. All Australia's major institutions of parliament, bureauocracy, education, etc., and even common language are modelled on British lines. A strong sentimental attachment to what many regarded as 'the home country' persisted for a long time. Against this, however, emerged the great Australian myth of the egalitarian classless society and along with it the belief that social class stratification was not reflected in language. While Australian social class structure permits more mobility than the British system and Australian English is in some respects more uniform than British English, both of these assumptions about classless in society and language are without empirical foundation (see, e.g. Encel 1970 on society and Horvath 1985 on language). Despite its strong links with the Commonwealth, since World War II and the entry of Britain into the European Economic Community, Australia has turned its attention more to the role it plays in the Pacific socioeconomic community. Earlier in their history the three countries, but in particular the United States and Australia, employed the melting-pot metaphor about ethnicity, which suggested that over time ethnicity and other primordial identifications would decline in importance and be absorbed or assimilated within the mainstream culture. This metaphor was employed 4at home' in Britain too, with regard to both indigenous minorities and newer ones arriving

14

SUZANNE ROMAINE

from the 50s onwards from former colonies. The 'melting pot', of course, assumes the dominance of the mainstream culture and implicitly promotes its continued hegemony. As far as language was concerned, the policy was a laissez-faire one: English did not need to be taught, because children would 'pick it up' as needed. There was never any question though that the dominant language was the only viable one. The antipathy to multiculturalism and multilingualism runs deep in the ethos of all three nations, though in Australia the signs are perhaps more promising of a reversal. For instance, President Theodore Roosevelt in 1918 said: We have room for but one language here and that is the English language, for we intend to see that the crucible turns our people as Americans, of American nationality and not as dwellers in a polyglot boarding house and we have room for but one loyalty and that is a loyalty to the American people. (cited in Bikales 1985) Attitudes in the United States in the 1980s, several decades after the Bilingual Education Act of 1968, have not in many respects progressed beyond this, despite the passing of the Bilingual Education Act in 1968. However, as Fishman (1981: 518) points out, this was an act against bilingualism rather than for it. In any case, public opinion turned against bilingualism and there has been pressure to curb bilingual education programmes and other facilities. President Ronald Reagan condemned the maintenance of languages other than English as 'un-American' and declared that 'it is absolutely wrong and against American concept to have a bilingual education program that is now openly, admittedly dedicated to preserving their native language and never getting them adequate in English so that they can go out into the job market' (New York Times, 3 March 1981). In 1983 he cut the Federal budget for bilingual education. Bullivant (1984: 68) has also remarked on the resentment among AngloAustralians of the amount of Federal money being spent on ethnic schools. In 1986 there were government cuts to multicultural education. Today in the United States more recent manifestations of these negative attitudes towards bilingualism can be found in the efforts of a group called US English, founded by ex-Senator Hayakawa to lobby for a constitutional amendment which would make English the official language of the United States. The English Language Act was passed in California and other states, which makes English the official language for public use. US English also seeks to repeal laws mandating multilingual ballots and voting materials. It welcomes members who agree that English is and must remain the only language of the people of the United States. A similar group in Canada called APEC (Alliance for the Preservation of English in Canada) has as its motto: 'One language unites, two divide'.

Introduction

15

Historically speaking, Australian attitudes towards multilingualism have been no less tolerant than those of their English-speaking compatriots in the United States and Britain. Antipathy towards bilingualism is often motivated by fear that bilinguals are using the other language to talk about secret things, as can be seen in this extract from a letter written to the Sydney Morning Herald (13 February 1981): Nothing annoys me more than two or more 'ethnics'jabbering away in their native language in the company of English-speaking people, particularly in a work environment. Is it really too much to ask them to observe simple politeness by refraining from resorting to their native language in the company of Englishspeaking persons? For some time now English monolinguals have been in a powerful position at home and abroad to impose their perspective on others as the only publicly valid one. Nevertheless, all three nations have within recent years replaced this image with pluralistic metaphors which depict ethnic differences as enriching forces to the mainstream society. There are few social scientists today who would claim that 'melting pot' accurately describes the outcome of ethnic contact in the past or what should happen in the future (see, e.g. Glazer and Moynihan 1963, and Gleason 1984). There is little evidence in any of these countries to suggest that ethnic ties are weakening. Instead they are being preserved, though perhaps in subtle ways or renewed in others. The emergence of language varieties such as Kriol and Aboriginal English as symbols of Aboriginal identity is a good example. Although for some speakers, these varieties replace the use of traditional languages, it by no means implies that an Aboriginal identity has been lost (see Harris, Sandefur, and Kaldor and Malcolm, this volume). Just because the visible signs of an ethnic identity, such as distinctive dress, language and food, may fade, this does not mean that ethnicity has faded. Just because a group has lost its language does not mean it has lost its culture or ethnic vitality. It simply takes on a new manifestation: new resources are deployed in the expression of the ethnicity. Among those resources may be new languages or varieties, or transfer of communicative practices from the old language to a new one (see Eades, this volume). The role which a distinctive and traditional language plays in validating a group's identity both for in-group and out-group members varies. The belief that having one's own language is criterial for ethnic distinctiveness may be used by a state and its mainstream population to deny the legitimacy of claims to special status and land rights made by groups who have shifted from their indigenous languages to the language of the majority. Such is the case in Tasmania, for instance, where the surviving mixed-race descendants of the Aboriginal population numbering around

16

SUZANNE ROMAINE

2,500 find themselves treated as non-persons by the conservative white population. The State Government recognises neither the ethnic identity of the survivors nor any of their claims to ancestral territory or sacred sites. As identity shifts, members often hold multiple identities which are not well described by labels such as Italian-Australian, Greek-Australian, etc. Within such labelled entities there are often many subgroups, each with their own linguistic and other allegiances, but who present a common, for example, 'Italian' or 'Greek' identity to outsiders. Cities whose census figures show them to be heterogeneous often turn out to consist of relatively homogeneous sub-cultures. Some ethnic communities and subsections of them may be more 'in step' with and even contribute to the linguistic evolution of the mainstream variety of English than others, as Horvath (this volume) shows. Because ethnicity transcends biological relations, it is cued socially through the use of symbolic cultural messages about in-group and out-group membership, as in code-switching (see McConvell, this volume). Another common thread in attitudes towards languages shared by the three major English-speaking countries is what Milroy and Milroy (1985) call the 'complaint tradition', that is, complaints about the supposed incorrect use of English. Milroy and Milroy trace its history back to medieval times. Ferguson and Heath (1981: xxvii) also note that many Americans have strong convictions that there is a correct way to speak and write English, even if they are not always sure what it is. They are often apologetic about their own pronunciation, grammar, and choice of words, and they approve the criticisms which appear in editorials, public speeches, and popular books against current low standards' in English. American insecurity is notorious. Quite possibly no other nation buys so many style manuals and how-to-improve your language books in proportion to the population.

While I have no precise comparative statistics, my impression is that concern for standards of 'correct English' is no less a public passion in Australia and the UK than in the United States. In 1989, Prince Charles angered school teachers by complaining that his staff could not write or speak English properly. A few weeks previously The Times Higher Education Suplement carried a front-page article in which several Oxford professors complained about the low standards of English used by students at Oxford University and suggested the possibility of introducing remedial instruction in English grammar. This comes interestingly in the wake of a government inquiry into the teaching of English chaired by Brian Cox and recommendations for tolerance of pupils' local varieties of English while stressing the need for Standard English for particular purposes as part of a new National Curriculum. An examination of the daily press will reveal that the complaint tradition is thriving in the UK and Australia too.

Introduction

17

A look at Australian media reveals a similar complaint tradition. Australian newspapers, like British ones, have complained not only about the inadequacies of local forms of English, but also about the increasing use of Americanisms. A recent editorial in the Sunday Mail (22 January 1989) claims that Australian speech is 'far away the worst in the English-speaking world'. Statements such as these echo those expressed by commentators on Australian English from the earliest days of its history. Thus, William Churchill, a member of the American Philological Society, who visited Australia in 1911 described Australian speech as 'the most brutal maltreatment that has ever been inflicted on the mother tongue of the great English-speaking nations' (cited in Delbridge 1990: 67). The Sunday Mail article, for instance, singles out as good examples of speech certain actors whose 'accents were and never could be anything but Australian (and heaven forbid that they should be)'. These he calls cultivated, but from what is said, they are clearly not imitations of RP. However, he singles out for criticism prominent individuals such as the Prime Minister, Bob Hawke, who is described as a brilliant fellow. No dolt gets to be a Rhodes Scholar and Prime Minister of Australia. Keating is no slouch. But would you allow them to make announcements in a lift? 'level three, four, foive . . . kay? ay? kay . . . sem, ayte, noine, . . . goan deown . . .' Merciful heavens. The lingual laziness which caused people to make noises like Hawke and Keating is pretty well endemic to both radio and television . . . it's been years since the Nine Network cricket commentators . . . spoke anything but pure pidgin. The author naively suggests that 'bad speech' could be corrected in a couple of generations by elocution classes in the schools. Interestingly, given the spate of attacks on Americanisms, he says he wishes Australians could be taught to copy North American forms of speech. Although most Australians have learned at school to take an anti-American stance in language, especially in spelling, Delbridge (1990: 73) observes that by mid1985 six of Australia's major urban newspapers used the American -or spellings for British -our ones, for example, color/colour. These cases may be signs of a greater tolerance for American English. Concern for correctness and purity is also found in some of Australia's new migrant language communities (just as it is in Britain, see Chana and Romaine 1984), where the threat of increased borrowing from English and code-switching are sometimes seen as undesirable developments. Conflicts arise between what Haugen (1977) calls 'rhetorical and communicative norms'. The former belong to a highly codified standard language, while the latter show a great deal of variability and may deviate considerably, particularly in immigrant communities undergoing language shift (see, e.g. Tamis, this volume, on conflict in the Greek community).

18

SUZANNE ROMAINE

The Australian 'speech community9 in global perspective

A number of the contributors raise the notion of 'speech community', a term which has caused considerable discussion and little agreement (see, e.g. Romaine 1982). Some, like Labov (1966), have stressed the sharing of both ways of speaking and attitudes towards them as defining criteria for membership in the same community. In what sense can Australia (or indeed, any country) be considered a speech community? The nation-state is usually taken to be a highly relevant unit of macro-sociolinguistic analysis, as this and the two companion volumes (Ferguson and Heath 1981 and Trudgill 1984) and many other studies of societal bilingualism (e.g. the papers in Fishman 1978) attest. It is clearly an artificial unit in many respects since there are about 30 times as many languages as there are countries. This means that most contain a proportion of bilinguals, who do not necessarily share the same speech patterns, let alone attitudes towards them. It would not be an exaggeration to say that the notions we have of a language and speech community are very much artefacts of European models of the world. Sutton and Walsh in their chapters show how social networks and linguistic communities do not overlap in the Aboriginal view of language. Different speakers of what a linguist would call one language may belong to geographically and politically distinct networks and have little contact. In order to make sense of the fact that people claim ownership of languages they never use, one has to understand traditional patterns of socioterritorial segmentation and communication networks along with marriage patterns, residence rights, etc. Thus, there is a major difference between the spatial distribution of languages when mapped according to their association with land-owning groups and when mapped according to their actual usage by members of land-owning groups. We can only speculate about how languages and patterns of language use would have evolved on the Australian continent without the intervention of British colonialism. The process by which inequalities in the distribution of land and other resources that must be shared by coexisting groups are transformed into the familiar hallmarks of linguistic inequality is possibly peculiarly western. The coming of Europeans to many parts of the New World imposed a western pattern of sociolinguistic stratification in which knowledge of the coloniser's language became both prestigious and essential. There can be no doubt that the spread of a few western metropolitan languages such as English, French and Spanish has not only drastically reduced the number of languages in the world, but also reshaped the structures of those that survive. These western European languages have themselves become more alike through borrowing. These factors have

Introduction 19 irrevocably altered the linguistic ecology of the globe. While they suggest an overall decline in linguistic diversity, these centripetal trends are continually being counterbalanced by the creation of new languages and varieties such as we have seen in Australia with the development of Kriol, new varieties of English, and new varieties of migrant languages. There is no evidence to indicate that English is becoming more uniform, either within the main Anglophone countries or elsewhere, given the continual increase in the number of second language speakers, who are continually creating new Englishes such as Aboriginal English. The chapters in part III show that there is little uniformity in the community languages. Australia provides linguists with a chance to observe the comparative effects of contact between English and other languages in a variety of situations, though virtually no attention has been paid to contact between Aboriginal languages. Much also remains to be done on the varieties of English spoken by migrants. I have tried, however, to trace commonalities of attitude towards language use which characterise the Anglophone majority element of the major English-speaking countries. The ethos of'one language-one state', although a significant force in the major English-speaking countries is, however, a more general European phenomenon, and not specifically Anglophone. It is no accident that modern linguistic theory, which has its origins in European structuralism, takes the monolingual in a homogeneous speech community as its starting point in dealing with basic analytical problems such as the construction of grammars and the nature of competence. In Europe it has generally been the case that language differences have been associated with distinguishable territories, and later the nation-states occupying those territories. Because of the identification of national entities with distinguishable territories, heterogeneity has tended to be limited to the frontiers and was for that reason local and peripheral, for example, the Basques in Spain and France, and the Celtic 'fringe' in the British Isles and France. Thus, 25 out of 36 of the European countries are officially unilingual. In most of them, however, there are minorities (both indigenous and non-indigenous), whose languages do not have the same rights as those granted to the official languages. The marginalisation of the languages and cultures of the European states can be seen as a form of 'internal colonialism' leading to political domination of the periphery and to its economic exploitation by a 'core' (see e.g. Hechter 1975). Most of the European countries have practised exclusionist policies with regard to their own indigenous minorities as well as to immigrants. At various stages in their history most of these nations have felt that minority groups were threats to the cohesion of the state and have therefore tried to eradicate both the speakers and their languages (see Romaine 1989).

20

SUZANNE ROMAINE

Dorian (1981: 16-17, 38) notes that as early as the fourteenth century Scotland was described in terms of two opposing peoples. In his 1387 account of Scotland John of Fordun depicts the Lowlanders as 'of domestic and civilized habits, trusty, patient and urbane, decent in their attire, affable and peaceful'. The Gaels, however, are characterised as a 'savage and untamed nation, rude and independent, given to rapine, easeloving . . . hostile to the English people and language . . . and exceedingly cruel'. These attitudes were used to justify the Highland Clearances of 1800-50 (forcible evictions of tenants, referred to euphemistically as 'improvements') and to anglicise Scotland. Dorian (1981: 37) also cites the analogy drawn by one historian between the fate of the Highlanders and Aborigines at the hands of the English. Thus, Fairhurst (1964: 2): Undoubtedly the population was moved very much against its will, from lands its forbears had cultivated from time immemorial, without redress and with no right of appeal; inevitably ugly incidents would occur in these circumstances, though there was no organized resistance. The operation seems more reminiscent of the treatment meted out to the primitive Aborigines of a remote colonial area in that

period. At least one sociologist (Greeley 1989: 3) has pointed out that the British occupation of Ireland for four centuries was certainly as 'imperialistic and colonialist as anything the great powers have done anywhere in the world; and if the word genocide cannot be applied to British policy in Ireland, then it has little legitimate use'. He goes on to say that 'the "troubles" in Northern Ireland are a residue of the genocidal 17th century English policy of replacing the "savage" Irish Catholic natives with plantations of Protestant "settlers"'. In many ways the implicit 'language policy' pursued by the British in Australia (and later by Australian governments) was very similar to the one pursued by the British Government at home with its own indigenous minorities. The custom of separating Aboriginal children from their parents in order to educate them in English had an earlier precedent in the United Kingdom in the Statutes of Iona dating from 1609 (see Romaine and Dorian 1981). The Statutes had the express purpose of separating Highland children from their native Gaelic language and culture and educating them in English in the Lowlands, where they would learn not only the dominant language, but would do so in an environment where their own culture was seen as barbaric. The Statutes may well be one of the first instances of legislation in the English-speaking world designed to promote linguistic and cultural assimilation. They required 'everie gentilman or yeaman within the said Illandis to put his eldest son (or daughter) to the scuillis on the lawland, and interteny and bring thame up thair untill they may be able sufficientlie to speik, reid and wryte Inglische.

Introduction

21

Once colonial expansion began, these attitudes were applied to other settings in the conviction that the English language was inherently superior to all other languages. One of the most extreme statements of British linguistic chauvanism is Macaulay's Minute on Education of 1835. He was a member of the Supreme Council of India and was involved in the debate over the medium of instruction in Indian schools. With regard to the question of what language should be used, he observed: The claims of our language it is hardly necessary to recapitulate. It stands preeminent even among languages of the W e s t . . . It may safely be said that the literature now extant in that language is of greater value than all the literature which three hundred years ago was extant in all the languages of the world together . . . The question now before us is simply whether, when it is in our power to teach this language, we shall teach languages in which, by universal confession, there are no books on any subject which deserve to be compared with our own. (cited in Sharp 1920: 110)

The arrogance of these claims is outrageous by any standard, but seems even more strikingly so in the Indian context, where languages such as Sanskrit have high literary traditions much older than those of any of the European languages. The fact that Aboriginal languages had no written form but only oral tradition no doubt made it easier for Europeans to dismiss them as 'primitive' languages, and even today a popular view persists that these languages have only a few hundred words in their vocabulary (see Walsh, this volume). A similar stigma was later attached to some of the varieties of English acquired by Aboriginal people; namely, they were seen as unable to speak anything but a corrupted version of English. These myths about the primitiveness and inadequacy of Aboriginal languages were used to justify a policy of eradicating them. In a similar way it was easy for Europeans to dismiss Aboriginal claims to land which was seen by European eyes as uncultivated and therefore up for grabs. The battle for land rights continues today (see Koch, this volume). In much more recent history, when foreign workers came to some of the nations of Europe, they were expected to assimilate and it was not for some time that a 'language problem' was acknowledged. Even now acceptance of multiculturalism and multilingualism in some of the member states of the European Economic Community has not progressed very far in spite of the acceptance of the 1977 Council of Europe directive on the education of migrant workers (77/486/EEC), which enjoins member states to teach the mother tongue of the country of origin to the children of migrant workers and to provide tuition in the language of the host state. The British and other European colonisers carried this ideology with them along with their language to the societies of New World, many of which were already multilingual. It is against this cultural legacy that the marginalisation of the American Indians and the Australian Aborigines in

22

SUZANNE ROMAINE

the newly emergent and largely monolingual Anglophone speech communities of the New World must be seen. It is a cultural fact that no Anglophone nation anywhere has exhibited enthusiasm for any kind of bilingualism other than transitional. Only in South Africa, where English speakers are in themselves a numerical but elite minority have they accepted bilingualism, and it is an asymmetrical bilingualism in which Afrikaans speakers are more likely to be bilingual in English than vice versa. Given this emphasis, it is perhaps inevitable that bilingualism has been regarded as inherently problematic and that it represents an undesirable mode of organisation for a speech community and the individual. It has not been until very recently that real prospects for largescale language maintenance have been possible in any of the major Anglophone countries. As Fishman (1981: 517) notes (in speaking of the United States), language shift has been 'quintessential^ replacive. English as a second language has become English as a first language and the latter, in turn, has become English as an only language.' It is impossible to judge at the moment what effect present policies of multiculturalism and attitudes towards language skills in public policy will have on the ecology of Australia's languages. The recent National Agenda for a multicultural Australia foreshadows a possible Multiculturalism Bill which both declares English the national language and protects the rights of other languages. Many sociologists and political scientists are sceptical about multiculturalism. Lepervanche (1980:25), for instance, says that the historical evolution of ideologies of pluralism in Australia can be interpreted as 4a series of ideological transformations in the recreation of hegemony'. The latest is multiculturalism. Bullivant (1984) detects bandwagonism and educational faddism in the new multicultural ideology. He notes that the Australian education system has repeatedly seized on educational philosophies and ideologies from overseas, particularly from the United States and Britain, some years after these policies were current, and then later abandoned them when they proved unsuccessful. Multiculturalism may be doomed to follow a similar fate. Bullivant also (1984: 71) observes that reliance on a naive form of cultural pluralism cannot alleviate the discrimination felt by minority groups. Structural and institutional pluralism must accompany it, but this will be resisted as much as possible by the dominant majority. One alternative is for ethnic groups to set up their own institutions, such as the Aboriginal schools discussed above. In general, though, such moves are regarded by the government as risky and not to be tolerated. The Australian Population and Immigration Council and the Australian Ethnic Affairs Council has stated (1979: 4) that: A major cause for concern is whether the creation of a network of ethnic organizations and the formalization of group differences will adversely affect

Introduction

23

national unity. . . This is a delicate subject, but also a crucial one. It would certainly be legitimate for Government in a multicultural society to prevent the formation of divisive institutions that threatened national security. A more recent warning by the Australian Council on Population and Ethnic Affairs went even further in emphasising the government's obligation to maintain control over threats to what it called the 'core' (a euphemism for dominant) culture (1982: 30): Inevitably there will be clashes between the core culture and elements of the minority cultures making up our society. . . However, where clashes occur, the core culture must prevail until it is modified by consensus or by appropriate authoritative action. However, in some cases of conflict, such as human rights and freedoms, society and governments should be insistent that the core culture prevail without modification. In such situations the Council would envisage the rejection of the offending element of the inconsistent culture. How far can the development of pluralist institutional structures go without leading to a complete separateness from the wider society? Activists for Aboriginal rights such as Marcia Langton do not accept assimilation as a viable way of surviving. She says (1981: 16-22): We have rejected the notion that we are assimilating into the European population and adopting white life styles. We are exploring our own Aboriginality and are finding that the white social scientists cannot accept our own view of ourselves . . . Aboriginal 'society' and 'culture' must be seen as 'complete, integrated and consistent systems relevant to their members - not merely as a truncated (or castrated) version of any other socio-cultural system'. It is all too easy to assume that cultural pluralism is unreservedly a good thing (Edwards 1981: 33). Some of the premises on which this assumption is based rely on static and simplistic conceptions of culture, society and ethnicity. Pluralism does not always foster the acceptance of other groups and their lifestyles if the groups are segregated. As we have seen, the general climate of tolerance of diversity has not been great in any of the Anglophone countries. At the same time, however, many 'small' languages are experiencing revival just at a time when they seem to be at the point of no return (see Fishman et al. 1985). The future

In so far as each language embodies a world view which is to some extent unique, linguistic diversity is to be encouraged. This is one important reason for taking positive steps towards language maintenance in Australia and elsewhere. Language maintenance contributes to both society and intellectual life. The richness of grammatical structure displayed in traditional Aboriginal languages has attracted the interest of both

24

SUZANNE ROMAINE

descriptive and theoretical linguists. As far as theoretical linguistics is concerned, the problems posed by languages such as Warlpiri and Guugu Yimidhirr for mainstream universalist models of grammatical structure, which are based to a large extent on the standard varieties of European languages, especially English, are instructive. If, for example, Hale (1983) is right in claiming that the syntactic structure of Warlpiri is nonconfigurational, this entails assuming that alleged universals such as the Xbar convention are suspended in languages like these, or that phrase structure is derivative. Likewise, if Levinson's (1987) account of anaphora in Guugu Yimidhirr is valid, then Chomsky has over-grammaticised what must have an intrinsically pragmatic content by virtue of basing his Binding Principles primarily on English, which has a very different system of anaphora. These brief examples show how our understanding of human language is enhanced by studying Aboriginal languages. Australia is a country which can no longer aspire to be a 'melting pot'. The language policy engenders hope that language maintenance is now recognised as being in the public interest. It is imperative for all of us to recognise languages as important natural resources that must be cultivated and not squandered. Despite the present-day linguistic diversity in Australia, we must not forget the many languages which have died nor those that are under threat today.

Parti Aboriginal and Islander languages

Overview of indigenous languages of Australia MICHAEL WALSH

1.1 Indigenous languages of Australia and their speakers

The term 'Australian languages' can be used to refer to the indigenous languages of Australia including the languages of the Western Torres Strait and, perhaps, Tasmania. Most of the land mass including offshore islands through territorial affiliations has been linked to speakers of an indigenous language variety. Australia therefore contrasts with nearby Papua New Guinea, where significant areas of land have been uninhabited and unaffiliated with language groups (Wurm and Hattori 1981). Australian languages are both numerous and diverse (see map 1.1). It has been estimated that around 250 distinct languages were spoken at the time of the first (significant) European contact in the late eighteenth century. Usually each language would have a number of dialects so that the total number of named varieties would have run to many hundreds (see e.g. Sutton, this volume). These varieties were spoken by a population of around 300,000 according to estimates generally accepted until fairly recently. However, Butlin (1983) suggests that previous estimates did not sufficiently allow for the devastating effects of introduced diseases such as smallpox: he therefore proposes an original population of around 900,000. A population of this size requires a rethinking of views on the relationship between language varieties and their speakers. Either there were many more distinct linguistic varieties atfirstcontact or there was a much larger population for each variety (3-4,000 on average for each distinct language). While the detailed study of dialectal and sociolectal differentiation in Australia has scarcely begun, we can be confident that quite small populations, say, 30-40 people, maintained distinctive linguistic varieties recognisably different in terms of their labels, their vocabulary and their rules for appropriate use in social interaction (see Sutton 1978, and this volume). If we assume a fairly homogeneous distribution of language variety-to-population across the continent, then it would follow that the linguistic diversity of the original population would have been prodigious.

/ *^

AV;

1000 km

0

114

(119 J 120 r 121 V 122

rV

500 miles

0

112 113 1 0 3 m

IUZ

b

98 97

137

/140V I /i43

«^y

^ v ^ /i45

I

8 V

127

138

136

139

135

133c

4

'^. 2 \ 2

24

^ 94

°

22

V 28

25

27

i

31 32

35

o>i*"Vv

36

89 9

84

\

° 88

83

87

81

133d 133a

\

80

82

150

iq

M2

91 86 85

133 b 13

p 146 /(/151 \ ) 152

s iL

10

92

142

144

ii

1

J

7

93

126

141

(6 &

^V 96

123

A . /

125

{ * 104J l _ k o ; c.

1001°i

V

c

/£ 124

-C 4

"* 111 1 0 7 / 115 110 108106/

wvf

\

40c

133f

37

133e 131 132

3

4 Oa

129

39

406

40

79

133i 128

133h

41

43

78

\i54

61

133g

\

I

I

155

y

^-~

^—-—^~\

—K

A ) ii\

V,

75 76

Map 1.1 Approximate location of languages referred to (from Dixon 1980)

Y

5049/

73 68b

62

I,

68c

68d

CO f 5 2/ 55/

6 4

54

65

rtp

51

53

68a

c

\

63 74

60/—

56/

)

J 44/

AJ

72

y^

\

38 {

130

/

/

Overview of indigenous languages of Australia

29

Key. A single number is used for each language; the location of dialects is shown only for Margany, Wemba-Wemba and Western Desert. Adnyamadhanha 128 Alawa 100 Alyawarra 84 Andegerebina 83 Anguthimri 5 Anindilyakwa 105 Anmatjera 85 Arabana 132 Aranda 82 Arandic sbgp. 82-6 Atampaya 3 Awabakat 49 Awngthim 5 Baagandji 72 Balyku 150 Banjalang 42 Banyjima 150 Baraba-Baraba 68d Bardi 140 Bariman Gutinhma 14 Bidyara 40c Biri 36 Biyaygirri 33 Bunaba 126 Bungandij 71 Burarra 109 Burduna 151 Daly River sbgp. 119-21 Darkinyung 50 Dhalandji 151 Dharawal 52 Dharuk 51 Dhuduroa 59 Dhurga 55 Dhuwala Dhuwa! 107

Diyari 129 Djapu 107 Djaru 138 Djiga 278c 108 Djingjli 97 Dyangadi 47 Dyirbal 30 Dyirringany 56 Eastern Torres Straits language 2 Gabi 38 Gad an g 48 Gamilaraay 61 Garadyari 143 Garawa 93 Gariyarra 148 Girramay 30 Gog-Nar 24 Gugada 133i Gugu-Badhun 32 Gugu-Wara 21 Gugu-Yalanji 22 Gulnay 30 Gumbaynggir 45 Gunbalang 115 Gundungura 53 Gunggari 40d Gunggay 29 Gungoragone 112 Gunibidji 111 Gunwinygu 114 Gunwinygu subgp. 113-15 Gtinya 40b Gupapuynu 107 Gureng-Gureng 37 Gurindji 137 Gurrama 147

Guugu Yimidhirr 23 Guwa81 Guwamu 40e Jab Wurrung 68e Jabugay 28 Jaitmathang 59 Jaja Wurrung 68f Jirrbal 30 Jirru 30 Kaititj 86 Kala Lagaw Langgus 1 Kalkatungu 89 Keramin 74 Kitja 127 Koko-Bera 19 Kolijon 69 Kukatj 26 Kuku-Thaypan 18 Kunggari 80 Kurnu 72 Kurtjar 25 Kuuku-Ya'u 9 Kuurn Kopan Noot 70 Lama-Lama 17 Lardil 95 Linngithigh 7 Luritja 133a Madhi-Madhi 68b Malak-Malak 120 Malyangapa 78 Mamu 30 Mangala 142 Mangarayi 99 Mantjiltjara 133c Mara 101 Maranunggu 121 Margany 40a

Marie subgp. 32, 36, 40,81 Maung 116 Mbabaram 27 Mbara 31 Mbiywom 8 Mpalitjanh 4 Mudbura 137 Muk-Thang 60 Murawari 41 Murrinh-Patha Nakara 110 Ngaanyatjara 133e Ngajan 30 Ngaliwuru 123 Ngamini 131 Ngandi 103 Nganyaywana 46 Ngarigo 54 Ngarla 145 Ngarluma 148 Ngarndji 98 Ngawun 35 Ngayarda subgp. 145-50 Ngayawung 76 Ngiyambaa 63 Nhanda 154 Nhuwala 149 North Kimberley sbgp. 124-5 Northern Paman sbgp. 3-8 Ntra'ngith 5 Nunggubuyu 104 Nyamal 146 Nyanganyatjara 133g Nyangatyatjara 133h

Nyangumarda 144 Nyawaygi 34 Nyigina 141 Nyungal 22 Nyungar 155 Olgolo 20 Oykangand 16 Pallanganmiddang 58 Peek Whurrong 70 Pintupi 133d Pitjan(tja)tjara 133f Pitta-Pitta 87 Rembarrnga 113 Ritharrju 106 Tanganekald 77 Thargari 152 Thawa 57 Tiwi 118 Tyeraity 119 Umbuygamu 13 Umpila 9 Ungarinyin 125 Uradhi 3 Waga 39 Wagaman 22 Wakoora 22 Walmatjari 139 WWambaya 92 Wangaaybuwan 63 Wangganguru 132 Wangkumara 79 Wari 30 Warlmanpa 136 Warlpiri 135 Warluwarra 90 Warndarang 102 Warnman 134

Warrgamay 33 Warumungu 91 Warungu 32 Wathawurrung 67 Wemba-Wemba 68a Wergaia 68c Western Desert 133 Western Torres Straits language 1 Wik-Me'nh 12 Wik-Muminh 11 Wik-Munkan 10 Wiradhuri 62 Worora 124 Wuywurrung 66 Yabula-Yabuia 64 Yadhaykenu 3 Yalarnnga 88 Yandruwanhtha 130 Yanyuwa 96 Yaralde 77 Yawuru (Yauor) 141 Yaygir 44 Yidiny 29 Yinggarda 153 Yinwum 6 Yinyjibarnrdi 147 Yir-Yoront 15 Yitha-Yitha 73 Yiwaja 117 Yolnu sbgp. 106-8 Yota-Yota 65 Yugambal 43 Yukulta 94 Yulbarija 133b Yuwaaliyaay 61 Yuyu 75

30

MICHAEL WALSH

Today many Australian languages are under threat. A variety of factors have contributed to their decline (see, e.g. McConvell, this volume). Some languages have died simply because their speakers have died. In other instances the languages are dying 4by inches': the essential link between generations of speakers has been broken (probably irretrievably) and now the language is losing not only its grammatical complexity and lexical richness but also its full range of functions in social interaction (see also Bavin and Shopen, this volume). It seems that as many as 50 languages have become extinct over the last 200 years. Perhaps 130 languages have less than 50 speakers and only remain in limited use by older speakers (Yallop 1982: 30). This leaves just 70 or so languages with viable communities of speakers and of these only about 25 have 250 or more speakers. Even apparently 'healthy' languages are subject to rapid shift. Evans (1987), for example, reports that the Kaiadilt of Mornington Island in the Gulf of Carpentaria have no fluent speakers under 45 years of age. Younger speakers retain active command of a small range of vocabulary but speak Kaiadilt with varying degrees of fluency. Varieties of English have taken over within 40 years of significant white contact. The variety of English will depend on the speaker's age, sex and background (in terms of education, work experience, etc.) as well as situation. Elsewhere indigenous lingue franche have emerged as a common means of communication for a community or region. A lingua franca can also contribute to language shift. At Wadeye (on the west coast of the Northern Territory) from 1935 missionaries brought together speakers of a number of mutually unintelligible languages. Over the past 50 years MurrinhPatha (see language number 122 in map 1.1) has become the lingua franca for the area and is now used as the medium for the local bilingual education programme. Younger inhabitants of Wadeye are nearly all fluent in M urrinh-Patha but have differing degrees of fluency in the language(s) of their parents. There are some indications that children are using their parents languages less than the lingua franca, Murrinh-Patha. The most widespread lingua franca in northern Australia is Kriol (see Harris, and Sandefur, this volume), an English-based Creole used among numerous Aboriginal groups who formerly spoke quite different Aboriginal languages. Kriol varies over the vast area in which it is used by having an input from the local Aboriginal language in a particular area. Nevertheless, Kriol is largely intelligible over this whole area. 1.2

Early studies of Australian languages

The earliest known study of an Australian language dates to 1770, when Captain James Cook along with his exploration contingent was delayed at

Overview of indigenous languages of Australia

31

a remote place on the east coast of Australia. Appropriately enough, this site in northern Queensland is near modern day Cooktown. A number of his party recorded vocabulary from the local inhabitants including a word for a strange hopping animal, which has passed into English as kangaroo, perhaps thefirstloan word from an Aboriginal language and very likely the best known. Other explorers took down wordlists but it was the focused zeal of missionaries that yielded the most detailed early accounts of Australian languages. One such was Edward Threlkeld who established a mission from 1825 at Lake Macquarie about 70 miles north of Sydney. In attempting to translate scripture into the local language, Awabakal (see language number 49 in map 1.1), Threlkeld was obliged to enter deeply into the language, producing a grammar and spelling book as well as gospel translations. Such missionary work was usually fairly short term and in any case not widespread. A small but significant number of amateur scholars cast the net wider. The most ambitious project of this kind was carried out by E.M. Curr, who had been a sheep farmer in Victoria in the latter part of the nineteenth century. Curr sent out questionnaires all over the country - to policemen, squatters, surveyors, magistrates, postmasters - to settlers of all kinds including missionaries. In this way Curr collected together hundreds of vocabularies published in a four-volume work (1886-7). To put this achievement in perspective it should be borne in mind that nothing approaching the size and scope of this work had been produced in the first 100 years of settlement and it was a long time before something of its magnitude was to appear in the second century. Often contributions to Curr amounted to 100 words or less. Tragically, this constitutes the bulk of our knowledge for some languages no longer in active use, for example, Pallanganmiddang spoken around Tangambalanga just to the east of Wodonga in Victoria (see language number 58 on Map 1.1). Curr's nationwide survey made more apparent the connections among Australian languages: languages spoken great distances apart shared vocabulary which was either identical or similar enough to be recognisable. In general, there was a pattern of indifference towards Aboriginal languages and their study. Capell (1971) distinguishes two periods of research: 'scientific' (from the 1930s onwards), and, 'pre-scientific' (prior to the 1930s). The history of research can also be seen as cyclical: an initial period of fairly widespread enthusiasm and concern for Aboriginal people followed by an age of indifference interrupted from time to time by renewed interest from amateurs or missionaries, finally to end with the onset of professional scholars in the 1960s (Dixon 1980: 12). Either scenario indicates long periods when virtually nothing was recorded for the majority of Australian languages. The last 30 years have seen a tremendous increase in the quantity and

32

MICHAEL WALSH

quality of research on Australian languages, though a few individuals produced good work prior to the 1960s, Capell, for example, notable among them. However, it is only from the 1960s that the institutional framework within Australia was able to foster research in a widespread way. In the late 1950s the Summer Institute of Linguistics established an Australian branch and began placing teams of linguists to study languages for literacy and teaching Bible translation. In 1961 the Australian Institute of Aboriginal Studies was founded and soon became the most important source for research funds. In 1965 the first autonomous department of linguistics was set up at Monash University and others have followed. Prior to this linguistics was treated as part of some 'kindred' discipline: Modern Languages; English; Anthropology etc. Anthropology had also been late to emerge as a separate discipline (1926 at the University of Sydney), but even then its intellectual heritage was from Britain so that linguistics for a long time maintained a fairly marginal status within anthropology (Henson 1974). Over the last 15 years increasing emphasis has been directed towards the role of Aboriginal and Torres Strait Islander people in the study of their own languages. In 1974 the School of Australian linguistics was established near Darwin. This institution seeks to train native speakers in techniques of linguistic analysis and applied linguistics to enable them to return to their own communities as literacy workers, translators, interpreters and key participants in bilingual education programmes (see also 1.10 below). As far as theoretical research is concerned, during the 1980s Aborigines have been actively involved as co-researchers in lexicographic projects at the Institute for Aboriginal Development and through the Lexicon Project at the Massachusetts Institute of Technology. It is this kind of collaborative research which presents the greatest promise for indepth analysis of Australian languages in the future. With the aid of intuition and introspection on the part of Aboriginal people, linguists can produce better descriptions. 1.3

Classification

The first comprehensive classification of Australian languages was published by Schmidt (1919). On the basis of similarities of grammar and vocabulary he set up a grouping of 'southern languages' which differed from 'northern languages'. Schmidt worked totally from materials gathered by others and these materials were often both meagre and poor in quality. Nevertheless, the careful compilation and comparison of early sources by Schmidt remains a valuable sourcebook for later attempts at classification. The next major classification was proposed by Capell (1956), who divided languages into two groups: prefixing and suffixing. In fact, the

Overview of indigenous languages of Australia

33

terms are somewhat misleading since the prefixing languages have suffixes as well. The classification is motivated mainly by typological considerations but there is still a good deal of overlap with Schmidt's northern-southern division. Hale (1964) put forward the third major classification, and was able to show that some of the supposed differences found in some languages in Central Australia and Cape York could be attributed to fairly recent historical change. Some of Schmidt's northern division was thus shown to group with languages in the south. Hale named this new grouping PamaNyungan after words for 'man' at the geographical extremities of the country: pama in Cape York and nyungar in the south-west of Western Australia. This Pama-Nyungan 'family' covers over four-fifths of the continent and shows considerable typological similarity. Map 1.2 shows the major classification schemes for Australian languages (after Dixon 1980: 20). The remaining fifth or so of the Australian landmass is an area of extraordinary linguistic diversity in terms of the number of different languages as well as in typological variation. Quite a few of these languages have been treated as language-isolates while the others tend to fall into small groupings of between two and a dozen languages (see Yallop 1982: 27-54, for a survey of languages giving their locations, classification and the numbers of speakers of the more populous varieties). According to the lexicostatistically based classification presented by O'Grady, Voegelin and Voegelin (1966), the non-Pama-Nyungan languages can be grouped into some 20 or so 'families' on a par with the geographically widespread Pama-Nyungan 'family'. This classification is not without its problems. To start with, the materials available for many of the non-Pama-Nyungan languages were fairly meagre at the time the classification was proposed. Relying merely on lexicostatistical counts can be misleading, as later research taking grammatical information into account has shown. Some scholars (e.g. Blake 1988) have now begun to question the traditional make-up of Pama-Nyungan. One can still accept Hale's basic split into Pama-Nyungan and non-Pama-Nyungan languages on typological grounds but in terms of genetic relationship neither set of languages can be shown to have a genetic unity. Most linguists agree that nearly all Australian languages derive from a common ancestor, proto-Australian. The task ahead is to build up a clearer picture of the nature of proto-Australian (Dixon 1970, 1980:225ff.) and to reassess language groupings on the basis of careful consideration of all the available sources for each language. A long-term project devoted to the latter aim is in progress at the Australian National University under Dixon's direction.

—

Schmidt's Northern/Southern division (1919) Capell's Prefixing/ Non-prefixing division (1956)

1

6

500miles

p* 121 ^ S " ~^ 1 2 2

J^

113 103 K^yYios

20

"98

{

Z' Z g

1 4 8

-\/^ 145

147

136-

/(/i51

35

36

86

84

:

^

^

> ' 88

83**:

"I

87

81

.

80

82

\

\153

37

130 4C a

129

132

41

43

78

133g

\

^—-—'N-> x^>

\

)

J

)

^~^r^

r J ^

J

42 / /

44/ .3

72

,

. 155

61

[ )

40e

128

133h \i54 \

38 39

40b ^Od

79

133i

^

1

40c

133f 133e 131

J

.• 28 j 31...•' •' 30\ 32 3 3 , N v ,

133d 133a

C

150 152

/

x

22

:

-5

P-N

91

135

85

1

/ 26

/

133 b

1 4 6

24

^ ^ ^ >v

93 _

139

144

95

°

17

J 16 18 2911 r 19 20

92

138

^ /

n

97

J^r'

^7i43**1*42

\ M

15

Pr

123^

''5y \z

125

»)/ r

1001°1 99

(£^124 /\Jk

/ 3S

,/vV/-^ 1

• * " • • rdi T Id INyUf iydi i/ !\Un~raiTlcMNYUl lUaM QlVlblOn (due to Hale, about 1961) 1000km 0

V

h Sf ]~J\

63

75

74

76

68b 68a 68c 6 8 d

?3 62

V71 \

68e 6 8 f

64

65

66

58 &b 59

^70^69 67 ^ 6 0 / —

53

5-7 5i/ 52

J

55/ 54

5

i

y

Map 1.2 Major classifications of Australian languages (from Dixon 1980) (note that Schmidt had inadequate data on northern languages, and on their locations. His line has been slightly repositioned in the light of modern knowledge). For key, see map 1.2

Overview of indigenous languages of Australia 1.4

35

Relations with languages outside Australia

Many have speculated about possible connections between Australian languages and other languages of the world. During the nineteenth century links were proposed with Malayo-Polynesian (Austronesian) and Papuan languages. Others suggested connections with African and Dravidian languages. The Dravidian proposal has some superficial appeal. There are some similarities in broad typological terms for morphology and syntax while the phonology shows surprising similarities. Like Australian languages (see 1.7 below), many Dravidian languages have a six-place nasalstop system with essentially the same points of articulation. Some physical anthropologists have also drawn attention to the striking physical resemblance between southern Indians and northern Aborigines (Abbie 1970: 216 ff.). In addition, some Aborigines travelling in India have been struck by the apparent similarities between Dravidian and Aboriginal languages. In fact, the proposed connection does not stand up to close scrutiny. A careful investigation turns up no more than chance resemblance without any sound basis on which a credible genetic connection could be built. A connection with Papua New Guinea seems likely given its proximity to the Australian mainland but this too proves to be groundless. The Torres Strait Islands between Queensland and Papua New Guinea form an island link between the two landmasses. The northernmost islands are in sight of the Papua New Guinea coastline. However, the islands divide up into two quite distinct types: Miriam, a Papuan language spoken in the east and Kala Lagaw Ya, an Australian language spoken in the west. Kala Lagaw Ya shows some unusual features for an Australian language such as the presence of/s/ and /z/ (perhaps borrowed from the language cloSe to the east in which fricatives are an expected feature), but it is otherwise clearly of the Australian type in terms of grammatical features (see Ford and Ober, this volume and Comrie 1981). Whatever the situation may have been in the past, from our present vantage point, the Australian language family is linguistically isolated. There may have been sister languages to proto-Australian somewhere in South-east Asia, but even if there had been, any genetic trace would have gradually eroded away by inevitable change in the daughter languages. People have been in Australia for at least 40,000 years (and much longer periods have been suggested). We know through archaeological work that there has been more than one influx of people of which the most recent happened about 4,000 years ago when the dingo (wild dog) was brought into Australia. Even if the ancestry of present day Aboriginal languages goes back only 4,000 years, it is unlikely that careful study could establish links with

36

MICHAEL WALSH

languages outside. In the absence of written records it would have been well nigh impossible to establish connections among some of the more distant members of the Indo-European family which derive from the ancestor language, proto-Indo-European, spoken around 5,000 years ago. We have written records for Indo-European languages which go back thousands of years. For Australian languages the greatest time depth for any language is a little over 200 years. But there are numerous other possible scenarios - also equally unprovable. The language(s) that came in with the dingo may have been absorbed, or displaced by existing languages and thus have disappeared without a trace. If that were so, proto-Australian would have to be older than 4,000 years but how much older is very hard to say: whether it took over from existing languages at some unspecified time in the past, or, whether it has withstood all encroachments and goes back in a continuous line for tens of thousands of years. Given our present state of archaeological knowledge, it is hard to disprove most claims about suggested invasions and migrations in the past. In more recent times there is growing evidence of contact with Austronesian languages from Indonesia. From towards the end of the seventeenth century traders from around what is today southern Sulawesi visited the shores of northern Australia to collect and process a muchprized commodity variously known as trepang, bech-de-mer or seacucumber. These Macassan traders set up seasonal camps staying on the northern Australian coast for months at a time and mingling with the local Aboriginal population. It even seems that some Aborigines travelled back to Indonesia with the boat crews, returning to Australia on later trading expeditions. This contact is demonstrated linguistically by a sizeable stock of Austronesian words in some Aboriginal languages of north-east Arnhem land (Walker and Zorc 1981). In addition, it may be that a Macassan-based pidgin developed for use not only between Aborigines and the boat crews but also for casual contact between Aborigines along the coast who did not have a language in common (Urry and Walsh 1981). 1.5

Language names

The names for forms of speech in Aboriginal Australia vary in interesting and perplexing ways. There are often numerous named varieties within a small area (see Sutton, this volume). This is one reason why it is not so easy to give precise estimates for the number of languages. Consider some of the names for language varieties that can be found around Wadeye (formerly Port Keats), a remote settlement of around 1,500 people on the west coast of the Northern Territory. The lingua franca for the area is Murrinh-Patha but a number of other languages are in use: Marri-Ngarr, Marri-Djabin,

Overview of indigenous languages of Australia

37

Magiti-Ge, Djamindjung. Murrinh-Patha is itself made up of three dialects: Murrinh-Kura; Murrinh-Diminin and Murrinh-Patha. In addition a particular individual can have an array of'language' labels such as the following: Murrinh-Kura Murrinh-Diminin Murrinh-Patha Murrinh-Thitay Murrinh-Yidiyi Murrinh-Kale Murrinh-Bathuk Murrinh-Kimul

'water language' 'gravel language' 'good language' 'sugarbag language' 'Port Keats language' 'mother('s) language' 'Bathuk language' 'Kimul language'

The same person may be heard to refer to Murrinh-Yuwan and Garama. The puzzle for the investigator is to determine which of these labels are simply alternates (in some sense) and which are clearly distinct forms of speech for the purposes of making estimates of numbers of 'languages', speakers, and so on. All these labels are descriptive in some way. The first two describe the nature of the speech designated. Murrinh-Kura, for instance, is regarded as 'light' and 'clear', while Murrinh-Diminin has been described as 'a bit rough'. Murrinh-Thitay refers to the sugarbag (a kind of wild honey) totem and is thereby an appropriate label for the speech of an individual who has that as one of his totems (in the same way, there are other designations such as Murrinh-Tek - 'black cockatoo language', Murrinh-Thinmel - 'seagull language', etc.). Murrinh-Yidiyi identifies the form of speech in terms of its main centre of population, Yidiyi, the Murrinh-Patha name for the township. Murrinh-Kale refers to the fact that the speaker's primary linguistic affiliation is shared by his mother (not something that can automatically be assumed in a community where there is such a mix of different languages). The last two examples refer to place-names in the Murrinh-Patha-speaking area. Bathuk is an important locality for the sugarbag totem and Kimul, a coastal site where the township was formerly located. This leaves only two labels unexplained: Murrinh-Yuwan and Garama. The former is the Murrinh-Patha name for its southerly linguistic neighbour, Djamindjung, while Garama is the Djamindjung name for its linguistic neighbour to the north, Murrinh-Patha. I have dwelt on the explanation of these names at some length precisely because the norm is to have no explanation at all, simply a bewildering array of names in which it is not at all clear whether one name is a hyponym of another, an approximate synonym or something quite separate. This has significant implications for the classification of languages in Australia. Elsewhere in Australia language names can reveal something of the nature

38

MICHAEL WALSH

of the language being referred to. In parts of central Australia, for instance, there are a number of related varieties which form part of the Western Desert language (a dialect chain spread over a vast area of some 500,000 square miles). These varieties include: Pitjantjatjara; Yankuntjatjara; Ngaanyatjara and Nyanganyatjara. Each name has a common suffix, -tjara - 'having'. Thus, Pitjantjatjara oe is for the most part morphologically conditioned. Some examples are seen in: Initial syllable of a form governing a preceding qualifier:

Kalaw Kawaw Ya

121

gimal loeg-a-l -H/H H] (combined with the principle of automatic downdrift) accounts for the intonation pattern. There is a further tendency for certain constituents to be the natural focus of their clause, and these form the following hierarchy: WH words and nay, particles, and negatives. Focus is otherwise determined by position as outlined above, though these tendencies may be overridden by speakers if they so choose. Some minor intonation rule adjustments are required for mono- and di-syllables, as the KKY canonic form for descriptive purposes is trisyllabic. Some sample derivations are as follows; focused elements are italicised. Ezoera Gabulpa moeginakaz manin H L accent assignment H HL L accent spread 1 H H H H H H H H L L L L accent spread 2 'Ezra gave the child to Gabu' This sentence may represent unmarked focus (as in answer to the question, 'What happened?') or marked focus (as in answer to, 'What did Ezra give to Gabu?). In the latter case, the response could be elliptically just moeginakaz.

130

KEVIN FORD AND DANA OBER

The tendency for penultimate focus can be overridden by speakers, as in: Ezoera Gabulpa moeginakaz manin H L accent assignment H L HL post-focus accent HHL HHL HHLL L L accent spread H H!H H H ! H HHLL L L adjustment rule 'it was Ezra who gave the child to Gabu' Within the noun phrase, word-order appears to be flexible. As noted, qualifiers normally precede their head, though they may follow, in which case focus or emphasis applies to the initial element. Between qualifiers the order is flexible, as in: nuy senaw garkaz ngaybiya minakay kikirilayg he that boy ABS I LOC more sick PRED 'that boy is more sick than me' cf. senaw nuy garkaz ngaybiya minakay kikirilayg 'that boy is more sick than me' ukasar kuykuthal moeginakaz - 'two tall children' two tall child cf. kuykuthal moeginakaz ukasar - 'two TALL children' The difference in focus here is reinforced by intonation and the last two sentences could be spoken in such a way as to bring any of the three words into focus. In particular complex structures (see 7.5) the symbol + is used to mark 'upstep', or a return to utterance-initial pitch level.

7.4.9 Non-finite clauses 7.4.9.1 Causative/purpose Verb forms add the dative suffix to the infinitive. Certain nominal arguments also take the dative; and the semantic range is wide. Some examples are: ngoey-mu-n ubi umay-pa lag-oe-pa moey-pa want dog DAT house DAT take DAT we POSS 'we want to take the dog(s) home' nga-th nuy-n thari-pa man-i-n I ERG he ACC run DAT make PR PF 'I made him run' ngath palay tharipa may-gi-nga I ERG they 2 run DAT make NEG PRED

Kalaw Kawaw Ya

131

'I didn't make the two of them run'. nga-w ubi garkoez-i-pa puy-pa poeth-a-y-pa I POSS want man DAT tree DAT chop DAT 'I want the man/men to cut the tree/trees down' While the non-finite is possible, the finite form in the next example is stylistically preferable when the subjects change: ngaw ubi kedha + garkoez-i-n puy-l patha-moey-n thus man ERG PL PL PR PF moeginakoez-i-w ubi ngay-a-pa nga-wgeth child POSS want I DAT I POSS hand matham-oe-y-pa hit INF DAT- 'the child wants me to hit myself Similarly, the next example is normally preferred to the previous one. moeginakoeziw ubi kedha + ngay ngaw geth matham-i-z PR PF

7.4.9.2 Predication with agent This structure involves an absolutive subject, possessive agent, and participial verb with suffix -zi, followed by the predicator -ngaj-mayl, as in: moeginakaz Ezoera-n gumi gasam-oe-ychild ABS POSS secretly take 'the child was taken secretly by Ezra'

zi-nga INF

7.4.9.3 Simple predication The suffixes, which attach to the verbal infinitive, are those exemplified with adjective qualifiers in 7.3.3.1 (with positive and negative, and singular and plural forms). An alternative negative is possible with lawnga. Additional uses include: ngay igil-i-l-nga - 'I am alive' I NOM life PRED ngay ukasar geth-a-l-ayg - 'I have two arms' I

two

arm PRED

ngi bokadhzoe-gi-g - 'you don't have any money' you money

7.5

NEG PRED

Syntax - non-simple sentences.

Multi-verb sentences in KKY are essentially paratactic. There are no subordinate verb forms, and there are no formal grounds for establishing a separate class of conjunctions. Use is made of the linking clitic -a 'and', which also acts as the noun phrase linker, and a range of sentence

132

KEVIN FORD AND DANA OBER

modifiers. There is a relative pronoun which declines for case and number (singular, dual and plural), thereby showing concord with its antecedent noun-phrase. There is an alternative finite relative construction which does not require the relative pronoun, as well as a non-finite relative construction. Sentence intonation reinforces the paratactic analysis of multi-verb constructions in that 'upstep' highlights the break between clauses. Five types of construction will be considered separately.

7.5.1

Clitic-conjoining

The linker is -a - 'and' which attaches to the final word of the preceding clause, as in: lawnga Ezoera woerab-a-l poer-a-y-gi-nga -a NEG

PL pluck INF NEG PRED and

+ nu-ngu ira - pa gumi ma-y-gi-nga he POSS in-law DAT secretly give 'no, Ezra didn't pick the coconuts and give them secretly to his in-laws' The linker is omissable only in noun phrase concatenations, as in: palay alay ipa - 'husband and wife' they 2 husband wife

7.5.2 Appositional complements Finite sentences can be set in apposition to the modifier kedha - 'thus', typically after verbs of saying, perceiving and questioning. Such complements can be statements, questions or direct speech, kedha is syntactically part of the controlling clause and is omissable; the appositional clause is marked intonationally by upstep; as in: nuy ngayapa kedha muliz + Ezoera Gabulpa he I DAT thus say PR PF DAT soekoeri maninu arrow ABS give IMM PA 'he told me that Ezra gave the sharp-pointed arrow to Gabu' (61) ngay Ezoeralpa kedha yapupoeybiz + nuydh ngabepa DAT ask PR PF he ERG who DAT dagul manin spear 'I asked Ezra who he had given the fishing spear to'

Kalaw Kawaw Ya

133

7.5.3 nanga clauses Temporal, potential conditional, and relative sentences are linked by the particle nanga, which may not be sentence-initial and which may be repeated in long sentences and/or for emphasis. Clause breaks are invariably marked by intonational upstep. nanga is intonationally the information focus of its clause. 7.5.3.1 Temporal clauses Typical temporal modifiers are mi-thunara and namuyth, as in: sena-w-kay nanga dhangal namuyth karay-pa that M FUT dugong up DAT uzar-ipa + buway-garkoez-i-n-kay senabi thunar-a go PR IMP harpooner ERG FUT that time ADV wap nithu-ypa harpoon cast PR IMP 'as soon as the dugong surfaces, the harpooner will throw the harpoon' 7.5.3.2 Potential conditions The particle nanga is used with either of the future time clitics kay and bat, as in: ngalpa-kay nanga dhangal gasamay-g kayb we PL dugong catch NEG today + ngalpa-kay ay-gi-gal kun-i-ya food NEG PL back COM 'if we don't catch a dugong today, we won't have any food to take back' 7.5.3.3 Relative clauses Finite expressions employ either of the formulae: mi- NP na(nga) and REL na(nga), as in: mi-moeginakoezin nanga dhugu pathan + nuy uzariz ERG log chop he go 'the child who cut the log went away' Ezoera soekoeri senawbi moeginakoezipa manin that (M) DAT give + wapi ngadh nanga pagan fish who ERG spear 'Ezra gave the sharp-pointed arrow to the child who speared th&fish'

134

KEVIN FORD AND DANA OBER

The non-finite relative has the following structure: NP + POSS ( = object of relative verb), Verb + INF + -zi ( = participial form), NP (with functions of s, A or p in its finite clause). Compare the following finite and non-finite equivalents. wapi mi-moeginakoezin nanga gasamoenu 4- nuy nungu fish ABS.child ERG catch IMM PA he he POSS loegoenu 'the child who caught the fish is at home' wapi-w gasam-oe-y-zi moeginakaz nungu loegoenu POSS

INF PC

'the child who caught the fish is at home' The participial expression may also be used predicatively, in which case the choice of predicative suffix (singular/plural pairs -g/ -gal and -ngaj -mayI) presupposes a subjective and objective force, respectively (see 7.3.3.1), correlating with the system of number concord and ergativity (see Comrie 1981), as in: senawbi moeginakaz wapiw gasamoeyzig that M child ABS fish POSS catch PRED 'that is the child who caught the fish' sethabi moeginakoezil wapiw gasamoeyzigal those

PL

PRED PL

'those are the children who caught the fish' senawbi moeginakaz baydamaw thoeydhayzinga 'that is the child the shark mauled' sethabi moeginakoezil baydhamaw thoeydhayzimayl 'those are the children the shark mauled' As expected, only the subjective forms {-gj-gal) are possible with intransitive verb forms, as in: soekoeri mi-moeginakoezin nanga Gabulpa arrow ERG manin + nuy mayoedhayzig/*mayoedhayzinga give he cry PRED 'the child who gave the sharp-pointed arrow to Gabu is the one who cried'

7.5.4 Irrealis conditions Counter-factual conditions are marked by the particle nay in the conditional clause and by the clitic -nay in both clauses. Intonational upstep marks the clause boundary. The particle nay occurs penultimately in its clause as the information focus, as in:

Kalaw Kawaw Ya

135

Ezoera-nay Gabulpa dagul-al nay ma-may-i-nu DAT

PL

PL

IMM PA

+ nunga apuwan-nay nuyn koey-m-a mathamoenu he POSS mother ERG he ACC big ADV beat IMM PA 'if Ezra had given the fishing spears to Gabu, his mother would have beaten him soundly' 7.5.5 Other ad verbally linked clauses Sentence modifiers which help to link finite clauses are se kasa (midh) 'although', kasa(kay) kedha-'bxxi\ kedhazungufkedha) - 'because', wagel'or else', lawnga - 'or (not)' and kurusipa - 'until'. Examples are: nungu apuwan nuyn mathaman + se kasa Ezoera Gabulpa dagulal mayginga 'Ezra was beaten by his mother even though he didn't giv§ the fishing spears to Gabu' ngaw ubi Dawinipa maypa + kasa-kay kedha ngay bokadhzagig 'I want to go to Darwin but I don't have any money' Notes 1. We have used the following abbreviations. locative LOC ablative ABL masculine M absolutive ABS NEG negative accusative ACC NOM nominative adjective ADJ P adverb patient ADV PA past augment AUG PART particle aversive AVS PC participle/participal comitative COM PF perfect dative DAT PL plural dual DU POSS possessive ergative ERG PRED predicate feminine F PR present future FUT REC recent HAB habitual REL relative imitative IMIT REM remote immediate IMM S subject imperfect IMP SG imperative singular IMPER TR instrumental transitive INS 1 downstep (the lowering of the following tone) intransitive INTR

136

KEVIN FORD AND DANA OBER

Appendix 1 KKY sample noun declensions NOM^j >ABS ACCJ ERG/INS POSS DAT ABL LOC COM IMIT PL

'foodstuff' 'Boekar(f)' boekar ayza boekarina boekar ayzapun boekarina ayzaw boekaraw ayzapa boekaroelngu ayzangu boekaroenu ayzanu boekaraniya ayzaya boekaradh ayzadh boekaral ayzapul

'Diwe(f)' diwe diwena diwe diwena diwew diwelngu diwenu diweniya diwedh diwel

'g'child'

'woman'

ngep

yoepkaz

ngepan ngepaw ngepapa ngepangu ngepanu ngepiya ngepadh ngepal

yoepkoezin yoepkowziw yoepkoezipa yoenkoezingu yoekkoezinu yoepkoeziya yoepkoezidh yoepkoezil

KKY sample verb conjunctions: verb dudup - 'drown* (tr/intr) Transitive Intransitive Pres. Perf.

SG.

dudupan

PL.

Pres. Imp erf.

SG.

dudupoepa

PL.

SG. DU.

duduparngu dudupoenu dudupoengu

PL.

dudupoedhin

PL.

dudupoedha

PL.

dudupar

PL.

SG.

(kay, bat)

DU. PL.

dudupay dudupoemar dudupoemoeyr

DU.

Fut. Simple

dudupidha dudupoemoedha dudupoemoeydha

DU.

Rem. Past Imp. SG.

dudupaydhin dudupoemoedhin dudupoemoeydhin

DU.

Imm. Past Imp. SG.

dudupayngu dudupoemoengu dudupoemoeyngu

DU.

Rem. Past Simple SG.

dudupima dudupoemoenu dudupoemoeynu

DU. PL.

Rec. Past Simple SG.

dudupayrngu dudupoemoerngu dudupoemoeyrngu

PL.

Imm. Past Simple SG.

dudupipa dudupoempa dudupoemoeypa

DU.

Pres. Hab. Rec. Past Imp.

dudupiz dudupoeman dudupoemoeyn

DU.

dudupoene

dudupidhe dudupoemoene dudupoemoeyne

Kalaw Kawaw Ya

SG.

Past Hab.

dudupupu

dudupipu dudupoemoempu dudupoemoeypu dudupoeyginga

DU. PL.

Negative Imper. Pos.

SG, DU, PL. SG.

dudupar

dudup dudupoemar dudupoemoeyr dudupayg

DU. PL.

Imper. Neg.

137

SG, DU, PL

Deictic system Proximate In view Not in view Locational Remote In view Not in view Locational

M

F

DU

PL

in

ina

inubi inuki

inabi inaki

ipal ipalbi ipalki

itha ithabi ithaki

senaw senawbi senawki

sena senabi senaki

sepal sepalbi sepalki

setha sethabi sethaki

'over there' Nominal Locational 'up there' Nominal Locational

pinungap pinungapki

pinangap pinangapki

pipalngap pipalngapki

pithangap pithangapki

pinuka pinukaki

pinaka pinakaki

pipalka pipalkaki

pithaka pithakaki

'down there' Nominal Locational

pinuguy pinuguyki

pinaguy pinaguyki

pipalguy pipalguyki

pithaguy pithaguyki

'up at the front Nominal pinupay Locational pinup ay ki

pi napay pinapayki

pipalpay pipalpayki

pithapay pithapayki

'down at the back' pinupun Nominal Locational pinupunki

pinapun pinapunki

pipalpun pipalpunki

pithapun pithapunki

The initial/?/- of the forms in the lastfiveseries is omissable. Examples of concord with the locational forms. a. ngath burum senawki wakaydhin I ERG pig ABS chase REM PA 'I chased a pig there' b. thana burum senawki wakaydhin 'they chased a pig there'

thana burum sepalki wakaymoedhin 'they chased two pigs there' d. thana burumal sethaki wakaymoeydhin 'they chased pigs there' c

ApA. KKY personal pronoun system Nominative Ergative subject

Accusative object

Possessive belonging to

Locativeaccompaniment Ablative from with

Dative to

Imitative like

1 I

ngay (intr.) ngath (tr.)

ngoena

ngaw (M) nguzu (F)

ngaybiya

ngawngu

ngayapa

ngawdh

2

ngi (intr.) ngidh (tr.)

Person

c

I N

a

U L

ngin

nginu

ngibiya

nginungu

ngibepa

nginudh

na (intr.) nadh (tr.)

nan

nanu

nabiya

nanungu

nabepa

nanudh

nuy (intr.) nuydh (tr.)

nuyn

nungu

nubiya

nungungu

nubepa

nungudh

1 we (exc.)

ngalbe

ngalbe

ngalben

ngalbeniya

ngalbelngu

ngalbelpa

ngalbedh

he&I we (inc.) you & I

ngoeba

ngoeba

ngoeban

ngoebaniya

ngoebalngu

ngoebalpa

ngoebadh

2 you (two) 3 they (two) 1 we (exc.) they & I

ngipel palay ngoey

ngipel palay ngoey

ngipen palamun ngoeymun

ngipeniya palamuniya ngoeymuniya

ngipelngu palamulngu ngoeymulngu

ngipelpa palamulpa ngoeymulpa

ngipedh palamudh

we (inc.) we & you

ngalpa

ngalpa

ngalpan

ngalpaniya

ngalpalngu

ngalpalpa

ngalpadh

you

ngitha

ngitha

ngithamun

ngithamuniya

ngithamulngu

ngithamulpa

ngithamudh

3 they

thana

thana

thanamun

thanamuniya

thanamulngu

thanmulpa

thanamudh

you

3 she

R

he

D U A

L

p

L U p A

2

ngoemudh

T

Kalaw Kawaw Ya Appendix 2 Basic wordlist (Swadesh 100) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

I

ngay

you sg

ngi

we

ngoey

this that who? what?

ina

not all

many one two big

long small woman man

person fish bird dog

louse tree seed leaf root bark skin flesh blood bone grease

sena nga

miza lawnga mura koeyma urapun ukasar koey kuykuthal moegina yoepkaz garkaz mabayg wapi uruy umay ari puy

kapu nis

sipi piya goengaw madhu kulka ridh idi

egg

kakur

horn tail feather hair head

—

ear eye

nose mouth

koewb baba yalbup kuyk kawra purka pati gud

43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84

tooth tongue claw foot knee hand belly neck breast heart liver drink v eat v bite v see v hear v know v sleep v die v kill v swim v fly v walk v come v lie down sit

stand give say v sun

moon star water rain stone sand earth cloud smoke fire ashes burn v

dhang noey awar ngar kulu geth maytha mudul susu ngoenakap sib

wan-i purth-a thoeydh-i im-a karngem-i ngulayg (Pred) uthuy yun-a um-e umamatham-a way-i palg-i mab uzar-i ay-a apiya yun-a than-u thar-i ma-

mul-i/thar-a goeyga moelpal thithuy nguki ari kula buthu boeradhar ziya thu muy

kunar muy punath-a

139

140

KEVIN FORD AND DANA OBER

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

path mountain red green yellow white black night hot cold full new good round dry name

yabugudh pad

kulkagoemul maludhgoemul bamidhgoemul goerabgoemul kubikub kubil komal gabu gudapoelam kayn kapu bokadh thepadh nel

List of particles

na(nga)

[temporal marker] [irrealis condition] midh [interrogative] nag [interrogative] waza [interrogative - positive orientation] se kasa midh [concession] [obligation] ngaru sike [possibility] matha [durative] 'thus' kedha kedhazungu kedha 'because' 'but' kasa-kay kedha 'like, for example' mathakedha minakoey [comparative] bu [superlative] dhapa [superlative] wagel 'or else' kurusipa 'until' lawnga 'or' nay

Kalaw Kawaw Ya 141 Appendix 3 Sample text Bayra ~- n adhi

(Danalgub-a-n um-a - y - zi-nga)

Baira POSS story ABS Danalgub POSS speak INF PC PRED

The story of Baira (as told by Danalgub) War thunar-a thana kulba moebayg-a-l kedha Other time ADV they NOM old person PL thither Once upon a time people used to go across to uzar-moey-n Sigabadhoer-pa burum-pa gO PL PR PF

DAT pig

lum-a-y-pa.

DAT hunt INF DAT

Sigabadhoer to hunt pigs. Umay man-i-n. Koey-goesar garkoez-i-l ladh-u-n Dog ABS take PR PF Big many man PL go PL PR PF They would take dogs, and quite a large group of men would go. Thana kal-a-nu burum-i-yalum-e-mi-n. They NOM back LOC pig COM hunt PL PR PF And they hunted for pigs. War moebayg-a-n na burum nanga matham-a-n nanga Other person ERG when pig ABS when kill PR PF when When the others killed a pig, + nuy-dh Bayra burum matham-oe - y - gi - nga. he ERG

Baira

pig ABS kill

INF NEG PRED

Baira failed to bag one. nuy nga - w dhugu. Senaw Bayra nanga That M Baira particular he NOM I POSS g'father That particular Baira was my grandfather. Thana ladh-u-n. Nu-ngu kasa gamu-kawba-as-i-n. They NOM go PL PR PF He POSS only body tired get PR PF They moved on, and he just grew weary. Yan burum-i-ya lum-i-z + war moebayg-a-n nanga In vain pig COM hunt PR PF other person ERG when He hunted in vain for a pig while the others burum koey-m-a matham-a-n pig ABS big ADV kill PR PF bagged many. Nuy kedha uzar-i-z -a 4- war wang gasam-a-n. He NOM thither go PR PF and other jungle reach PR PF So he went to another part of the jungle.

142

KEVIN FORD AND DANA OBER

Nuy siki woelmay. Kedha nanga nuy He NOM there keep going PR PF Thither when he NOM (irreg. vb) He was walking around there and when he nag-i-z + nuy-dh mabayg im-a-n, look PR PF he ERG person ABS see PR PF looked round, he saw someone. Nuy goengaw adhal me-pa. Thabu koey za. He NOM skin ABS outside be PR IMP Snake ABS big thing ABS His skin was lying beside him. It was a big snake. Wa dhawdhay moebayg. Mabayg adhal nipa. Yes mainland person ABS Person ABS outside sit PR IMP Yes, it was a mainlander, sitting outside his skin. Nuy-dh woesar

koba-nga purth-a-ypa.

He ERG wallaby ABS raw PRED eat PR IMP

He was eating a wallaby raw. Nuy nu-be-pa kedha + 'koeymeg-a ngi friend Q you NOM He NOM he DAT thus He said to Baira, 'Friend, are you here, too?'

kay-a here Q

Wa ay a +ngoe-ba ina woesar koba-nga purth-a-ypa. Yes come you + 1 DU this wallaby raw PRED eat PR IMP Yes, come! We two can share this raw wallaby meat.

8 Understanding language shift: a step towards language maintenance PATRICK McCONVELL

8.1

Facing facts and facing theories

If nothing is done about it, almost all Aboriginal languages will be dead by the year 2000. Even the two most likely survivors, the Yolngu languages of north-east Arnhem Land and the Western Desert language may not last long beyond that date. Most of us who have worked for some time in the field of Aboriginal languages would agree with statements like this. However, if we were asked to show why we thought a particular language was going to die, we would often not be able to give a very coherent account of our reasoning. Nor would any two researchers necessarily come up with the same kinds of answers about how and why a language dies. In recent study of a dying language in Australia, Schmidt (1985) was unable to give any general theory of language death that would fit the many different linguistic and sociocultural features of the different languages that have died or are dying. She points out that linguists can predict neither when nor what types of changes will occur in language contact situations generally, despite some decades of impressive work on the subject. She also notes that sociocultural factors are more important than linguistic factors in determining whether a language survives or not. I believe that a beginning has been made in constructing a sociocultural theory of language death. In this discussion I shall be concentrating on recent accounts of why and how a language lives and dies. I shall not be describing, therefore, the history of massacre, forced movement, and institutionalisation of Aboriginal people, nor the periods of prohibition and denigration of their culture and languages by schools and other bodies which are necessary background for the understanding of the present situation of Aboriginal languages in Australia (see Hudson and McConvell, 1984; and Fesl 1988). I shall focus on the present situation and how theories of language shift can contribute to language maintenance.

144

PATRICK MCCONVELL

Theories of language shift (the change-over in a community from speaking one language to speaking another) have strong implications for the practice of language maintenance. If we have wrong ideas about how and why people change from one language to another, we are not likely to find the right ways of stopping or reversing the process. Also, whether they say so or not, people who advocate a particular way of maintaining a language are usually guided by assumptions about language shift. If the ideas that lie hidden behind their practice are not brought out in the open, they cannot be tested to see if they hold water. 8.2

Language shift

Language death and language shift can be regarded as two ways of looking at the same type of process. Most studies by and large consider only what is happening to one of the languages in the language situation. What is needed is a framework which looks not at individual languages but at entire bilingual or multilingual situations, seeing the functions of each language as fitting together to make a whole. Changes in one language's use are seen as related to changes in the use of other languages, and not necessarily related to a particular end-point. The future development is to be predicted and not assumed from the state of the whole situation and the theory. Here I will draw on three approaches to the study of language shift and death as they apply to findings from my research on the maintenance of Aboriginal languages: the domain theory, the interactional-variationist theory, and the adaptation theory. 8.3

The domain theory

The domain approach to language maintenance may be traced back to a conception of the relationship between 'stable bilingualism' and separation of the two languages into distinct domains of social interaction (sometimes termed 'bilingualism with diglossia'), attributed to Fishman. On the other side of the coin is the hypothesis that lack of strict separation of domains is a sign of 'transitional bilingualism'. In published work, Fishman (1972: 115) seems to soften statements about this correlation. A stronger form of the hypothesis has been advanced by Di Pietro (1970: 19) under the title 'Universal Number One': The presence of multilingualism in a speech community depends on the association of each language involved with specific domains of social interaction . . . a perfect balance of multilingualism in which, say, English and Spanish would be used equally as well for all domains of interaction is highly transitory and represents the step just before a new stage of monolingualism in one or the other language.

Understanding language shift

145

This way of putting it really talks about only the two extremes of bilingualism, both of which are relatively rare: where there is strict separation of domains (and where bilingualism is stable, according to the hypothesis) and where both languages are used equally in all domains (highly unstable, according to the hypothesis). Thus, this view does not tell us much about intermediate states, where one language is chosen for some domains but both languages are used in others. This hypothesis is presented as a correlation at one point in time: it does not predict when, how or why bilingualism shifts historically from one state to another. There are several ways in which this model is inadequate. It does not account for the social functions of the codes, and in particular does not explore the social meaning of the codes in code-switching, which has been shown in many studies (Blom and Gumperz 1972; Gumperz and Hernandez-Chavez 1972; Scotton and Ury 1977; Gal 1979). A number of people working in the domain model or similar frameworks still regard code-switching in the same domains, particularly within sentences, as random or unorganised by definition, in the tradition of Weinreich (1953: 73). This lends to a negative view of code-switching, and code-switching is blamed for language shift (Harris 1977). Scotton (1982) has argued, on the contrary, that code-switching may be a positive force, helping to maintain multilingualism. The concept of domain itself is not very well defined. Domains have traditionally been seen primarily in terms of place or setting, but there is an increasing amount of evidence that place or setting has very little or no role in choosing which language to use in most bilingual situations (Denison 1971; Gal 1979; McConvell 1986). What appears to be an effect of setting may result from a combination of the other two factors which are usually referred to in defining domains: the identity of participants, and the topic of discourse. Where more than one language is used in the same domain (with the same people and talking about the same sorts of things) this hypothesis predicts language shift. In such a model, 'expressive' or 'metaphorical' code-switching between languages used by the same people talking about the same subject is a sign of impending shift. We can certainlyfindexamples where stable bilingualism, often of very long standing, is reportedly correlated with separation of language use into domains, for example, Paraguay (Rubin 1961) and Village India (Gumperz 1969). On the other hand, there are places where 'expressive' code-switching is found in communities which appear to be undergoing language shift, for example, Eastern Austria (Gal 1979; Gumperz 1982b) and Northern Australia (McConvell 1986). However, there are equally plenty of examples of 'expressive' code-switching in stable bilingual situations: Northern Italy (Denison 1971), East Africa (Scotton and Ury 1977), French Canada,

146

PATRICK McCONVELL

Belgium, the Philippines, etc. The existence of the common correlation of code-switching not structured by domain with stable bilingualism is enough to force us to qualify or change the historical prognosis that codeswitching leads to code shift. 8.4

The interactional-variationist theory

The interactional analysis of the use of different types of speech has revealed that much of the variation in language which might previously have been dismissed as random is in fact quite systematic, and is a vehicle of social meaning. Speakers switch to a different language, or use a different dialect form, not necessarily because they are speaking to a different person or about a different subject, but because they want to express a feeling or point of view about something which is going on. Although such codeswitching seems to be used in a number of different ways, its meaning can usually be tracked down to the basic meaning 'them' versus 'us' (Gumperz and Hernandez-Chavez 1972), that is, expressing identification with one or other social groups and the values associated with them. There is clear evidence that bilingual code-switching and other types of linguistic variation are used by Australian Aborigines in this way, too (Haviland 1982; McConvell 1986). This function of code-switching occurs both in language shift situations and apparently stable bilingualism. So does it tell us anything about the process of language shift? The most interesting hypotheses in this area are those which try to relate codeswitching between languages to style shifting or choice of registers in a single language. It has been widely acknowledged that the choice of language in bilingual situations plays a similar functional role to the choice of style or register in a single language (Gumperz and HernandezChavez 1972). Recent work on language shift seems to show that functional equivalence of style switching and language switching means that in the historical change of language shift, one can replace the other. Gumperz (1982b: 52) reports that in the younger generation of a bilingual community in Austria the skills of stylistic manipulation in German are replacing the reliance on switching between Serbo-Croat and German characteristic of the older peopler. Gal (1979) shows that style switching and language switching are complementary to each other. Style switching is used between speakers where only one language is habitually used, whether it is the old language Hungarian, or the new language German, while language switching is used where two languages are habitually used between speakers. This suggests that, over time, style choice in the old language gives way to language choice, which in turn gives way to style choice in the new language. I have some doubts about whether the usage studied by Gal is strictly

Understanding language shift

147

complementary, even in the Austrian situation she investigated. In my own work on Gurindji code-switching (McConvell 1988) both style shifting (in the sense of alternation between a local and more widespread dialect as the term is used by Gal) in both Gurindji and Kriol, and language switching between them occur in the same conversations, used by the same people. However, we could take a slightly weaker position and say that collapsing of style or register distinctions is a feature of the old language of the younger generation of speakers in language shift and attempt to relate this to the language switching behaviour of those of their parents' generation. Thus both loss of style/register distinctions in the old language and increasing use of them in the new language by the younger generation could be symptoms by which we could recognise language shift in progress perhaps a more reliable indicator of language shift than the presence of language switching alone. Gal (1979) herself does seem to regard the presence of conversational language switching as an accurate predictor of language shift, like the domain theorists discussed in 8.3. This view seems to arise from certain properties of her model. The theoretical problems arise mainly when we try to extrapolate from that to make it a model of a diachronic process. As with the domain theory, this model predicts that once the process of change in functions of the languages had started, it would continue spreading through the social groups involved inexorably until the new language became categorically accepted throughout the community (see e.g. Bailey 1973; Labov 1972; and Bickerton 1975, on variationist models of language change). Language change and changes in the use of languages do not always follow this pattern. Different changes can be generated from different groups and spread along different paths. We have seen that variation, such as code-switching, need not always immediately give way to a categorical pattern but can remain fairly stable for a lengthy period. Moreover, as Gal (1979: 154) points out, a pattern like the one shown in the chart need not necessarily mean that a historical change has taken place. It could be that change in use of the languages is a cyclical phenomenon related to the age of individuals: Either speakers regularly change their patterns of language choice as they age, so that in each generation young people use more German and then switch progressively to using more Hungarian as they grow older, or . . . people within a generation retain their patterns relatively unchanged throughout life, but each generation systematically differs from the preceding ones, so that old people's patterns constitute a historically older pattern that is being replaced as the older generations die, by the newer patterns of the young. It is by no means unknown for young people in a group to speak differently from their elders, and later to adopt the older people's way of

148

PATRICK McCONVELL

talking. In fact, Gal is simplifying the question by posing it as a simple alternative like this. Among other possible explanations of the pattern on the chart is that, formerly, people spoke varieties of one language throughout their lives, but that more recently the sociolinguistic conventions have changed so that young people speak mainly the new language, then later in their lives speak the old language. This kind of scenario might have some appeal in some of the Australian Aboriginal situations that I know. It is known that many Aboriginal groups either have or had a distinct 'Baby Talk' variety of the language that was used to address children sometimes up to the age of seven or eight. Where it is now customary for older people to address children in Kriol, the new language, as among the Gurindji and Kija, one could speculate that Kriol has taken on the functional role that the Baby Talk form of the language had before. I have seen examples of Gurindji children who during their school years appeared to talk nothing but Kriol, but who begin to talk more Gurindji in their late teens as they are accepted as members of the adult group. I would not argue that the position of Kriol is wholly a life-cycle matter, nor would I deny that language shift seems to be well underway in such Aboriginal groups. I am merely making the point that the cycles and the historical process can become mixed in a potentially confusing way. This also raises other issues which the simple variationist model cannot handle adequately. One of these concerns the role of passive knowledge of a language. Clearly Gurindji youngsters who on the face of it speak only Kriol are also building up a passive knowledge of Gurindji, and when the time and circumstances are right, may use that knowledge actively. In the meantime, they will continue to speak Kriol even when spoken to in Gurindji. This is an example of what Gal calls 'unreciprocal' use of a language: it is also common for Oberwarter young people to carry on such conversations in two languages with older people: the old person talks Hungarian and the young one, German (Gal 1979: 110-11). I argue below that such asymmetries may be among the key features that point to the progress of language shift. 8.5

The adaptation theory

Bavin and Shopen's observation of linguistic change among school students in Warlpiri communities made them suspect that language shift may be a danger, even for this reputedly 'strong' language (Bavin and Shopen 1985b, and this volume). Shopen has been advocating language adaptation or language engineering as a language maintenance measure for Warlpiri. This strategy is based on the theory that rapid cultural change is

Understanding language shift

149

responsible for language shift, because traditional Warlpiri is increasingly unable to talk about all the new things and ideas that are being introduced from the modern world (see Poulson et al. 1986). In a sense, this theory is a variant of the domain theory, and the phrase 'contemporary domains' is used for the areas such as shops, European law, sport etc. in which adaptation is going on. The emphasis is on topics, rather than setting or the identity of people spoken to, as determining the language used. Furthermore, according to this theory it is not so much simply the association of a topic with a social group and its values which makes a speaker choose a particular language but the inadequacy of the old language to express certain concepts which forces the switch to English. This is plausible, and certainly I have heard Aboriginal people give lack of adequate words for a particular topic in the other language as a reason for choosing to speak in a particular language (Hudson and McConvell 1984: 43). Also, it brings out the question of the enormous cultural differences between traditional Aborigines and Europeans reflected in their languages and cultural differences of a type and scale not found between two European languages in bilingual situations such as the one Gal looks at in Austria. Certainly, these cultural differences and resulting difficulties in talking about certain things belonging to one culture in the language of another do motivate some instances of language choice. In my experience, this theory may not explain even half the language choices in Aboriginal bilingual situations, no matter what the people say or believe about their language behaviour (a notoriously unreliable guide among any group of people). Why does the person telling off the people at a picture show talk English, when he could perfectly well do it in his old language (S. Harris 1984: 136)? Such usages relate to the social functions and social meanings of the languages (e.g. the use of a superordinate language as an authoritative language), and not to particular cultural differences between groups. This is also why there is remarkable similarity in such uses of language in bilingual situations throughout the world. It may be that language engineering will not provide the language maintenance strategy we are looking for, because language shift could happen anyway, for other social reasons that have been overlooked by the adaptation theory. If one believes in the diachronic domain theory, one would see language engineering as creating the very conditions for language shift, by making the old and new languages share the same domains. I do not think the domain theory is necessarily correct about that, but I do fear that language engineering could spend so much effort altering the old language and culture that it would become a tool of cultural assimilation.

150

8.6

PATRICK MCCONVELL

A better theory?

All these theories have something to contribute to a better, more useful, theory. I shall now discuss how they could be integrated and made more realistic. Firstly, I agree that the expression of either solidarity with or distance from certain social groups is probably the most important function of language choice in bilingual situations and probably also the factor which above all other determines whether language shift takes place. However, this is not the whole story, and for this reason I have written of three functions of language choice in bilingualism (McConvell 1986). Thefirstis the basic communication function. This has to do with relations between bilinguals in the speech community and outsiders who know only one of the languages in the community. Under these circumstances, bilinguals are free to open a communication channel with the outsider by speaking the language they know, or to erect a communication barrier by speaking in the language they do not know. Both these choices can have great functional value. It is of considerable importance if this choice becomes restricted. This basic communication function may be of particular importance for us in distinguishing between different language situations in Aboriginal Australia. It seems to be a fairly well-observed rule in many communities that English or the nearest thing to it that people can manage should be spoken in the presence of English-speaking whites. This is consciously articulated by Aboriginal people (Hudson and McConvell 1984; Arkwookerum, Woolla 1982) and often reinforced by teachers, white coworkers, etc. demanding the same thing. In the same places, no one really expects Whites to learn Aboriginal languages. By contrast, in north-east Arnhem Land and Western Desert communities, there seems to be no such rule requiring people to speak a form of English in the presence of Whites, and there is a strong expectation that outsiders will try to learn the local language. The presence of a rule restricting basic communication choice may well be a symptom of language shift ahead. On the other side, some modern developments could lead in the opposite direction. In communities using two-way radios traditional languages are now being used sometimes on the radio to maintain confidentiality. The second function is social. It refers to language choice between bilinguals who share the same language repertoire. A language is chosen in order to identify the speaker, listener or both with a particular social group. This identification may be always the same between a pair of people, or it may be variable, in which case code-switching in the course of one conversation takes place, expressing social meaning. The third function is cultural. This has to do with culturally-based

Understanding language shift

151

function language Basic yes/no choice Communication Function

Figure 8.1 Model of language choice in bilingualism choices between languages which make available contrasting world-views and systems of knowledge. Choice of a language may reflect either the basic communication function, or one or both of the other two (social and cultural functions), as in figure 8.1. In this perspective language shift can be seen as resulting from the loss of the functions of bilingualism, that is, the loss of functional choices between languages, and in the case of the social function, their likely replacement by choices of style within one language. Likewise, language maintenance can be seen as the retention of these choices, or their replacement by other choices between the languages which still serve the major functions in different ways. We should perhaps speak of the maintenance of bilingualism, rather than maintenance of a language. This theory of language shift has implications quite different from the domains and adaptation theories. Survival of the old language depends on the number and importance of functional choices the bilingual situation can maintain. The applied versions of the adaptation and domains theories, on the other hand, would both tend to reduce one or other of the functions of language choice. Engineering the old language to emulate the new one is likely to reduce the value of the cultural function in language choice. Similarly, eliminating socially meaningful code-switching between languages as advocated by domain theorists in Aboriginal bilingual education (Harris 1977) would tend to reduce the value of the social function of the bilingualism, as Scotton's work (1982) also implies. From this start, which I shall call for convenience the functional choice theory of language shift, we have to look more deeply at the actual process of language shift, how it progresses from stage to stage, and (importantly for language maintenance) whether this progress is inevitable, or can stop at particular stages. Here the interactional-variationist theory can be of great help. It is necessary, however, to remove the historical fatalism from

152

PATRICK MCCONVELL

the theory, which says that once the shift has started it continues in one direction until it reaches the end-point, monolingualism. This part of the theory is not justifiable on theoretical grounds, nor is it supported by evidence, since there are numerous groups who mix languages in single conversations, yet have maintained bilingualism over many generations. It is important to recognise when and how an entire bilingual situation becomes, as it were, tilted, then overbalances into a rapid shift into monolingualism. At this stage, we do not yet have a full theory of 'tip' (Dorian 1981: 51). All three of the functions, basic communication, social and cultural, would normally be involved in combination in determining the course of such events. Ultimately, though, it is in the area of the social function where the drama of language shift and language maintenance is played out. New patterns, or redistributions of language choice functions, spread through groups within bilingual Aboriginal communities in much the same way as that described by Gal for Oberwart. Given the individual and family differences which can be ascribed to the influences of different social networks (Gal 1979), it is not particularly surprising that family differences in degree of language shift also exist in small fairly homogeneous communities like Daguragu and Turkey Creek, where I have worked. These communities also show a pattern of age distribution of languages such that the young speak Kriol, the middle-age group speak either the old language to some people and Kriol to others or code-switching, and only the very old speak the old language most of the time. Consistent 'unreciprocal' uses of the two languages in these Aboriginal language shift situations seems rare, although it happened frequently between different traditional languages, and still does, where Kriol does not intervene as a lingua franca. Code-switching between the old and new languages has become widely used in many situations among the middle-age group but, unlike in Oberwart, it seems that switching between local varieties of the old language, for the older ones, and different lects of Kriol, nearer to and farther from standard English, for the young, can also occur in the same conversation. All these types of switching have social meanings, which I relate to what I call a social arenas configuration. Figure 8.2 shows the social arenas configuration I use to explain code-switching at Daguragu, among the Gurindji (McConvell 1988). The nesting pattern of the languages/varieties indicates that the more inclusive groups properly include the more exclusive ones, for instance a member of the Wanyjirra local group also considers himself a member of the Gurindji speech community, and a Cattle Station Aboriginal. This particular type of social arenas configuration, which is probably found associated with a large number of Aboriginal languages currently undergoing language shift, is not itself to blame for the shift. It is similar to

Understanding language shift

153

KRIOL Cattle Station Aborigines GURINDJI Speech Community LOCAL DIALECTS Local Origin Groups

Figure 8.2 Social arenas relevant to codeswitching at Daguragu the one associated with trilingualism in Sauris (Denison 1971), which, according to historical records, has been, until recently at least, very stable. However, there can be an asymmetry built into this type of model of language use which tends to favour the larger, more inclusive, languages. A speaker of a more exclusive language, for example, Gurindji, cannot speak to a person identified as a speaker of a more inclusive language, for example Kriol, in the more exclusive language (even though the listener may understand it); nor would, say, a Kriol speaker be expected to attempt to talk Gurindji to a person identified as a Gurindji speaker, under this rule. Clearly if this rule is combined with a rule that defines children as a new language (Kriol) speaking group, as appears to be happening in a number of places, then acquisition of the old language by the younger generation would be difficult and probably not encouraged. Thus, language shift would be hard to avoid. Kriol has a variety of styles based on a 'heavy-light' (basilectalacrolectal) continuum (see e.g. Sandefur, this volume) which could easily be adapted to various 'expressive' functions related to in-group and outgroup identity, and thus take over many of the social functions currently served by switching between languages. This is likely to happen in stages, so that the use of certain traditional language words and phrases is retained to flavour the Kriol for specific expressive effect, without people actually being able to speak the old language properly or even understand it. 8.7

Learning to be bilingual

Young Aboriginals in the areas I know sometimes do have quite extensive passive knowledge of the old language, but are unable to use it normally

154

PATRICK McCONVELL

because of strong sociolinguistic rules against its use. Use of Standard Australian English is similarly prohibited by strong peer-group pressure. Part of the reason for the force of this prohibition may stem from the social meanings of the old language created by code-switching. These social rules about language use have exerted a strong psychological influence on the younger generation, perhaps as strong as the oppressive attempts to wipe out Aboriginal languages by white people in the past. An illustration of this recently was the reaction of Turkey Creek children to a film of Pitjantjatjara people at Ernabella. They watched the adults speaking the Pitjantjatjara language to each other with some interest, and without strong reaction, but when children came on the screen speaking Pitjantjatjara, they were greeted with gales of almost hysterical laughter by the Turkey Creek children, who normally speak Kriol rather than their traditional language, Kija. From this and similar incidents I have concluded that the rule against speaking the old language is not primarily a question of the basic communication function, that is the competence of the children, or of the cultural function, but of the social function, with strong psychological backing supporting the identification of speaking the old language with old people and their ways. We need to be aware of the kind of acquisition process which children in unstable bilingual situations go through, both in terms of learning the languages themselves, and in learning the sociolinguistic rules and social meanings of the codes used. We need not assume that these are acquired intact, but are undoubtedly reanalysed and reinterpreted by the children. Although I have looked mainly at the social determinants of language shift, it is also possible that the complexity or difficulty of learning a particular language could affect its chances of being learned under adverse circumstances and, therefore, its chances of survival. Slobin (1982) compared acquisition in different languages and suggested that certain types of grammatical encodings of categories in a language make particular aspects of a language easier to learn, for example, one-to-one mapping of that form and meaning. Many Australian languages are very far from this ideal, particularly the prefixing ones such as Tiwi and Kija, in which fusional irregular non-syllabic morphemes abound (see also Bavin and Shopen, this volume). Aborigines sometimes wonder whether the children use Kriol simply because it is easier. Winnie Budbaria at Turkey Creek said: 'English is straight out, our languages are all back-to-back'. Does this mean that language shift is harder to resist for such languages than, say, for Arrernte, which follows many of these ideal features - and, incidentally, has survived remarkably well in the urban environment of Alice Springs? Another related investigation is children's acquisition of a prefixing language. Do children simplify the language in ways that might be expected from these observations about learnability?

Understanding language shift

155

While several studies have begun to be made on the linguistic aspects of young people's speech in Australia, I do not know of any ethnographic study of the development of communicative competence (cf., e.g. Schieffelin's 1989 on the Kaluli). Ron Day, a Mer Islander who grew up in the period when language shift to Torres Strait Creole was going on on the island, has described to me what happened to him and his peers as they went through a process of feeling rejected by their elders and by the teachers for not speaking either Meriam Mir or English 'correctly'. The children then embraced the pidgin as their own. Similar processes have gone on and are going on with young Aborigines today, and are leading to language shift, a result that Topsy Chestnut (Hudson and McConvell 1984: 37) feels sure they will regret. Young people don't care about the language, but when they get older they feel sorry about it. That's why we want to keep the languages. Until we understand these processes of language shift better, we shall find it hard to help people like Topsy keep their languages.

Part II Pidgins and Creoles

Overview of the pidgin and Creole languages of Australia PETER MUHLHAUSLER

9.1

Introduction

The aim of this chapter is to provide a general picture of the pidgins and Creoles spoken in Australia. Outside its scope are descriptive and historical accounts of individual languages as well as Aboriginal English, the history and structure of which differ in a number of significant aspects from that of English-derived pidgins and Creoles, and are dealt with in other chapters of this volume (see Kaldor and Malcolm, Eades, and Koch). The greatest need in an overview of this kind seems to be clarification of basic terminology. As pointed out by Sandefur (1985b), 'the terms "pidgin", "creole" and "Aboriginal English" have been used with a great deal of ambiguity in recent years'. This ambiguity reflects an insufficient understanding of the phenomena at hand as well as the continuing influence of folk labels for varieties of speech used by non-mainstream Australians. One view, in particular, that has continued to survive even in quite respectable publications is that there are underdeveloped languages with few or no abstract terms, and that pidgin and Creole languages are corruptions of true languages. Hence we find Pidgin English referred to as a 'quaint and macaronic jargon' or 'English perverted and mangled by the natives' or, from a different perspective (Strehlow 1947: xviii) 'English perverted and mangled by ignorant whites, who have in turn taught this ridiculous gibberish to the natives and who then affect to be amused by the childish babblings of these "savages".' Such contempt for pidgins and Creoles has led to a long tradition of misunderstanding and prejudice as well as a deplorable lack of large-scale studies of these languages. Crowley and Rigsby (1979: 154) remark: When most Australians speak of Tidgin English' or 'Pidgin', they generally think of something they also call 'broken English', which is a language variety that no one takes seriously. Pidgin is a sort of simplified English and its simplicity is believed to

160

PETER MUHLHAUSLER

reflect the lesser mental capabilities of its darker skinned speakers. Such misconceptions are dangerous because they serve to rationalise European ethnocentrism and they perpetuate racist stereotypes. The terms 'pidgin' and 'creole' are, however, technical terms used by linguists to refer to structurally and contextually definable linguistic systems: 1 A 'pidgin' is a contact language used among people who have no other language in common. It is a second language for those who use it, and its use is restricted to a limited number of situational contexts, such as trading (e.g. Chinese Pidgin English), plantation life (e.g. Queensland Kanaka Pidgin) or military operations (e.g. Korean Bamboo Pidgin). Since a pidgin is used in a small number of contexts, it has a smaller lexicon and a less complex grammar than languages with native speakers. Grammar and lexicon are derived from many sources, including the superimposed European language, local languages and universal grammar. 2 If parents of different linguistic backgrounds use a pidgin to communicate at home, their children will grow up speaking this pidgin as their first language. If this happens in many households of a community, the nativised pidgin can become the language of a new speech community. This process is known as 'creolisation'. A creolised pidgin or 'creole' is structurally more complex than a second-language pidgin as it has to meet all the communicative requirements of native speakers. The structural complexity of a Creole is comparable to that of other languages. Creolisation illustrates the human capacity to 'create' language. Far from being an imitation of their parents' pidgin, a first-generation creole can be structurally quite different. I have observed first-generation creole speakers of Tok Pisin in Papua New Guinea who spoke a language that was so different from their parents' pidgin that the latter could not follow the creole when spoken by the children among themselves. When communicating with their parents these children switched to the former's less developed variety. Pidgins and Creoles have to be regarded as dynamic rather than static systems. Often, creolisation is preceded by a long history of structural expansion of a pidgin, so that the difference between first and second language is not very great. In other cases, creolisation occurs before a sophisticated pidgin has had time to develop. In such cases the importance of universal grammar is considerable. In the development from a rudimentary pidgin to a full creole a number of stages can be distinguished It must be remembered that the stages intermediate between the jargon stage and the creolisation stage are optional and not found in all creole life histories. I have given below a rough outline of the structural and functional properties of pidgins at various stages in their development.

Overview of the pidgin and Creole languages of Australia Structural properties One or two-word sentences, very small lexicon, simple sound system, great individual variation.

Functional characteristics Used for communication in limited referential domains, e.g. trade, labour recruiting, military

Stable pidgin (Chinese Pidgin English)

Simple sentences as well as some complex ones, social norms concerning linguistic correctness.

Used for communication in a fixed number of domains, for social control and, to a small extent, selfexpression.

Expanded pidgin (e.g. Tok Pisin in Papua New Guinea)

Complex grammar, development of a word-formation component, increase in speech tempo.

Used in almost all domains of everyday life, for selfexpression, wordplay, literature, instrumental in providing cohesion in heterogeneous groups.

Stage Jargon (e.g. Thai-English)

161

In addition to varieties arranged on such a developmental continuum (which could represent speakers belonging to different age groups), one finds varieties of pidgin and Creole English that are in some ways closer to English. Anglicised versions of pidgins and Creoles develop in communities where English is the superordinate language. The two main reasons for their development are the facilitation of vertical communication and prestige factors. By 'vertical communication' I mean communication between people of different social status, such as that between Englishspeaking overseers and pidgin-speaking labourers on a plantation. As very few Europeans bother to learn a pidgin properly, the indigenous pidgin speakers make use of whatever knowledge of the standard language they have in communicating with English speaking outsiders. This ad hoc anglicised variety is the only type of 'pidgin' that most Europeans ever get to hear.

162

PETER MUHLHAUSLER

The implications for the study of Australian pidgins are far-reaching. As pointed out by Troy (1985), the documentary evidence we have of varieties such as New South Wales Pidgin English tends to reflect such anglicised jargons as were in use among Europeans and Aborigines rather than the more stable varieties used when Aboriginal speakers from different language backgrounds communicated. A second reason why Pidgin English can become more anglicised involves prestige. If the speakers of a pidgin are constantly told that their language is bad and if they are made to believe that English is the only way to full social acceptance, they will try very hard to acquire as much of the standard language as they can in an informal learning context. Being able to use standard English expressions in one's pidgin may enhance a speaker's social standing. It is important to keep in mind that replacing a pidgin construction with one borrowed from English does not mean that the language can suddenly express more. All that takes place is a kind of restructuring. Present-day forms of Aboriginal English may well be heavily reconstructed earlier pidgins and Creoles. Leaving aside the historical details of this question, I will provide briefly some criteria for distinguising a Creole from a variety of Aboriginal English. The principal criterion one can appeal to here is that of targeted learning: whereas pidgins are second languages developed in the absence of full access to a target language and Creoles first languages learnt without a full target, Aboriginal English is more in the nature of an interlanguage, that is, a variety of English oriented towards the norms of a standard variety, and sharing many of its underlying forms. There tends to remain a structural gap between pidgins and Creoles on the one hand, and their related standard languages on the other. This distinction can become blurred when a pidgin or Creole is heavily restructured during intensive contact with its lexifier' language. One must also consider the adequacy of pidgins and Creoles. It is often argued that pidgins and Creoles are deficient when compared with languages such as English, in that one cannot talk about certain aspects of the world or that speaking such a language promotes muddled thinking. Proponents of this view ignore a number of important linguistic and sociolinguistic findings, including the following. 1 A language is simpler than another language if it is more regular (and thus more learnable), without any loss in communicative potential or, as it is called, referential adequacy. In the following examples Northern Territory Kriol encodes the idea of'person' in a uniform fashion, where English relies on different lexical forms: gitamen klebamen

'guitar player' 'smart man'

Overview of the pidgin and Creole languages of Australia pailitmen songmen raithenmen wailmen

163

'pilot' 'singer' 'assistant' 'wild man'

2 A language is called impoverished if it is unable to express what is relevant to its speakers' needs. Since a pidgin cannot fulfil all of its speakers' communicative needs, it is impoverished. Since pidgin users can express all these needs in their first language, however, this does not reflect any inferiority on the pidgin users' part. Someone who uses a pidgin for trade purposes only is like a scientist who knows enough of a foreign language to communicate in his/her narrow field of specialisation. No language has a word for everything. Sandefur and Sandefur's Kriol dictionary (1979a), for instance, contains a number of entries which are translated by longish paraphrases in English. Examples are: nyapaja soptri bush rod bundin

kinship term for an old man who is in 'banjimen' relationship a kind of wattle tree, used for fish poison and medicine Aboriginal foot walking track water lily seed at certain stage of development, when it is brown

There are considerable differences in the density of semantic fields across languages; the English kinship system looks very impoverished, for instance, when compared to the system of kinship terms found in some Australian pidgins and Creoles. 4 By referential adequacy we mean the ability of a language to refer to aspects of the 'real' world. The referential potential of many western languages is great because of the existence of a vast number of specialists' sub-languages. Whilst such sub-languages may express a multitude of technical concepts, specialist vocabulary for the description of cultural activities of other societies may be missing. 5 The fact that the lexicon of a language contains a very large number of entries does not mean that all members of the language community can actually use them. There are hundreds of terms for parts of the body which are known to medical experts only; the average speaker of English can handle about 70, as many as the speaker of a Creole derived from English. 6 Languages may also differ in expressive adequacy, that is, in their power to convey the same information in different ways. Most pidgins, particularly in the early phases of their life, lack such stylistic devices. Extended pidgins and Creoles, on the other hand, often exhibit a rich inventory of registers and styles. In the absence of detailed studies, questions of adequacy have to be

164

PETER MUHLHAUSLER

treated with greatest care. In particular, the presence or absence of certain structural or lexical devices cannot be taken as a direct index of speakers' ability to communicate as the total communication process includes shared knowledge of the world, intonation, gesture and many other components. What seems uncontroversial, however, is that a negative image of one's language tends to promote feelings of inferiority. The past history of Australia is full of examples where languages and their speakers were degraded, mistreated and misunderstood. In the absence of any reliable criteria for assessing the overall quality of different communication systems we are well advised to treat pidgins and Creoles with respect, thus enhancing the positive feelings towards them that have begun to develop among their speakers. 9.2

An inventory of pidgins and Creoles found in Australia

From what has been said so far it should have become clear that counting pidgins is not an easy job. Roughly speaking however, we can distinguish the following principal traditions: (i) East Coast Aboriginal Pidgin, stemming from the early colonial days and initial contacts between whites and Aborigines, (ii) Western Australian Pidgin English (see Miihlhausler, forthcoming), (iii) Chinese Pidgin, brought by Chinese migrants in the late nineteenth century, (iv) Melanesian Pidgin, spoken by the many thousands of black workers in the Queensland sugar cane industry, (v) Macassarese trade pidgin used in pre-European days between visiting trepang fishermen and Aborigines as well as Aboriginal lingua franca in the Northern Territory, (vi) pidginised forms of Malay used in the pearling fisheries around Broome and Thursday Island, particularly between 1900 and 1930. (vii) possibly unrelated to (v) and (vi), Plantation Pidgin Malay and later creolised forms of Cocos and Christmas Island, (viii) the Pitcairnese Creole English transported to Norfolk Island, (ix) Bass Strait English, possibly reflecting an earlier Tasmanian pidgin, (x) pidginised or simplified Aboriginal vernaculars used as mission and intergroup lingua franca such as Gunwinygu or Pitjantjatara. The status and viability of these languages differs greatly. Generally speaking, the most healthy are those that have the support of large Aboriginal communities such as Northern Territory (Roper River) Kriol or Torres Strait Broken. Others such as Kanaka English, Broome Pearling Pidgin or Macassarese Pidgin are virtually extinct.

Overview of the pidgin and Creole languages of Australia

165

9.3 The history of research into Australian pidgins and Creoles There are numerous gaps in our understanding of the pidgins and Creoles of Australia directly related to the scarcity of scholarly investigation of them. Interest in the pidgins and Creoles of Australia is a very recent phenomenon. The bibliography of Reinecke et al (1975) lists a mere 45 publications, most of them concerned with other matters but containing text samples. The most comprehensive account of any variety is Sayer's (1944) 'haphazard book' (Reinecke et al. 1975: 588), which gives data from Northern Territory Pidgin English based on the author's first hand experience. Data from other varieties are very uneven. Banfield (1908) for instance, gives a massive body of data for Aboriginal Pidgin English of Dunk Island; other areas such as South Australia or Victoria were virtually unaccounted for. Baker's summary of what was known about Aboriginal Pidgin English in the 1960s (e.g. Baker 1953, 1966) adds little to Sayer. Conspicuous by its absence is any treatment of creolisation, a fact which led an overseas writer working with secondary sources (Bauer 1975: 140), to the conclusion: 'Since Australian Pidgin English was replaced rapidly by substandard English, no tendencies towards creolisation have been observed among Australian Aboriginals.' The existence of creolised varieties was established in the late 1970s by Rigsby and Crowley for Torres Strait Broken and Cape York Creole and by Sandefur for Northern Territory Kriol. Projects concerned with Aboriginal pidgins and Creoles were virtually non-existent before 1975, mainly because whatever English-related form of speech was spoke by these people was simply classified as 'English'. Consequently, some varieties were surveyed as part of dialect surveys, the most important being Flint's Queensland Speech Survey in the 1960s. One of Flint's collaborators in particular, Dutton, investigated a number of non-standard forms of speech which were later classified as pidgins and Creoles. Dutton has continued to survey such varieties in Queensland and his work offers the greatest time depth and continuity of research to date. From the mid-1970s onward a number of institutions in Australia began to get involved in (typically small) projects on pidgins and Creoles, a consequence of the world-wide upsurge of pidgin and Creole studies as well as a number of political changes affecting the Aboriginal community. Whilst it is impossible to deal with this work in an exhaustive fashion (a more comprehensive account can be found in Sandefur 1983), there are certain institutions which warrant mention. The Australian Institute of Aboriginal Studies provided grants both for smaller projects such as my research in North Queensland (Miihlhausler 1979) and the Torres Strait as well as for a more substantial fellowship to

166

PETER MUHLHAUSLER

study Torres Strait Broken, an award held by Shnukal between 1980 and 1983. The Institute also acts as a repository of language materials. The study of Australian pidgins and Creoles was made an official research emphasis at the Department of Linguistics, Research School of Pacific Studies of the Australian National University in the mid-1970s and a considerable amount of work has been carried out by staff and students since. Particular mention should be made of Dutton's work on Canefields English and Queensland Aboriginal English (Dutton 1980, 1983; Dutton and Miihlhausler 1984) and Tryon's work on Northern Territory Kriol. A number of articles concerned with these languages have been published in the departmental publication Pacific Linguistics and a beginning has been made in the compilation of a map of Australian pidgins, Creoles and lingua franca for the forthcoming Atlas of Languages for Intercultural Communication in the Pacific, a joint venture between members of this department and linguists at Oxford University. Some badly needed research on the history of Pidgin English in New South Wales is currently being conducted by Troy whilst Hosokawa has made some exciting findings about the language contact situation around Broome (Western Australia). Work on Kriol in central and Northern Australia is also carried out by the Department of Linguistics, The Faculties, Australian National University, in particular by Koch. The University of Queensland was one of the early centres of research on pidgins and Creoles and there is considerable interest in both the Anthropology Department (e.g. Rigsby and Sutton 1982; J.W. Harris 1984) and among linguists attached to other departments (e.g. Shnukal). Work tends to concentrate on Cape York and the Torres Strait. Members of the Department of Anthropology of the University of Western Australia (in particular, Kaldor and Sansom) have provided much-needed studies of both pidgins and varieties of Aboriginal English in Western Australia. Details on a number of projects concerned with nonstandard English in Western Australia are given by Sandefur (1983). In 1987 Muhlhausler carried out a survey of Pidgin English in the south-west and the Pilbarra under the auspices of this department. This very sketchy survey of research at Australian universities immediately reveals some major gaps: no work has been undertaken to my knowledge by universities in Tasmania, South Australia and Victoria on the pidgin and Creole varieties spoken within these states. The scale of operations at the institutions mentioned is also very moderate and largescale systematic surveys such as those begun at the Australian National University are still very much the exception. Generally speaking, research into these languages at Australian Universities remains patchy and haphazard and fails to do justice to their growing social and educational importance.

Overview of the pidgin and Creole languages of Australia

167

A number of other researchers such as Sandefur, Hudson, and Sharpe at institutions outside the universities have worked more systematically on these questions. The Summer Institute of Linguistics, whose concern is primarily with mission work, education and Bible translation and only secondarily with linguistics, has nevertheless produced an impressive body of descriptions and lexicographical and sociolinguistic data. Some work, particularly work of a practical nature is also carried out under the auspices of the School of Australian Linguistics at Batchelor and the Kimberley Language Resource Centre at Halls Creek. The question of Kriol as an educational medium is a dominant one. Other much-needed research is the systematic documentation of the linguistic history and the sociohistorical context of Australian pidgins and Creoles. The models provided by Troy (1985) for New South Wales and J.W. Harris (1984) for the Northern Territory can be easily adapted to other parts of Australia. What is surprising is the sheer amount of historical information still recoverable. Government and mission archives in Australia and the UK are full of materials in and on these languages, as are travel books, newspapers and scientific writings. Work with present-day and former speakers can be very rewarding, as has been shown, for instance, by Dutton (1980) in the case of Canefields English, by Hosokawa with the contact vernaculars of Broome, and by Miihlhausler for Nyungar Pidgin English. One hopes that the accelerating pace of research can be maintained and that the next few years will see a significant increase in detailed accounts of the pidgins and Creoles of Australia.

9.4

The origin and development of Australian pidgins and Creoles

The development of pidgins and Creoles in Australia is generally seen as a direct consequence of contacts between Aborigines and outside groups, such as South-east Asian traders in the north or Europeans in most parts of Australia. Whether or not indigenous pidgins were used in pre-contact days is not known, but is generally regarded as unlikely. Instead, widespread biand multilingualism, together with dialect and language chaining, enabled Aboriginals to communicate efficiently across considerable distances. Still, in view of the recent findings of numerous indigenous pidgins in neighbouring Papua New Guinea, where their existence had been doubted in the past, it might be wise to postpone judgement on this question. The general pattern of pidgin-creole development is repeated several times in the more recent post-1788 history of Australia. It proceeds in the following stages: (1) A relatively unstable jargon develops between Europeans and an Aboriginal group

168

PETER MUHLHAUSLER

(2) Stabilisation and genuine pidgin development occurs as this jargon is used as a means of intercommunication by Aboriginal groups who either had not communicated previously or else who reserve this language for talking about things European (3) Some expansion takes place as European modes of life become more pervasive (4) Creolisation occurs where permanent non-traditional settlements are established. With regard to the situation around Port Jackson, for example, we find that 4a barbarous mixture of English with the Port Jackson dialect is spoken by either party' (Collins 1798). The relatively high Aboriginal content in its lexicon (more than 50 per cent in the data collected by Troy 1985) reflects the relative powerlessness of the coloniser in the very early period. Troy (1985) argues that a much more stable pidgin appears to have been spoken further inland along the traditional communication network of the Aboriginals. However, data for this variety are very hard to come by as the typical reports are by English speakers who availed themselves of the less stable jargon. Disease, 'dispersal' and assimilation prevented the development of a Creole in Port Jackson (Sydney) and indeed New South Wales. A very similar picture is found elsewhere. Thus, Urry and Walsh (1981) comment on considerable differences in the proficiency with which Macassarese Pidgin was spoken. Whilst direct linguistic evidence is lacking it seems reasonable to suggest that the coastal varieties were less stable as they involved contact between visiting outsiders and single Aboriginal groups whereas inland, Macassarese was used as an intergroup lingua franca. Again, in the absence of major disruptions following the Macassarese visits, no Creole community appears to have developed. Creolisation typically is the result of a major break in the continuity of transmission of languages from one generation to another. Thus, it is found in conditions such as social breakdown, displacement, slavery and mass migration. Creolisation can occur only if a sufficiently large new speech community is involved, and members of such a community adopt the new language as an act of identity. In the case of Australia, the conditions for creolisation were met in a number of areas, particularly in the far North. In the case of Roper River Kriol, as described by J.W. Harris (1984), creolisation began at the beginning of this century. A new mission community was created at Roper where Aboriginals from various language groups came together. Because of the small numbers of speakers of each language, language decline and loss set in and the traditional vernaculars were replaced by pidgin English as the primary language. The expatriate missionaries tended to com-

Overview of the pidgin and Creole languages of Australia

169

municate with the Aboriginals in pidgin English or simplified English. Children were separated from their parents through the dormitory system. Of these factors the last is probably the most important, as the dormitories provided a centre for the development of a new society, neither traditional nor western whose medium of expression was Kriol. Contemporary evidence suggests that the pidgin English of Roper changed from a relatively rudimentary language into a much more complex one in a short span of time, incorporating the many innovations developed by the children themselves. Mission dormitories also appear to have played an important role in the emergence of creolised English in the Cape York Peninsula but not in the Torres Strait (Shnukal, personal communication). Creolisation of other Australian pidgins was prevented by adverse social conditions. Thus, the rapid decline in numbers of Aboriginals, their dispersal and their status as marginal dwellers at the outskirts of white settlements in the South promoted the adoption of a non-standard Australian English. Queensland Kanaka English, despite favourable conditions between about 1900 and 1906, did not become a Creole because of the large-scale repatriation of South Sea Islanders after 1904 and the dispersal of the black communities. To what extent plantation Creoles developed on Christmas and Cocos Islands to the west of Australia is not known, though it seems likely that creolised forms of pidgin Malay are spoken there. Creolisation, like pidginisation, appears to begin at small foci of development and then spread over wider areas. Thus, creolised varieties of Northern Australian Pidgin are found in most parts of Northern Australia and the processes begun around the turn of the century continue. The dynamics of this remain ill-described and indeed ill-understood. Most researchers have concentrated on small geographic areas (Sandefur and Sandefur's 1979b survey of the Kimberleys is a notable exception). The picture that emerges from most descriptions is abstract, static and suggestive of self-contained pidgin and Creole languages. In reality, the linguistic situation in the various areas of contact tended to be in considerable flux and susceptible to many influences. Thus, in each case researchers will have to establish whether a pidgin tradition is an independent development or a continuation of an earlier tradition, what mergers and divergent developments occurred, and the extent of variation for a given time and place. With regard to the first factor, we find that pidgin development in Port Jackson was not entirely independent. Thus, early visitors to Australia had had experience with the reduced English-Polynesian jargons spoken in Tahiti and the Marquesas and, in the case of some individuals, Africa and the West Indies. It is significant, for instance, that most of the early London Missionary Society mission personnel had worked in Tahiti prior to their

170

PETER MUHLHAUSLER

coming to Australia. In the case of another pidgin, Queensland Kanaka (Canefields) English, a number of the white overseers had worked on pidgin speaking plantations elsewhere, whereas many of the black recruits had a smattering of broken English. Pidgin English in the Torres Strait again has very strong outside connections. It was brought there mainly by South Sea islanders, both teachers and mission helpers and later time-expired workers from the canefields. It soon became the language for discussing non-traditional matters by the indigenous population and in some areas intermarriage with the South Sea islanders resulted in creolisation (see Shnukal 1983a and this volume). Added to this is the influence of the pearl-lugger Malay spoken on board the numerous pearling vessels ploughing the sea around Thursday Island. Finally, the Macassarese Pidgin of the Northern Territory is another example of a pidgin imported from outside Australia and, more recently, the Pidgin German of some industries in Victoria (see Clyne 1977a) has been added. We have insufficient evidence to address a number of issues. One question in particular, that of the Pidgin English spoken by Chinese immigrants in the nineteenth century remains unsolved. The few examples I have been able to trace appear to exhibit some similarities with Canton Pidgin English, but I have also seen evidence that many Chinese did not speak any form of English on arrival in Australia. Whether they picked up their English from other pidgin speaking groups such as the South Sea Islanders or whether they acquired Chinese (Canton) Pidgin English from other Chinese remains to be established (see Baker 1987). Independent development is established for a number of places, though some of the varieties identified did not have a very long life. A particularly shortlived variety was spoken around Port Essington in the Northern Territory (see Harris 1984) and the Moreton Bay (Brisbane) pidgin identified by Dutton (1983). Others, for example, varieties spoken by the Afghans, remain to be identified. An important issue is the extent to which the earliest variety, Port Jackson Pidgin, can be established as the ancestor of all other Englishbased pidgins in Australia. It is certainly true that this variety was diffused extensively by settlers. Aborigines and, in particular travelling stockmen and workers on the transcontinental telegraph line. This diffusion did not extend to the south half of Western Australia until the 1890s, by which time an independently formed Nyungar Pidgin English had already established itself (see Miihlhausler, forthcoming). Words from the original Port Jackson Pidgin are found in Northern Territory Kriol, for instance: benjy'abdomen, stomach', bogey - 'wash, swim', cobra - 'head', yarraman 'horse', and jidan - 'camp, stay' (J.W. Harris 1984). On the other hand, there is interesting circumstantial evidence that a partial relexification of

Overview of the pidgin and Creole languages of Australia

171

Macassarese Pidgin was also involved in the development of Kriol. The complexity of the history of Kriol is found also with the varieties of pidgin and Creole around Broome currently studied by Hosowaka and in the history of Queensland Canefields English. Mergers, divergence and new starts are continuous factors in the histories of such contact languages. A direct result of these factors is a considerable amount of variability in each of these languages. Variability is organised in terms of three dimensions: 1 First vs. second language, for example, both Kriol and Torres Strait Broken exist as first and second languages, and the history of creolisation can still be recovered by moving from location to location. Thus, in the Torres Strait, Torres Strait Creole replaced Pidgin English on Erub in the 1890s, on Moa around 1910, on Masig between 1920 and 1930 and on other central eastern islands in the 1940s (see Shnukal 1983a: 179). 2 White vs. black varieties: The White varieties and the varieties spoken by Blacks with or in the presence of Whites tend to be less systematic and considerably more anglicised (light) than the varieties spoken among Aboriginals. 3 Style, for example, 'light' (English influenced) and 'deep' or 'heavy' varieties of pidgin and Creole have been reported for most of the languages under discussion (see Sandefur 1984a). The myth that pidgins and Creoles are monostylistic is certainly not applicable to the Australian situation and existing accounts fail to do justice to the stylistic and other types of variation found in these language. 9.5

Further research issues

Many questions remain unanswered, though it is hoped that they are at least answerable in principle with further research. 1 What is the relative position of the varieties spoken in the Torres Straits, Cape York Peninsula and in the Roper River areas? Are we dealing with a continuum or can one identify separate traditions? Some interesting groundwork has been carried out by Rigsby (1984). 2 What is the link between Queensland Kanaka English and other pidgins and Creoles? J.W. Harris (1984) has cast doubt on links with Kriol, but there are very strong links with the Torres Strait varieties which need to be investigated more closely. 3 Were there separate traditions in Tasmania, South Australia and Victoria? 4 Have there been instances of creolisation in the south? 5 What languages were used by non-European immigrant groups such as the Afghans, 'Manilamen', and more recently, the Timorese?

172

PETER MUHLHAUSLER

Finally, when determining the number of Australian pidgins and Creoles, researchers should not forget that such languages are often spoken across national boundaries. Thus Pidgin English of the Torres Strait has been used extensively on the south coast of Papua and Norfolkese is also spoken on Pitcairn; Canefields English has very strong links with Bislama and Solomons Pijin, and Macassarese was probably similar to the Malayo Portuguese spoken in the Dutch Empire between Capetown and China. A task less formidable than identifying Australian pidgins and Creoles is tracing the origins and diffusion of typically Australian lexical items and constructions. Some useful beginnings have been made by Sandefur (1984a) and J.W. Harris (1984) and more detailed work on this problem is currently being carried out by Philip Baker and me. There are also educational and political questions. When the Aboriginal Languages Association was formed in 1981, Aboriginal English, pidgin and Creole were recognised in its constitution as modern Aboriginal languages, a major change after years of ridicule, neglect and misunderstanding. Rather than being regarded as obstacles to progress, dyseconomies and something to be ashamed of, an increasing number of their speakers seem to be developing more positive attitudes. The situation is reminiscent of the Creole situation in Belize, described by Le Page and Tabouret-Keller (1985), where the development of a strong sense of community goes hand in hand with the development of a narrowly focused and stable Creole. It seems possible that the wide scatter of Australian pidgin and Creole grammar and lexicon will become similarly narrowed down and focused with the increasing identification of Aboriginal speakers with this language. The focusing has also been helped by active promotion of Creole by organisations such as Summer Institute of Linguistics and the establishment of bilingual school programmes in the Northern Territory. First findings suggest that this programme has been quite a success. Murtagh (1979: 54), for instance, concludes that there are 'very definite trends towards the superiority of bilingual schooling over monolingual Englishonly schooling for Creole-speaking students with regard to oral language proficiency in both mother tongue, Creole, and second language, English . . . students schooled bilingually show progressively greater success at separating the two languages than their counterparts schooled monolingually.' This increasing ability to separate the two languages (English and Creole) which bilingually-schooled students have shown and which appears to be explainable only in terms of the two languages being taught as separate entities in the classroom, constitutes a powerful argument for the introduction of bilingual education to other schools where similar conditions obtain. Continued success will depend on adequate funding, on-going research

Overview of the pidgin and Creole languages of Australia

173

and the extension of the use of pidgins and Creoles in the schools of Western Australia and Queensland. At the same time, there seems an urgent need for in-service training aimed at eliminating misconceptions still prevalent about these languages among the majority of teachers. Bilingual education tends to concentrate on teaching in Creoles rather than about Creoles. Given the special situation of these languages, it would seem useful to develop syllabuses telling students about their nature, origins and social role. The aim of such courses would be to reinforce the growing favourable attitudes towards pidgins and Creoles among Aborigines and at the same time they would help to counteract the somewhat paradoxical developments that have occurred in the most recent past. As Shnukal, writing on attitudes towards Torres Strait Broken (1983b: 3), says: Ironically, at a time when the Creole is decreolising, that is, becoming linguistically more like English, the Creole is being rediscovered as a marker of ethnicity and separateness from Europeans among Islander children.' The discrepancy between linguistic and social developments encountered in education and language attitudes has further repercussions in other areas, particularly language rights issues. Unless languages such as Kriol or Torres Strait Broken are officially recognised as separate languages, their speakers will not enjoy rights to translation and interpretation enjoyed by speakers of other minority languages in Australia. As illustrated by Koch (1985 and this volume), miscommunication in land rights claims is very frequently due to the fact that Aboriginals speak 4a non-standard form of English', and/or a pidgin or Creole lexically related to English. Similar misunderstandings are common in many, if not most, dealings between Aboriginal speakers of pidgin or Creole English and white officials and fellow citizens. It seems time to address the question of how many social and economic inefficiencies and injustices could be remedied by assigning these languages a proper status. 9.6

Conclusions

The political and social fate of Australian pidgins and Creoles has been such that only the English-based varieties have survived. Their survival in turn will depend on whether Aboriginal pidgin and Creole Englishes are recognised as equal but separate systems or substandard systems parasitic upon English. Language planners, educators and politicians will have to address many issues if the pidgin and Creole question is to be meaningfully integrated with a wider Australian national language policy (see Ozolins, this volume).

10 Queensland Kanaka English PETER MUHLHAUSLER

10.1

Introduction

This chapter will provide a brief sketch of the sociohistorical background and general structure of Queensland Kanaka English1 (see also Dutton 1980; Dutton and Miihlhausler 1984; and Miihlhausler 1986). I will also explore how this language relates to some crucial aspects of pidgin and Creole linguistics, in particular Hall's (1962) model of the life-cycle of these languages and, secondly, the role of universals in its formation and development. The origins of Queensland Kanaka English date back to the period between 1864 and 1904 when more than 60,000 Melanesians were brought to Queensland to work on the sugar plantations and in other rural industries. The recruiting pattern for the Queensland plantations was similar to that for the other Pacific plantation centres, that is, a continuous shift from south to north, beginning with the Loyalties in the 1860s and ending up in the northern Solomons of New Guinea in the 1880s. It thus seems likely that some kind of Loyalty Island Pidgin English formed the foundation of Queensland Kanaka English. As the recruiting grounds shifted, considerable numbers of non-speakers were blackbirded. These non-speakers in all likelihood modelled their pidgin on existing varieties. Recruiting from previously uncontacted areas ceased around 1885, after which date the majority of recruits could speak Pidgin English, many having been recruited for a second time, after serving a term of contract in Queensland itself or on some other plantation in the Pacific. In the years after 1900 no new recruiting grounds were opened up and the Pacific Island Labourers Act of 1901 meant the end of the system of recruiting Melanesian labourers for Queensland. In the years between 1900 and the repatriation of most workers in 1906, the vast majority of Melanesians on the plantations were old hands and time-expired men with a considerable knowledge of the stable Kanaka English spoken on the

Queensland Kanaka English

175

Queensland plantation. About 50 per cent of the 9,327 workers employed in 1901 into this category (Corris 1973). The historical context suggests a period of strong outside influence followed by increasing isolation, particularly in the last years on the plantation. Throughout the period English persisted as a target language. In the initial years of small private plantations and employment of Kanakas in the pastoral industry this influence must have been particularly strong, whereas in later years, on the large coastal sugar plantations employing several hundred Melanesians, the social environment for the development of separate pidgin norms was more favourable. The final years of Kanaka English were characterised by an increase in 'wokabauts', that is, the Melanesian practice of travelling long distances in the hope of reemployment on another plantation. This undoubtedly promoted the diffusion of local innovations and hence, greater uniformity of the pidgin. 10.2 Kanaka English and the pidgin-creole life cycle

Hall (1966: 126ff) discusses an important difference between normal languages and pidgins, namely that the latter possess life cycles. Thus, 4a pidgin language usually comes into existence for a specific reason, lasts just as long as the situation that called it into being, and then goes quickly out of use'. Moreover, 'a pidgin acquires a longer lease of life only by becoming the native language of a group of speakers' (becoming creolised). Queensland Kanaka English, with its life span of only about 50 years is an excellent illustration of this principle. There were two main reasons for its coming into being. One was the desire, on the part of the plantation owners, to be able to transmit orders to their workforce efficiently without letting them become part of the acrolectal English speaking society. The second was the need of the multilingual workforce to have a common means of verbal intercommunication as well as their desire to emulate the prestige target culture. Rapid crystallisation of a pidgin language in the 1870s was promoted by the fact that first generation plantation workers already spoke some form of reduced English, as well as the ghetto character of the Melanesian living quarters and the geographic remoteness from the speakers' home territory. The social and geographical boundaries that defined the movement of the Melanesians in Queensland also determined the extent to which their pidgin language could stabilise and expand. The development of stable grammatical norms independent of English was promoted by factors such as the relative distance of acrolectal English and the common Oceanic substratum of its speakers. The fact that the language remained lexically and structurally relatively reduced reflects the practice of employing workers for relatively short contact periods only and the limited range of

176

PETER MUHLHAUSLER

Table 10.1. Number of male and female Melanesians in Queensland Dates

Males

Females

Total

1868 1871 1876 1881 1886 1891 1901 1911 1921

1,536 2,255 4,938 5,975 9,116 8,498 8,380 1,404 1,350

7 81 170 373 921 745 380 335 537

1,543 2,336 5,108 6,348 10,037 9,243 8,760 1,739 1,887

functions and domains served by Kanaka English for most of its life cycle. Unlike its daughter languages, Solomon Islands Pijin and Vanuatu Bislama, Kanaka English never reached the stage of an expanded pidgin, at least not for a significant length of time. The continued life of a pidgin language, as pointed out by Hall, depends on the continuation of favourable environmental factors of creolisation. There was growing public pressure for a White Australia policy after 1900 which resulted in a decision to repatriate all Melanesian workers between 1904 and 1906. This move put a virtual end to the expansion of Kanaka English. Its main raison d'etre, that of serving as a lingua franca on the multilingual plantations, ceased. However, contrary to the intentions of the Queensland Government and contrary to what is believed by many white Australians, not all Melanesians were sent back between 1904 and 1906. The numbers that remained are estimated at between 1,500 and 2,000. Few of those who did remain continued to work on the plantations, and the former principal reason for the use of Kanaka English simply was no more. For some larger Melanesian groups around Mackay and Ayr, it continued as a marker of Melanesian identity for one, and sometimes, two generations. For those Melanesians that were dispersed among the White and Aboriginal communities of Queensland this second function of establishing identity was of little relevance. The viability of a pidgin and its chance of becoming creolised very much depend on the relative distribution of males and females among its users. Schneider (1934: 86) provides the information in table 10.1. The fact that less than ten per cent of all recruits were female is indicative of the obstacles encountered by those workers who wanted to lead a normal family life on the plantations. The small proportion of females, together with the practice of employing labourers for limited short periods only, prevented the emergence of stable immigrant communities where nativised (creolised) varieties of Kanaka English could develop. As can be seen from

Queensland Kanaka English

177

the figures, significant changes in this proportion occurred only after the majority of the Melanesians had been repatriated. Moreover, at least around Mackay, a reasonably sizeable Melanesian community had established itself, the members of which communicated mainly in Kanaka English. Nevertheless, no creolisation appears to have taken place. Whilst the parents in many Melanesian families spoke to one another in Kanaka English, thereby providing the preconditions for this language to become a Creole, they discouraged their children from using it. Instead they urged them to learn 'proper' English at school or from their older siblings. These negative parental attitudes towards pidgin were reinforced by the fact that most Melanesian children went to White state schools where both teachers and peers looked down on Kanaka English. Jourdan (1983) has studied the decline of Kanaka English in Mackay, one of the main centres of the sugar industry. It appears that, after a brief period of bilingualism and bidialectism extending for no more than three generations, a total shift to English monolingualism occurred. The case of Queensland not only illustrates how changing external conditions lead to the death of a pidgin, it also shows that creolisation does not necessarily occur when children grow up in pidgin-speaking families. In fact, the same phenomenon was observed with Samoan Plantation Pidgin where, instead of creolising their parents' pidgin, children shifted to a more useful language spoken in the wider community. Whilst Kanaka English thus died out in Queensland itself, it continues to be spoken, in a changed and more developed form, in the areas where the labourers were originally recruited. Both Solomon Islands Pijin and Vanuatu Bislama are in many ways continuations of the Queensland tradition of Pidgin English. 10.3 Some theoretical implications

The study of Queensland Kanaka English can throw light on a number of theoretical issues in pidgin and Creole studies, in particular those encountered in the debate concerning Bickerton's (1981) bioprogram hypothesis of creolisation. There are, for instance, a number of striking similarities between the Creole situation in Hawaii and Queensland. In both instances we are dealing with an influx of foreign workers into a settler's colony, that is one about to be dominated by White Anglo-Saxon modes of behaviour. This is quite different from the colonial setting encountered in Surinam or Jamaica. In both Hawaii and Queensland formal education was provided in the acrolect for children born on the plantations. Thus, bilingualism in the Creole and the acrolect was the norm rather than the exception. The full impact of the schools on Hawaii Creole English still remains to be documented. Queensland and Hawaii were also part of a number of Pacific

178

PETER MUHLHAUSLER

trade networks in which reduced forms of English and other trade languages were used. Another similarity is the presence of immigrant groups other than those employed on the plantations. However, there remain a number of differences, the most salient being the number of plantation workers, who stayed on permanently and the size of the Creole community. It appears that these two factors, crucial to the viability of a separate social group, at the same time determine the extent to which an independent Creole can emerge. Creolisation, is not, as suggested by Bickerton (1986:4), a matter of acquisition by individuals, but a result of special extralinguistic factors. Thus, in spite of very similar superstrate and substrate languages, the linguistic solutions adopted on the plantations of Hawaii and Queensland differed because those social factors conducive to the creation of a novel Creole language were lacking in Queensland. The few Kanakas who grew up speaking Kanaka English as their first language did not produce, as far as can be ascertained from the linguistic data collected, a qualitatively different Creole. Such qualitatively different systems are only beginning to emerge in the varieties of Kanaka English that were taken back by workers returning to the Solomons and the New Hebrides (now Vanuatu). It appears that the complexities that have emerged in the grammars of Bislama and Solomon Islands Pijin originated to a very significant degree among adult urban speakers in the process of constructing a new social and linguistic identity. Earlier isolated urban dwellers do not appear to have produced such linguistic elaborations, again confirming the principle that creolisation is dependent on social conditions. A second theoretical issue, the discussion of which has involved frequent reference to Kanaka English, is the role of substratum influence in pidgin development. Keesing (1988) and Miihlhausler (1987) have concentrated on the emergence of pronominal marking or resumptive pronominalisation in Kanaka English and its daughter languages. Whilst Keesing emphasises the close similarities between pronominal marking in the most developed forms of Kanaka English and the Oceanic languages spoken natively by its Melanesian users, I attempt to demonstrate an independent development of this construction, ranging from no marking in the early varieties of the language to increasing grammaticalisation in later stages. The independent development hypothesis is given additional support by a number of important differences between pronomial marking in Kanaka English and its Oceanic substrate languages. At the moment, the issue remains unsettled as there are considerable problems with the quality of the data and their interpretation. A similar case study could be undertaken for other constructions, such as the marking of relative clauses by means of we (from English where) which either developed independently in Kanaka English and many other Creoles or can be traced back to a common source. Because of the well-known historical links between the plantations of Queensland

Queensland Kanaka English

179

and many parts of the Pacific and Australia, this issue of monogenesis vs. polygenesis of pidgins and Creoles can be studied particularly well with reference to Kanaka English. 10.4

Conclusions

In comparison with other pidgins and Creoles in Australia and the Pacific, relatively little has been written about Queensland Kanaka English, unlike, for instance, languages such as New Guinea Tok Pisin. The dearth of information about this language is out of proportion to its importance as a crucial stage in the development and diffusion of Pidgin English in the south-western Pacific, and its consequent importance for the theory of pidginisation and creolisation. Whilst only a handful of active speakers survive, there is ample scope for archival research and for comparative work with its successor languages. Note 1. Kanaka is the term by which the Melanesians of Queensland refer to themselves (cf. canaque for the indigenous New Caledonians).

11 Torres Strait Creole ANNA SHNUKAL

11.1

Introduction

On ten of the 18 inhabited islands of Torres Strait, lying between Cape York Peninsula and the mainland of Papua New Guinea, and in the Cape York communities of Bamaga and Seisia, where Torres Strait Islanders predominate, the traditional languages are being or have been replaced by an English-based Creole. Its speakers call this language Broken (from 'Broken English'), Pizin (from 'Pidgin English'), BigThap ('BigPiece') or, among some younger speakers, Blaikman ('Blackfellow') or Ailan Tok ('Island Talk'), but it is never called Langgus, a name reserved for the indigenous languages and English. Here I shall refer to it as Torres Strait Creole (TSC). Today, the children of Erub, Hammond Island, Masig, Mer, Moa, Purma, Thursday Island, Ugar, Waraber and Yam (see map 11.1) speak TSC as theirfirstlanguage, although in some communities the children can still understand their traditional language. On Masig, Mer, Moa, Purma and Waraber only elderly speakers of the traditional languages remain. The Creole has been the first language of the majority of inhabitants of Erub, Ugar and St Paul's Community, Moa, for four generations. The number offirst-languagespeakers of the Creole in the Torres Strait islands and in the Cape York settlement of Bamaga is probably around 3,000 out of a total Islander population of about 4,000. However, since the end of World War II there has been a steady outflow of people from the Strait and today a majority of Torres Strait Islanders, some 12,000 to 14,000 people, live permanently or temporarily on the mainland. As the lingua franca of all Torres Strait Islanders, including the majority now living in Bamaga and the coastal cities and towns of Queensland, the Creole may now have as many as 12,000 second-language speakers. Because of the prestige of English as the official language of Australia, the language of education, of administration and of the media, TSC and

Torres Strait Creole

PAPUA

NEW

GUINEA

Map 11.1 Torres Strait showing inhabited islands

181

182

ANNA SHNUKAL

English are in a diglossic relationship on those islands where TSC is the majority language, with English as the 'high' and TSC as the iow' variety. There as signs, however, that the domain boundaries are beginning to weaken, as the Creole has taken on the functions of pan-Islander ethnicity and identity among younger people. While many Islanders and Europeans consider the Creole to be merely a 'corrupt' or 'ungrammatical' form of English, this view cannot be sustained on either sociolinguistic or linguistic grounds. Few Europeans speak the Creole well, and the two languages are not mutually intelligible when spoken by native speakers at normal speed. For most Islanders, English is felt to be an alien language, a product and shaper of European, rather than Islander, culture. Linguistically, TSC more closely resembles the indigenous Torres Strait languages in its phonological, semantic and pragmatic structure than it does English. Nevertheless, some 85 per cent of the vocabulary of TSC has been borrowed from English, often with substantial phonological and semantic change (see 11.3 below). There are two main dialects of the creole, eastern and western, which reflect differences between the two major indigenous language areas of Torres Strait and different periods of creolisation. The eastern dialect is spoken as a first language on the three eastern islands, Erub, Ugar and Mer, and has been heavily influenced, both phonologically and lexically, by its indigenous substrate Papuan language, Meriam Mir (or Miriam). The western dialect is spoken on Moa, on Hammond Island and on all the central islands, although there are minor phonological and lexical differences among them. Like the eastern dialect, the western dialect has also been greatly influenced by its indigenous substrate language, Kala Lagaw Ya (called Mabuiag in early texts), which belongs to the Australian PamaNyungan language family (see Ford and Ober, this volume). Moreover, TSC exhibits characteristics of a post-creole continuum, a chain of varieties influenced by the superstrate language. English represents literacy, educational achievement and socioeconomic advancement, and has greatly influenced the younger generation of Islanders, the first to have been educated at secondary school by English-speaking Europeans. Generally speaking, the Creole varieties spoken by younger Islanders are phonologically, syntactically and semantically more like English than those spoken by their parents and grandparents. Nevertheless, all speakers of TSC share a common core of linguistic features and these will be briefly discussed in the latter section of this paper. 11.2 History TSC developed from the varieties of Pacific Pidgin English that were brought to Torres Strait by the Pacific Islanders and Europeans who

Torres Strait Creole 183 established the beche-de-mer, trochus and pearling industries from the late 1840s.1 Thus it is historically related to the varieties of Kanaka Pidgin English spoken in the canefields of northern Queensland (see Miihlhausler, chapter 10, this volume) and to three modern Pacific Creoles, Papua New Guinea Tok Pisin, Vanuatu Bislama and Solomon Islands Pijin. The rapid spread of the Creole was due to two main factors: its use as a lingua franca among Torres Strait Islanders speaking two mutually unintelligible languages, and the widespread belief that it was English. Although religion, trade, reciprocal visiting and name exchange had long linked the two groups of Islanders, they had no common language. With the commercial exploitation of the region's marine resources a lingua franca quickly arose among the newly arrived Europeans, Pacific Islanders (both Polynesians and Melanesians), Papua New Guineans, Australian Aborigines, Filipinos, Indonesians and Japanese and it was soon spoken by the Torres Strait Islanders who worked with the newcomers. This lingua franca was Pacific Pidgin English. By the late 1890s a majority of the children on the eastern islands of Erub and Ugar had immigrant fathers and Meriam Mir-speaking mothers and they began to use their fathers' pidgin as their primary language, thus creolising the pidgin. About ten years after its creolisation on Erub and Ugar, the pidgin again creolised independently on St Paul's Anglican Mission on Moa Island, under similar circumstances but with the western island language as its major substrate. For political, cultural and economic reasons the most influential men in these communities were Pacific Islanders, who were perceived to have a special relationship with the Europeans and became cultural mediators between them and the Torres Strait Islanders. Their children chose the pidgin as their primary tongue, partly to emphasise that they were different from, and superior to, other Torres Strait Islanders, and partly because they believed that they were speaking English which, as the 'white man's language', was more appropriate to the new order that was being established in the Strait. It became the first language of children born during the 1920s on Masig and Yam, of those born after about 1930 on Waraber, Purma and Hammond Island and of those born after World War II on Mer and Thursday Island. From 1871 Pacific Islander pastors were deposited by the London Missionary Society on several of the islands of Torres Strait, as part of the planned conversion to Christianity of the peoples of Papua New Guinea. These men were extremely influential in reshaping the cultural, social and religious life of the Torres Strait Islanders and exercised what appears to have been despotic control on the islands. Although the islands were annexed by the then colony of Queensland in 1897, the Government Resident, who was responsible for the adminis-

184

ANNA SHNUKAL

tration of the islands, interfered very little in island affairs. In 1904, however, the Islanders came under the Queensland Aboriginals Protection and Restriction on the Sale of Opium Act of 1897, and their lives became increasingly controlled by the state government. European administratorteachers, placed on the largest islands to supervise the Islanders in almost all matters, began to displace the Pacific Island teachers as the principal authorities. They discouraged the use of indigenous languages and encouraged parents to speak to their children in (Pidgin) English to improve their schooling. From the second decade of the century Torres Strait Islanders began to be trained as assistant teachers and were soon being sent to communities other than their home island, often across the traditional language barrier. Because they did not share the community's traditional language, they had to use the pidgin in everyday conversation as well as in their teaching, thus increasing its prestige and the general belief that it was English. At the same time, trained Islander police and clergy, also men of some authority and influence, were deliberately being posted away from their home islands, which further reinforced the Creole's use and status. It was also during the period between the two world wars that a panIslander consciousness began to develop out of the previous linguistic, political and cultural divisions. It was a response in part to Christian ideology, in part to the shared experience of government control, and in part to the first east-west marriages, contracted by the descendants of the Pacific Islanders, who sought partners across the linguistic barrier in order to 'keep the blood'. As external circumstances weakened the traditional east-west barrier, they also created and strengthened pan-Islander ties which were symbolised by the newly-creolised pidgin. Thus, its function as the shared language of Torres Strait Islanders can be viewed as an extension of that of its ancestor pidgin as the lingua franca of the marine industries of Torres Strait. World War II saw the beginning of the end of isolation of the islands from the rest of Australia. Some 700 men joined the Torres Strait Defence Force and for the first time served alongside Europeans in a situation of equality. It was then they realised the Creole was not English. With all Europeans evacuated from the islands, Islanders again became responsible for the provision of their own local government, education, health and religious services. When the war ended, they presented the government with a series of demands: Australian citizenship, freedom of movement, control of wages, access to alcohol, better pay, housing, education and jobs. After many years most of these demands were granted and island councils were freed from state government control. Within a few years, however, people began to leave the outer islands in

Torres Strait Creole 185 search of improved services. They settled on Thursday Island (which had been barred to them before), the new communities of Bamaga and Seisia, and in the cities of mainland Queensland, many of which now have large Islander communities. Here, also, as whenever Islanders come together, the common language is TSC, although among the children in the cities English is becoming the dominant language. In this they are no different from the children of other non-English speaking families in Australia (and particularly in Queensland), where languages other than English usually disappear within a generation (see e.g. Clyne, this volume). 11.3 Structure of Torres Strait Creole2 11.3.1 Phonological structure

This brief phonological description of TSC sets out the core sound structure of the basilectal varieties and discusses some of the sound changes which occur when lexemes are borrowed from the major contributing languages. The phonemic differences between the eastern and western dialects are also noted. There is also considerable variation among speakers as they 'shift' their speech towards more English-sounding norms in formal situations and in the presence of Europeans. TSC has 24 contrastive sounds: 15 consonants, 5 vowels and 4 dip thongs. Table 11.2 presents the phonetic values for the symbols of TSC orthography. Vowel length is not distinctive in TSC and words may begin and end with any of the consonants, vowels or diphthongs above. Any sequence of two or more consonants that begins an English word can also begin a word in TSC. Words may have a maximum of four syllables. Any borrowed word of more than four syllables is reduced, e.g. prizareta ~ 'refrigerator', Early English borrowings may add an epenthetic vowel to initial or final consonant clusters, silip - 'sleep', or anis - 'ants'. TSC has three degrees of stress: primary, secondary and tertiary. With few exceptions borrowed words keep their original stress, although stress pattern of individual words may differ according to speaker's age and island origin. The eastern dialect has an additional phoneme, eu, found only in two words of Meriam Mir origin, beuger- 'booby' and euski - 'to squat'. The western dialect has three additional phonemes, a diphthong ui occurring in words borrowed from Kala Lagaw Ya and a voiced/voiceless pair of interdental stops. There has been some phonological simplification of words borrowed into the Creole from its superstrate and substrate languages. In English-

186

ANNA SHNUKAL

Table 11.1 Phonological inventory of TSC Consonants

Stops Nasals Fricatives Liquids Semi-vowels

Bilabial

Dental

P,b m

t, d n s, z 1, r

w

ng

y Vowels Front

Centre

i e

High Mid Low

Velar

Palatal

Back u 0

a Diphthongs Front

Mid Low

Centre

ei

Back oi

ai, au

derived vocabulary the major changes are the substitution of'pure' vowels (i.e. monophthongs) for English ones, alveolar fricatives for English affricatives and stops for non-alveolar fricatives. In vocabulary derived from the indigenous languages final diphthongs tend to become monophthongs and the final vowel, as well as some internal vowels and consonants, may disappear. It is now impossible to know whether the few sound changes observed in lexemes borrowed from Meriam Mir are the result of nativisation or are long-standing dialect differences between Erub and Ugar. In lexemes borrowed from Kala Lagaw Ya, however, the phonemic long vowels of that language have merged with their corresponding short vowels in TSC and the final voiceless vowel has disappeared, for example, thaapi- 'piece' becomes thap. The vowel oe has generally become / or a, e.g. kibim - 'black trevally' from koebim. The diphthongs oey and oew have become / and 0, respectively, for example, ginau - 'Torres Strait Pigeon' from goeynaw, and word-initial ng is beginning to be replaced by n, for example, nar - 'lower leg' from ngar. 11.3.2

Morphology

This section and the next briefly describe the major features of TSC morphology and syntax. There is some morphophonemic variation

Torres Strait Creole

187

between dialects, but only the core shared features of the morphological system will be discussed here. Like most Creoles, TSC has little inflectional morphology. Most words are monomorphemic, even those of bimorphemic origin, for example, wanem - 'what' from what name). The derivational component is beginning to increase, which is a general tendency in the development of Creoles. Transitivity and causativity are linked in TSC, both being marked by the addition of the bound inflectional verb suffix -e (originally from him) to verbs of English origin. Thus, the verb, raw- 'run', is intransitive in No ran diskain! - 'Don't run around like that!', but transitive (or perhaps causative) in Rane em! - 'Chase him!'. There are seven derivational bound verb suffixes, all of which are derived from English verb particles and which add a spatial dimension to Englishderived verbs. They are: -ap - 'up', -aut 'out', -baut - 'about', -daun 'down', -op - 'off', -raun - 'around' and -wei - 'away'. The transitive/causative suffix may attach to -ap, -aut and -daun, as in asmape'hoist', mekraute - 'decipher' and belidaune - 'put face down'. There are three bound personal pronoun suffixes, dual -tu, plural -pla and reflexive -selp; a bound noun suffix, -wan - 'one', which derives nouns from adjectives, for example, slaikwan - 'a loose one'; and four adjective and adverb suffixes, -kain, -said, -taim and -wei, which express approximation, location, time and manner, respectively. Compounding is still a productive process in TSC, for example, baikso'backache' from 'back sore', but reduplication, apparently a productive process in the early pidgin, is no longer used. There remain, however, at least nine reduplicated, 'intensified' adjectives, derived from English nouns, such kalakala - 'multi-coloured' and spotspot - 'spotted'. 11.3.3

Syntax

TSC is an SVO language, and the basic sentence division is between subject noun phrase and predicate verb phrase, indicated by the predicate marker i (probably from he). Although there are subjectless sentences in TSC, there is none without a predicate. There is less syntactic variation in the Creole than phonological and lexical variation. Syntactic variation is largely dialectal, the result of two (and sometimes three) contributing systems still seeking an equilibrium. In the western dialect, for example, competition between the Kala Lagaw Ya derived repetitive aspect marker, lak (from laka), and the English derived repetitive aspect adverb, gen, has word order implications, since each occurs in a different position in the sentence, e.g. Ai lak go go - 'I'll go again' and / delait gen - 'It was daylight again'. Some of the variation, however, appears to be the result of universal trends in Creole development. Similar

188

ANNA SHNUKAL

implications for word order arise from the recent movement into the verb phrase of aspect markers, like the dubitive mait and the completive pints, which were previously sentence initial and final, respectively, e.g. Yu mait bin luk em - 'You may have seen him' and Bel i pinis go -The bell has (already) sounded'. In this section, however, only the core of the shared system will be discussed. 11.33.1 The noun phrase Noun phrases consist of an obligatory noun or pronoun and four optional elements. The order of these elements is: determiner, quantifier, adjective, noun/pronoun, preposition phrase. Determiners may be articles (singular and plural definite, singular indefinite and generic), demonstrative adjectives (singular and plural, proximal and distal), or possessive adjectives (identical with the personal pronouns, except for the first-person singular form, mai). As for quantifiers, there is both a cardinal and an ordinal series of numerals, and six general quantifiers, all based on English, but with some semantic shift, for example, lelbet - 'a few', olgeda - 'all', plenti 'many', pulap - 'plenty of, tumas - 'too many'. Attributive adjectives are not inflected for case, number or gender. They may occur singly or in a string, as in smol siki gel - 'small cheeky girl'. Nouns do not inflect for case, number or gender. Case is marked, as in English, by word order for the core syntactic relations, and preposition phrases for peripheral relations. Number is optionally marked by the choice of singular or plural determiners and gender by the adjectives man 'male' or oman - 'female', for example man ata - 'grandfather' and oman ata - 'grandmother'. There are four classes of pronouns: personal pronouns; indefinite pronouns; demonstrative pronouns; and interrogative pronouns. The subject and object personal pronouns have singular, dual and plural forms, with an inclusive/exclusive distinction made in the first-person plural pronouns. Case is not marked except in the first-person singular, with ai and mi as subject and object, respectively. The 14 indefinite pronouns are based on those of English, with the exception of olgeda (from altogether) meaning 'everybody, everyone, everything', an old form currently being replaced by eribodi and ebriting. The four demonstrative pronouns are formed by adding the noun suffix -wan to the demonstrative adjectives; and there are 13 interrogative pronouns which begin interrogative sentences.

11.33.2 Like the noun phrase, the verb phrase contains an obligatory element as well as four optional elements. The obligatory element is the verb or

Torres Strait Creole

189

predicate adjective. The order in which the elements occur is: predicate marker, modal, tense marker, aspect marker, for example: Ol i mait bin pinis strete yu. 3pl PM mood tense aspect correct you. They may have already corrected you. There is some evidence that all predicate adjectives were stative verbs in the pidgin and that they are currently being reanalysed as adjectives, possibly under the influence of English. There is no present tense marker in TSC but past and future tenses may be optionally signalled by the tense markers bin and go, respectively, e.g. Mary i bin graule em - 'Mary quarrelled with her' and Demtu go luk pamle bio demtu - 'They will visit their family'. Modality is expressed by six modal verbs (all except one borrowed from English modal verbs, but with some semantic shift), three modal particles, and several modal adverbs. The aspect system consists of six core aspect markers: kip (ITERATIVE); nomo (CESSATIVE); oltaim (HABITUAL); pinis (COMPLETIVE); stat (INCEPTIVE); and stil (CONTINUATIVE), as well as three adverbs and a set of verbal constructions, whose primary function is to convey aspectual meaning. 11.3.3.3 The preposition phrase Peripheral intrasentential grammatical relations are expressed in TSC by preposition phrases, consisting of preposition 4- noun phrase. The four basic prepositions in TSC, which may have been the only prepositions in the pidgin, are lo (LOCATION) from 'along', go (GOAL) from 'go', kam (SOURCE) from 'come' and bio (POSSESSION) from 'belong'. There are an additional eleven complex prepositions of location consisting of a noun + lo, for example, andanit lo - 'under, underneath, below' and antap lo - 'on, above, on top of. Notions of location and direction are central to the grammar of TSC and marked in various ways; of the 23 prepositions currently in use, 17 mark location or direction. The other six prepositions express TIME (apta - 'after', bipo 'before' and til 'until'), ACCOMPANIMENT (ene - 'with'), MANNER (olsem - 'like'), INSTRUMENT (gad - 'by means of), BENEFIT (po - 'for'). 11.3.3.4 Simple sentences A simple sentence is made up of two distinct parts, an optional subject and an obligatory predicate. Simple sentences express one of three processes: verbless sentences express relations; subjectless sentences express states; and 'full' sentences express events. There are five major simple sentence types. Relations of identity are expressed by verbless equational sentences, in which both subject and predicate are noun phraes, e.g. Mislam i boi bio Kemuel - 'Mislam

190

ANNA SHNUKAL

is Kemuel's son'. Relations of description are expressed by verbless descriptive sentences, in which the predicate is an adjective, adverb or preposition phrase, e.g. Kaikia i redi nau - The meal is ready now'. States of affairs are expressed by subjectless impersonal sentences, like / ren naw- It's raining' and existential sentences, for example, Igadkap yaThere is a cup here', where the predicate marker is obligatory. Events are expressed by 'full' verbal sentences, sentences with both a noun phrase subject and a verb, for example, At bio mi i zam - 'My heart leaped'. Yes-no questions have the same structure as the sentences discussed above, but are spoken with rising intonation: I gad sos lo dis strit?- 'Is there a church in this street?'. Information questions begin with an interrogative pronoun or interrogative adjective, all of which are based on English interrogatives, although again there has been some semantic shift. Two examples are Wanem yu bin mekem?- 'What did you do?' and Amaspilam dapota i karem? - 'How much film does the camera hold?'. Imperative sentences have an optional second-person pronoun followed by an obligatory main verb, for example Yu prese tos po mi! - 'Hold the torch on for me!'. There are two negative particles in TSC, no and neba (the latter used only in sentences with past tense meaning), both of which precede the verb as well as any tense, aspect or modality markers, for example, Dalassa neba luk nating deya - 'Dalassa didn't see anything there' and Em no bin sing 'She didn't sing'. Some mention has already been made of the importance of location and direction in the Creole but the complex rules governing the use of obligatory locational and directional particles can be only briefly discussed here. The proximal or distal location with respect to the speaker of subjects of stative verbs and of the 'lexically simple' motion verbs, kam - 'come' and go - 'go' must be indicated by the preverbial deictics, ya (from 'here') and de (from 'there'), respectively, for example, Gemai ya slip - 'Gemai is sleeping (close to the speaker)' and Trakta de kam - 'The tractor is coming (distant from the speaker)'. Moreover, the direction in which the subjects of 'lexically complex' verbs of motion are travelling must also be specified. The use of go specifies that the subject is travelling away from the speaker, as in Em bin ran go sanbis - 'She ran to the shore', whereas kam specifies movement towards the speaker. Homomorphic adverbs are similarly used to indicate direction of movement, as in Sakem kam! - 'Throw it (over here)!'. 11.33.5 Complex sentences Coordinate sentences are linked by the coordinate conjunctions, ane 'and', bat- 'but', instedor matha- 'instead of, but. . . instead' and o- 'or', for example, Mitu Kathy bin dans ane Akabu bin sing - 'Kathy and I both danced and Akabu sang'.

Torres Strait Creole 191 Sentences may be subordinated either by juxtaposition or use of complementisers. There are five TEMPORAL complementisers, apta- 'after', bipo - 'before', sun - 'as soons as', til- 'until' and wen - 'when, after, as', three CAUSAL complementisers, bikos - 'because', daswai - 'therefore, for that reason' and prom - 'because', two CONDITIONAL complementisers, ip 'if and orels - 'or, or else', and two PURPOSE complementisers, po - 'to, in order to' and slong- 'so that, in order that'. Temporal clauses introduced by wen and conditional clauses must take the future tense marker go, if they refer to future events, for example, Yu go raite letapo mitu wen yu go libi Purma - 'You'll write to us after you leave Coconut Island' and Yu go zam tumas, leg bloyugo brok- 'If youjump around too much, you'll break your leg'. In addition, interrogative sentences may become embedded questions, for example, Ai no sabe wiskain ol man i stop - 'I don't know how people live'. All sentential noun phrases, whether their grammatical function, may be relativised by using optional we, for example, Dat stori we yu bin spik i prapa paniwan - 'That story you told was very funny' and Ai lukraun mai klos ai bin lusem - 'I looked for my dress (that) I had lost'. 11.3.4 Lexicon The lexicon of TSC comes from many sources, chiefly English, Meriam Mir and Kala Lagaw Ya, and it continues to borrow words from English and from Kala Lagaw Ya. English-derived vocabulary constitutes about 85 per cent of the lexicon, with another 14 per cent (in each dialect) from the regional indigenous languages and a further one per cent from Japanese, Malay-Indonesian and various Pacific island languages. The major contribution of the indigenous languages, as might be expected, has been in the domains of flora, fauna, traditional life, place names, kin terms and body parts, e.g. abal- 'pandanus', arti- 'octopus', bud - 'mourning period', Maizab Kaur - 'Bramble Cay', aka - 'grandmother'. However, English and, to a minor extent the Pacific languages, have also contributed to these domains, with the result that there is considerable synonymy in the Creole. The choice among synonyms is governed by situational context and by a concern for stylistic effect, as well as by sociological characteristics of the speaker, such as island of origin, gender and age. Moreover, in addition to the phonological changes outlined earlier, the borrowed words often undergo semantic change. This is most apparent in English-derived words and is the source of some misunderstanding between Islanders and Europeans. It would appear that English words with related meanings have been chosen to fit the preexisting semantic categories of the island languages. For example, the word traik (from track) in TSC has slightly different meanings in the eastern and western

192

ANNA SHNUKAL

dialects. In the east it means 'sole of the foot' and leg refers to both 'lower leg' and 'foot (excluding sole)'. This parallels the semantic distinction made in Meriam Mir, but differs from the use of traik and leg in the western dialect. There, leg refers to the part of the body below the knee to the ankle (ngar in Kala Lagaw Ya) and traik means what 'foot' does in English, the part of the body from the ankle down. A comparison of borrowed vocabulary from English and the two indigenous languages suggests that English lexemes tend to acquire further related meanings, whereas the indigenous words lose meanings they once had. A word like kese, for example (from catch) now means 'catch', 'take', 'hold', 'grasp', 'seize', 'arrive at', 'hit', 'understand' and 'embrace', while tereg, from Meriam Mir tirig meaning both iawyer-cane' and 'teeth', now has only the first meaning in the Creole. Today, lexical innovation comes primarily through borrowing from English. However, new lexemes are sometimes created by compounding and phrase and sentence construction. New noun phrases are generally of the form: adjective + noun, such as big win - 'cyclone' and smol haikai - 'snack'; or noun-h Wo - 'of + noun, such as maut bio krik - 'estuary' and taim bio plan - 'planting season'. Sentences which express as a complex idea what is, in English, a compound word, includepauli singaut - 'cockcrow' and san i kamap - 'sunrise'. New verb phrases are sometimes formed by expressing as separate words the component meanings of Latinate English verbs, such as meke kam mo gud'improve' and pute insaid- 'insert'. These phrases are called by Islanders 'breaking down English', a process which is the origin of the most common name for the creole, Broken. 11.4 Future developments

It is only recently that TSC has gained a measure of acceptance as a language separate from English. Many Islanders and Europeans do not accept this view, however, largely due to the close linguistic relationship between Creole and English and the absence of codified norms. Moreover, few Europeans have heard basilectal varieties. TSC was without its own orthography until a few years ago and only three modern linguistic descriptions have so far been published.3 Since English is the only written medium of communication for most Islanders, even those whose mother tongue is an indigenous language, they tend to use English orthographical conventions in writing creole. All three Torres Strait island languages had, until the 1970s, been ignored or stigmatised as 'uncivilised' by the European administration. Schoolchildren were officially discouraged from using their indigenous languages in the schools and playgrounds. It was not until the 1980s that

Torres Strait Creole 193 TSC was recognised as a language separate from English and discussed in language classes at the Thursday Island and Bamaga High Schools. The increasing interest in Torres Strait languages led to the introduction of an official English as a Second Language programme on Thursday Island, but it should be pointed out that unofficial defacto bilingual programmes using both English and an island language had long been implemented by Islander teachers in the primary schools, often with the encouragement of their European principals. Another form of official support has come from the Australian Broadcasting Commission, which in 1985 assisted the transmission of an Islander broadcasting service from Thursday Island. There is now widespread and enthusiastic support for broadcasts in the indigenous languages as well as TSC. In fact, the attitude of Torres Strait Islanders towards the Creole appears to be changing. Originally believed to be English, the language gained prestige through its association with Europeans and their 'allies', the Pacific Islanders. This prestige was reinforced by its connection with Christianity, education and the newly emerging Torres Strait Islander elite. Then, as Islanders began to become more integrated into the broader Australian economy and society, it came to be seen as a substandard variety of English and even as an excuse for denying Islanders their just demands as Australians. Creole's low status among older Islanders and its connection with the Pacific Islanders has meant that the people of the western islands, Mer and Bamaga, feel threatened by its spread. The traditional languages there are becoming weaker, and while many Islanders could have accepted the loss, provided that the replacement language assured them and their children an equal place alongside whites in Australian society, they have come to realise that this is not the case. They blame the Creole for keeping them second-class citizens. Conversely, TSC today is being rediscovered and revalued by younger Islanders as a marker of identity, ethnicity and separateness from white Australians. Ironically, this is occurring in the context of decreolisation. Many Islander children who are fully bilingual in both English and TSC know very well the differences between the two. For them the Creole is an addition to their linguistic repertoire, which is not available to nonIslanders. Moreover, they cannot be made to feel ashamed of their competence in English since they speak it well. Whether the Creole remains a viable language in its ownrightor whether, as happened to Kanaka Pidgin English in Queensland within one or two generations, it becomes an ethnolect of Australian English, is difficult to predict (see Dutton 1980; Muhlhausler 1981,1985; Dutton and Muhlhausler 1984; and Shnukal 1985b). Certainly, its future will be decided by social and cultural factors, the most important of which is the maintenance of a

194

ANNA SHNUKAL

distinctive Torres Strait Islander identity in an Australian state which remains antagonistic to the expression of linguistic, social and cultural differences. 11.5

Conclusion

The Creole's phonological, syntactic, semantic and discourse structure require further study, as do its regional and social varieties and the continuing influence of superstrate and substrates in its development. Another issue is the influence of TSC on the speech of young urban Queensland blacks. Finally, our understanding of Australia's language ecology would be broadened by an analysis of the critical factors involved in the shift by Torres Strait Islanders away from their indigenous languages towards the Creole and by an account of the developments arising from the Creole's new functions as a codified language, increasingly being used in education and the media, and as an ethnicity marker among young Islanders. Notes 1. An account of historical and sociocultural influences on the creolisation and spread of TSC can be found in Crowley and Rigsby (1979), Shnukal (1983) and Shnukal (1985a). 2. The general framework here is based on Crowley and Rigsby's (1979) description of Cape York Creole. Although the phonology of the two varieties is rather different, the morphological and syntactic structure is very similar. 3. Apart from a brief discussion and word list of the early pidgin in Ray (1907), there is an account in Dutton (1970) of the 'informal English' spoken in Torres Strait, based on data from several adolescent boys, a description by Crowley and Rigsby (1979) of Cape York Creole, a variety spoken by Aboriginal second-language speakers from Bamaga and some grammatical and lexical information in Shnukal (1982a). Shnukal (1988) is an introduction to the language, its history, development, phonology, grammar and lexicon.

12 Kriol - the creation of a new language JOHN W. HARRIS

124

Introduction

Kriol is an English-based Creole, widely spoken by Aboriginal Australians in the 'Top end' of the Northern Territory and adjacent regions. Emerging early this century but derogated or ignored until just over ten years ago, Kriol has since become better documented. It is now the subject of considerable interest and debate, due largely to its expanding use in education and in the Christian church (see Harris and Sandefur 1985b; J.W. Harris 1986b). Kriol is of more than local interest because it is probably the latest Creole to be comprehensively described linguistically (Sandefur 1979; Sandefur and Sandefur 1979a; and Sandefur, this volume), sociolinguistically (Sandefur 1984b), and historically (J.W. Harris 1984, 1986a). It is also of importance because the circumstances in which Kriol arose differ from those in the plantation and post-plantation societies from which so many other Creoles have emerged. I outline briefly the history of Kriol in four broad stages: its pre-European background, the eras of pidgin genesis and stabilisation and the modern period of creolisation and extension. 12.2 The pre-European context

Speech communities consisting of large numbers of overlapping language communities are typical of Aboriginal Australia (Rigsby and Sutton 1982; and Sutton, this volume), particularly in areas such as North Australia where favourable natural environments are able to support large numbers of people organised into small groups (White 1978: 48). Along the Northern Territory coast and hinterland, multilingualism was the cultural norm. Furthermore, for several centuries prior to the European invasion of the region, Aboriginal people of the Northern Territory coast had a close

196

JOHN W. HARRIS

trading relationship with South-east Asia. The trepang trade was largely based in Macassar (in modern Suluwesi) from where a largefleetof around 50 praus, with multilingual crews, arrived each December and remained until April (MacKnight 1976). The Portuguese-Malay trade language of the archipelago was adapted in North Australia to become the 'Macassan' Pidgin of communication, not only between Aborigines and traders but also, eventually, between distant Aboriginal groups (Urry and Walsh 1981). English, then, did not come into a pristine situation of societal monolingualism but into a complex cultural region in which there had already been developed skills of cross-cultural communication. 12.3 The rise of pidgin Englishes

From the sixteenth century until the early nineteenth, European explorers and slavers made fleeting contact with the northern coast. Their only linguistic heritage consisted of a few Portuguese loanwords (J.W. Harris 1986a: 117). In 1824 the British established the first of three-short lived military garrisons, Fort Dundas on Melville Island. Language contact during its five violent years was minimal. The second garrison, Fort Wellington, was established on Raffles Bay in 1827. Its first year was as violent as its predecessor but the appointment of the remarkable Collet Barket as commandant in 1828 led to an amicable relationship between the garrison and the local Aboriginal people. A contact language based on both English and the local vernacular began to emerge but its development was terminated by the abandonment of the garrison in 1829 (Harris 1985: 160). The third garrison, Victoria, on Port Essington lasted from 1838 to 1849, which, under circumstances of reasonable rapport between the settlers and the Aborigines, was a sufficient length of time to enable significant pidgin development to occur. Initially, the Port Essington people attempted to speak to the British in the 'Macassan' Pidgin, the language of their interaction with foreigners. Earl (1842: 140) reported: You ask for vocabularies. I am in the mostridiculousperplexity about them. After having collected many words, I found that I was making a vocabulary of a horrid patois of the Macassan dialect: in fact, nearly all the words the natives used when speaking with us are Macassarese . . . Early in 1840, McArthur reported that the Aboriginal people could speak some English but that no Whites spoke their language. By 1841 Stokes (1846 (II): 357) described the contact language of Port Essington as 'such few words of broken English as were then in use at the colony'. After a few more years it was becoming frequently recorded that some Aboriginal

Kriol - the creation of a new language

197

people spoke what was recognisably English (e.g. Keppel 1983: 158; Sweatman in Allen and Corris 1977:44). The Aborigines' contact language had changed in two years from the attempt to speak the 'Macassan' Pidgin, which was not generally understood by the British, to the use of an Englishbased pidgin understood by both groups. There is some evidence that what may have been taking place was a relexification of the 'Macassan' Pidgin (Harris 1985: 165-6). After an unsuccessful attempt (1864 to 1866) to establish a township at Escape Cliffs on the mouth of the Adelaide River, permanent European settlement was finally achieved at Darwin in 1869. Relationships between the settlers and the local Larrakia people were generally friendly and an English-based pidgin emerged rapidly. Ashwin (1932: 92) was able to describe the Larrakia people in 1872 as 'Port Darwin natives . . . semicivilised ones, which understand pigeon English'. Wildey, visiting Darwin in 1873, commented that 'some speak English fairly' (1876: 118). There is, however, little recorded data from this early Darwin period, making it difficult to assess the extent to which the Darwin pidgin may have been influenced by nautical jargon or by pidgin models from South Australia, where most of the early settlers originated from. Typically, the model presented to the Larrakia people was a non-standard English with pidgin elements, as exemplified by the use of terms such as savvy and piccaninny and baby-talk words such as turn turn for food. Important models were also provided by Port Essington people, who were frequently employed as interpreters. Searcy (1909: 36), for example, reports: 'These boys spoke good English, Macassar, and . . . their own language . . . I always took Port Essington boys with me on my trips.' This 'good English' was the Port Essington Pidgin English, as the very large number of quotations of Searcy's writings demonstrate. Thus, Searcy (1912: 172, in reference to 1882): ' " N o more, me think it", was Boom's reply. "That fella snake been piccaninny, him bin come alonga wood from over there. Blackfella alonga Port Essington say Macassar man have to catchem wood, no more good fella sit down along island"' [No, I don't think so. That was a young snake, he came from the wood over there. Port Essington Aborigines say the Macassans have to gather wood, there's no more good (wood) on the island.] It is deceptively easy to emphasise the European men's communication with a few Port Essington men and several prominent Darwin men, because these interactions are documented, and to ignore the much more frequent communication between European and Aboriginal women in which much linguistic innovation must have occurred. Thus, Sowden (1882: 144): 'The young lubras [black women] are engaged by the English as washerwomen and charwomen, and they work very well, I am told. Indeed, altogether they are almost indispensable to the whites.'

198

JOHN W. HARRIS

The discovery of gold in 1872 changed Darwin from a tiny European outpost to a bustling multi-ethnic frontier town. In a decade or so the nonAboriginal population jumped from 200 to 8,000, many of whom were Chinese. There were not many mining camps inland from Darwin where communication with Aborigines demanded a contact language. The pidgin English of Darwin was almost certainly transported inland as a model. As Sowden (1882) clearly shows, Chinese Pidgin English was spoken by Chinese to Europeans. It could hardly have failed to influence the developing pidgin of Darwin and the mining camps. This can be seen in Searcy's (1912: 127) example: 'by-em-by plenty fella policeman takem one piecee blackfella along bush' [Later, many policemen will take one Aboriginal into the bush]. In this sentence, spoken by a European in about 1883, fella, a Southwestern Pacific pidgin item, and piecee, a Chinese pidgin English item, both perform the same function as a suffix to a prenominal. This variable use of alternate forms no doubt characterised the pidgin for many years in this region of complex linguistic mixing. The Overland Telegraph and the goldrush opened up the Northern Territory to exploitation by the cattle industry. Thus the pastoral frontier, which had reached Queensland in the 1850s, extended westwards into the Northern Territory in the 1870s along the valleys of the Roper and Victoria river systems and into the Kimberley region of Western Australia. These were violent years, but the Europeans were there to stay. Despite the valiant efforts of the Aborigines, their resistance was mercilessly put down. Important sites of pidgin development in the pastoral region were the frontier townships of Roper Bar, Borroloola, and Katherine, and those cattle stations where, either by choice or coercion, Aboriginal people were 'made quiet' (Merlan 1978). Thus, one Aboriginal account observes (Jess James Garalgnanjag and Maudie Mangui, as told to Jay Read, in Read and Read 1978: 74): Then people come quiet now from after that feller bin get shot now . . . white man bin settlem down blackfeller now. . . might as well him bin just give up . . . and we might as well sit down . . . that when they bin havem that station. Of these sites, the cattle stations were finally the most important. The 'settling down' of Aboriginal people at the stations institutionalised contact between linguistically different peoples who had thus far been scattered and discontinuous. In this way such contact became a daily fact of life. The cattle station pidgin, or, more accurately, the Elsey Station version of it, has been immortalised in Gunn's books (1905,1908), the major source of specific linguistic data. Thus, Gunn (1905: 54, 78): You plenty savey. Me no more savey yabber stick. I think you close-up savey white fellow paper-yabber. Jimmy.

Kriol - the creation of a new language

199

[You do understand. I don't understand message sticks. I think you nearly understand European writing.] Him Maluka. Him track belong Maluka. Him bogey last night. [That's Maluka. Those are Maluka's tracks. He bathed last night.] Historical studies of the extension of the pastoral frontier from New South Wales through inland Queensland and into the Northern Territory (J.W. Harris 1986a: 184-214), reinforced by lexical analysis of the pidgins (J.W. Harris 1986a: 260-300; J.W. Harris 1988), strongly indicate that the major lexical source of the pastoral industry pidgin was South-east Australian Pidgin English (SEAPE). This pidgin began in the late eighteenth century as a contact language around Port Jackson. It was carried throughout much of Australia and was the invariable model presented to newly encountered Aboriginal groups. Favenc (1904, quoted in Baker 1966: 313) writes: The pidgin talk which is considered so essential for carrying on conversation with a blackfellow is mostly of very old origin . . . As the whites pushed on and on amongst new tribes, nothing was taken from the local dialects to add to the general pidgin stock, but the original was carried along.

12.4 The stabilisation of the pidgins

The single most important factor in the stabilisation of pidgin Englishes in the Northern Territory was their convergence towards a common standard, a phenomenon typical of the Pacific region (Miihlhausler 1985). The fact that the various pidgins, due to a common ancestry, were not dissimilar to start with, helped this merger. The convergence was very largely effected by the greatly increased mobility of the population. Aboriginal people often accompanied Europeans to places far from their own country. Many Aborigines were sentenced to periods in Darwin's Fanny Gaol where they mixed with people from other parts of the Northern Territory. Europeans were also very mobile. Stockmen moved from station to station, policemen spent time in different postings and government officials of various kinds moved about a great deal. These people took the pidgin with them, abandoning its local idiosyncrasies when necessary and preserving its commonalities. A typical example was Lionel Gee, government surveyor, goldfield warden and magistrate who, in his various roles, travelled extensively in the Northern Territory in thefirstdecade of this century. Among his later reminiscences, Gee (1926: 15) observed that the pidgin, although mostly lexically English, was not semantically English: 'though the words used are quite English, some of them are put in such a different sense that it takes a little time before

200

JOHN W. HARRIS

a new chum can get the hang of them and talk understandably with his black brother.' Not only was Gee able to use pidgin English wherever he went, but he and people like him were responsible for the convergence of pidgins towards a common standard. It is clear that by the beginning of this century, the pastoral industry pidgins had merged into one widelyunderstood lingua franca, best referred to as Northern Territory Pidgin English (NTPE). If one compares the pidgin English spoken in Elsie Masson's kitchen in Darwin in 1913 (Masson 1915) with that spoken in Jeannie Gunn's kitchen at Elsey Station in 1902 (Gunn 1905), or if one compares the pidgin English spoken in Borroloola Courthouse in 1902 (Spencer 1928) with that spoken in Darwin Courthouse in 1913 (Masson 1915), it is evident that all are examples of the same language. 12.5 The creolisation of the Pidgin

The history of Kriol well illustrates both the relationship between a Creole and its pidgin ancestor, and the important principle that creolisation occurs as a result of rapid social change and the demand for a primary language in a newly-emerged community. The first place in the Northern Territory where pressure was placed on the pidgin to expand to become the primary language of a new community was the Roper River Mission (now Ngukurr), where creolisation began to occur shortly after 1908 (see J.W. Harris 1986a for details). The invasion of the Roper River region by Europeans had commenced with the construction of the Overland Telegraph in the early 1870s. Huge cattle drives were then undertaken as the pastoral frontier moved from Queensland into the Northern Territory. Cattle stations were established in the 1870s and 1880s and a small township emerged at Roper Bar, the shallow crossing used by European drovers, miners, settlers, cattle thieves and anyone else who had to cross the Roper River travelling north or south. These were violent years and a great deal of aggression was directed at Aboriginal people in the region. As one of the early missionaries, R.D. Joynt (1918: 7) wrote, hundreds, had been 'shot down like game'. The massacre of Aboriginal people in a 'war of extermination' was widespread and continuous throughout the whole of the pastoral frontier. Initially, the battle was not entirely one-sided. The Aboriginal people of the Roper River region had gained themselves a reputation for fierce and concerted resistance to the European invasion of their lands (e.g. Willshire 1896: 7; Waters 1913:101-4). The abandonment of most of the cattle stations in the region in the 1890s was attributed to their efforts. Any possibility, however, of Aboriginal people being able to preserve their traditional cultural integrity was drastically ended at the turn of the

Kriol - the creation of a new language

201

century when the London-based Eastern and African Cold-storage Company acquired massive tracts of unleased or abandoned land to carve out a pastoral empire from the Roper River north into Arnhem Land (Bauer 1964: 157). Purchasing all the major stocked and viable cattle stations along the western Roper River, they began moving cattle eastward. The Company had no intention of allowing Aboriginal resistance to hinder this huge project. Determined to exterminate them, they employed gangs of up to 14 men to hunt all inhabitants of the region and shoot them on sight. With the police and other authorities turning a blind eye, the hunting gangs of the Company staged an unprecedented, systematic campaign of extermination against the Roper River people. They almost succeeded. This near annihilation of the Aboriginal people of the region produced the first factor necessary for the genesis of a Creole; sudden and drastic social change and the accompanying severe disruption of normal language transmission. The second requirement for the genesis of a creole is a new community. Challenged by the plight of Aboriginal people, the Anglican Church was determined to establish a mission, and choose a site on the Roper River itself (Cole 1968). The mission commenced in 1908, and was perceived as a haven of refuge by the scattered people of the region. By 1909 some 200 Aboriginal people gathered there. They were the remnants of the Mara, Wandarang, Alawa, Ngalakan and Ngandi people together with the easternmost Mangarayi people and the southernmost members of the Rembarrnga and Nunggubuyu. As Barnabas Roberts, an Alawa man who came to the Mission as a young boy, once said: (personal communication to John Sandefur), 'If the missionaries hadn't come, my tribe would have been all shot down'. The eight groups spoke separate and distinct languages. As is typical of Aboriginal people, the adults were multilingual. Although they had not lived permanently in such close proximity before, in their traditional lives they had met regularly for ceremonial and other purposes. Over the course of a lifetime, these people had normally become fluent speakers of each other's languages. The children, however, were not yet multilingual. Approximately 70 children attended school at the mission. There they were forced into contact with other children whose languages they had not yet had time to learn. They were the new community and they needed a primary language. Whereas their parents could communicate with other adults by speaking Alawa or Mara or whatever, the children could not. What they had in common was the English pidgin used between Aboriginal and European people (NTPE) and the English they were hearing in school.l With this limited input, it was this younger generation who, in the course of their lifetime, created the Creole, manipulating the lexical resources available to them and drawing on linguistic universals to create a

202

JOHN W. HARRIS

language which catered for all their communicative needs. Ralph Dinggul writes of the linguistic situation in Stones in Bilingual Situations, a set of readers produced for use in Kriol/English bilingual programs (n.d.: 1): La Ngukurr melabat garrim eitbala langgus. Wen naja traib wandim tok la dis traib, jei tok mijalb garrim Kriol. Jad impotan langgus im Kriol. Olabat gan sabi bla wanim olabat toktok. [At Ngukurr we have eight languages. When another tribe wants to talk with this tribe, they communicate with one another using Kriol. The important language is Kriol. They can understand what they are talking about.] A similar creolisation process occurred elsewhere in later years in such places as the Kimberley cattle stations and around the World War II army camps in the top end of the Northern Territory. 2 These Creoles have spread and merged into one widely spoken language in use from far western Queensland, through parts of the Northern Territory and into the north of Western Australia. The language is now called Kriol. It is, however, important to distinguish between creolisation in the past and the spread of Kriol today. Whereas many thousands of people speak Kriol as their primary language, thousands of others from surrounding areas speak it as a second language. Among some of these groups there is a tendency for Kriol to assume near-primary language status, particularly among older children. This is very evident, for example, at Numbulwar. There is, however, an encouraging reverse trend for young adults to use less Kriol and to use Nunggubuyu for adult communication. In the years before serious attention was given to Creoles, such languages were usually dismissed as marginal, inadequate and improper (see Muhlhausler, chapter 9, this volume). Kriol itelf has been called 'English perverted', 'ridiculous gibberish' and 'lingual bastardisation' (see e.g. Turner 1966: 202, Strehlow 1947: xix, and Baker 1966: 316). Now, however, that Kriol has been objectively studied and described, it is seen to be a Creole in the tradition of all the great regional Creoles of the world. Although children had a vital role in its beginnings, it is now the language of adults, a fully-fledged language, the vehicle of communication of a large and increasing number of Australians. In common with many creole-speaking people throughout the world, Kriol speakers were themselves often convinced of its inferiority. This is, however, changing rapidly. A growing sense of community identity and political action are becoming linked with Kriol speech in the minds of many young people. Bilingual education programmes in Kriol and English have been particularly important, especially the outstandingly successful programme at Barunga school. In this context, it would be difficult to exaggerate the critical importance of the Kriol Bible translation programme. An increasingly indigenous

Kriol - the creation of a new language 203 Christian church is a significant institution in the larger Kriol-speaking communities. A critical component of their growing pride in their distinctiveness and unique identity is the possession of the Bible in Kriol. Thefirstedition, containing selected books, was published in 1985. A larger volume containing 30 per cent of the Bible appeared in 1988 and the whole Bible will be completed within the next decade. Notes 1. Langford-Smith (1935: 136) says that he was unable to communicate with a 16-year-old girl who suffered from leprosy and lived with her parents and was never part of the school and dormitory peer group. Speaking neither English nor Pidgin English, she knew only her home language. 2. An important additional factor in these creolisation sites was that existing Kriol speakers were often part of these new extended speech communities. This meant that there was already a Creole model and it is difficult to determine the extent to which new creolisation took place or Kriol was simply adopted with local dialect differences developing.

13 A sketch of the structure of Kriol JOHN SANDEFUR

13.1 Introduction This chapter focuses on the main distinguishing structural properties of Kriol, with some mention of the variation that exists between dialects (see Sandefur 1986 and Sandefur and Harris 1986). Other brief general descriptions can be found in Graber (1986b), Sharpe (1985) and Steffensen (1977). Sociohistorical aspects of the language are discussed by Harris (this volume).1 13.2 Phonological structure The phonological structure of Kriol is very complex, due both to the origin of the language and to the effects of continued contact with English, the socially dominant language from which it was in part derived (for details of the orthographic structure of Kriol, see Sandefur 1984a). The phonological system consists of a continuum of sounds which extend from what is referred to in the eastern dialects as hebi - 'heavy' Kriol to lait - 'light' Kriol with a span in between being prapa - 'proper' Kriol (see Sandefur 1982a, 1982b, 1985b, 1986). The extreme heavy end of the continuum reveals a subsystem that is virtually identical with that of traditional Aboriginal languages. There are, for example, no affricates, no fricatives, no contrastive voicing with stops, no consonant clusters within a syllable, but five points of articulation for stops and nasals. The extreme light subsystem, in contrast, includes virtually all the contrasts which occur in English. A significant difference between the two subsystems, however, is that while the heavy subsystem 'eliminates' all the non-Aboriginal contrasts of English, the light subsystem does not eliminate the nonEnglish contrasts of traditional languages. The phonemes of Kriol are shown in figure 13.1.

A sketch of the structure of Kriol 205 Consonants stops

t d

p

b

rt rd

tj j

affricates

tj

fricatives

j sh

f

b nasals laterals rhotics semi-consonants

High Mid

m

th th

s s

n 1 rr

w

Front i e

Low

rn rl r r

Simple vowels Central

k g

h

s ny ly

ng

y

Back

u 0

a Diphthongs

ai ei oi au ou

low mid mid low mid

central to high front front to high front back to high front central to high back back to high back

Figure 13.1 Phonemes of Kriol In the formative stage of Kriol, when Aborigines borrowed words from English, they reproduced the words in Kriol with the traditional language sounds. As they remained in contact with English, however, they began restoring some of the original English sounds to the borrowed words. As this restoration has taken place, the older pronunciations for the most part have been supplemented with new pronunciations rather than being replaced. Thus, a word in Kriol today may have several pronunciations along the continuum (e.g. balijiman-blijiman-plisman - 'police'). As one moves along the continuum, sound changes generally follow an implicational pattern consisting of several stages rather than yielding to the English sound in one step (see Sandefur 1979). Some words, however, move from heavy to light in one step (e.g. heavy brog - 'frog' moves directly to light/rag). More often than not, however, the transition takes several steps, for example, heavy mawuj - 'mouth' becomes mawus before becoming maus andfinallymouth. Words composed of sounds which are common to both subsystems remain constant throughout the continuum, for example, mani - 'money' is mani regardless of position on the continuum. With the exception of the extreme heavy and light variations of some words, most Kriol speakers control virtually all pronunciations in their active everyday speech. No Kriol speaker employs a consistently light

206

JOHN SANDEFUR

pronunciation. There are, however, some Kriol speakers who tend to have consistently heavy pronunciation. These are mostly mother-tongue speakers of a traditional language who speak Kriol as a second language and who speak no English. With few exceptions, every stream of Kriol speech will contain some words with heavy 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 for a given word. 13.3 Morphological structure

Most of the morphological structure of Kriol is associated with the verb. The Kriol verb consists of a stem and three orders of suffixes, all of which mark either transitivity or aspect (see Hudson 1983a, 1983b; and Sandefur 1979). There are two suffixes in the first order, a transitive marker and a progressive aspect marker. Although the two morphemes cannot co-occur, it is possible to have a transitive verb with a progressive aspect, in which case the progressive aspect marker takes precedence, with transitivity being indicated by the context or the semantics of the verb stem. The form of the transitive marker is -Vm. There is a tendency for the vowel to harmonise with the last vowel of the verb stem (e.g. kuk-um 'cook', kil-im - 'hit', tjak-am - 'throw'), although in many cases, it is unstressed, and neutralises to an indistinct central vowel. Transitive verbs are normally marked, but the suffix can be omitted from verbs which have an intrinsically transitive stem, provided the object is explicitly stated and the resultant stem is an acceptable phonological shape, for example, Aibin rid-im det buk. - Aibin rid del buk. - 'I read the book'. When thisfirstorder suffix is omitted, second and third order suffixes are also omitted, for example, Imbin binij-im-ap ola daga.-Imbin binij ola daga. - 'He consumed all the food.' The first order progressive aspect suffix -in has two forms, -in and -ing. These variants are conditioned by rhythm and style rather than by vowel harmony like the otherfirst-ordersuffix variants. In the western dialects the progressive aspect suffix can alternate between first and second order positions when it occurs in intransitive verbs derived from English phrasal verbs, for example Jing-in-at-jing-at-in - 'singing out, calling'. Progressive forms are more common amongst young people than older people. The first-order progressive aspect suffix is related to the third-order iterative suffix -bat. There is an overlap of meaning (continuous or durational aspect), and an interweaving of distribution and co-occurrence. Both -in and -bat can be suffixed to the same verb, with a low central vowel optionally inserted between the two to avoid the resultant alveolar-bilabial

A sketch of the structure of Kriol

207

consonant cluster {e.g. silip-in-bat-silip-in-a-bat - 'sleeping'). The iterative meaning of-bat is more common than the durative meaning and can refer to either a repetition of action or to a plurality of participants. Iterative and durative aspect can also be indicated by reduplication of the verb, with extreme iteration or duration indicated by multiple reduplication to the order of six to eight times, usually accompanied by a rise in pitch which is sustained throughout the reduplication (see also Steffensen 1979a). Extreme duration can also be indicated by lengthening a vowel (usually the final vowel) of the verb and simultaneously raising the pitch and sustaining the high pitch on the lengthened vowel. There are ten suffixes in the second order, all of which are derived from English prepositions and are based on a spatial dimension. Four of them also include aspectual meaning: (i) -ap refers primarily to action performed in the vertical dimension or to spatial motion towards a physical goal, for example, Imbin galim-ap la top. 'He climbed to the top.', and in its aspectual meaning refers to an action carried out to its fullest extent, for example, Deibin binij-im-ap olabat mani. - 'They spent all of their money.' (ii) -bek primarily refers to an entity returning towards another entity or place, for example, Imbin go-bek la kemp. - 'He went back home.', and in its aspectual meaning refers to retaliation or reciprocation involving two entities with the assumption of a previous action that invoked the reciprocation, for example, Aibin owum-bek im. - 'I gave him a gift in return (for the gift he gave me).' (iii) -dan primarily refers to action performed in a downward direction, for example, Imbin juk-dan. - 'He ducked down.', and in its aspectual meaning refers to an event which causes the cessation of one state while simultaneously causing a change to a different state, for example, Det modiga bin breikdan. - 'The vehicle stopped working.' (iv) -at in its aspectual meaning refers to an action that is performed until a goal or a change of place or state is reached, for example, Melabat bin kipgon raidap melabat bin kam-at langa kemp. - 'We continued until we arrived home.' The morphology not associated with the verb is limited to adjectives and some adverbs. Adjectives normally occur with a suffix {-wan, -bala and less frequently -baga) that nominalises the adjective, thus enabling the adjective to stand in a nominal position in certain syntactic constructions, for example, Aibin lukjigi-wan sneik.-Aibin luk jigi-wan. - 'I saw a poisonous snake.' Some verbs can also be nominalised by these suffixes, for example, stilimbat - 'stealing', stilimbat-bala - 'thief, although their occurrence is not very common. There are also several suffixes that can be added to a noun, adjective or

208

JOHN SANDEFUR

verb to form adverbs. These include, for example, -said to express location, for example, rib-said - 'beside (a person)', -wei to express manner, for example, brabli-wei - 'correctly, sincerely, and -taim to express temporal relations, for example, kambek-taim - 'when we come back'. 13.4 Syntactic structure The two central syntactic relationships of Kriol, subject and object, are identified by word order (see Sandefur 1979, but especially Hudson 1983a). The subject is always overtly marked in the clause and is identified by its position as the closest nominal or pronominal element preceding the verb, for example, Det olgaman silip. - T h e woman is asleep.' The object follows a transitive verb, for example, Det olgaman kukumbat daga. - 'The woman is cooking food.', although with ditransitive verbs it can be separated from the verb by another argument. Unlike the subject, the object does not have to be overtly marked in the clause, for example, Imbin kukumbat. - 'She cooked (the food).' An exception to the post-positioning of the object is its movement to the beginning of the clause for topicalisation, for example, Ola daga, deibin binijimap. - 'They ate all the food'. Most other syntactic relationships in Kriol are handled by the five prepositions. Blanga indicates a benefactive/purposive relation, for example, Aibin gibit im mani blanga daga, en imbin baiyim blanga melabat. - 'I gave him some money for food, and he bought some for us.' It has three other forms, blanganda (which tends to be archaic), bla (which is in free variation with blanga), and ba (which is not very widespread and tends to be used mainly by young children). Langa indicates a locative relation, for example, Imbin bogi langa riba. - 'He swam in the river.' It has four other forms, langanda (which tends to be archaic), la (in free variation with langa) and nanga and na (used in some eastern cattle station dialects). This preposition is also used to indicate the syntactic relation of the traditional category of indirect object, for example, Imbin gibit ola enijing langa mi. 'He gave all the things to me.' Brom indicates an ablative relation, for example, Olabat bin kaman brom deya. - 'They came from there.' It has two other forms, the heavier form burrum (used in the longer established dialects) and the lighter form from (which tends to be the norm in the western dialects). Garram indicates an associative relation, for example, Melabat kaan go garram yumob. - 'We cannot go with you.' It has two other forms, garra (in free variation with garram) and gat (restricted in its function and used mainly in the western dialects). It can occur in both verbal and verbless clauses. Fo indicates a purposive relation, for example, Deibin hambagam mi fo daga. - 'They pestered me for food.' It is used mainly in the western dialects, and in most constructions blanga can be used in its place, for example, Deibin hambagam mi blanga daga.

A sketch of the structure of Kriol

209

In addition to their syntactic function of relating the noun phrase to the verb, three of the prepositions {blanga, fo and garram) can function within a prepositional phrase modifying the referent of the phrase. When used adnominally, the benefactive/purposive and purposive prepositional phrases indicate a possessive relationship, for example, Det dog blanga mi bin binij. - 'That dog of mine died.'. While the preposition normally precedes the noun in the prepositional phrase, in the western dialects the preposition in certain contexts can be postposed, for example, Jonfo dedi bin kaman. - 'John's father came.' The purposive prepositional phrase is not very common in the eastern dialects. The associative prepositional phrase indicates a descriptive meaning that could be glossed as 'having' or 'comitative', for example, Det men garram ola blad bin go la hospil. - 'The man who was bleeding went to the clinic' Verbless clauses are used in Kriol to describe states which are perceived as existing at the time of speaking. If reference is made to a state that existed before the time of speaking, the completive verbal auxiliary bin can be used without a verbal element, although a verbal element is normally included. If the reference is to a potential or future state (indicated by the verbal auxiliary garra), then a verbal element is obligatory and the verbless clause becomes a verbal clause. Verbless clauses in Kriol consist of a subject and a predicate and can be divided into nine types (after Hudson 1983a) according to the composition of these two constituents:

NP + ADJECTIVE, e.g. Dis ti prapa swit. - 'This tea is really NP + NOMINALISED ADJECTIVE, e.g. Dis daga brabli gudwan. ~

sweet.' 'This food is very good.' This type of clause is much more common than the preceding one. NP + NP, e.g. Det blanbo gudwan bif - 'The fruit bat is good meat.' NP-f-LOCATIVE (with the locative being a locative phrase, ablative phrase or intrinsically locative lexeme), e.g. Det haus la ai pleis. 'The house is on a rise.', Melabat brom Ropa. - 'We are from Roper.', Olabat tharrei. - 'They are over there.' N P + [BLANGA + NP], e.g. Det bout blanga olabat dedi. - 'That boat belongs to their father.' NP + [GARRAM + NP], e.g. Dis bifgarram blendifet. - 'This meat is (nice and) fatty.' NP + [da + NP], e.g. Dis da king. - 'This is the best one.' The particle da is probably derived through the reanalysis of the English article and is heard mainly in the speech of young people in the western dialects. PRONOUN + [GARRAM + NP], e.g. Dei garram bigmob tetul deya. 'There are lots of turtles there.' The pronoun is restricted to third-

210

JOHN SANDEFUR

person plural in the eastern dialects and third-person plural or singular in the western dialects. PRONOUN + TEMPORAL, e.g. Im dinataim. - 'It is noon.' The pronoun is restricted to third-person singular, and the temporal is usually identified by the suffix -taim. When topicalisation is applied to verbless clauses, the phrase to be topicalised is moved to the beginning of the clause and the subject identified by a pronoun following the topic and preceding the predicate, for example, Ola biginini, olabat bulap. - 'The children are all full.' The subject can also be moved to the end of the clause for lesser focus, in which case a subject pronoun is placed preceding the predicate to maintain the obligatory subject-predicate order, for example, Im sofwan, det bif. - 'The meat is tender.' When tense or aspect is added to a verbless clause, in most cases the clause has to be transformed into a verbal clause (see also Graber 1986a). Verbs used for this purpose include the copula bi, git, go - 'become'; the intransitive verbs sidan - 'sit', jandap - 'stand', stop - 'stay', leidan - 'lay'; and the transitive verbs abum - 'have, possess' and gedim - 'obtain, acquire', for example, Im bos. - 'He is boss.' and Im garra sidan bos. - 'He will be boss.'; Im kwait. - 'It is quiet.' and Imbin stap kwait. - 'It is quiet.' Most of these verbs differ in regard to the verbless clause types in which they may occur. Stap is widely used in the western dialects, whereas sidan is used in the eastern dialects. The copula bi in the western dialects is almost obligatory with the future tense (especially in the speech of children) and optional with the past tense, whereas in the eastern dialects it is optional with the future tense and rarely occurs with the past tense. 13.5

Lexical structure2

When lexemes are taken from English into Kriol, especially heavy Kriol, the neutralisation of phonemes allows for a great deal of potential homophony. Much of this potential homophony is avoided, however, for several reasons: many English words have never been taken into Kriol, in many cases the concept is expressed by an alternative word that presents no potential homophony, and in other cases a grammatical device (such as a transitive suffix) avoids homophony, for example, bak- 'to buck' is limited in transitivity to avoid homophony with bakam - 'to fuck'. As indicated in 13.2, words borrowed from English have normally undergone significant pronunciation changes. In addition, many borrowed words undergo grammatical changes as well as significant semantic changes. Almost all Kriol verbs are derived from English ones, which are not marked for transitivity. The borrowed verb is treated as intransitive

A sketch of the structure of Kriol 211 and a transitive form is derived by suffixation of the transitive marker - Vm, for example, blidin - 'to be bleeding', blidin- im - 'to cause to bleed'. Many Kriol verbs have both an intransitive and a transitive form, although there are some intrinsically transitive verbs which require obligatory marking and have no intransitive counterpart, for example, totjim - 'to shine a light on', as well as some intrinsically intransitive verbs that have no transitive counterpart, for example, jandap - 'to stand'. Some Kriol verbs are derived from English words other than verbs, for example, nidulum - 'give an injection to' from 'needle, which has involved a shift in the word class. More significant socially, however, is the semantic shift that often takes place in a borrowed word. It is significant in that it frequently causes crosscultural miscommunication because the typical English speaker, not realising that a shift has taken place, interprets Kriol speakers on the basis of English etymology rather than Kriol semantics, for example, in the context of writing, libum from the English 'leave it' means to 'delete'. Often when the meanings of English borrowed words have been extended (for an example of lexical expansion, see Steffensen 1979b), their references have been made to parallel the semantic range of traditional Aboriginal language words, for example, kukwan from English 'cooked one' means 'ripe' as well as 'cooked'. Of course, not all Kriol words are derived from English. Many are borrowed from traditional Aboriginal languages. Most such words are restricted to the dialect of Kriol spoken in the geographical area of the particular traditional language, although some such words have gained wider currency, for example, munanga - 'person of European descent', and a few are used throughout virtually all of the Kriol language area, for example, lambarra - 'father-in-law'). As with words borrowed from English, those borrowed from traditional languages may undergo grammatical changes. For example, the main meaning-bearing morpheme from a traditional language compound verb may take on full verbal status in Kriol and be inflected for aspect, for example, dilaj (from Walmajarri 'persistently demand one's own way') as in Wan boib in dilajbat langa is mamifob mani. -'One boy pestered his mother for money.' Unlike words borrowed from English, however, words borrowed from traditional languages seldom undergo pronunciation changes or significant semantic shifts. 13.6

Conclusion

I have highlighed some of the distinguishing structural properties of Kriol and commented briefly on some of the variation that exists between dialects. Other aspects of the structure of Kriol have not been mentioned

212

JOHN SANDEFUR

but are discussed in Graber (1987a), Hudson (1983a) and Sandefur (1979). Much more research and analysis needs to be done, especially in the area of discourse structure (see Graber 1987b). Notes 1. I am indebted to Joyce Hudson from whose research I have drawn heavily in the preparation of this paper, especially her 1983a volume. 2. The only published Kriol dictionary presently available is Sandefur (1979). Noreen Pym has undertaken further compilation of the Kriol dictionary and a computer printout is available from the Summer Institute of Linguistics, Darwin.

Part III Transplanted languages other than English

14 Overview of Immigrant' or community languages MICHAEL CLYNE

14.1

Demographic situation

There are some 75 to 100 'immigrant' or 'community languages' in use in Australia, the exact number depending on one's criteria for the term 'language'. The term 'community language' is employed to denote the socalled 'immigrant' languages used within the Australian community to emphasise the legitimacy of their continuing existence. These languages have been part of the Australian scene since early in the history of white settlement. However, there has been discontinuity between the earlier and contemporary use of community languages. The 1986 Census data provides the most detailed large-scale corpus of information on home language use. Of the Australian population 13.6 per cent reported employing a language other than English at home. (However, there are many who now use a community language not in their own but in their parents' homes). This rose to 18.5 per cent in Victoria, and was lowest in Queensland (5.9 per cent) and Tasmania (3.6 per cent), the states least affected by postwar migration. The community languages most widely used in the home were Italian (415,700 speakers), Greek (277,500), Serbo-Croatian (140,500), Chinese (139,100), Arabic (199,200), and German (111,300). Other languages with over 50,000 regular users were Dutch, French, Polish, and Spanish. With the large number of South-east Asian refugees since 1976, Vietnamese has also now joined this group of languages. A language which narrowly misses this category is Maltese, and Macedonian may have been excluded due to underclaiming. Australia has no single 'majority minority language' of the significance of Spanish in the United States. Precise numbers of speakers are difficult to estimate since different speakers designate their language in different ways. Serbo-Croatian/Croatian/Serbian/Yugoslav is a case in point (see Doucet, this volume). There are concentrations of different languages in different states. Most

216

MICHAEL CLYNE

of the largest languages (Italian, Greek, German, Maltese, Polish, Dutch, Serbo-Croatian) are employed by more speakers in Melbourne than in Sydney, as is Macedonian, one of the languages most widely spoken by state-school children in Melbourne. Arabic, Spanish, Portuguese, and Chinese are far more highly represented in Sydney. Only in Western Australia do the numbers of Italian speakers exceed the total numbers of speakers of the next four languages (in this case, German, Serbo-Croatian, Dutch and Greek), and they outnumber their total by 30 per cent. Compared with other states, Western Australia has few Greek speakers but relatively large Serbo-Croatian and Macedonian speech communities as well as over-proportionate numbers of Portuguese, French and Burmese speakers. From table 14.1 it can be seen that the rank ordering of the languages varies between states, but the most widely used community language in all states except Tasmania is Italian. In Tasmania it is Dutch, in the Australian Capital Territory it is Serbo-Croatian, and in the Northern Territory, Greek, which is exceeded only by all the Aboriginal languages taken together. Greek comes second in Victoria, New South Wales and South Australia, while German has second place in Queensland, Chinese in Western Australia, and Italian in the Australian Capital Territory and Tasmania. Within cities, there are concentrations of particular languages in certain areas. As these are largely due to arrival times and resettlement patterns, there are unique incidences of ethnolinguistic co-settlement (Turkish/ Vietnamese,Maltese/Serbo-Croatian,Macedonian/Italian,Khmer/Greek). Table 14.1 also shows the number of users of languages other than English by state. 14.2 Assimilation and language shift

The mass immigration programme launched in 1947 was embedded in an assimilation policy and general attitudes not conducive to linguistic pluralism (see Martin 1978; Clyne 1982: 17-21). Although community languages were maintained in the home domain and in those domains determined by the ethnic communities themselves (e.g. ethnic social life, religion, ethnic press, ethnic part-time schools, certain aspects of the transactional domain - continental shops, cafes, guest houses), there was little public support for community languages. Interpreting and translating services were inadequate and there were no employment officers, social workers and teachers with any knowledge of languages spoken by the migrant groups. Few local libraries stocked books in community languages. Ethnic newspapers had to publish sections in English (until 1956). Television established itself as a monolingual English medium, and radio stations were severely restricted in the amount of transmission permitted in

Overview of 'immigrant' or community languages

217

Table 14.1. Home-use of languages other than English: total numbers for first 14 community languages and Aboriginal languages (1986 Census data) NSW

Vic.

Qld

WA

SA

Tas.

113,203 178,097 26,115 43,590 48,179 1,590 Italian 96,652 128,562 10,491 5,864 28,622 1,508 Greek Serbo-Croatian ") Croatian |\ 57,957 49,496 7,160 11,662 12,529 697 Serbian J1

Yugoslav Chinese Arabic German Spanish Polish Vietnamese Dutch Maltese French Macedonian Aboriginal languages

66,974 88,475 35,324 42,783 21,362 25,506 15,852 24,086 20,256 15,131 806

40,443 24,515 32,665 18,556 22,920 21,680 18,278 30,535 14,803 24,090 374

ACT Aust.

NT

1,040 3,951 415,765 2,747 3,026 277,472 248 4,506 140,575

9,696 11,389 5,833 820 1,965 204 69 1,590 1,446 2,415 992 14,526 8,206 14,190 1,999 228 4,770 3,231 2,114 391 185 4,889 7,295 9,044 1,499 427 193 5,560 5,046 6,158 447 9,820 8,272 6,473 2,013 632 1,839 29 2,076 26 547 6,807 5,375 2,994 804 7 861 3 373 4,801 15 23,138 5,017 8,420 2,970

1,980 139,100 473 119,187 2,654 111,276 1,888 73,961 1,444 68,638 1,286 65,856 1,026 62,181 283 59,506 1,204 52,790 344 45,610 50 40,790

languages other than English. Legislation against bilingual education enacted during World War I was still in force. French, the dominant modern language taught in Australian schools at the time, was not spoken by many migrants, and there were discriminatory measures practised against 'native speakers' in examinations in other languages, for example, German and Italian (Clyne 1982). The period of assimilationist policies and attitudes is reflected in high language shift between ethnolinguistic groups. Of the large groups whose statistics were processed by the Australian Bureau of Statistics in 1976, Greek-Australians display the greatest ethnolinguistic vitality in the sense of Giles, Bourhis and Taylor (1977) (see Tamis, this volume), and Dutch-Australians the least (see Pauwels, chapter 15, this volume). Among groups known to be successful in language maintenance include Latvianand Ukrainian-Australians. There is also variation between states in the rate of language shift, probably as a result of state policies, with Victoria and South Australia exhibiting lower percentages for most languages, and Queensland, Western Australia and Tasmania higher ones (Clyne 1988b). There is also variation in domains such as education, social welfare and libraries which were to some extent satisfied by multicultural policies. This can be seen in table 14.2. These include the gradual introduction of community languages into primary schools, the provision of Saturday

218

MICHAEL CLYNE

Table 14.2. Language shift (per cent, to the nearest whole number) 1986 Census Birthplace of speaker or of ethnic parents Germany

(1) (2) (2m)

Greece

0) (2) (2m)

Italy

0)

Malta

(2) (2m) (1) (2) (2m)

Netherlands

0)

Poland Yugoslavia

(2) (2m) (1) (1)

NSW Vic.

Qld

WA

SA

Tas.

NT

ACT Aust.

41 73 84 5 9 36 11 30 66 26 62 84 49 85 99 17 8

42 83 90 9 20 56 16 44 80 44 73 94 49 86 99 20 18

44 72 87 8 22 54 11 36 76 55 73 92 45 82 99 40 12

38 58 85 4 8 40 10 28 71 35 70 89 47 85 99 13 14

43 76 87 5 5 60 23 43 89 74 84 94 47 84 99 17 25

49 75 88 5 8 41 19 43 81 68 89 91 60 84 100 30 33

44 70 84 6 12 50 16 38 80 47 74 93 51 86 — 15 10

39 70 85 4 7 41 9 24 69 22 53 86 49 87 99 16 9

41 73 85 4 9 41 11 29 71 26 59 87 48 85 99 16 10

(1) Overseas-born (2) Australian-born of two parents of same country of birth. (2m) Australian-born of one 'ethnic', one Australian-British or Irish-born parent.

classes within State Education Departments in languages not widely taught in day schools, the recognition of all (or most) community languages as matriculation subjects limited bilingual education programmes in some areas, government subsidies for part-time ethnic schools, a nationwide Telephone Interpreter Service, ethnic and multilingual radio stations, multicultural television, and a reflection of the reading public in local library holdings, as well as some public and company notices in appropriate languages other than English (see also Ozolins, this volume). It has not yet been possible to examine the effects of new policies and attitudes on language maintenance and shift. The change in the wording of the census language question from regular use (1976) to home use (1986) renders comparisons between 1979 and 1986 impracticable. Similarly, it will be difficult to assess the consequence of ad hoc government cuts to funding in multicultural education and allied areas experienced in the 1986 Federal Budget. As this has put an extra onus on state funds and policies, it is likely that differences between the states may be widened (see 14.2). Comparisons between the language shift rates of different ethnolinguistic groupings suggest both clear-cut and ambivalent factors which can lead to either language maintenance or language shift, depending on their combination with other factors (Kloss 1966). Among the clear-cut factors

Overview of 'immigrant' or community languages

219

in the Australian context are exogamy, pre-migration language maintenance experience, cultural similarity to the dominant group, and the role of language as a core value in the group's cultural values system (Smolicz 1979, 1981). Among the ambivalent factors are: 1 Numerical strength of the group: large groups can afford extensive language maintenance efforts, but may be 'swamped' more easily by the dominant language and culture than are closely-knit smaller groups. 2 Class, education and prior knowledge of English: working-class background, low educational levels and little or no English can lead to either a reliance on thefirstlanguage, or an all-out effort to master English to the exclusion of the first language, to achieve social mobility. 3 The political situation in the homeland: refugees either reject the language of the oppressive regime, which is their ownfirstlanguage, or they feel responsible for its projection from the regime in the isolation of their exile. 4 Ethnic religious dominations: these promote the first language if services and social functions are conducted in that language, but promote language shift if English is adopted as their language as is the case, for instance, in the Reformed Church of Australia. 5 Linguistic similarity of the ethnic language to English: in cases such as German and Dutch, this has probably contributed to language shift, because it has eased the English acquisition process, but this does not apply to Italian. On the other hand, there has been a large language shift from Maltese, a Semitic language unrelated to English (Clyne 1982: 35). In all these instances, other factors codetermine language maintenance and language shift. A factor still requiring some cross-linguistic study is the role of dialect vs. standard language. Bettoni (1981 and this volume) demonstrates a latergeneration gravitation towards Standard Italian. Among migrants from diglossic situations, the High language tends to be supplemented by English. However, Pauwels's (1986a) study of first language maintenance in Swabians and Limburgers yielded no substantial quantitative differences in language maintenance between standard and non-standard speakers in each of the groups although there was some variation between the groups. Limburgers displayed dialect loyalty but used English to other Dutch speakers whilst Swabians, who have a continuum of varieties available to them, employed more standard German variety to non-Swabian German speakers. On the whole, language shift decreases with age in the first generation and increases with age in the second. Female speakers tend to shift slightly less than males. This applies especially to Northern European ethnic groups in thefirstgeneration and to Southern European ones in the second

220

MICHAEL CLYNE

generation. The shift from the mother's language is generally less than from the father's language if the other parent is a monolingual English speaker. However, in the second generation of mixed (non-anglo) parentage, the differential between language shift from the two languages depends on relative ethnolinguistic vitality (Clyne 1982: ch. 2). The strength of language maintenance institutions such as ethnic schools and churches helps determine the vitality of the Greek language in Australia, as Tamis (this volume) shows. There is anecdotal evidence that some second-generation Australians are reverting to their ethnic languages in order to pass them on to their children. Partners in some mixed marriages are raising their children bilingually on the principle of each parent speaking 'their' language to the children, and some part-time ethnic schools are now having to focus on second language teaching and on the activation of passive language skills because of the background of their pupils. 14.3 Multiculturalism and the changing functions of community languages

The widespread acceptance of multiculturalism and, by implication, multilingualism, has been associated with a new national identity in Australia (Ozolins 1985; Clyne 1986a). Kouzmin (1988) has shown that in the second generation of Russian-speaking groups, the symbolic function of the language replaces the communicative one, which is dominant in the first generation. This is probably a universal of community languages in the Australian context. Within the policy of multiculturalism, community languages have, to some degree, been 'mainstreamed' (demarginalised). For instance, they are taught in some schools to pupils of all ethnic backgrounds, and there are some bilingual programs based on a modified version of the Canadian partial immersion model (e.g. Clyne 1986b; Swain and Lapkin 1982) offered to children with or without a home background in the language. 'Mainstream' churches which had previously regarded themselves as 100 per cent Anglo-Saxon (Anglican) or which had established congregations using languages other than English originally as a transitional measure (Uniting/ex-Presbyterian-Methodist; Roman Catholic) are now starting new ethnic parishes on a permanent footing and proudly designating themselves as multicultural and multilingual (see e.g. Archbishop's Commission 1985; Houston 1986). Some communities have reconstituted or are reconstituting themselves, and language is proving an important factor. The multiplicity of Italian regional clubs are partly coordinated and partly supplemented by supraregional Italian ethnic organisations, employing lingua popolare or Standard Italian as a medium. The various groupings and vintages that make up

Overview of 'immigrant' or community languages

221

the Australian-German speech community (including prewar refugees from National Socialism, Palestinian-Germans (Templars), postwar Germans, Austrians, Swiss, and ethnic Germans from eastern and central Europe), all of whom have a different relationship to the German language, cannot be regarded as a single cohesive ethnic group (Clyne 1988b). However, for certain purposes such as welfare matters, the language brings them together as a supra-ethnic German-language community. Kouzmin (1988) points out that recent Russian-Jewish refugees employ Russian as a means of communication and cultural continuity, but not as a language of symbolic identification. This function is assumed by Hebrew. Similarly, an important factor detracting from language loyalty among German-speaking Jews in Australia (in addition to the events of the Third Reich and the 'anti-foreign language' and especially anti-German attitudes prevailing in Australia during World War II) has been the reconstitution of the Australian-Jewish community into a more cohesive and independent ethnic group and the redefinition of Jewish ethnicity in the Diaspora since the Six-Day War. Zionism has become a core value to Liberals and Orthodox alike, and an attachment to but not necessarily the use of Hebrew is its symbolic marker. The Liberal Jewish community, based largely on German- speakers and ther descendants, have now not only adopted Zionism but also started day schools, following their opposition to full-time Jewish schools. Klarberg (1983) has shown how the seven Jewish day schools, each linked with some ideological direction within Judaism are all teaching Hebrew but how they have all developed different language policies towards this instruction. By the third generation, in many families any loyalty to the language of the grandparents is transferred to the symbolic 'ethnic language' of the Jewish (as opposed to German-speaking western European or Yiddish-speaking eastern European) 'ethnic community' which has, to a large extent, overcome the old east-west divisions. Fishman et al. (1985) have shown that identification with a language and positive attitudes towards it do not necessarily coincide with its maintenance. The same applies to the later generations of a number of speech communities in Australia. 14.4 Relations with metropolitan countries

One of the important resources for continuity and renewal that Australian community languages can draw on is contact with their origin homelands and/or (other) countries in which their language has official or national language status. Nevertheless, this is not a prerequisite fof language maintenance. The very opposition to the political regime in the homeland or to the country which has deprived it of its nationhood provides a strong incentive to refugee groups to keep up their language. They see it as their

222

MICHAEL CLYNE

responsibility to protect the language and consider unacceptable language changes that have taken place in the country of origin. Some refugee groups, for example, Hungarians and Poles, were divided as to how much support they should accept from the homeland, for example, in resource materials for ethnic schools. In other communities, for example, Latvian and Lithuanian, there has long been a political 'softening' in the second generation, which has absorbed Australian 'multicultural' ideology and sees that contact with, and material from the homeland (including films) as essential for language maintenance. There are some languages that have lost their heartland and/or which have never enjoyed 'majority status'. These include Armenian, Romany and Yiddish. (Perth in Western Australia can boost both a Romany parttime ethnic school and 1-hour weekly radio programmes in the language). Yiddish and, to a lesser extent, Armenian, derive some support in the form of educational, literary and film materials from the much larger speech communities in the United States. Economic migrants, for example, Italians, Dutch, and post-1970 arrivals from Yugoslavia, are able to take advantage of cheaper and faster travel. Their children can now visit their parents' former homelands regularly, and attend school there, and elderly (non-English-speaking) relatives come to Australia on visits. Books and videos can be readily imported, and benefit the processes of linguistic renewal and language maintenance in adults and children. Representatives of foreign governments, such as the Goethe Institut and the Istituto Italiano di Cultura, provide high culture activities in the ethnic language which in refugee groups have to be organised internally. Some refugee groups, for example, Hungarian and Estonian, developed links with similar communities in other English- and nonEnglish-speaking countries of migration through their worldwide umbrella organisations. These offered mutual support in curriculum development for ethnic schools, provided information based on homeland events, and facilitated the organisations of international cultural festivals. At such meetings, the ethnic language is used out of necessity as well as volition, being both object and lingua franca. A similar situation has developed recently for some economic migrant groups such as the Italians, who are now well represented in various European countries as well as in Australia and North and South America and who are organised internationally through bodies such as FILEF (Federation of Italian Migrant Workers and Their Families). Speakers of pluricentric languages (i.e. whose languages are used by several nations, such as Dutch, German, French, Spanish) all of whom recognise different norms, and identify differently with the language, present something of a problem. 'Mincmties', such as Austrians, French-Canadians, Flemings, and Chileans are not always happy to be

Overview of 'immigrant' or community languages

223

dominated by the much larger and vocal 'majority' speakers of their language. The 'minorities' are, in a sense, multiple group members. Some people in such situations eventually discard one of their layers of identity, for example, 'German-speaking', 'Yugoslav', 'Lebanese', and keep another, for example, 'Swiss', 'Croatian', 'Arab'. The tension between the different subgroups within a speech community is sometimes based on different migration vintages, for example, Czechs who came in 1949 or 1968, Hungarians who came in 1949 or 1956-7, Serbs who migrated in 1949 or the 1970s, or different religions, for example, Muslim or Christian Lebanese. Among eastern Europeans, the more recent vintage was exposed to 'Socialist' changes in the language which alienate the earlier group. Among other things, this adds further potential for creativity. Within some language groups, particular institutions or individuals such as newspaper editors, clergy and ethnic school teachers, are entrusted with the task of keeping the level of language high, and especially of'keeping the language pure'. There does not appear to be any community language spoken in Australia that is free of English influence. The nature and extent of this influence depends on the functions of the languages as well as on the networks in which they are employed. Despite valiant attempts to declare an 'Australitalian' the Italian language of Australia (Andreoni 1980; Leoni 1981), there is little uniformity in any community language today (Campbell 1979, 1980). Variation occurs between individuals, families and groups according to their experiences and everyday communication needs through English and the other language as well as for idiosyncratic reasons (Clyne 1982: Ch. 4). There are particular sociolinguistic, communicative and stylistic functions of 'mixed varieties' which are characteristic of the speech situation in Australia as opposed to the country of origin, for example, citing key-words that had been used in English in bilingual interaction; as a more informal or familiar way of expressing oneself; within specific networks of which 'mixed variety' is a symbol; to emphasise the Australian context; for humorous reasons; to facilitate the use of synonyms. Links with language developments in the homeland ensure that fossilisation and large-scale loss and restructuring do not occur at the group level. In some cases, speech communities in Australia are more reluctant to accept transference from English than is the homeland society. In homogeneous rural settlements which retained a community language over several generations but almost completely lost contact with the country of origin, such as the German-speaking community of Tarrington (formerly Hochkirch, Western Victoria), the following characteristics have stood out in contrast to the greater variation andflexibilityin the German of recent migrants: stabilised lexical and semantic transference, a regular pattern of integrating lexical transfers from English, including the stabilised assign-

224

MICHAEL CLYNE

ment of genders to transferred nouns, and stabilised use of archaisms and sortie neologisms (Clyne 1968). This stabilised situation does not apply to the same extent in the more recent North Queensland Italian settlement of Giru (Campbell 1979,1980; Bettoni 1981), whose transference patterns are similar to those in the urban centre of Townsville. Nevertheless, there are common tendencies within each language, especially in the rules for integrating English transfers and in syntactic (and, to some degree, morphological) transference. The former rules are often also operative in the country of origin; the latter may sometimes be seen as more advanced indicators of a drift inherent in the language (Sapir 1949). The common tendencies within a language can facilitate crosslinguistic comparisons which take into account both sociolinguistic factors and the structural properties of the languages (see Clyne, chapter 16, this volume). 14.5 The development of research methodology

Australian language contact research has drawn heavily on overseas (especially American) models and methods. Much of the initial descriptive linguistic work on immigrant languages (e.g. Clyne 1967, 1972a on German, 1977b on Dutch; Kaminskas 1972 on Spanish; Kouzmin 1973 on Russian; Bettoni 1981 on Italian) was influenced by Weinreich (1953) and Haugen (1956). Early methods of interviewing employed were both direct (eliciting) and indirect (leading questions). Terminology was adapted and there was a gradual extension of the framework to incorporate pragmatic and psycholinguistic perspectives. With the publication of Fishman (1966), sociology of language perspectives took a more prominent place in Australian research conducted by both sociologists (e.g. Harvey 1974; Smolicz and Harris 1976; Smolicz 1979), and linguists (e.g. Clyne 1976b, 198 la, 1982; Garner 1985; Pauwels 1980,1986a). This included the study of language attitudes promoted by the evolution of the social psychology of language (e.g. Pauwels 1985a, and this volume; Rado 1976; Lewis, Rado and Forster 1983; Giles, Rosenthal and Young 1985; Callan, Gallois and Forbes 1985; Gallois and Callan 1981, 1985, 1986). Such studies have focused largely on questions of language and identity. Linguistic conclusions have been verified by psycholinguistic experimentation (e.g. Clyne 1972b) as well as sociolinguistic data collection (e.g. Clyne 1980) and tests of awareness of deviation, perception of'Australianness' and attitudes to transfer phenomena (Pauwels, this volume). More recently, Bettoni (1985, 1986, and this volume) has adopted the language attrition paradigm. At present, linguistic, sociolinguistic, psycholinguistic, sociology of language, social psychological and demographic methodologies are all employed and combined in research on immigrant languages in Australia.

Overview of 'immigrant' or community languages

225

The languages that have been described in detail are those mentioned above, together with Serbo-Croatian (Doucet 1984 and this volume) and Greek (Tamis 1985a and this volume). In addition, there have been studies of aspects of French (Ludwig-Wyder 1982), Hungarian (Endrody 1971), Japanese (Masumi-So 1984), Macedonian (Hill 1980), and Polish (Sussex 1980). Work is in progress on Finnish in Australia (Hannele Hentula, Turku) and on the Sinhala grammar of Sinhalese children in Melbourne (Dipemala de Silva). Oksaar (1975) has been conducting a longitudinal project on Estonian in the United States, Sweden and Australia, but to my knowledge, no findings have been published on the Australian data. The similarity offindingson lexicon across languages and between Australian and overseas studies has led scholars away from producing more and more word lists of lexical transfers and their motivation, to a greater interest in grammatical and sociolinguistic questions. Among issues that have occupied the attention of Australian researchers are the effects of exogamy on language shift (Pauwels 1985a) and bilingual language acquisition based on the principle of 'one person, one language (Saunders 1982; Dopke 1988), and first-language reversion and secondlanguage attrition in elderly bilinguals (Clyne 1977c, 1981b). 14.6

Gaps in research

The demographic situation and ethnolinguistic mix invite a somewhat more intensive study of the broader communicative complex, including pragmatic transference, cross-cultural communicative conflict and breakdown, and interactional networks - the kinds of areas developed by Neustupny (1978) and research by him and his students in the context of Japanese-Australian relations (Neustupny 1985a, 1985b; Bolitho 1976; Marriott 1984). This is potentially a very important area, especially considering the dependency of access to power on successful mastery of English communication rules (Pride 1985). There has been a dearth of studies on communication in ethnic and mainstream Australian institutions (see Clyne and Manton 1978 on meeting routines). The questions of multiple group membership, identity and its linguistic manifestations (see 14.4) have also been under-researched, despite their significance in Australia. They have been examined in relation to speakers from areas with strong regional identity (e.g. Pauwels 1986a). McNamara is examining Israelis in Australia, their relation to the Jewish community and their maintenance of Hebrew. As will be apparent from 14.5, only a small number of the 75-100 languages used regularly in Australia have so far been described. They are nearly all Indo-European and do not offer a wide range of typological characteristics (see Greenberg 1978; Lehmann 1977). It would be most

226

MICHAEL CLYNE

desirable for morphological and syntactic transference and change and the grammatical integration of lexical transfers to be examined in a variety of languages, including Albanian, Arabic, Chinese, Khmer, Maltese, Tamil, Tongan, Turkish, and Vietnamese. This would facilitate conclusions on linguistic and sociolinguistic universals of language contact. On the other hand, no systematic studies have been made of the transference patterns, code-switching and language shift of migrants from other Englishspeaking countries, for example, Americans, people from different parts of English, Irish, Scots, and Sri Lanka (Ceylon) burghers. Claims concerning first-language reversion and second-language attrition in the elderly have been made almost entirely on the strength of cross-sectional data. The next step would be to conduct a large-scale longitudinal study over 15-20 years to examine development in individuals. This would be costly and difficult to organise, but virtually the only way for progress to be made in this area. At the same time it would be highly desirable for research to be conducted on the activation of passive skills in a language and on the reactivation of lost skills. Since first and second language attrition and passivisation are such common phenomena in Australia, such research would be very feasible here. It could help answer a number of questions. What is the relation between language acquisition, attrition and reactivisation? Is it easier to acquire active skills in a second language or to activate passive skills in that language? How does such activation occur? Due to the rapidity of language shift and the limited number of community languages so far described in Australia, it has not yet been possible to show the role of each generation in language change although some advances have been made for Italian, German and Dutch (Bettoni 1985, and this volume; Bettoni and Gibbons 1988; Clyne 1972a, 1985, and this volume; Pauwels, chapter 13, this volume). 14.7

Prospects

As Fishman (1985: 515) has stated, predictions on language maintenance and shift are impossible since we do not have knowledge of future immigration movements or 'ethnic revivals'. However, it is possible to identify some major issues that are likely to affect language maintenance and shift. They are: 1. Immigration intake policies and settlement patterns - number, source countries, and language and education background of intake, and settlement and dispersion; 2. Policies towards multiculturalism - the type and extent of institutional

Overview of 'immigrant' or community languages

227

support for multiculturalism in general, and National Language Policy in particular; 3. New technologies - subscriber television, home video equipment and personal computers will make it at least theoretically possible for Australians to enjoy equal access in the media domain to community languages and to English. In effect, the size and initiative of the speech community in Australia and the socioeconomic status of the individual, family or group will determine that some languages will benefit from the new technologies more than others. Undoubtedly, continuing immigration of speakers of languages already represented provides an impetus for language maintenance but it is not essential. (Ukrainian and Latvian, whose speakers are nearly all products of the 1949-51 'Displaced Persons' wave, have been maintained far more than German, which should have benefited from continuing migration, and Czech, which has been brought to Australia by three vintages of migrants). On present indications (Commonwealth 1986), immigration rates are likely to be stepped up again to avoid a population decline, and intake policies will, on the whole, probably benefit language communities already represented in this country. Many languages are already in an advanced stage of language shift. It may be that large-scale immigration of new non-English-speaking migrants from the country of origin could reverse this. Even if it is too late to reverse the shift to English in some communities, many other languages, especially those of the more recent migrant groups, will ensure the continuation of the multilingual nature of Australian society.

15 Dutch in Australia: perceptions of and attitudes towards transference and other language contact phenomena ANNE PAUWELS

15.1 Introduction1 The original preoccupation with the description of the linguistic performance of immigrant bilinguals in their two languages in an attempt to discover universal linguistic features of language contact has often meant that questions regarding a bilingual's perception and awareness of contact phenomena have not received a lot of attention. Usually studies of language contact in the Australian context have done little more than mention the fact that bilingual informants, when questioned about their speech patterns, realised that some mixing of languages occurred (see e.g. Clyne 1967; Pauwels 1980; Bettoni 1981). The same studies often also recorded that this awareness of mixing one's languages triggered off a whole array of feelings in the bilinguals ranging from indifference or resignation to annoyance or even acute embarrassment. Recently, with the upsurge of studies of language attitudes, researchers have become interested in establishing to what extent bilinguals' perceptions of 'language mixing'2 have an impact on the process of language maintenance or language shift (see e.g. Chana and Romaine's 1984 study of the attitudes to mixing of Panjabi English bilinguals in Britain). That is, does the perceived linguistic quality - pure or mixed - of the ethnic language affect its status, usefulness and so on with respect to English? I will explore this issue in relation to Dutch in Australia. 15.2 Previous studies A notable exception to the earlier ad hoc approach to the bilingual's awareness of contact phenomena is an experiment conducted by Clyne (1975). He tested the verbal and non-verbal reactions of ten German-English bilingual immigrants towards a German language sample

Dutch in Australia

229

containing a great number of integrated lexical and semantic transfers. The experiment revealed that there was some correlation between subjects' own use of transfers and their linguistic awareness of them: the two subjects whose own speech contained most instances of transference showed a lack of awareness of the transfers. On the whole, however, the transfers were detected and, for some informants, they became the source of great amusement because there was ambiguity or a double meaning across languages. The most thorough account to date of attitudes towards language mixing in the Australian immigrant context is the study by Bettoni and Gibbons (1988). By means of the matched guise technique the authors examined the attitudes of Italo-Australians towards a range of language varieties, including two Italian dialects, Regional Italian, Light (English-Italian) Mixture, Heavy (English-Italian) Mixture, and English, that characterise the Italo-Australian speech community (see Bettoni, this volume). The Light Mixture is more typical of first-generation Italian migrants, while the Heavy Mixture is linked to the second generation. However, they found that their informants did not associate positive feelings with either Mixture. Bettoni and Gibbons summarised the attitude towards the Heavy Mixture as one of hostility. The Light Mixture fared slightly better as it attracted mainly intermediate scores. It seems to be regarded as the unmarked or neutral choice in the Italo-Australian speech community: it is neither liked nor disliked. Positive feelings were associated with English and Regional Italian. The authors conclude that in the ItaloAustralian community puristic attitudes toward language prevail. Positive ratings are given to the 'pure' varieties, whereas English interference in Italian is mostly judged negatively. This may accelerate the shift towards English. 15.3 Attitudes towards Dutch as expressed by Dutch immigrants in Australia

Of the larger non-English-speaking migrant groups in Australia, the Dutch rank fourth in number. The Dutch record the highest rate of shift to English in both the first and second generation. The 1976 Census revealed that an estimated 43.5 per cent offirstgeneration Dutch migrants no longer claimed to use Dutch regularly (compared with 3 per cent for Greeks, 6.3 per cent for Italians, 27.8 per cent for Germans and 20.2 per cent for Poles). Language shift in the second generation is reported to be as high.as 80.8 per cent for the Dutch group compared with 10.0 per cent for Greeks, 18.5 per cent for Italians and 62.8 per cent for Germans (see also Clyne, chapter 14, this volume). Previous studies on Dutch language maintenance (e.g. Harvey 1974; Clyne 1977b; Smolicz and Lean 1979; Pauwels 1980) have all

230

ANNE PAUWELS

shown that the Dutch do not seem to be greatly concerned with the maintenance of their language in Australia. Although linguistic similarity between English and Dutch as well as cultural similarity between the two groups have provided some explanation for the high rate of shift in the Dutch community, the apparent lack of interest in maintenance has most often been linked to the fact that the Dutch language is not a core value in their cultural value system. Smolicz (1981) uses the term 'core value' to refer to those specific aspects of a culture which are seen as central to a group's life and values. My research has attempted to establish whether the apparent lack of interest in the fate of Dutch is matched by feelings of indifference towards the quality of Dutch spoken/used in Australia.3 15.4 The Dutch language in Australia

As is the case with other immigrant languages spoken in Australia, there is no stable or uniform Australian variety of Dutch (Clyne 1982). However, there are a number of phenomena that mark the speech of manyfirstand/or second generation Dutch-English bilinguals, for example, a range of lexical, semantic, morphological, syntactic, phonological and other transfers, instances of triggering (i.e. internally conditioned switching, see Clyne 1967) and marginal passages, that is, unlimited switching between two languages in both directions at the grammatical and lexical levels, promoted by a large number of compromise forms, and a phonic pattern from the first language, (see Clyne 1982: 108; and Hasselmo 1961). Although there are some inter-generational differences in the use of certain types of contact phenomena, these may be less clear-cut than those separating the generations in other language groups. The high degree of structural similarity between English and Dutch has meant that certain phenomena associated with the speech of second-generation bilinguals of other language groups, for example, the German group, are already present in the speech of first-generation Dutch bilinguals (see Clyne, chapter 16, this volume). On the basis of speech samples of manyfirst-and second-generation Dutch (Clyne 1977b; Pauwels 1980) it can be claimed, however, that marginal passages are a typical phenomenon of (elderly) first-generation speakers who have little competence in English. With regard to transference patterns, it can be cautiously stated that semantic transfers of any kind are more likely to mark the speech of firstgeneration than of second-generation bilinguals. Phonological transference (Australian accent in Dutch), certain types of morphological transference, for example, grammatical gender and plural markers, and syntactic transference, for example, generalised use of 'hebben'- 'to have' as auxiliary for past compound tenses, are more likely to occur in the

Dutch in Australia 231 speech of second-generation bilinguals. It is not surprising that the characteristic linguistic features of the second generation coincide with many so-called developmental errors (or interlanguage features) made by second-language learners of Dutch, as the former learn Dutch in similar conditions, where they are dominant and fluent in a language other than Dutch. Lexical transfers and syntactic transference involving word order (SVO instead of SOV) are present in the speech of both generations. 15.5 The investigation 75.5.7 Aims I will discuss one aspect of a current project on the Dutch language in Australia in which I examine in depth questions of awareness of language contact phenomena, that is, transference and code-switching, at sentence and discourse level, as well as the attitudes towards such phenomena. The central focus of this investigation is the degree of awareness of different types of transference (as presented in isolated sentences) among a group of first generation Dutch immigrants and the exploration of their attitudes towards these phenomena. I try to establish whether the type of transference/contact phenomenon has an impact on the degree of awareness expressed by bilinguals as well as whether attitudes towards language contact phenomena differ according to the type of phenomenon involved. I also examine to what extent awareness of and attitudes towards transference are interrelated. 75.5.2 The test sentences Eight sentences each illustrating a particular type of transference were presented to a group of 25first-generationDutch informants to test their awareness of transference phenomena and to register their attitudes towards them. In each case the transfer is underlined. Below the glosses I have noted the nature of the transfer and given the 'correct' Dutch version. 'Correct' in this context means conformity with the Standard Dutch norm propagated by grammars/dictionaries. Of course, in some instances the deviation could be corrected in various ways. Each of these were seen as correct. 1. Er staan paarden in de paddok. 'There are horses in the paddock'. Lexical transfer: integrated for gender, graphemic integration. Correct: Er staan paarden in de wei. 2. We groeien tomaten in de tuin. 'We grow tomatoes in the garden.'

232

3.

4.

5.

6.

7.

8.

ANNE PAUWELS

Semantic transfer of the verb 'to grow'. Correct: We kweken tomaten in de tuin. Ik heb in Frankrijk geleefd voor over een jaar. 'I have been living in France for over a year.' Loan translation involving a semantic transfer {to live) and syntactic transference (use of present perfect instead of present simple+order of adverbial phrases) Correct: Ik woon al longer\meer dan een jaar in Frankrijk. Het heeft zo hard geregend dat de straten staan helemaal blank. 'It has rained so much that the streets are flooded.' Syntactic transference: SVO instead of SOV. Correct: Het heeft zo hard geregend dat de straten helemaal blank staan. Ik heb het meer interessante boek gekozen. 'I chose the more interesting book.' Morphological transference: periphrastic expression of the comparative, possibly under the influence of English. Correct: Ik heb het interessantere boek gekozen.4 Hij heeft van zijn fiets gevallen. 'He has fallen off his bike.' Syntactic transference: the generalised use of 'hebben' - 'to have' as auxiliary for past compound tenses. Correct: Hij is van zijn fiets gevallen. Dat is de langste kanaal. 'This is the longest channel.' Morphological transference: wrong gender. Correct: Dat is het langste kanaal. Ik loopte snel naar de winkel. 'I quickly ran to the shop.' Morphological transference: 'lopen'- 'to run' is a strong verb, treated here as weak. Correct: Ik Hep snel naar de winkel.

15.5.3

The tests

(a) Awareness test I was interested not only in the question of the extent to which the informants were aware of the deviation in the presented sentences but also whether they perceived the deviations as the result of Dutclb-English language contact in the Australian context. The first aspect 'degree of awareness', called the A-test, was rated on a 3-point scale: 1 = Aware of the deviation and capable of providing a correct version. 2 = Aware of the deviation but unable to provide a correct version. 3=Unaware of the deviation.

Dutch in Australia

233

With regard to the second aspect, perception of Australianness of the deviations, called the B-test, informants were requested to tick one of the following four alternatives: I associate this way of speaking/writing mainly with 1 = Adult native speakers in the Netherlands. 2 = Adult native speakers of Dutch in Australia. 3 = Non-native speakers (i.e. learners) of Dutch. 4= Children of Dutch immigrants in Australia. (b) Attitudinal test Informants were asked to rate the acceptability of the sentences along a 5-point scale: •1

=

2= 3= 4= 5=

approve of the use of this sentence in spoken and written Dutch, accept the use of this sentence in spoken and written Dutch, approve of the use of this sentence in spoken Dutch only, accept the use of this sentence in spoken Dutch only, disapprove of the use of this sentence in spoken and written Dutch.

15.5.4 Administering of the tests Because these tasks formed part of an interview in which informants were asked to provide other information as well as perform other tasks, the administering of the attitudinal test did not follow directly after the awareness test. This 'spreading' of the tests helped considerably in avoiding 'automatic' replies by the informants (usually caused by being exposed to the same task for too long) and any influence of the tests on each other. 15.5.5 The informants On the basis of their knowledge and use of Dutch, 25 first-generation bilinguals who had taken part in other aspects of the project were selected for this investigation. All informants had to be native speakers of Standard Dutch and had to use Dutch regularly (i.e. daily) in Australia. They had to have been exposed to education in the Dutch language for at least eight years in the Netherlands. The minimum period of residence in Australia acceptable was set at ten years (in 1985). These restrictions were imposed on the group of informants to avoid too much heterogeneity within a relatively small sample. There were 13 women and 12 men in the sample. Five informants were under 35, ten informants were between 36 and 45, six informants were between 46 and 55 and four were older than 55. A control group of Dutch nationals (i.e. Dutch native speakers residing permanently in the Netherlands) was introduced to ascertain the effect of

234

ANNE PAUWELS

living in a language contact situation on awareness of the deviations in the sentences. This group closely resembled the Australian-Dutch group with respect to sex and age distribution, level of education and variety of Dutch used. 15.6

Results

15.6.1 Awareness of deviations in the sentences

Both the Dutch-Australians and the Dutch nationals were subjected to the first awareness task, that is, the A-test to detect deviation from the Standard Dutch norm in each of the sentences and, if possible, to provide a correct version. The Dutch nationals had no problem whatsoever in perceiving deviations in the sentences and were very competent in providing a correct version. All sentences were recognised as deviations from the Standard Dutch norm by all informants and, with the exception of item 5 (periphrastic expression of the comparative), all were able to provide a correct version for each of the sentences. The Dutch-Australian group was far less uniform in its response patterns. Although the degree of awareness that the sentences represented deviations from the Standard Dutch norm was still high in this group (83.5 per cent), some informants had problems in providing a correct version (22.5 per cent). This seems to suggest that an immigrant's competence to recognise certain types of deviations is not greatly affected by living outside the original speech community for a prolonged period of time, but that a lack of (regular) contact with that community may lead to difficulties in providing a correct version. Item 7 (wrong gender) and item 4 (word order), item 6 (generalised use of 'hebben' - 'to have') and item 8 (verb morphology) were recognised and corrected by 22,23,24 and 25 informants respectively. Sentences that were recognised as containing a deviation but caused problems with regard to providing the correct version included the lexical transfer (item 1), the semantic transfer (item 2) and the loan translation (item 3): 36 per cent could not provide a correct version for item 1, 40 per cent were unable to supply a correct version for item 2, and 56 per cent could not do so for the loan translation. The semantic transfer and the periphrastic expression of the comparative recorded the highest scores for lack of awareness. In each case ten informants were unaware of any deviation. 15.6.2 Perception of the 'Australianness' of the deviation

The task measuring the informants' perception of the 'Australianness' of the deviations was administered slightly differently to the two groups.

Dutch in Australia

235

Because it was assumed that the informants from the Netherlands would be unfamiliar with the specific situation of Dutch in Australia, the alternatives 'adult native speaker of Dutch in Australia, and 'children of Dutch immigrants in Australia' were replaced by 'adult native speakers of Dutch residing in an English-speaking country' and 'children of native Dutch speakers residing in an English-speaking country'. With the exception of item 5, most Dutch informants attributed the presented features to be characteristic of the speech of adult non-native speakers. That adult native speakers of Dutch residing abroad were not readily implied as likely producers of the 'errors' is probably partly due to the lack of (regular) contact that exists between Dutch nationals and Dutch expatriates and partly due to the lack of experience Dutch nationals have with language contact. The only item seen as attributable to the Dutch native speakers abroad is the lexical transfer of the first sentence. Dutch-Australians seemed to recognise lexical transfer as typical of the Dutch spoken by adult native speakers in Australia. Although about half of the Dutch-Australian group saw the semantic transfer (item 2) and the loan translation (item 3) as characteristic of Dutch speakers in Australia, a substantial number of informants (10 and 7 respectively) related the items to native speakers in the Netherlands. So-called developmental or language learner errors, for example, wrong gender, generalised use of 'hebben'- 'to have' and wrong verb morphology, tended to be seen as characteristic of either non-native speakers of Dutch or children of Dutch immigrants in Australia. 15.6.3 Attitudinal test In passing judgement on the sentences, informants were asked to consider the sentences as those of adult native speakers of Dutch. Although the 5point scale system allowed for some differentiation with regard to type of language use, that is, speech or writing, further refinements, such as specification of register, could not be incorporated in this scale system. The two groups also reacted quite differently with regard to the attitudinal test. Again the Dutch group was much more uniform in its answers than were the Dutch-Australians. Only two of thefivealternatives were used by the Dutch group, namely: (4) I accept the use of the sentence in spoken Dutch only. (5) I disapprove of the use of this sentence in spoken and written Dutch. Item 4 (SVO instead of SOV) and item 5 (comparative) were seen by the majority of Dutch nationals (22) as acceptable in spoken language. The acceptance of item 4 may be due to interpreting word order deviations as the result of wrong sentence planning, a feature typical of the spoken

236

ANNE PAUWELS

language. The acceptance of item 5 probably reflects innovative tendencies in the Dutch language in the Netherlands, possibly under the influence of English. Although the overall impression gained from the answers given by the Dutch-Australians was also one of disapproval of the items (70 per cent of all answers had a (5) score), quite a few items were given a (1) or a (2) score by some informants. The judging of items 6, 7, 8 was very uniform: they were almost unanimously rejected (i.e. disapproved of in spoken and written Dutch) by 25, 24 and 25 informants respectively. Informants were divided over the acceptability of such phenomena as the lexical transfer (item 1), the semantic transfer (item 2) and the loan translation (item 3). These are the ratings given to these items. lexical transfer (1) 2 informants (2) 6 informants (3) 3 informants (4) 2 informants (5) 12 informants

semantic transfer (1) 5 informants (2) 6 informants (3) 0 informants (4) 3 informants (5) 11 informants

loan translation (1) 5 informants (2) 6 informants (3) 0 informants (4) 0 informants (5) 14 informants

The item with the lowest disapproval score, that is, a score of 5, was the periphrastic expression of the comparative: only ten informants disapproved of it in speech and writing. 15.7 Links between awareness and attitudes

In order to establish whether there was an association between the degree of awareness and the approval rating of the phenomena, tables examining the contingency between the tests were produced.5 Contingency table 15.1 analysing links between the first awareness test (the A-test) and the attitudinal test revealed a strong association. Informants who were aware of, and able to correct the deviation, that is, vertical score 1, were more likely to disapprove of the use of the item in speech and writing, that is, horizontal score 5. Lack of awareness of the deviation, vertical score 3, tended to be associated with complete approval of the item, horizontal score 1. If informants were aware of the deviation but could not correct it, then they were slightly less strong in their approval rating: they accepted rather than approved of the item in speech and writing. Of greater interest to this investigation is the possible link between the informants' attitudes towards the items/deviations and their perception of the likely source of the deviation, that is, the B-test. Although the association between the B-test and the attitudinal test is not of the same magnitude as that between the latter and the A-test, it is still appreciable. This is shown in table 15.2.

Dutch in Australia Table 15.1. Contingency between the A-test and the Attitudinal test Attitudes

1 A w a 2 r e n 3 e s s All

1

2

3

4

5

All

4 11.59

8 14.64

3 2.44

3 7.93

104 85.40

122 122.0

5 4.28

9 5.40

0 0.90

4 2.92

27 31.50

45 45.00

10 3.13

7 3.96

1 0.66

6 2.14

9 23.10

33 33.00

19 19.00

24 24.00

4 4.00

13 13.00

140 140.00

200 200.00

Chi-square= 52.764 with D.F. = 8 Tabled X .01, 8 = 20.1 p 20.1 Explanation of symbols: Horizontal scores (1), (2), (3), (4), (5), cf. attitudinal scores. Vertical scores (1), (2), (-3), cf. scores for A-test.

Table 15.2. Contingency between B-test and attitudinal test Australianness

1 A 2 t t i

j

t 4 u d e j

All

1

2

3

4

All

8 3.70 4 4.68 0 0.78 7 2.54 20 27.30

8 6.27 15 7.92 2 1.32 3 4.29 38 46.20

2 3.70 2 4.68 2 0.78 1 2.54 32 27.30

1 5.32 3 6.72 0 1.12 2 3.64 50 39.20

19 19.00 24 24.00 4 4.00 13 13.00 140 140.00

39 39.00

66 66.00

39 39.00

56 56.00

200 200.00

Chi-square=41.042 with D.F.= 12 p 26.2

237

238

ANNE PAUWELS

If an item was seen as typical of a Dutch native speaker in the Netherlands, horizontal score 1, the item would be given full approval for use in speech and writing vertical score 1. Items that were associated with the children of Dutch immigrants in Australia, horizontal score 4, were disapproved of, vertical score 5. If an item was attributed to an adult native speaker of Dutch residing in Australia, its use was found acceptable in speech and writing, vertical score 4. It could be expected that items which carried full approval would be linked to native speakers of Dutch in the Netherlands and vice versa. A more interesting observation was that those items attributed to adult native speakers of Dutch living in Australia did not attract negative ratings. The lexical transfer, the semantic transfer and the loan translation which were recognised as being deviations (although not always corrected) were seen by their most likely producers, that is, the first-generation informants, as acceptable both in speech and writing. The items identified as typical of the second generation did not fare so well on the approval scale: they attracted strongly negative ratings. 15.8 Discussion The overall impression gained from the test results is that first-generation Dutch immigrants are, on the whole, aware of language contact phenomena characterising the speech of first-generation Dutch speakers in Australia. They do, however, disapprove strongly of the use of the contact phenomena which are linked to the second generation. Living in a contact situation seems to have some effect on the degree of awareness that the items/phenomena presented are deviations from the Standard Dutch norm. Dutch immigrants are less likely than Dutch nationals to be able to correct an item. On the basis of the data obtained here, it is not possible to pinpoint the cause of this difference. It seems most likely that the failure to provide the correct version in some cases may be due to difficulties of recall. Psycholinguistic testing could possibly clarify this finding. There is also a difference between the Dutch-Australians and the Dutch nationals in their attitudes. The latter tend not to differentiate between the different items in their judgements. This may be related to the fact that the items are seen as typical of non-native speakers of Dutch. Learners of Dutch are expected to make errors which need to be corrected and which are normally not condoned by native speakers. Passingjudgement on items ascribed to fellow native speakers is a far more delicate task (see the attitudes towards item 5 in the Dutch group). Dutch-Australians can be said to still hold rather puristic attitudes towards language contact phenomena: 70 per cent of all answers recorded disapproved of the

Dutch in Australia

239

deviations presented in the sentences. However, their judgements of approval/disapproval are more differentiated than those of Dutch nationals in relation to the type of phenomenon involved. Although the attitudinal test revealed that the lexical and semantic transfers and the loan translation attracted both approval and disapproval scores rather evenly, Table 15.2 indicated that these phenomena were seen as acceptable (in speech and writing) when recognised as coming from adult native speakers in Australia. Thus, there seems to be some evidence that the first-generation Dutch informants are not overly concerned that the Dutch language characteristic of their generation and situation deviates from that of native speakers in the Netherlands. Perhaps this attitude may be cautiously summed up as one of indifference towards the linguistic quality of their Dutch. This feeling of indifference is not, however, extended towards phenomena involving grammatical deviations from the Standard Dutch norm which were associated with the speech/language use of the second generation in Australia. There seems to be some similarity to the observations made by Bettoni and Gibbons (1988) that the overall attitude towards the speech characteristic of the second generation ItaloAustralians is one of hostility. In the case of the Dutch in this investigation, the overall attitude is one of disapproval. The exploratory character of and the extremely limited data base for this investigation do not allow a valid prognosis regarding the relationship of this issue to the process of shift observed in the Dutch community. However, some speculation at this stage is not misplaced. It cannot be claimed that the Dutch as language users per se are possibly less concerned with correctness in their language than other language users. Both Dutch nationals and Dutch immigrants hold rather puristic attitudes towards correctness. Nevertheless, the latter do not seem to worry too much that the Dutch used by the first generation in Australia no longer adheres to the Standard Dutch norm. This may be a linguistic reflection of their lack of concern with the maintenance of Dutch beyond the first generation. The strong feelings of disapproval towards the deviations linked to the second generation may also affect the language behaviour of the first generation towards the second generation. Because the latter are no longer regarded as 'native' speakers but as 'learners' of Dutch (on the basis of their errors), the first generation may prefer to switch to English to communicate with their offspring. Notes 1. I would like to thank Monash University for giving me financial support (Monash Special Grants) which enabled me to carry out this research, and Kathy Diamantopoulos for assistance with the data processing.

240

ANNE PAUWELS

2. I use the term 'language mixing' to refer to all types of linguistic phenomena resulting from language contact. 3. So far two papers have resulted from this project: one (Pauwels 1985a) deals with a comparison of the attitudes of Dutch-Australians and Dutch Nationals towards levels of correctness in Dutch, and the other (Pauwels 1986b) is a matched-guise study examining the attitudes of Dutch-Australians towards language mixing. 4. In Dutch the superlative could also be used when two items are compared with each other. 5. Contingency tables were produced for each individual item as well as for all items. Here, only the all-items tables will be analysed.

16 German and Dutch in Australia: structures and use MICHAEL CLYNE

16.1 Introduction The German and Dutch speech communities in Australia appear outwardly similar. Both are socioeconomically well established (Derrick, Pyne and Price 1976: 31), relatively large, and fairly assimilated. According to the 1976 Census (adjusted figures, Clyne 1982: 12), Australia then had 170,644 regular users of German (54,824 in Victoria) and 64,768 regular users of Dutch (20,606 in Victoria). The languages and cultures do not diverge markedly from those of the dominant Anglo-Australian group and, incidentally, resemble each other. In comparison with some other communities, such as the Greeks, the German-Dutch similarities hold true but in comparison between them, the German and Dutch speech communities show some marked differences. I shall be focusing here on two aspects: the structure of the German and Dutch languages as used by postwar immigrants and their children and patterns of language use and maintenance of the two languages. I will explore possible interrelations between structure and language use, taking into account migration history and community dynamics and demographic factors (notably settlement patterns). 16.2 Language maintenance and community dynamics Every study so far conducted in Australia on language maintenance and shift, whether small-scale and detailed or large-scale and superficial, has found that, of all the speech communities in Australia, the Dutch speakers have experienced the most rapid shift to the use of English only, both within the home and elsewhere (see, e.g. Harvey 1974; Clyne 1977b, 1982, and chapter 14, this volume; Pauwels 1980; Smolicz and Harris 1976; ABS 1976, 1983). While the Dutch in the Netherlands are one of the most

242

MICHAEL CLYNE

multilingual peoples in the world, immigrants from that country have, by their indifference to their first language, produced the most monolingual second generation of any non-British ethnic group in Australia. The Dutch, in contrast to the Germans for example, tend to speak English to their children (Clyne 1967, 1977b). In a survey of parental attitudes in different ethnic groups, Smolicz and Lean (1979: 20) have shown that Dutch-Australians disagree more with the learning of ethnic languages than any of the other groups, including 'British-Australians'. They have not founded ethnic Saturday schools to teach their language to future generations (with the exception of a small Saturday school in Sydney). They have not taken advantage of recent language policy initiatives, for example, in Victoria, to ensure that their language is introduced into state primary schools as a regular part of the curriculum by qualified supernumery teachers. On both counts they are unique among major ethnic groups. Due to the fragmented nature of Dutch society at the time of their migration, with compartmentalisation based on religious denomination, the Dutch are not a very cohesive ethnic community. Diglossia and the use of dialects may detract more from the maintenance of Dutch than from that of German. This problem has been studied for Limburgs and Swabian by Pauwels (1986a). While almost all the Dutch in Australia came to this country in the 1950s, German speakers have had a more diverse settlement history. A number of vintages (including, e.g. descendants of nineteenth century settlers in former closed rural enclaves; refugees from Nazism who came in the 1930s, the Templars, Swabian Pietists from Palestine, who were interned in Australia during World War II; postwar immigrants who have arrived at various times since 1951 from Germany, Austria and Switzerland, as well as ethnic Germans from central and eastern Europe) constitute what is a very heterogeneous speech community, but certainly not a single ethnic group (Clyne 1988c). The German speakers are the most dispersed of all immigrant speech communities within all metropolitan areas of Australia, with different German-speaking groups preferring different areas, as in the case of Jewish refugees, Templars, postwar German immigrants, ethnic Germans. Also, each 'national' group, for example, Germans, Austrians, and Swiss, has its own clubs (see Clyne 1981a), and some groups, such as the Templars, have their own infrastructure, promoting language and cultural maintenance. Many of the clubs are not attracting the support of the ethnic young, and membership of non-German speakers is causing language shift in the organisations. The German-medium churches, too, are having to adopt bilingual strategies to serve both young and old. There are some organisations, for example, welfare and lobby groups, serving all German speakers. The fact that German is both an international and an Australian

German and Dutch in Australia

243

Table 16.1. Language shift in the 1st and 2nd generations: German and Dutch (based on regular use of English only, ABS 1976)

German Dutch

1st generation

2nd generation (both parents from same country)

27.79% 43.55%

62.28% 80.79%

community language means that it is well represented in education, at the secondary and tertiary levels and more recently in the primary schools. There are also strong attempts in some cities to maintain the language inter-generationally through Saturday schools. Melbourne alone has nine German language part-time ethnic schools. Nevertheless, German is only moderately well maintained; though the contrast with Dutch is still marked. At present, a high degree of exogamy appears to limit the tenacity of both languages but it is still more likely for a non-German spouse to know or learn German than for a non-Dutch spouse, Dutch. The effects of new policies and attitudes and the patterns among the most recent immigrants have yet to be examined. (A project on this topic is currently being undertaken by Anne Pauwels.) Table 16.1 shows the contrast between the Dutch and German communities with regard to language shift. Because language shift is so high among the Dutch, the tendency towards first-language reversion in the elderly is felt all the more dramatically (Clyne 1977c). Also, in some subgroups there are 'bilingual networks' in which Dutch and English are used by the same people, with much stabilised transference of lexical items as well as code-switching. This is very prevalent within parishes of the Reformed Church which consider themselves to be religious rather than ethnic communities and conduct almost all their services in English, although almost all their members are of Dutch origin. There are few similar networks among German speakers, one being that of German and Austrian prewar refugees within the Sydney and Melbourne Jewish communities. Both German and Dutch language maintenance has benefited from cheaper air travel resulting in visits to the country of origin or from elderly relatives from abroad. 16.3 Features of language structure common to Dutch and German

My data on language structures are based on similar taped interviews conducted with matching groups of 200 German-speaking and 200 Dutchspeaking postwar immigrants in Victoria and their children. (One difference, however, is that the Dutch speakers were taped about seven years

244

MICHAEL CLYNE

after the German speakers and were consequently older. See Clyne 1967, 1977b.) Among the common features were the following: (1) Lexical transfers, i.e. the transference of lexemes (form and meaning) especially in the work domain, e.g. customer, office, college, period, milk, bar, shop, shoppen, supervisen, and in the social and recreational domain and that of Australian institutions and experience, e.g. beach, breakfast, drover, gum-tree; (German) television (Dutch televisie) watchen, relaxen, settlen. (2) Phonological integration of lexical transfers, e.g. English Building becomes [blldlrjk] in both Dutch and German English fence and sum become Dutch [fons] and [sem], respectively. English changed and front become German [tfe: njt] and [fRont], respectively. (3) Compromise forms between English and German/Dutch, i.e. forms representing a convergence between English and Dutch or German, e.g. [de: z] (from Dutch [de: z] and English [ 6i: z] - 'these'); ['di.alekt] (from Dutch [dia'tekt] and English ['daislekt]) - 'dialect'; [hai'drant] (from German [hy'drant] and English ['haidiant] - 'hydrant'; and [Ji: f] (from German [Ja:f] and English [JI:p] - 'sheep'). (4) Semantic transfers, i.e. the transference of the meaning of an English word to a homophonous or partly synonymous German or Dutch one, e.g. Dutch smal, German schmal- 'narrow' used for klein - 'small'; Dutch stil, German still - 'silent' used for Dutch nog, German noch - 'still'; Dutch weten, German wissen - 'to know facts' used for kennen - 'to know people'. 163.1 Syntactic transference The two types of syntactic transference in both the German and Dutch corpus are generalisation of SVO word order, for example, German: Und dann meine Mutter macht uns das Essen alles zurecht. (Standard German: Und dann macht meine . . .) 'And then my mother gets our meal ready'; and Dutch: Maar als wij praten in het Hollands, ze verstaan drommels goed. (Standard Dutch: Maar als wij in het Hollands praten, verstaan ze . . .) 'But when we speak in Dutch, they understand darned well.' Discontinuous constituents are also brought closer together, e.g. German: Er fahrt wieder zuruck zu Koln. (Standard German: Er fahrt wieder nach Koln zuruck.) 'He went back to Cologne.', and Dutch: Ze heeft gestudeerd op Hobart University. (Standard Dutch: Ze heeft aan de Universiteit Hobart gestudeerd.) 'She studied at Hobart University.' Interlingual and inter-generational differences in the distribution of the two types will be dealt with below.

German and Dutch in Australia 16.3.2

245

Code-switching

In both languages there are two main causes of code-switching. It is motivated by situational factors, such as domain, topic, interlocutor, and interaction type. Certain words of ambiguous language affiliation also trigger off a switch from one language to the other. The actual words vary from individual, but the main categories are: (1) Lexical transfers Dutch: Ze zijn gedeeltelijk [they were partly] waitress and the others are staff, and German: Im Augenblick les' ich eins, das handelt von einem alten [at the moment I'm reading one which deals with an old] secondhand dealer and his son. (2) Homophonous diamorphs, i.e. morphs in the two languages that sound similar. Dutch: Mijn idee is, dat het [my idea is that it is] is parks and pubs, and German: Das ist das Cafe [that is the cafe] near dem [the] Oriental Restaurant. (3) Proper nouns Dutch: Dat heb je [you have that] in Melbourne too, and German: Dieses Bild ist [this picture is] in Melbourne, Collins Street, is if! (4) Compromise forms Dutch: Die hadden allerlei contractors, die kwamen meest uit voor twee jaar en [they had all kinds of contractors who mostly came out for two years and], being the only Australian, well, hij kon geen Engels spreken, hij moest [he couldn't speak any English, he had to learn Dutch], and Das is [Is] [that is] taken round the coast. Triggering can occur in anticipation of, as well as subsequent to, the trigger-word, e.g. Dutch: Het is vlak, maar toch wat [it is flat, but still what]/yeah, / cant think of the word, hilly, and German: Ja, ich arbeit' im australischen Heer als Dozent in einer der Militarschulen [yes I work in the Australian army as a lecturer in one of the military schools] at Balcombe. Most of these phenomena are common to the community languages so far researched in Australia (Clyne 1982:93-115) and probably to all languages in contact with English (see Haugen 1956, 1973).

16.4 Differences in structure between Dutch and German Of somewhat greater interest are the differences between Dutch and German.

246

MICHAEL CLYNE

16.4.1 Gender Postwar German-speaking immigrants assign lexical transfers to the gender of a semantic equivalent or, to a lesser extent, to that of a morphemic equivalent, for example, das Breakfast [ party franks little frankfurts • saveloys V savs 4 cocktail saveloys cocktail savs

i

1

•

Nariel Creek* li

V ICTOR I A

WOevonport .

Map 20.1 Regional distribution of Cheerio ~ Frankfurt etc.

* Launceston

N E W TH

S O U T H

W A L E S

AUSTRALIA

Clare- .Mintaro

Darlington Po Coleambally U rayBri 9e

g^Str"aln¥bynT L

f

Parrakie

Lameroo Oaklands ^|D_ , Tocumwal Finley

T

^E.Henty Culcairn.olo ~i •Walla Walla Corowam .„ Jindera-(5-

Rutherglen Tr-AlburyDD|oD gWodonga

Map 20.2 South-East regional distribution of Dink ~ Dinky ~ Donkey

298

PAULINE BRYANT

New South Wales/Victoria/South Australia border area. Ideally, four people would have been interviewed in each place, one man and one woman under 40, and one of each over 40. In practice, in some places fewer than this number were interviewed, particularly in small country towns visited at weekends. On average, three people were interviewed per town in 55 towns. Most (73 per cent) of the informants met the length of residence criterion of having lived in their town from the time when they started until when they left school. Indeed, most had lived there all their lives. Thus, their usages are considered current. Most of the informants fell into two socioeconomic groups, shopkeepers and shop assistants, but a few were farmers, housewives, etc. Informants were interviewed individually, using the same photographs that had been used in the Canberra verification surveys, and the interviewerfilledin their answers on the same type of form. Map 20.2 displays the results for one item investigated in thefieldsurvey. The item refers to the practice among children of one giving another a ride on the bar of a bicycle. It can be used as both a noun ('Andrew is giving Matthew a dink') and a verb ('Andrew is dinking Matthew'). The hollow symbols represent the terms used in the second of the Canberra verification surveys, and the black symbols of the same shape represent the same terms found in the field survey. The Canberra verification survey covered the whole of Australia, but the results from that survey are shown only for the area covered in the field survey. Moulamein in southern New South Wales appears to be the southern limit of 'double', which is used in central and northern New South Wales and Queensland. The term 'dub', which is used in Temora, Cootamundra and Junee, is also used to the east and north-east of these towns in an area not covered in the field survey. In Victoria 'dink' is used almost without exception. It is also used in the South Australia town of Mt Gambier and in the southern part of New South Wales. Both these areas have close ties with Victoria, being originally settled from that State and having strong trade links with it. In South Australia the picture is more complex. 'Dinky', which is the principal term used east and south of Adelaide, is also used in southern Western Australia. Thus it is the distinctive term for the centralsouth and south-west of the country. 'Donkey' is used only in a confined area of South Australia, mainly to the north and west of Adelaide, with some overlap in the 'dinky' and 'donkey' areas. Along the Victoria/South Australia border north of Mt Gambier, there is no influence across the State border, in spite of the proximity of some of the towns, for example, Pinnaroo (South Australia) and Murrayville (Victoria). Bryant (1989b) contains more detailed results from the final survey. Thefieldsurvey results correlate well with the recollected usage from the Canberra-based surveys, though with the comparatively small numbers of informants for each item in each area no statistical test of correlation was

Regional usage in the lexicon of Australian English

299

attempted at that stage. Some out-of-area usages were given by the current speakers surveyed; for example, Victorian dink used by some South Australian speakers in areas with no links with Victoria. Discrepancies which in the Canberra-based surveys had been attributed to faulty recall now appeared to be at least partly genuine out-of-area usage which is apparently characteristic of Australian English. 20.5 The final survey

The verification surveys had confirmed the existence of the four regions suggested by the pilot survey, and had better defined their extent. The regions were now defined as the South-West (roughly the southern part of Western Australia), the South-Centre (roughly South Australia), the South-East (roughly Victoria, Tasmania and the southern part of New South Wales), and the North-East (roughly Queensland and the part of New South Wales not included in the South-East; see Bryant 1989b). The tasks of the final survey were to delineate the borders of the regions more precisely, and to do as comprehensive a coverage as possible of the rest of the country. From the 186 items tested in the verification surveys, 72 items went into the final survey. Of the remaining items, the variation in naming was mainly synonymy, for example,pillow case "pillow slip. These items had been reported as having regional variation, and so deserve further investigation at a later stage. Thefinalsurvey was conducted in two stages, as the verification surveys had been,firstin Canberra, then in thefield.It began with a coarse-grained coverage of the country, conducted in Canberra as before, but this time other venues around Canberra were also used, as well as a network system of contacts. Most of these informants gave recollected usage as in the Canberra verification surveys. However, two additional sources of informants were available in Canberra for the final survey. The first was the Australian Bicentennial Castrol World Rally in Canberra in March 1988 to which participants travelled from all over Australia. The second was the Visitors' Information Centre, which for many interstate visitors is theirfirststop in Canberra. At both these venues, speakers who were current users of their home-town language were interviewed. From all these sources, using both recollected and current usage, an adequate coverage of most of Australia except the States furthest from Canberra was achieved. Obviously areas nearest to Canberra provided the greatest number of visitors, so that New South Wales and Victoria, which in any case are the most populous States, were best represented. This was important for definition of the border areas of the South-East region. The trial surveys had shown that these were the most complex borders, so it is

SOUTH

AUSTRALIA NEW

r* * r!\:

• • •

Slide Slippery dip Slippery slide Limit of 'slide' -• Southern and eastern limits of 'slippery dip'

Map 20.3 Slide

SOUTH

WALES

302

PAULINE BRYANT

fortuitous that the majority of the informants in the final survey came from that part of the country. By the end of the Canberra-based stage of the final survey it had become obvious that these borders were even more extensive and more complex than previously thought. Near the junction of New South Wales, Victoria and South Australia for some items; further east, it was not always between the Murray and Murrumbidgee Rivers as the trial surveys had suggested, but for some items as far north as the Lachlan River; and nearer the coast, the Great Dividing Range appeared to have a barrier effect. These findings were the result of covering more towns in the final survey than in the verification surveys. The second stage of the final survey consisted of four field trips to define these borders more exactly. Thus, in these areas a fine-grained survey of the type conducted in the field verification survey was carried out. The informants interviewed on these field trips were current users. In both parts of the final survey, Canberra-based and in the field, a total of 845 informants was interviewed, 325 of them current users, and 540 recollected users. 20.6

Preliminary analysis of the final survey

Some work was done on a sample of 18 items from the final survey to analyse out-of-area usage (see Bryant 1989b for a discussion of how areas, and therefore out-of-area usage, were defined). Of current users, 1.5 per cent used out-of-area terms, compared with 3.5 percent of recollected users in the same areas, that is, there was a 2 per cent error rate among recollected users. Even this figure may be on the high side as some responses classified as out of area may turn out with further work to be part of minor sub-areas. It is safe to say that the error rate for recollected usage is 2 per cent at the worst. Because the total number of current users was higher in the final survey than in all the trial surveys combined, (325 compared with a total of 243), it was possible to eliminate the 2 per cent error rate by basing the findings principally on current usage, with recollected usage providing supplementary information. Of the 42 items included in the field verification survey, 36 were among the 72 used in the final survey. As 116 of the informants from that survey had provided current usage, their data were included in the final survey. This gave the following:

Regional usage in the lexicon of Australian English Field survey Final survey Total

Current 116 325

Recollected use Not included 520

441

520

303

At present, the data from the final and field surveys are being analysed, starting with the South-East region. Map 20.3 shows the distribution in the South-East region of the regional term for one item, based on current usage only from the final and field surveys. Heteroglosses (see Chambers and Trudgill 1980: 103-5, and Bryant 1989b for a discussion of heteroglosses as used in this analysis) for the northern limit of the South-East term slide, and the southern and eastern limits of the North-East/South Central term slippery dip are shown. 20.7

Conclusion

A number of issues posing methodological problems have arisen in the course of this project. Three points, one specific to Australia and two of wider applicability, are of particular interest. First, out-of-area regional usage in Australian English is a phenomenon which confuses the identification of regions and needs further investigation. Second, the use of recollected usage, in the verification surveys, is remarkably reliable, with at worst 2 per cent of responses being incorrectly recalled. Third, the pictorial questionnaire technique, developed for use with groups of informants and also used with individuals, has advantages not found in other methods of data collection.

21 Finding a place in Sydney: migrants and language change BARBARA HORVATH

21.1

Introduction

When migrants enter their new country, they are immediately faced with the sometimes daunting task of finding a place to live. Studies of urban settlement patterns record that migrants often are concentrated in certain parts of cities; Little Italy and Chinatown in New York City may be among the better-known ethnic communities in the world, but they are far from unique. Ethnic neighbourhoods give migrants a place in the new country where they can speak their own language and obtain the goods and services they need to maintain to some degree the way of life they grew up with. Ethnic communities are not entirely separate entities, however, built on the edge of town. They are created within the confines of the host community, occupying areas once exclusively the domain of the host community or areas vacated by other, often more upwardly mobile ethnic minorities. Initially, the hosts may be attracted by some of the innovations brought in by the migrants, notably the food and the festivals, but it is rare for the hosts to hold the ethnic communities in high regard. Nevertheless, whether the host is attracted or repelled, infindinga place for themselves, migrants bring about quite often dramatic changes in the host community. Equally as daunting for migrants, but not as consciously undertaken, is the task offindinga 'place to speak' within the host speech community. The ethnic neighbourhood can never totally recreate the conditions of the home country. Cultural patterns cannot remain unchanged, the migrants cannot fail to join the speech community of the host. It is not a matter of simply learning to speak the host's language; the migrant must find a sociolinguistic niche to occupy. In locating themselves sociolinguistically, they also have the same potential for changing the character of the speech community as they have for changing its social geography. The role of migrants as agents of language change should be a particularly fascinating one for sociolinguists. However, all too often

Migrants and language change

305

migrants, particularly first-generation migrants, are defined out of the field of interest; they are not considered to be members of the speech community because they do not speak the language as native speakers. This approach to defining a speech community is reminiscent of the early dialect geographers who wanted to include only those members of the speech community who best represented the speakers of an earlier form of the dialect and accordingly choose primarily rural, elderly folk for their studies, for example, Kurath's study of English in the New England region of the US. Sociolinguists interested in language change in progress have concentrated on urban communities where the actual processes of language change in progress can best be seen. It is not often, however, that the role of the migrant is focused upon. Too often only native speakers of the language are included in the study, and this may well overlook important sources of change within the speech community. Sydney is a multilingual speech community, largely as a result of immigration since the end of the Second World War. The most numerous groups of migrants from non-English speaking backgrounds from the late forties and fifties are the Greeks and Italians, but many other European language groups came during this period. New languages are continually added to the speech community as Australia accepts migrants and refugees from countries suffering political upheaval, for example, Spanish from Latin America, Arabic from Lebanon, and Vietnamese and Chinese from Vietnam. All of the languages are replenished by speakers who enter under the family reunion scheme which is part of the current Australian immigration policy. Support for multilingualism and multiculturalism is the overt policy of the Government (see Ozolins, this volume); the Government funds a radio station in Sydney, for instance, which broadcasts in 60 languages. English, as would be expected, occupies a privileged place in the multilingual policy. All speakers are encouraged to speak English and there is an extensive adult English as a Second Language programme sponsored by the Government. Few can choose to ignore English because it is the language of government, most of the media, education, and work. In this chapter I will examine how migrants and their children enter into the English speech community, that is, how they begin to fit into the sociolinguistic patterns that are already well-established in the community, and what effect, if any, they have on those patterns. 21.2 Australian English: the 1940s Fortunately we have available good descriptions of Australian English (AE) at the time just prior and during the early phase of the major migrations of non-English speaking people to Sydney and the rest of

306

BARBARA HORVATH

Australia. Although there are earlier writings to be found about AE, the academic study of AE began in the 1940s with the work of Mitchell (1946). Along with his colleague, Delbridge, he was responsible for the first major description and empirical study of AE (Mitchell and Delbridge 1965). Since the forties there has been a steady flow of research on AE, with studies of phonological and lexical variation accounting for the better part of the published work. From the very beginning of this research tradition, the variability within AE has been recognised as of major interest. Mitchell and Delbridge have found very little disagreement with their division of AE into three major types, Broad, General and Cultivated. Although one might quibble that the labels embody too much subjective evaluation, they do correctly suggest that Broad has the least overt prestige, Cultivated the most, and General occupies a middle position. The focus of the study of AE from the start has been on the variable pronunciation of a subset of the vowels. Some descriptions of consonant variation exist and researchers usually associate particular variants of the consonants with the varieties of AE as described by Mitchell and Delbridge. However, variation in the vowels, (iy), (ey), (ow), (ay), (aw), and (uw) (see table 21.1 for their phonetic realisations) has traditionally been regarded as the main differentiators of the three varieties of AE. Mitchell and Delbridge (1965a: 15) reported the following distribution for the three varieties: Broad General Cultivated

34% 55 % 11 %

21.3 Australian English: the present The description here summarises some of the results of a sociolinguistic survey of the Sydney speech community more fully reported in Horvath (1985). In this survey linguistic data was collected from 177 speakers using the typical sociolinguistic interview developed by Labov and his associates; the social characteristics that were built into the sample included ethnicity (Anglo-Celtic; Italian; and Greek); social class (lower working class; upper working class; and middle class); age (teenagers and adults their parents' age); and sex. 213.1

Determining the sociolects

The Sydney speech community was divided into sociolects by using a principal components analysis of the vowel variables (iy), (ey), (ow), (ay) and (aw). Table 21.1 gives the phonetic description of the variants. The

Migrants and language change

307

Table 21.1. The phonetic variants of selected AE vowels

(iy) (ey) (ow) (ay) (aw)

Cultivated

General

Broad

Ethnic Broad

[31] [AI]

[3" I]

—

[81]

[AVI]

[am]

[AVUI]

[ai]

[ a " i]

[an]

[XD]

[D I] [aeb]

[a" •] [am] [DI] [ED]

phonetic description basically follows that of Mitchell and Delbridge (1965: 80-4) and Bernard (1970), except that some additions had to be made for an adequate description of the variation found in this set of speakers. An additional variety had to be identified and in keeping with the AE tradition in giving labels to varieties which give some sense of social interpretation, we have labelled this variety Ethnic Broad (EB). It is 'broader than broad' in two senses: (1) if one acknowledges a relationship among the variants of the vowels such that the direction of change is from the Cultivated to the Broad, then the EB variant represents a further step in the change pattern, and (2) it has even less prestige than Broad. In fact, whereas it could be argued that Broad AE enjoys covert though not overt prestige, Ethnic Broad enjoys neither. In addition, since many of the speakers spoke English with a decided accent, we needed to be able to distinguish those vowels which were clearly part of the AE system from those that were probably the product of interference from the first language. We therefore included a category called Accented to carry this distinction, but did not characterise it phonetically. By using only linguistic data and excluding social information, we are able to identify groups of speakers who share similar patterns of variation in their pronunciation of the selected vowels. Once the patterns of variation have been determined, the social and linguistic distribution within each group can then be described; we define these groups as sociolects. Each sociolect, then, consists of patterns of linguistic variation and patterns of social variation. The principal components analysis yields two major subdivisions among the speakers in the sample; these subdivisions have been labelled the core and the periphery of the speech community. Further subdivisions can then be made: two sociolects in the periphery and four in the core. The appropriateness of the labels will become apparent once the distribution of the vowels and speakers is described.

308

BARBARA HORVATH

21.3.2 The periphery Of the 177 speakers, 47 are identified as being in the periphery. The distribution of each of the variants for each of the vowels for the two sociolects are shown in the histograms in figure 21.1. The primary difference between the two peripheral sociolects is the degree of accented variants; Sociolect 1 is much less integrated into the AE speech community and Sociolect 2 is more integrated, where integration is a measure of the proportion of vowels in a peripheral sociolect which are clearly part of the AE system. The social characteristics of these two sociolects is also of interest. All of these speakers are adults and are either Italian or Greek; it is not true, however, to say that all of the Italian or Greek adults are in the periphery. In fact, ten are in the core speech community. The major social characteristic separating the two sociolects is ethnicity: 62 per cent of Sociolect 1 are Greeks and 71 per cent of Sociolect 2 are Italians. Mitchell and Delbridge categorised speakers to arrive at their determination of the numbers of speakers of the three varieties. For a number of reasons, however, we categorised individual occurrences of the vowels by the speakers. As we can see from the histograms, if we were to label Sociolect 1 'Accented' we would be failing to notice the occurrence of substantial numbers of variants that were not accented. However, in order to compare our study with the Mitchell and Delbridge study, all of the vowel data have been aggregated in table 21.2 and the overall percentage of each variety for the two sociolects is shown. Clearly the major differences between the two is not only in the amount of accented vowels but also the use of the Ethnic Broad, Broad and General variants. Notice that there is almost an equal amount of Cultivated variants in the two sociolects. We will return to discuss the interpretation of these differences after we have examined the core sociolects. 21.3.3 The core The core consists of 130 speakers; the distribution of the variants for the five vowels is shown in histograms in figure 21.2. The most obvious differences between the core and the periphery are the absence in the core of any Accented variants; surprisingly there are no Ethnic Broad variants. This principal components analysis shows this relationship between the core and the periphery as a definite split, hence the labels. On the other hand, the relationship among the four sociolects is quite different; that these sociolects form a continuum rather than clearly defined separate dialects is shown in the decrease of the number of Broad variants as one goes from Sociolect 1 to 4 and the increase in the number of

Migrants and language change

309

Table 21.2. Percent distribution of vowel variants in the peripheral sociolects Vowel variant

Sociolect 1 (%)

Sociolect2

Cultivated General Broad Ethnic Broad Accented

18.32 3.82 1.04 4.07 72.74

17.89 11.33 7.83 20.56 42.39

(iy) ' (ey) ' (ow) ' (ay) ' (aw)

(iy) ' (ey) ' (ow)' (ay) ' (aw)

SOCIOLECT 1

SOCIOLECT 2

Figure 21.1 The linguistic structure of the sociolects in the periphery Cultivated variants from Sociolect 1 to 4. The General variants peak in Sociolect 2 and 3. Once again we see clearly that there is no variety in which only one of the variants is present and all the others absent. In fact, there was not a single speaker in the sample who used only one variant of any of the vowels studied. The social characteristics of the four sociolects are quite complex. The

310

BARBARA HORVATH SOCIOLECT 1

100%

S0CI0LECT2 100%9080 70 6050 40 30 2010-

(iy) • (ey) ' (ow)

(ay) ' (aw)

0

SOCIOLECT 3 100%

C'

(iy) ' (ey) ' (ow)' (ay) ' (aw) SOCIOLECT 4

100%9080 70605040302010-

(iy) ' (ey) T (ow) 1 (ay) ' (aw)1

0

(iy) F (ey) ' (ow)' (ay) ' (aw)

Figure 21.2 The linguistic structure of the sociolects in the core

Migrants and language change 100%

100%

80

80

60

60

40

40

20

20

311

0 1

2 3 GENDER

4 Sociolect 1

2

3 AGE

100%

100%

80

80

60

60

40

40

20

20

0 1 2 3 4 Sociolect 1 SOCIOECONOMIC CLASS

2 3 ETHNICITY

4

Figure 21.3 Distribution of the social characteristics across the core sociolects

graphs infigure21.3 show the importance of sex differentiation. Sociolect 4 at the Cultivated end of the continuum consists entirely of females, and males dominate Sociolect 1 at the Broad end. I will examine the ethnic dimension of this variable further. Middle class speakers increase from the Broad to the Cultivated end and working class speakers decrease. However, it is age and ethnicity that are the most interesting variables in addressing the question of how it is that migrants find a place in the host speech community. First of all, it is important to recognise that the core is over-represented by teenagers, despite the fact that we began with an evenly distributed sample, that is, 30 speakers each in three ethnic groups and two age groups. (In fact we ended up with only 27 Italian adults, which totals 177 speakers instead of the expected 180.) This accounts for the fact that teenagers dominate all core sociolects except for Sociolect 4. There are also more Anglo-Celtics among the 130 core speakers since all 60 are in the core and there are only 30 Italian and 30 Greek teenagers plus the seven Greek and three Italian adults. The most relevant patterns on the ethnicity graph are the dominance of Anglo-Celtics in the Broad sociolect and the Greeks in the Cultivated.

312

BARBARA HORVATH

Table 21.3. Percent distribution of Anglo-Celtic speakers over the three varieties of AE Sydney

Broad General Cultivated

Mitchell and Delbridge

Adults

Teenagers

34 55 11

27 67 7

20 76 3

Table 21.4. Percent distribution of Greek and Italian speakers over the three varieties of AE Sydney

Broad General Cultivated

Greeks

Italians

3 87 11

6 91 3

Let us begin to assess how migrants enter the speech community by examining each ethnic group in turn. First of all, let us compare the Anglo-Celtic group in the core with the Mitchell and Delbridge distribution across the three varieties, as shown in table 21.3. The adult speakers in the Sydney study should be more or less from the same generation since Mitchell and Delbridge collected their data from teenagers in about 1960. For the sake of comparison, core Sociolects 2 and 3 have been combined. We can take the distribution of Anglo-Celtic adults across the three varieties to be roughly equivalent to the patterns found by Mitchell and Delbridge. There are a number of differences in the way data was collected and analysed in the two studies so it would be a mistake to take such a comparison too seriously. The differences between the Sydney adults and teenagers is important; whereas two-thirds of the adults are in the General category, three-quarters of the teenagers are. Both the Broad and the Cultivated varieties would appear to be losing speakers from this ethnic group. If we now add the Greek and Italian teenagers as well as the ten Greek and Italian adults who are in the core, as in table 21.4, we can see whether the distribution patterns are the same as the Anglo teenagers. If so, this would indicate that once in the core the members of these ethnic groups are indistinguishable from the members of the host speech community.

Migrants and language change

313

Table 21.5. Percent distribution of all ethnic groups across the four core sociolects Ethnic groups Anglo-Celtic

Greek

Italian

Sociolect

Adults

Teens

All

All

1 2 3 4

27 30 37 7

20 53 23 3

3 38 49 11

6 58 33 3

(Broad) (General) (General) (Cultivated)

The move toward the General variety is even more noticeable among the other two ethnic groups, with the Italians being almost entirely within the General sociolect. Quite interesting is the very low representation of both groups in the Broad category compared to the Anglo-Celtic teenagers. Also interesting is the number of speakers among the Greeks who are in the Cultivated variety; they represent the highest percentage in this variety of all the ethnic groups. In fact, the differences between the three ethnic groups can be shown to be even more distinctive when we separate the two sociolects which we combined to compare the results with the Mitchell and Delbridge findings, as in table 21.5. The Italians (which includes all the teenagers and three adults) and the Anglo-Celtic teenagers are alike in their patterning except that the Italians are not represented in Sociolect 1 to the same degree. However, the Greeks show themselves once again to be quite different; well over half of them are to be found at the Cultivated end of the continuum compared to only about a quarter of the Anglo-Celtic teenagers and just over one-third of the Italians. This pattern alone would be justification enough to claim that the Italians and the Greeks remain distinctive within the core community. However, when we examine once again the sex variable we see an even more striking difference. The male/female distinction is a very important one for the Anglo-Celtic speakers, for both adults and teenagers, but for the Greek and Italian teenagers, it is not. 21.4 Inter-generational variation

According to variationist accounts of language change such as Labov (1966) a) language change proceeds quantitatively; b) the process of change is linguistically and socially structured;

314

BARBARA HORVATH

c) social evaluation is the motor of change; d) inter-generational variation is a surrogate for change over time: each generation moves a change along 'one step'. What needs further discussion within the variationist paradigm is the role of the host community in urban settings where there is a high influx of migrants. It is possible to interpret the failure to include migrants in sociolinguistic studies as a presumption that it is the host community that sets the parameters for how a change will proceed. Migrants from nonEnglish speaking backgrounds, or rather their children, become Aussies by losing their accents and falling in step linguistically with their social counterparts, that is, in this case with Anglo-Celtic Australians of similar sex and social class backgrounds. As we have seen so far, the second-generation Italians and Greeks remain distinctive to some degree. One pattern that is definitely not followed is the distinction that Anglo-Celtics make between males and females which is a distinction long noted in AE studies (Horvath 1985; Mitchell and Delbridge 1965; McBurney quoted in Ellis 1887). Let us look at the inter-generational variation between the three ethnic groups and then take up the central question of how migrants enter into a speech community and what effect they have on that community's patterns of sociolinguistic variation. The differences between the Anglo adults and teenagers is more or less the expected one; there is a gradual shifting away from the highly marked ends of the continuum and a movement toward General with the sex differences being maintained. One of the major differences between the Greek adults and teenagers, though dramatic, is expected: the teenagers do not speak with a Greek accent. The Greek adults are also very low on the use of either the Ethnic Broad or Broad variants of the vowels and so are the teenagers. The Greek adults, both in the periphery and in the core, have a distinct preference for the cultivated end of the continuum and so do the teenagers. These two ethnic groups, the Greeks and the Anglo-Celtics, demonstrate the expected linguistic patterns. The younger generation moves the pattern of vowel variation along 'one step'. The Italian teenagers, however, do not meet the expected intergenerational pattern, except that they too do not speak with an accent. Although the use of Ethnic Broad and Broad vowels make up over a quarter of the adults' linguistic behaviour and there is also a fair representation in the Cultivated variety, the teenagers move away from both of these vareties and are to be found almost entirely in the two General varieties. If we assume that the development of the Ethnic Broad variety moves the Australian English pattern along 'one step' (as I have argued elsewhere), then the Italian teenagers are taking two steps backwards and in so doing jump ahead of their Anglo counterparts.

Migrants and language change

315

We have already seen that the second-generation Greeks do not simply fall in line sociolinguistically with the supposed Anglo-Celtic trendsetters; the Anglo-Celtics are moving away from the Cultivated end of the continuum and it is the Greeks who are breathing new life into it. I have identified three intergenerational patterns: 1) Anglos follow the previous generation as expected; 2) Greeks follow the previous generation but this is not expected because instead they should follow their Anglo counterparts; 3) Italians do not follow the previous generation nor do they behave like their Anglo-Celtic counterparts; instead they jump ahead of the Anglo-Celtic teenagers by moving not one but two steps away from the ends of the continuum.

21.5 The speech community and change in progress The two generational model of change that emerges from the analysis of the patterns of variation in the study of the Sydney speech community is shown in figure 21.4. In the first generation, migrants form satellite or peripheral speech communities which may be attached to the host speech communities at different places. Recognising that there is variation in the first generation migrants, nevertheless, we can say that by and large the Italians 'attached' themselves at the Broad end of the AE continuum and the Greeks at the Cultivated end. However, the Italian first generation can be described as being more integrated into the AE speech community than the Greeks for two reasons. Firstly, overall they use more vowels that are clearly part of the AE system. Secondly, they add a new variant to the AE vowel continuum. The first-generation Greeks, by contrast, use a very small proportion of AE vowels and the majority of the AE vowels are Cultivated. In the second generation we see that all of the speakers are fully within the host speech community, that is there are no more accented vowels and, importantly, no EB variants of the vowels. Ethnicity, however, remains a distinctive social variable for this second generation. They do not, for instance, maintain the linguistic importance of the sex variable and they have a different distribution over the vowel continuum. We can also see what impact this influx of the children of migrants has had on the host community. The very highly stigmatised variant EB is clearly avoided by the children of Italian migrants; the children move right away from the Broad end of the continuum. The development of the EB variant may well have acted as an impetus for the Anglo-Celtic teenagers to move away from the Broad end of the continuum as well. Although Broad has been overtly stigmatised almost from the beginning, it has been maintained largely as a male-dominated variety with covert prestige. The Cultivated end of the continuum appears always to have had a minority of

316

BARBARA HORVATH

FIRST GENERATION

SECOND GENERATION

ACCENTED ENGLISH

THE AUSTRALIAN I ENGLISH CONTINUUM

Figure 21.4 The sociolinguistic relationship between migrants and their hosts: an intergenerational model

speakers. Without the newcomers to this sociolect, the Greeks, it may well have withered away. This might be called the 'cultural cringe' variety. It is regarded by Australians to be the closest to British Received Pronunciation and therefore the most acceptable and the one with overt prestige. The rise of nationalism which is nowflourishingin Australia, so that England is no longer regarded as 'home' as it was widely thought to be in the 1940s, has resulted in a more positive social evaluation of the General, though not the Broad, variety. The Greeks, however, having no particular associations, either positive or negative, with England, simply regard the Cultivated end of the spectrum as the dialect of AE having prestige. The Australians, on the other hand, have redefined what it means to speak 'real Aussie English' as General. In the discussion so far it is the speech community that has played a key role. In the past the notion of the speech community has not been particularly well theorised. Studies of New York City, Philadelphia, Washington, DC, Norwich and Belfast more or less assumed that the people living in the confines of the city could be regarded as members of the speech community, unless of course they were clearly a subculture within that city, such as the Blacks of Harlem or Washington, DC where they were

Migrants and language change

317

considered a separate speech community. In the study discussed here it was necessary to begin by questioning the assumption that everyone who lived in Sydney belonged to the same speech community. That questioning has led to a better understanding both of how migrants enter into a new speech community and of how they affect the patterns of variation and change in the host community. I began this chapter with reference to how migrants situate themselves geographically in the host community. It is interesting to find in the geographical literature corroboration for the notion of integration into a speech community as well as specific corroboration for the notion that the Italians and the Greeks differ with respect to the degree of integration into the AE speech community. In the social atlas of Sydney Horvath and Tait (1984) use an index of residential segregation which measures the extent to which the geographical distribution of a subpopulation differs from that of the total population. An index of 0 means that the subpopulation is perfectly integrated and an index of 100 indicates total segregation. In Sydney the least integrated ethnic group is the Vietnamese, the most recent migrants, with an index of 79; third generation or greater Australians have an index of 19. The Greeks in Sydney have an index of 55 on this measure and the Italians 45. Sociolinguists may well have gone too far in drawing a distinction between themselves and dialect geographers. Perhaps it is time to reconsider the relationship between sociolinguistics and social geography; migrants trying to find a place in Sydney certainly ignore neither.

22 Gender differences in Australian English ANNE PAUWELS

22.1

The study of sex/gender1 differences in language

Prior to the emergence of 'feminist linguistics' in the mid-1970s the exploration of the differences in linguistic behaviour of the sexes had featured primarily in anthropological and later in sociolinguistic studies. Anthropologists observed some phonological, grammatical and lexical contrasts in a range of 'exotic' languages where the sex of the speaker or addressee determined the choice of the linguistic form. Such differences were usually referred to as sex-exclusive differences (for a more detailed account, see, e.g. Bodine 1975; Brouwer et al. 1978; and Coates 1986). Sociolinguists studying urban dialects of European languages presented evidence of sex-preferential differences at the phonological, syntactic and prosodic levels (for a selective survey of English language studies, see Coates 1986). Despite various shortcomings of explanatory and methodological nature in earlier anthropological and sociolinguistic studies with respect to the gender issue (see Cameron and Coates 1985; Coates 1986; and Milroy 1987), such studies provided and still provide a substantial database for the analysis of gender differences in language within a feminist linguistic framework. Over the past 15 years psycholinguists, pragmalinguists, conversational analysts, etc., have also increasingly become interested in a more serious study of the differences and similarities in the linguistic behaviour of the sexes (see Thorne, Kramarae and Henley 1983). In addition to the study of isolated language variables, for example, gender differences in pronunciation, attention is also being paid to differences in the communicative competence of men and women, in their speech styles, for example, politeness phenomena, topic choice, turn-taking behaviour, interruptions, questions and directives, and to the acquisition of gender-specific speech styles. Another major strand in feminist linguistic research concerns the

Gender differences in Australian English

319

representation of the sexes in the language system, that is, the relationship between ideology, gender and the language system. Many languages are found to exhibit in their structures a gender bias (against women in favour of men). This androcentrism in language is often experienced negatively by female language users. In relation to this issue, the following aspects are usually examined: areas of sexism/androcentrism in language, effects of androcentrism on language users and proposals for eliminating gender bias in language. 22.2 The study of gender differences in language in Australia Language and gender/sex research in Australia focuses on similar issues as the ones outlined above and also shows similar trends in development. The multicultural and multilingual nature of Australian society allows the issue to be studied in relation to a multitude of Aboriginal/indigenous and immigrant/'transported' languages as well as cross-culturally. Sexexclusive markers and speech styles can still be observed in some Aboriginal languages (see e.g. Dixon 1980). The role of women and men in the process of language contact and language shift to English is being examined for indigenous and non-English speaking speech communities from sociolinguistic, social psychological and anthropological perspectives (see, e.g. Callan and Gallois 1982; Callan, Gallois and Forbes 1983; Evans 1986; Pauwels 1987b; Schmidt 1985; Troy 1987; and Horvath, this volume). The study of sex differences at the phonological and grammatical levels and the analysis of sexism in Australian English have so far gained most attention. More recent developments in Australian language and gender research include the study of gender specific speech styles 2 and of the linguistic change occurring as a consequence of feminist language planning as well as analyses of the reactions towards such change (e.g. Davies 1987; and Pauwels 1987a). The focus of this chapter will be on gender differences in phonology, grammar, prosody and lexis in Australian English. The data for this survey have come mainly from sociolinguistic studies of Australian English carried out over the past 20 years. (See Pauwels 1987c for other aspects of language and gender in Australia.) 22.3

Gender differences in Australian English

Data on gender differences in the use of Australian English have derived predominantly from wider variationist studies focusing on phonological and grammatical variation in a more general manner. Unfortunately, Australian data on sex-specific speech styles and on speech strategies used by women and men in various linguistic environments are very scarce. This

320

ANNE PAUWELS

is partly due to the fact that discourse and conversational analytic studies have so far emphasised the differences related to social class or ethnicity more than to those related to gender. Gender differences focus more prominently in recent studies adopting a systemic linguistic approach. This survey shall be limited tofindingsregarding gender differences in language below and at sentence level. Generally speaking, Australian findings on the gender distribution of (morpho-)phonological and some syntactic variants closely resemble the findings of American and British studies. Women are found to use the prestigious or standard variants of certain phonological and grammatical variables more than men. This should not come as a great surprise as the studies from which thesefindingshave sprung have all employed more or less the same methodology and the situation of women in these three societies is comparable. Thefirstmajor (pre-sociolinguistic) study to look at gender and language was that of Mitchell and Delbridge (1965), in which the speech of Australian adolescents was investigated. The vowel variants of/i/, /u/, / ei/, /ou/, /ai/ and /ao/ were regarded by Mitchell and Delbridge as the main differentiators of the Australian English varieties: Cultivated, General and Broad Australian. Cultivated Australian was regarded as the variety closest to British Received Pronunciation (RP) and Broad Australian as the variety furthest removed from RP. Male adolescents were found almost exclusively in the Broad and General categories of Australian English, whereas female adolescents were more evenly distributed over the three categories, as in table 22.1. A recent large-scale study examining variation in Australian English as spoken in Sydney (Horvath 1985) revealed that Anglo-Australian female teenagers and adults were more likely to employ the cultivated or general variants of the vowels than their male counterparts. Small-scale studies on /ai/ (Scott 1982) and /si/ (Constantinidis 1982) also found that girls and women used more prestigious (cultivated) variants than boys and men. Consonant variation has received less attention in the literature and has been linked less strongly with gender. Horvath (1985), however, found that [f] for [0], /h/ deletion and the flapped variant of/t/ were all closely associated with male speakers, while the aspirated /t/ was a feature of female speakers. Furthermore, the palatalisation of /t, d, s, z/ was recorded by Horvath as more frequent among male than female speakers. Postvocalic /r/ variants showed little gender link. The phonological variation of the verb morpheme -ing ([in], [in] or [an]) was also found to be associated with gender (see, e.g. Bradley and Bradley 1979; Horvath 1985; Ingram et al 1985; Shnukal 1982b; and Shopen 1978). Females consistently used the [in] pronunciation more than males. The variants of-ing as in thing, something, that is, [in] and [ink] were less clearly

Gender differences in Australian English

321

Table 22.1. Gender and major differentiators of Australian English varieties

Girls (53%) Boys (47%)

Broad

General

Cultivated

10% 24%

33.5% 22%

9.5% 1%

Source. Adapted from Mitchell and Delbridge 1965: 88.

gender-linked. Shnukal (1982b) though found some differences: men used the [ink] variant more, but this was not significant. Horvath (1985: 102) even noted that 'the [ink] variant stands out among all the other variants in that it is the only one that cannot be said to be either a male or female speech characteristic'. The tendency for female speakers to use more standard-like variants has also been confirmed at the grammatical level (see, e.g. Eisikovits 1981; and Shnukal 1978). Eisikovits (1981), for instance, investigated a set of grammatical variables among 40 working-class adolescents in Sydney. These included past tense and past participle marking of the lexical verb, the use of perfective have, the be or get passive, the use of multiple negation, subject verb agreement, is/are, was/were, dorit\doesrit, variant nominal constructions, me/my, and some aspects of clause syntax, for example, the use of what clauses to express comparatives, as in She does more than what Karen does. She discovered that the boys not only were more likely to use the vernacular variants than the girls but, as they grew older, also increased their use of such variants. Girls, on the other hand, showed a marked decrease in the use of non-standard/vernacular variants between the ages of 13 and 16. Differences between the sexes in this area of language use seem to vary somewhat among the varieties of English. Studies of grammatical variation in American English (see, e.g. Fasold 1972; Feagin 1979; and Wolfram and Christian 1975) did not reveal marked gender differences. Cheshire's study of Reading English (1982), however, did show gender differences similar to those found by Eisikovits in Sydney. Explanations for observed gender differences in the use of certain phonological and grammatical variables have been rather minimal. This is probably related to the nature of the projects from which most data have come: 'quantitative' or 'correlational' sociolinguistic studies provide descriptions of the relationship between linguistic and speaker variables, for example, age, social class, and sex, in order to obtain insights in the processes of language change. The question why a particular variant is closely associated with a certain group of speakers is of no immediate relevance to the main goal of such studies. If attempts at explaining

322

ANNE PAUWELS

women's greater use of standard/prestigious variants and men's greater use of vernacular/non-standard variants are made, these usually refer to the notions of status and solidarity put forward by Trudgill (1972). Women's greater sensitivity to prestigious linguistic forms (hence their greater use of such variants) can be attributed to their inferior social status in society. Since they do not have the same avenues as men to acquire status in society, for example, through education, employment, income, etc., women perceive the use of standard language forms as a way of acquiring status. Men's use of vernacular variants can be seen as expressing solidarity especially among the working class. Such forms are maintained as they may hold covert prestige for their users. The validity of such an explanation has recently been challenged on a number of grounds. As far as methodology is concerned, the model of social class membership used in variationist studies is unsatisfactory for the classification of women. Also, if marking prestige motivates women, why don't female partners of men in high-status occupations simply imitate the language of their male partners (for further criticisms, see, e.g. Cameron and Coates 1985; and Coates 1986). It seems more fruitful to seek explanations for persistent gender differences in Australian English by referring to the notion of social network (see e.g. Milroy 1980 and Cheshire 1982). Despite some problems with the calculation of the network strength score for women (see e.g. Cameron and Coates 1985), the social network approach seems to have a greater explanatory potential as it attempts to explain differences by referring to the controlling influence of the networks with which men and women are associated. Sociologists, historians and other commentators on social mores and societal structures in Australia (e.g. Conway 1971; Encel et al 1974; and Summers 1975) have repeatedly remarked upon the persistence since early colonial days of extreme division and differentiation between the sexes in this country. This rigid division of the sexes undoubtedly has an effect on the types of social contacts/networks women and men have and possibly on their linguistic behaviour. Also the strength of such cultural stereotypes of Australian men and women as the 'Ocker'3, the 'Damned Whore' and 'God's Police' (see Summers 1975) should not be underestimated in affecting the perception of what constitutes appropriate linguistic behaviour for the sexes in Australia as well as the actual linguistic behaviour of at least Australian adolescents (see Eisikovits 1981). 22.4

Gender and intonation

According to Lakoff (1975), the use of a rising terminal contour in responses to questions, to which only the speaker knows the answer, is characteristic of female speech. She interprets this 'question intonation' as

Gender differences in Australian English

323

an expression of uncertainty, that is, the speaker is seeking approval or confirmation of what she is saying, or deference. Quite a few studies set out to verify the gender specificity of this intonational pattern (e.g. Conley et al. 1978; Edelsky 1979; McConnell-Ginet 1978a; and O'Barr and Atkins 1980), but failed to come up with clear evidence in favour of this claim. For instance, Edelsky's experiment showed that the use of the rising terminal was equally infrequent in female as male students. O'Barr and Atkins' study of courtroom language showed that the use or non-use of rising intonation in declarative contexts is more a marker of speakers' social status (powerless or powerful) than of their sex. Australian research on intonation has focused on the increasing use of the High Rising Terminal (HRT) (also referred to as Australian Questioning Intonation (AQI)), which seems to be a relatively recent innovation in Australian English. Its occurrence is not restricted to responses to questions but can also be found in other contexts, for example, in narratives and descriptive speech (see Guy and Vonwiller 1984; Guy et al. 1986; and Horvath 1985). The social distribution of the HRT in Australian English indicates that female working-class adolescents are the most frequent users of this rising intonation. Yet, the difference in use between women and men is not very great: 60 per cent for women as opposed to 40 per cent for men, (Guy and Vonwiller 1984). There is no clear evidence from these Australian data for the gender specificity of this feature. As far as the interpretation of the HRT is concerned, Lakoff's suggestion that it signals deference and/or insecurity found some confirmation in the context of the interview situation. Australian working-class teenagers used the HRT frequently with female middle class (adult) interviewers. In other contexts, the HRT had other meanings, for example 'signalling the speaker's request for a heightened participation of the listener' (Horvath 1985: 132). 22.5

Gender and words

Examinations of the Australian lexis which made up a substantial part of philological and linguistic research well into the 1960s,4 provided very few details of possible gender differences. This is due, at least partly, to a male bias which characterised many of these studies. Idiomatic expressions, colloquialisms, slang, etc., which characterise the language of predominantly male groups, e.g. shearers, football fans, or are associated with male activities became the subject of careful examinations in a series of reports by the Australian Research Centre. The folklinguistic belief that colloquial and slang use of one's language is a predominantly male prerogative undoubtedly also played a role in this preoccupation with examining men's use of slang. As a consequence, the data on women's use of such

324

ANNE PAUWELS

phenomena in Australia is extremely limited. Keesing's (1982) collection of the slang of Australian women Sheilaspeak, and families, Familyspeak, is to my knowledge the only one to date to record female use of slang. The collection which is based on the author's own use and observations of women's slang as well as on some reports by other women, clearly shows that Australian women not only have a rich array of specifically female slang terms at their disposal but also use a good deal of them in their daily lives. Keesing furthermore points out that the meanings associated with specific slang expressions which are used both by men and women may be gender-specific, especially with regard to sexual terminology. For example, the meaning attached by men to the colloquial use of the verb Ho scrape' is 'to have sexual intercourse'. For women, colloquial 'to scrape' is synonymous with 'having a curette/an abortion'. Male bias is further present in the fact that such specific female interpretations go largely unnoticed in collections of Australian slang. 22.6 The politics of address The issue of address, which is among the primary resources for the realisation of social (including gender) relations, is being examined in depth by Poynton (1981, 1982, 1985). Working within a systemic linguistic framework, Poynton 5 distinguishes between approximately 36 categories of terms of address, including kinship and honorific titles, different types of names, solidary, approbatory and derogatory terms and analyses them in relation to the tenor dimensions of power, contact and affect. According to Poynton, the most striking observation about the terms of address in Australian English and their use is that the whole system is fundamentally skewed or asymmetrical, especially in relation to gender. She (1981: 10) notes: 'for only six out of the thirty-six categories which have been distinguished are the same (or parallel) terms available for both males and females, to be used in the same contexts and with the same, or similar meanings (apart from the sex difference itself)'. Address between males and females in public contexts is often asymmetrical/non-reciprocal reflecting power differences. A male superior may address his female secretary by either her first name or an endearment, but she will have to address him as Mr . . . or by his first name. There also seems to be a difference between men and women in the use of first names. If full forms of first names are not used, for example, Robert, Christine, James, Catherine, adult males will be addressed mainly with a monosyllabic truncated form of the full name, for example, Rob/Bob, Jim. Adult females, on the other hand, are more frequently addressed by the diminutive (mainly -y suffix) of their name, e.g. Chrissy/Chrissie, Cathy. Although diminutives are commonly used to

Gender differences in Australian English

325

children of both sexes, boys often come to see them, sometimes at an early age, as girls' names and reject them (Poynton 1985: 82). It seems that 'women, like children, can be addressed in public with conventionalised intimate forms to a greater extent than is permissible to men' (Poynton 1985: 82). Poynton ascribes this practice to the greater contactability of women (and children) in western cultures. This is further manifested in the practice of men (often complete strangers) addressing women in public by a variety of endearments, approbatory or derogatory terms. Asymmetry also characterises some categories of address. Men seem to have at their disposal a greater range of terms with which to address both men and women. Especially in the area of solidary terms and derogatory names, women 'miss out'. There are no female equivalent terms for mate or sport. Whereas men have a large range of derogatory terms available for denigrating women as women, women do not have access to a large set of words for denigrating men as men. Many of Poynton's findings regarding the 'gender' politics of address in Australia confirm the observations of overseas researchers (see, e.g. Kramer 1975; McConnell-Ginet 1978b; Stanley 1977; and Wolfson and Manes 1980).

22.7

Conclusion

In this chapter I have shown that the linguistic behaviour of Englishspeaking women and men in Australia closely resembles that of men and women in other parts of the English-speaking world. Women are more likely than men to use standard variants of certain phonological and grammatical variants. There is also some evidence for a genderdifferentiated use of slang and colloquial expressions. The system of terms of address in Australian English has been shown to be skewed, especially in relation to gender. Recent advances in language and gender research in Australia include the collection and analysis of data on gender differences in female and male discourse and conversational patterns and on the acquisition of gender specific language behaviour. Notes 1. Although, technically speaking, the term 'sex' has been used to refer to biological characteristics of the sexes and 'gender' to refer to social and cultural phenomena, the terms 'sex' and 'gender' are often used interchangeably. In relation to the study DTlinguistic behaviour of women and men, the term 'sex' is often preferred, because 'gender' is already a technical term in linguistics. In this article, the terms 'gender' and 'sex' will be used concurrently to describe the differences between men's and women's speech. 2. Anne Thwaite (University of Sydney) is currently working on a project analysing the speech strategies of women and men in a variety of contexts.

326

ANNE PAUWELS

3. Delbridge (1977) gives the following description of 'Ocker': 'good-hearted uncultured Australian working man, usually in early manhood, devoted to football and beer, with an apparently strong preference for male social company, a contempt for culture, a low estimation of women (who should be kept in their place), a cheerful aggressiveness in speech, and a markedly broad Australian accent.' 4. The aim of most lexical studies was to study the occurrences of words and phrases which are characteristically Australian in form. 5. Although Poynton acknowledges the importance of prosodic and grammatical features in the analysis of terms of address, she concentrates on the lexical realisations.

Part V Public policy and social issues

23 National language policy and planning: migrant languages ULDIS OZOLINS

23.1

Introduction

Explicit concern for an official language policy in Australia is of recent origin. It has been a focus of attention for language professionals, policy makers and others for only the last decade. However, language policy has had a longer history in Australia, and during the learly part of this century, when previously multilingual elements in the Australian population were overwhelmed by an assertive Australian monolingualism (see Clyne, chapter 14 this volume, and Walsh, this volume), language policy was one of many methods of control of both Aboriginal and migrant populations (see Fesl 1988). In the case of migrant languages, extensive controls over their institutionalisation evolved not only from general social antipathy towards alienness, but also more specific worries over alien languages and populations in wartime. Clyne (chapter 14, this volume) has summarised the pre-World War II restrictions on 'foreign languages' in school systems, school curriculum, and publishing of newspapers. Libraries stocked almost exclusively English language publications, and interpreting services were nonexistent. Following in the spirit of these policies, restrictions on the use of foreign languages in broadcasting were introduced in 1952. These policies were based upon a particular view of the nature of (white) Australian society, and a belief that monolingualism was essential for social cohesion. With such social and numerical dominance of Australian English, there was neither need nor provocation to make declarations of national language policy. Significantly, however, these policies and perspectives on language were to change, sometimes dramatically, over the 40 years of postwar immigration. Post war language policy in Australia is situated within the wider context of shifting government policy in a range of ethnic affairs issues. The federal government's constitutional responsibility for immigration has always

330

ULDIS OZOLINS

encompassed a policy reach into post-arrival services, defined at times narrowly (e.g. migrant reception, settlement services, direct language services for new arrivals). However, at other times, it has been defined quite broadly to include the total range of policies that affect migrants in Australia, and interaction between migrants and host culture. Compared with some, Australian governments have always been extremely interventionist in their approaches to particular social groups. Both Federal and State governments are heavily involved in direct provision of services, and the detailed regulation and control of other ostensibly private areas. This interventionist stance has brought its own kind of response: a growing mobilisation on ethnic affairs and demands for new sets of government interventions and services. An emphasis on centralisation and government delivery of service could also in time be exploited for a call for national perspectives on language policies. I shall briefly trace the evolution of ethnic affairs policy, and the growing language issue, in the decades of postwar migration, particularly with reference to migrants of non-English speaking background (NESB). 23.2 Migrants, language and social policy

Martin (1978) has outlined three gradually evolving phases in the reception given to the massive postwar migrant influx. The first phase, from the beginning of the migration programme to about the mid-1960s, is characterised by a belief in the essentially unproblematic assimilability of the migrant. Martin (1978: 27) describes the prevailing attitudes towards migrants as being 'lucky to have found a home in Australia, coming from the tensions and economic desolation of postwar Europe: they were essential to economic growth and they were assimilable'. This assimilation perspective had definite consequences for language. Firstly, in order to guarantee assimilation, an extensive English as a Second Language programme was instituted for adult migrants. This consisted at times of English classes in staging camps in Europe, shipboard classes, classes at migrant reception centres in Australia, continuation classes and English lessons by radio. The methodology adopted, the direct method, or 'Situational English', specifically shunned bilingual approaches to language learning. In view of the prevailing social theory of rapid assimilation, course organisers were encouraged to teach linguistically and nationally mixed classes to prevent back-translation and first-language use (Pitman 1952). Yet while this programme was instituted for adults, it was considered that no provision needed to be made for NESB children. It was expected that they would pick up the language quickly and easily from their Australian peers. Official reports loudly attested to the wisdom of this

National language policy and planning

331

policy, the most notable being conducted for the Immigration Advisory Council by Judge Dovey in 1960. This study surveyed school principals throughout Australia, and reported that migrant children 'overwhelmingly' settled well into Australian schools, were academically achieving as highly as their Australian classmates, and rarely causing behavioural problems in school. The very few migrant children who were not doing well at school were identified by the report as those who persisted in speaking their first language at home. Dovey implored parents to speak English at home for the educational good of their children (Australia, Commonwealth Immigration Advisory Council 1960). The Dovey report was replete with methodological inconsistencies and questionable relations between evidence and conclusions (Price 1960), but it stands as a very significant reflection of the linguistic and social attitudes towards migrants at the time. Migrant groups, confronted by such a set of language policies, were cautious but quietly insistent about maintaining their languages and establishing language maintenance institutions. Despite the restrictions, newspaper publishing in languages other than English (LOTEs) was undertaken by all migrant groups of any size and concentration. Most migrant communities, who were more directly concerned with language maintenance for the future, also established and continued to support parttime, after-hours schools for their children. These were run (usually voluntarily) by members of the community and presented a curriculum of largely language study, perhaps with some cultural or religious aspects as well. These particular schools were little noticed and even less understood by Australian public institutions. Gradually they acquired the generic name of 'ethnic schools'. Despite scarce resources, their continued existence (and later subsequent growth in numbers) demonstrated the determination of these groups to maintain their language into the next generation. The second phase Martin identifies is a period from around the mid1960s to the early 1970s when perceptions of migrants changed as a result of the growing evidence of severe social and economic disadvantage of the migrant population. Whereas they had tended to be seen as a group happily benefiting from the opportunities provided by Australian society, surveys began to show that migrants (then largely from Southern Europe) had distinctly low prospects for mobility. Considerable ethnic stratification was found in the workplace, with migrants concentrated heavily at the bottom of the occupational hierarchy. Moreover, their children, far from being the unproblematic population described by Dovey, were now concentrated in overcrowded and under-resourced inner-urban schools and often did poorly at school. The migrant, and migrant education in partidular, were now beginning to be defined as a problem (Martin 1978).

332

ULDIS OZOLINS

Immigration ministers of the late 1960s, Snedden and Lynch, recognised these albeit unintended consequences of the migration program and officially buried the rhetoric of assimilation and promoted instead a concept of integration. Australia's highly visible migrant communities had clearly not been absorbed into Australian society as previously envisaged. Snedden sought to bring about policy changes in line with this change of rhetoric. He directed his department's attention to failings in services to new settlers and urged attention particularly to educational matters, where he had found an almost complete lack of official data on the performance of migrant children. Sensing the beginning of political restlessness among the migrant communities themselves, Snedden and Lynch initiated developments in several areas of language policy (Martin 1978). Interpreting services were identified as an area requiring government provision and control, to ensure both adequacy of service and minimal standards for practitioners. Thus, planning was undertaken for the Federal Government's Emergency Telephone Interpreter Service (see 23.5). With the Immigration (Education) Act 1971 the Federal Government sponsored thefirstwidescale and systematic teaching of English as a Second Language to migrant children, funding English as a Second Language teachers, teacher training and later some capital costs. This constituted a deliberate break from previous ideas that migrant children would pick up English without formal tuition. Moreover, although the methodology adopted by this Child Migrant Education programme was again the direct method (with at this stage no bilingual element), the programme was to have an influence on language issues far beyond that of English as a Second Language teaching alone. The raising of the question of the educability of NESB children made it possible, as we shall see, for other definitions of 'migrant education' to be raised and debated. Language teachers also began during this period to mention language maintenance as a possible objective of language teaching in schools. This represented a departure from previous views of foreign language education, which held that foreign language learning was of intellectual/cultural value (especially for the elite), but at the same time shunned the languages spoken by migrants in Australia. The case of Italian (the language of Australia's largest NESB population) is instructive here (see also Bettoni, this volume). While there were moves to establish Italian as a legitimate school subject in earlier decades, these earlier advocates were careful to place language maintenance issues distinctly second to traditional arguments of the cultural and intellectual benefits of language study. Italian was thus primarily advocated on similar intellectual/cultural grounds that had underpinned the traditional place of French and German. For some teachers, the advent of a migrant population added a certain potential for learning of particular languages, as in this suggestion from a 1956 Victorian modern language teachers' conference:

National language policy and planning

333

in districts where there is a large migrant community, the language of that group could be taught in schools. For example, in an area where there are a number of Italian migrants, Italian could be taught, not of course for the benefit of the Italian children, but for that of the Australians who would have the opportunity of using a foreign language actively, of appreciating a foreign culture and thereby helping in the assimilation of migrants into the community' (Babel, no. 3, 1950: 33). Language teaching would retain its traditional clientele and purposes, but now would be made more alive by the presence of a population that could be practised on. The language would not, 'of course', be taught for language maintenance purposes. However, by the 1960s language maintenance aspects also began to be stressed by advocates of Italian (McCormack 1964). In 1964 Clyne took up McCormack's suggestions regarding Italian and generalised this to speak of the need for the wider study of what he called 'migrant languages' (a new term), not least for the direct benefit of recognising and respecting migrant cultures and languages, and promoting bilingualism. Clyne (1964) in particular argued against the contemporary view that bilingualism would be harmful to children, or would serve to retard assimilation of migrants into Australian society. The third stage of response to the migrant presence in Australia Martin dates from the early 1970s and characterises as being the development of perspectives of multiculturalism. This concept of multiculturalism has provoked a plethora of interpretations, with its ideological force being of far greater consequence than its detailed denotation. Several different strands have contributed to this phenomenon, and battles over the desirability or meaning of multiculturalism reflect these varying antecedents. Some, such as Grassby, Minister for Immigration (1973-4) in the Whitlam Labour government, stressed recognition of the value of migrant cultures, of the migrants' contribution to Australia and the development of a unique national identity marked by diversity within the greater unity of the 'family of the nation', and the benefit of welcoming cosmopolitan elements into Australian culture. Grassby (1973) was also one of the first public figures to stress the value of maintaining one's first language. Ethnic groups themselves became more active over this time. For example, a widely representative Migrant Workers Conference in Melbourne in 1973 (followed by similar events in other cities) spelled out a set of demands relating to industrial, welfare, legal and educational issues, including the release of an important Statement on Migrant Languages and Cultures, which set out language needs encompassing the teaching of English, support for language maintenance, and the aim of language learning for all Australians (Migrant Workers Conference 1973). A followup Migrant Education Action conference in Melbourne in 1974 further detailed language education demands. These moves also led to the development of representative community structures for ethnic groups,

334

ULDIS OZOLINS

with the establishment of Ethnic Communities Councils in Sydney and Melbourne in 1974-5, who from this time maintained a conspicuous presence in policy discourse on ethnic affairs. The Liberal Party, upon return to government in late 1975, explicitly declared its ethnic affairs interest by reconstituting the old Department of Immigration as the Department of Immigration and Ethnic Affairs. In 1977 it set up the first comprehensive inquiry into the wide range of postarrival services for migrants. This inquiry, under its Chairman Galbally, saw multiculturalism as a leitmotif that could guide policy in the various areas of welfare, education, communications, direct language services and other policy areas that impinged upon the well-being of migrants, not only in their initial period of settlement but also in their longer-term residence in and contribution to Australia (Review of Post-Arrival Programmes and Services to Migrants, 1978). Its rhetoric was not that of cosmopolitanism or migrants rights, but rather the development of an attitude of multiculturalism, stressing social cohesion and migrant participation in mainstream Australian society. In programmatic terms it stressed the coordination of post-arrival services, including language services, that had grown in an ad hoc manner, and brought disparate policies under a unifying rationale. The report was published in ten major community languages as well as in English. While multiculturalism remained a hotly contested issue by critics both of the left (Lepervanche 1980; Jakubowicz 1986) and right (Knopfelmacher 1982; Chipman 1985), the years since Grassby have seen significant development in aspects of language policy under this rubric of multiculturalism. I shall now look in turn at broadcasting policy, education and language services, to see the growth and conflation of some of the interests that would lead eventually to a national language policy. 23.3 Broadcasting policy and languages other than English

It is the area of broadcasting which has perhaps most controversially brought issues concerning languages other than English (LOTEs) into politics, given the heavy degree of government regulation and involvement in all aspects of Australian media. With restrictions on broadcasting in foreign languages throughout the 1950s and 1960s, such broadcasting was marginalised and by the early 1970s was generally heard on only a few, usually rural, radio stations. Broadcasting policy, however, was to be an area of keen interest for the Whitlam Labour government (1972-5). Firstly, it considerably expanded and diversified existing broadcasting modes and outlets, created a new Public Broadcasting sector and an FM sector, and gave new licenses under already existing powers. Migrant groups were to take advantage of several of these developments.

National language policy and planning

335

Secondly, in order to counter what it saw as the Australian Broadcasting Commission's (ABC) elitist and culturally outdated role, it created new radio stations for the ABC to assume new roles. The new Melbourne station, 3ZZ, was specified as an access station, attracting a variety of social groups with often non-mainstream points of view. With their relatively good organisation and access to a constant audience, ethnic groups became one of the dominant presences on 3ZZ, eventually occupying over half the air-time of the station (Dugdale 1979). Finally, following the activism of Grassby, the government set up its own radio stations using migrant languages. These stations, 2EA in Sydney and 3EA in Melbourne, had at first a primarily information role, to inform migrant groups about new government programmes, for example, health insurance, Social Security, learning English, etc., but they also gradually developed more general programming in LOTEs, but with less direct community involvement than did 3ZZ. The Fraser Liberal government (1975-83) followed up these initiatives with some innovations of its own. It closed the politically uncomfortable 3ZZ, but expanded the role of the EA stations, and slowly allowed a greater community input into the appointment of broadcasters and programme coordinators. The EA stations were entirely devoted to LOTE broadcasting, with programme time allocated roughly in line with the proportion of speakers in the population, but with a minimum time allotted to even the smallest groups. A small news unit supplied material that could be used by the LOTE programmes (Special Broadcasting Service 1979). With increased commitments to maintaining and developing broadcasting in LOTEs, both major political parties at the 1977 elections promised 'ethnic television'. This initiative was strongly supported by the Galbally Report and became a politically highly visible and much contested issue from that time. The Government sought tofindan appropriate mechanism to control ethnic broadcasting and, with the ABC reluctant to take this on, it established the Special Broadcasting Service (SBS) in 1978 with responsibility for the EA stations and powers to develop 'multilingual television'. There were few models anywhere in the world, however, that gave guidance as to what 'ethnic television' would be. As with 'ethnic radio', would this mean a television station catering largely to particular ethnic groups and involving them in programming? Over the years 1978-80, whqn the television station did finally appear, a subtle change took place in its description. The station was defined no longer as 'ethnic', but as 'multilingual' (the more neutral definition in the legislation). Finally and most publicly, as 'multicultural television', it appeared in a form far removed from that of ethnic radio. Beginning transmission in Sydney and Melbourne in October 1980, Multicultural Television Channel 0/28

336

ULDIS OZOLINS

provided programmes largely from overseas. The majority of these were in LOTEs and subtitled in English by the station's own subtitling unit. An English-language news and current affairs service was also provided, often presenting issues from aspects of multicultural life in Australia. A small production unit made and contracted local programmes. For the Government, 0/28 provided both a highly innovative and well-received experiment in broadcasting, and a policy headache, with virtually continual battles over the appropriate forms of control over the station and its programming format (Leong 1983). While the station has attempted to cater to a variety of tastes and interests, some of its programmes have been accused of being too similar to those of the ABC. The ABC at one stage even coveted SBS television, which under ABC control was envisaged as a Channel 4 type station including multicultural and LOTE programmes among other quality programmes. From another direction, the commercial television networks which had provided a few hours of programmes including some LOTEs (variety/travelogue programmes) gave these up at the end of 1984, arguing that these needs were now adequately catered for by SBS (Patterson 1986). SBS thus often found itself in a political cross-fire. For the migrant communities, Channel 0/28 represented a highly significant recognition of the importance of other cultures and languages in Australian society. At the same time, the lack of community input into the organisation made them critical of SBS. By the early 1980s, Ethnic Communities Councils felt that the EA network was inadequate for the variety of radio needs among ethnic communities, and moved to make greater use of public broadcasting outlets. Because of quite significant shifts from time to time in wider broadcasting policy, broadcasting in LOTEs has struggled to maintain a clear place for itself politically and bureaucratically, even if it has achieved a by now significant presence in the broadcasting spectrum. In 1986 the Hawke Government decided, ostensibly on grounds of cost, to give control of SBS to the ABC, a decision that provoked a furious reaction. Once more SBS became the subject of party politics: under intense pressure, and not wanting to cede the running of ethnic affairs to the Liberal Party once more, the Hawke Government spectacularly reversed this decision on the eve of the 1987 elections, and moved after that to legislate for SBS as a permanent body. This move has been the first promise of a secure future for SBS after a decade of precarious existence. While this has ensured the survival of SBS into the 1990s and the loss of some of its marginal status, new government imperatives of privatisation and increasing economic rationalism have also imposed the need for SBS television to find sponsorship (as distinct from advertising) to supplement its federally provided budget. SBS, an initiative justified on grounds of

National language policy and planning

337

multiculturalism, is now inextricably caught up in wider mainstream issues of public and private broadcasting. 23.4 Education policy, multiculturalism and migrant languages After the highly political and visible area of broadcasting, education was the most significant area of public policy to be affected by changing responses to the migrant presence. Martin (1978) documents a diversity of definitions of just what 'migrant education' should be from the very beginning of Child Migrant Education in 1971. The problems identified in the 1960s of poor school performance and low self-esteem of migrant children now suggested a variety of responses to practicing teachers. She (1978: 125) points to five distinct areas that developed out of concern for migrant education: the teaching of English; bilingual education; teaching of community languages, multicultural education; and ethnic schools. In turn, each of these areas has seen accelerating development and policy shift over the last decade. I shall leave aside Martin's category of 'multicultural education*, which dealt mainly with non-language issues. The teaching of English programmes have during this time remained the single largest commitment by Federal and State governments to the education of migrants, and provision of effective English teaching has always been insisted on by ethnic organisations. This aspect of migrant education was never established in Australia under the rubric of'bilingual education', and always maintained its essentially monolingual methodology. The sheer diversity of languages among the migrant population, with no one clearly dominant linguistic group, was a compelling argument for this choice. Yet the assessment of outcomes of Child Migrant Education has continually been scanty and guarded. In a series of reports reviewing the achievement of English as Second Language programmes and their desired future development in the early 1980s, Campbell (1984) drew attention to the inadequacy of current provision. Most radically, he pointed to shortcomings in English as a Second Language methodology even among specialist teachers and facilities, and the need to take more seriously the learner's first language. He called for greater utilisation of bilingual methodologies and care for the continuity of conceptual development of all students (Campbell 1984). Bilingual education has, pace Campbell, usually been recognised in Australia as an enterprise distinct from English as a Second Language teaching, even where the aim of bilingual teaching has been essentially transitional. To date, while there have been some isolated bilingual experiments in relation to migrant languages, the most comprehensive programmes have developed since 1973 in relation to Aboriginal education, where language maintenance has usually been an important

338

ULDIS OZOLINS

objective (see McConvell, this volume). Bilingual education involving migrant languages evolved first out of concern for the educational disadvantage of migrant children. It was tried in some schools of high migrant density in the early 1970s (Claydon 1973, 1975). Rado's work over the last decade has developed bilingual materials for these programmes, and her field research on bilingual programmes has shown language maintenance to be a goal valued by the participants themselves which leads to more successful overall educational outcomes (Rado 1975; Rado and Foster 1984). More recently, some schools have tried bilingual education for language enrichment, for example in the Bayswater project in Melbourne, where bilingual education in English and German has been used by a number of primary schools with only a small minority of German-speaking children (Clyne 1983). A small number of schools have also introduced bilingual methods for teaching newly arrived children (especially Chinese and Vietnamese) or longer-established groups who value language maintenance, for example, Greeks. While these relatively scattered experiments in bilingual education have in some cases been impressive, reviews of multicultural education have stressed the minimal impact that bilingual education has had in terms of influencing education practice and reaching significant numbers of children (Cahill 1984; Australia, Department of Immigration and Ethnic Affairs 1986). However, the effect of these programmes has extended well beyond their still limited reach in numbers of children affected. It has become a serious proposal in policy discourse, raising often controversial questions of methodology and educational priority. The teaching of community languages has become perhaps the most prominent area outside of English as a Second Language when considering the diversity of'migrant education'. It has been an area in which the policy imperative of multiculturalism can perhaps be most clearly exemplified in providing an element of migrants' own culture as part of school curriculum. Initiatives stemming from Grassby's time as Minister for Immigration led to the formation in 1974 of the Mather Committee to investigate explicitly the teaching of migrant languages in schools. Its report noted the general decline of language learning in secondary education, and the paucity of language learning in primary schools. Only about 8 per cent of primary schools provided such programmes, with some whole systems of primary education providing no language teaching at all. The report also noted that the range of languages taught was still extremely limited, and this had direct consequences for NESB children wanting to study their own languages. Only 1.4 per cent of NESB primary and 10 per cent of NESB secondary students were studying their first language. At this time, more primary-age school children studied their language in ethic schools than in their primary schools (Australia, Committee on the Teaching of Migrant

National language policy and planning

339

Languages in Schools 1976). The report made wide-ranging recommendations on teacher-training, the need for greater availability of languages at the higher secondary and tertiary levels, recognising the importance of language maintenance for NESB populations and the value of learning migrant languages as a general educational objective. While the report was not directly acted upon by the government, all of its major concerns were taken up by a variety of other institutions in the late 1970s. From 1976 onwards there was increasing activism on this front from the Schools Commission, a body charged with administering Federal funds to the various school systems. Following Gallbally's recommendations, the Schools Commission funded a Multicultural Education Programme (MEP) and adopted a guideline that half of the funds of this programme should be directed towards the teaching of community languages. Teacher training issues were also addressed from the mid-1970s, and for the first time there was training for language teachers in primary schools (Whitton 1975), as well as for bilingual teachers in secondary schools (Lewin-Poole 1977). The range of languages offered to secondary students grew to include all the widely-spoken languages in Australia. State education departments also became more involved in the field in the late 1970s, initiating curriculum development and in some cases specifically appointing community language teachers to develop programmes. As a result of pushing by ethnic communities and language teachers, the range of languages offered for final end of school examinations expanded until it gradually encompassed most migrant languages. In its 1980 review of multicultural and migrant education, the Australian Institute of Multicultural Affairs (AIMA), an independent policy advice body, recommended the expansion of language offerings in tertiary education and argued the need to develop these languages at the highest level to give them status and give their students a chance of educational continuity (Australian Institute of Multicultural Affairs 1980). Finally, the early 1980s saw the development in several States of comprehensive policies on language education, to which I shall return shortly. Finally, Martin points to the increased salience of ethnic schools in educational discourse. These had previously been conducted almost in secret by ethnic communities, and at times were regarded with suspicion by the regular education systems for distracting pupils from their day schools and above all from the learning of English. These ethnic schools were, however, treated sympathetically by the Mather committee, the Schools Commission and subsequent bodies, all of whom called for increased cooperation with this sector and attention to resource needs. AIMA recommended in 1980 that these schools should receive a Federal grant of A$30 per student per annum in the light of their massive contribution to language learning (AIMA 1980). Norst's (1982) study attested the great

340

ULDIS OZOLINS

vitality of these schools and in particular, their rapid growth in almost all language groups in the previous decade. To make a radical break from previous perceptions, Norst also recommended that these schools be renamed 'Community Language Schools', an interesting retreat from the description of'ethnic'. This was treated coolly by some of these schools, which were concerned with teaching not only language but also cultural and in some cases religious aspects which they felt could not be expected to be handled appropriately by the regular school systems. 23.5 Language services in Australia Language services were developed considerably from the early 1970s, both in terms of direct service provision, and regulation and planning. Following pressures to respond to migrant settlement difficulties at this time, the Government's first significant step was to provide an innovative service through the Emergency Telephone Interpreter Service. The service opened in February 1973 to provide a facility for interpreting (originally in eight languages) over the telephone in broadly defined situations of'emergency', for example, medical, police, critical welfare and hazard. The system allowed for conference calls and, as its equipment and range of languages improved, a considerably more sophisticated system of hook-ups and referrals. The 'Emergency' was dropped from its title in 1974, and thus, the Telephone Interpreter Service (TIS) became a general interpreting resource for community, as well as business and commercial needs. To supplement its telephone work, TIS also provided an on-call service of personal attendance by interpreters in situations of community interpreting (medical, legal) which could be booked in advance. The provision of this general interpreting service, which by the mid1980s was responding to nearly 200,000 calls per year, also provoked increased interest in specialist interpreting services, for example, legal, health, mental health and education. Throughout the late 1970s and early 1980s a number of specific services were established, often with federal assistance as a consequence of the Galbally report. Regulation of the interpreting profession was considered after the initial services were established and were finding difficulties in attracting suitable staff. Worries over the quality of interpreting in self-help situations and through private agencies also continued to exist. The Committee on Overseas Professional Qualifications, charged with determining the suitability of overseas-trained personnel for education and employment in Australia, considered the problems of interpreting. It noted that while interpreting was a recognised profession overseas, there was no system of training, accreditation or professional recognition in Australia. There was an urgent need to create a profession that not only met obvious interpreting

National language policy and planning

341

needs vis-a-vis NESB migrant communities, but also provided for Australia's external interpreting needs. The Committee proposed a gradation of interpreting levels and recommended accreditation and training procedures to ensure qualified practitioners (Australia, Committee on Overseas Professional Qualifications 1978). The first training courses were established concurrently with the Committee's deliberations in 1974-5, and in 1977 the Federal Government established the National Accreditation Authority for Translators and Interpreters (NAATI) to take on formally the work of establishing and regulating the profession. N A ATI's focus has been on accreditation and training for the whole range of languages in Australia, including Aboriginal languages. Its strategies included establishing a testing programme for candidates at various levels, getting recognition on the part of government and private bodies of the need for qualified practitioners, establishing minimal criteria for employment, and increasing training facilities to supply qualified interpreters. Language needs have also been met in the public service by the wider use of bilingual staff, formalised at the federal level through the granting of a language allowance for officers who use a LOTE in contact with the public in their work. This is linked to NAATI accreditation levels, thus providing an essential backup service to regular interpreting services. Historically, Australia has also suffered from a severe shortage of language resources for diplomacy, trade and foreign relations. For example, Australian embassies particularly in Europe and Asia have been heavily reliant upon the hiring of local nationals to perform interpreting and some public contact duties, a point that has been often raised in parliamentary debates on the diplomatic service. This has constantly raised the question of why the rich linguistic resources brought to Australia through immigration have not been put to effective use in Australia's international relations. As there was general reluctance to post the few diplomatic staff born in other countries to that country (even if they spoke that particular language), and given the generally low level of recruiting of speakers of other languages, the problem remained an enduring one for the diplomatic service, despite the incentive of a language proficiency allowance (Hall 1959; Australia, Senate Standing Committee on Foreign Affairs and Defence 1979). Apart from the incentive of the language allowance, the diplomatic service did make use of some overseas and Australian language schools, including from time to time the intensive language training facility at the Royal Australian Air Force at Point Cook. Based upon similar models of military language academies overseas, this facility was created in World War II to meet critical language shortages, particularly in Japanese, Russian and South-east Asian languages. It continued in peacetime to meet

342

ULDIS OZOLINS

the changing language needs of defence and, increasingly, other governments departments (Turner 1983). Also in the 1950s and 1960s the Federal Government expanded its own Australian National University's facilities in Asian and other major international languages. Concerns about external uses of languages were important in moves to establish a national language policy. With increasingly diversified and in particular increasing Asian migration in the late 1970s and 1980s, the distinction between 'migrant' languages and 'strategic' or 'Asian' or 'foreign' languages began to be less clear cut and came increasingly to be questioned. Australia had to husband all its language resources, for both internal and external uses. 23.6 Moves for a National Policy on Languages

Over the past decade, Australia's distinctive contribution to language policy has been the evolution of a National Policy on Languages. Reflecting many of the policy concerns already mentioned, the move for such an overarching policy also brought with it new coalitions of language interest groups, and to some extent moved concern for language into a broader context and away from migration issues discussed so far. Moves for a national policy arose in the late 1970s. A number of language interest groups began to push for greater coordination of disparate policies, and a much better organised set of language professional groups and ethnic and other organisations began to identify common concerns and push for government response. The foundation meeting of the Applied Linguistics Association of Australia (ALAA) in 1976 called for a national language policy, and a joint submission from ALAA and the Australian Linguistics Society in 1978 requested the government to establish a committee to advise on language policy. Individuals such as Ingram (1979) and Quinn in a series of editorials in Babel, also actively promoted this cause (Ozolins 1984). The professional language associations at this time had several motives in becoming interested in language policy at this level. First, there was widespread recognition of the continuing decline of language teaching in schools, despite the newfound interest in migrant languages and the persistent efforts to promote the teaching of Asian languages in schools. Secondly, the various language professional groups also found a common forum to meet and exchange views in ALAA, which brought together groups concerned with diverse issues, for example, Aboriginal languages, second language learning, English as a Second Language, and language services. This diversity of interests was to have a strong influence on the ultimate scope of the national policy. Finally, the linguists moved closer to the political arena by recognising joint interests with the ethnic com-

National language policy and planning

343

munities, with whom some prominent individuals, for example, Clyne, had had a long involvement. This united front of language professionals and a political lobby group such as the Federation of Ethnic Communities Councils (FECCA) came together most decisively in 1981, when at their annual congresses both organisations placed national language policy as top priority, and organised a detailed campaign to convince the Federal Government to act. FECCA lobbied the Government and set up a series of forums through its State organisations on language policy. ALAA organised among the language professional associations, and a joint body was formed - the Professional Language Associations for a National Language Policy (PLANLangPol), which in parallel to FECCA's activities also organised forums to discuss issues for a national language policy. The work of these coalitions and their successful lobbying has been discussed elsewhere (Ozolins 1985). The language lobby focussed particularly on the Federal Department of Education, concerned as it was with aspects of many of the issues canvassed by the groups, as well as the Department of Immigration and Ethnic Affairs. The Department of Education serviced the Senate Standing Committee on Education and the Arts, whose Chairman, Baden Teague, was well disposed towards language policy, and there was general endorsement of this issue as one suitable for the committee. The Senate committee took on this issue in May 1982, and the Department of Education issued a comprehensive background paper Towards a National Language Policy' at the same time (Australia, Department of Education 1982). There were several striking features of this concerted move for a language policy that gave the eventual Senate inquiry and subsequent developments their peculiarflavour.Most of all, these moves for a national policy covered aspects of all languages used in Australia, and importantly, this also meant English. While in the late 1970s the main focus for policy had been on addressing issues to do with languages other than English, the increasing cross-fertilisation of interests and the ubiquity of Englishrelated concerns gradually led to a change in what the envisaged policy would encompass. Also important here was the strong desire to take the national policy away from minority concerns and bring it into the mainstream of public life, which necessarily meant dealing with English. Thus adult literacy, the status of Australian English, English as a mother tongue, racism and sexism in language, as well as English as a Second Language, became part of the scope of the Senate Committee's inquiry. This made the Australian national policy initiative very different to such initiatives as the President's Commission on Foreign Languages and International Studies in the United States, which was oriented largely to foreign language teaching interests.

344

ULDIS OZOLINS

The Senate Committee received submissions from some 230 parties, held 20 days of public hearings, and deliberated for over two years before finally reporting in December 1984. The submission process itself was of great importance, giving an exhaustive panorama of the state of various language issues at the time. While many submissions pursued their sectional interests, most stressed the need for comprehensive policies. There was a general view that a national policy must look adequately at all language needs in Australia, and be mindful of the many diverse social functions played by language. The Committee's report, A National Language Policy (henceforth ANLP) was favourably disposed towards a great many of the issues placed before it. Because it was cautious in recommending items of large expenditure, it concerned itself largely with establishing principles for further action and indicating areas of greatest need. It recommended that a national language policy have the objective of guaranteeing the following to all Australians (ANLP: Recommendation 1): competence in English maintenance and development of languages other than English provision of services in languages other than English opportunities for learning second languages Beyond this, the report was rather uneven in its consideration of the various areas. Surprisingly perhaps in the light of the importance of English in the background to the inquiry, its sections on English were relatively incomplete. It deferred to much other recent work on English as a Second Language, but in the teaching of English as a mother tongue it showed little feeling for the issues, and perhaps lacked expertise in the area. The report was not, and could not hope to be, Australia's equivalent to the Bullock Report (1975). It was, however, more forceful on issues of adult illiteracy among English mother tongue speakers (ANLP: chapters 3-6). The most impressive and certainly most impassioned section of the report was on Aboriginal languages. Acutely aware of the uniqueness of these languages, and their current fragile situation, the report urged a comprehensive range of initiatives, especially the development of bilingual education and the better representation of Aboriginal interests in media planning. The central role of community consultation in all aspects of Aboriginal languages was repeatedly stressed (ANLP, chapter 8). The section relating to migrant languages, particularly chapter 11 on Teaching Languages Other than English, was by comparison rather mild and cautious in its conclusions, reflecting the considerably greater degree of division and differences of priority within the field on specific areas of policy and methodology. Thus, while it generally supported moves to increase language learning in schools and recognised the importance of

National language policy and planning

345

language maintenance to ethnic communities, it was more equivocal on specific measures such as proposals for bilingual education. Here, submissions and representations to the Committee had revealed important differences in attitudes to the desirability and practicality of bilingual programs. Goverment departments and agencies tended to be selective in their support, seeing it as possibly suitable for Aboriginal and recently arrived students, especially adolescents, and seeing the aim of such programs as essentially transitional. Some dismissed the possibility of effective maintenance programmes, but this was strongly contested by other submissions, particularly the Ethnic Communities Councils and practitioners from the few bilingual maintenance programmes that had been established in Australia. In relation to external language needs, the report noted considerable complacency among government bodies and businesses in Australia, who in submissions to the Committee had tended to play down the issue of language and were reluctant to identify needs and goals. The report warned this was a short-sighted view that would harm Australia in its relations with the outside world. The Senate Committee report, however, had little immediate follow-up, and delays in its release and subsequent response meant that many language activists devoted themselves to other areas of language policy that seemed to bring more fruitful response. However, the concern for a broadly based national language policy also affected the way these issues were now addressed at State levels. In Victoria, for example, languages gained increased prominence through the Labour government's appointment of supernumerary teachers to teach community languages in primary schools, as a direct result of its multicultural education election pledges. Also in Victoria, a working group headed by Lo Bianco produced a document in 1985, The Place of Languages Other Than English in Victorian Schools', that outlined needs across the primary and secondary sectors for language teaching, and set up principles of operation of the expanding number of programmes in primary schools. It stipulated, for example, that at least 3 hours per week needed to be devoted to language study if courses were to be worthwhile, so that language programmes would not be token nods towards multicultural policy, but legitimate school subjects demanding commitment (Victoria, State Board of Education and the Ministerial Advisory Committee on Migrant and Multicultural Education 1985). Similar policies were developed in other States, and an active National Advisory and Coordinating Committee on Multicultural Education provided policy initiatives at the Federal level as part of the Multicultural Education Programme (National Advisory and Coordinating Committee on Multicultural Education 1987). In these State moves it was important that the issue of language teaching

346

ULDIS OZOLINS

was defined as the teaching of languages to all pupils. It was constantly stressed that the teaching of languages was not a concern for NESB pupils and their parents alone. The Victorian 1985 policy document made the important symbolic break of moving from reference to 'community languages' to the broader languages other than English', the term used in all subsequent documents (Lo Bianco 1989). Alongside these State initiatives, language interest groups continued to push for a response to the Senate inquiry. Responding to this pressure, Education Minister Ryan in 1986 appointed Joseph Lo Bianco to advise what steps should be taken to realise a national policy on languages. The Lo Bianco report (Australia, Department of Education, Lo Bianco 1987) was a focused, implementation-oriented document compared to the more philosophical Senate Committee inquiry. Lo Bianco tried to establish clear guidelines for priority initiatives and monitoring of future language policy. It accepted the direction of much of the Senate Committee's recommendations, and devised four overall strategies to guide language policy (Lo Bianco 1987: 70): the conservation of Australia's linguistic resources the development and expansion of these resources the integration of Australian language teaching and language use efforts with national economic, social and cultural policies the provision of information and services in languages understood by clients. The Lo Bianco report attended to implementation strategies for all the areas covered in the Senate Committee report, but also added particular emphases of its own. In education, while arguing the importance of all languages and supporting the efforts needed to maintain and develop them, it took a strong stand on priorities and identified nine 'languages of wider teaching' that would be targeted for major curriculum development, teacher training and concentration of resources. The nine languages were Standard Chinese, Indonesian/Malay, Arabic, French, German, Greek, Italian, Japanese and Spanish. Lo Bianco also gave a stronger emphasis than the Senate Committee report to external language needs, as reflected in the mix of priority languages, pointing out that many of the languages important externally to Australia were also important in social life in Australia. As a national monitoring and policy development organisation, Lo Bianco recommended the establishment of an Advisory Committee on Australian Language Policy. The report was also careful to take account of Australia's system of federalism and included statements from all States and Territories on their own language policies, as well as recognising State initiatives in recommending that under the proposed Australian Second

National language policy and planning 347

Language Learning Programme, States develop major curriculum and material projects in certain languages on behalf of other States. Lo Bianco was careful to stress that while language policy was an integral part of multicultural policy, it was not exclusively a matter of multiculturalism alone. He presented language policy as an interest in its own right, with importance both to Australia in its internal affairs as well as relations with the rest of the world. It was this faith in the integrity of language policy per se that was to be tested in subsequent developments. Whatever Lo Bianco's intentions, however, the report was born in a period of intense political conflict, and the subsequent developments in language policy have been hostage to political concerns in ways perhaps little envisaged previously. The report served, in turn, as both a political saviour for the Hawke government, and later as an almost unwanted obstruction to the Government's own views on language. The report was released at a time when the Government, through a series of cutbacks in spending in its 1986 budget, had cut English as a Second Language programme, axed the Australian Institute of Multicultural Affairs and the Multicultural Education Programme and, as already mentioned, moved to amalgamate the SBS with the ABC. This created a furious reaction in the community, and slowly over the next six months the Government reversed most of these decisions. When it called an election for May 1987, Prime Minister Hawke finally agreed to continue SBS as a separate entity, and released the Lo Bianco report as an important element in his government's ethnic affairs policy. This effort was aimed specifically at winning back the vote of ethnic groups - an ironic turnabout for a policy which Lo Bianco had tried to de-ethnicise. The government promised to implement the National Policy on Languages, and fund it for A$28 million for a period of three years, and upon its re-election moved to substantiate this promise. The government set up an advisory committee, but with one notable change that distorted Lo Bianco's original intention. The body set up was the Australian Advisory Council on Languages and Multicultural Education (AACLAME). Multicultural education was tacked on to offer the Government a way of recovering from the 1986 cuts (Ozolins 1988). AACLAME started operation in late 1987, and had under its umbrella five major programmes: the Australian Second Language Learning Programme, the National Aboriginal Languages Programme, the Adult Literacy Action Campaign, the Multicultural and Cross-cultural Supplementation Programme, and the Asian Studies Programme. After the 1987 elections and the founding of AACLAME, however, Government rhetoric began to change quite dramatically, with both Hawke and Education Minister Dawkins (responsible for AACLAME) very publicly promoting Asian languages, and, in the case of Dawkins,

348

ULDIS OZOLINS

expressing little interest in any other languages. With economic arguments very much to the fore, this stance on the part of leading politicians also reflected a growing split between Asian language interests and those of other languages. This led to a deterioration of the previous coalition among language interest groups and common interest in a broad national policy. The situation became so openly polarised that Hawke was compelled to reassure FECCA at its 1988 Congress that the Government would not make 'artificial distinctions' between trade languages and others. However, with Hawke and Dawkins promoting exactly this distinction elsewhere, it seemed to be the case that ethnic affairs interests relative to languages would be promoted at election times, and economic interests the rest of the time. This pattern was to be repeated at the 1990 elections, with the narrowly returned Hawke Government promising, as part of its ethnic affairs policy, the continuance of a national policy on language. With the conservative parties seemingly losing interest in this issue and in ethnic affairs generally by this time, the Hawke Government was able to capitalise on these issues in a closely fought election: indications of long-term commitment, however, are more difficult to ascertain. AACLAME thus had to battle to assert itself against a relatively unsympathetic Minister and Government. Despite this, AACLAME continued to speak for language policy, particularly through its publication Vox', and by vigorously pursuing a number of issues in its broader role in policy advice. In 1989, part of its budget was allocated to create a Languages Institute of Australia, to be run on a collaborative basis utilising expertise in language research, language testing and language technology in Australia's major tertiary institutions, with a central secretariat in Melbourne. AACLAME itself, however, with funding guaranteed for a three-year period only, faces an uncertain future. This survey has shown the very strong development of language policy in Australia in many areas, and the patterns of response to migrant languages over four decades of mass migration. Migrant languages have established themselves in Australia as living languages but also importantly as politically active languages. While the vitality of these languages must primarily be a question of the efforts of the particular language communities themselves, the development of detailed policy in so many areas that affect their public and institutional use ensures that language policy will remain a critical intervening variable in their maintenance, change or decline. Future generations will be dependent upon the effectiveness of these policies to ensure that these languages continue to have adequate status and support in the Australian context.

24 Social class differences in the lexicon DAVID CORSON

24.1

Introduction

In this chapter I will present evidence to show that there is little difference between secondary-school children from various social backgrounds in England and Australia in terms of access to high status and specialist lexes of the language. The written language of Australia and its specialist oral counterpart has developed in parallel with the written language elsewhere. Turner (1966: 67), for example, writes: 'the coincidence of English and Australian written idiom is not a natural phenomenon, but is cultivated by literature in its widest sense'. For much of Australia's history, most books and even most magazines encountered by children were written and produced in the United Kingdom. More importantly perhaps, the examples of diction found in those sources were routinely pointed to as exemplars of 'good English' to be adopted in schools and in later life. Increasingly too in the twentieth century, Australian English made contact with other branches of English. Ramson (1972: 44) points to the lexicon of the scientist as the greatest single influence on the vocabulary of English: 'And it matters little whether the scientist is English, American or Australian. The specialist vocabulary of his particular branch of scientific enquiry acknowledges not regional but professional boundaries.' Later Ramson highlights the immediacy in contemporary communication between the different parts of the English-speaking world. So completely do occupational vocabularies today cut across regional boundaries, that the distinctively Australian element in our vocabulary has become proportionately far smaller than it once was. Linguistic stratification, even across continents and oceans, has increasingly become a function not of region but of one or another sociological variable. In particular, social class, as determined by occupation and income, has become a vital factor in setting the boundaries of shared linguistic features. This has not always been the case. Pre-Industrial Revolution social relations, in countries of the Old World, were vertical

350

DAVID CORSON

ones. Peasants, as a result of their shared experiences and situations, had more in common with those above and below them in the social hierarchy than with their peers elsewhere, and this commonality extended to the language available to them. In this chapter I explore differences in social group lexes drawn from Australia and England. My aim is to demonstrate that Australian and English social group lexes have a great deal in common. It will also be possible to show that, at least in respect of the high status specialist lexicon of the secondary school, there is little difference between English English and Australian English. In presenting this comparison I shall draw heavily on work assembled for another purpose (Corson 1985). These studies were originally conducted for reasons of educational research and were specifically aimed at isolating a key linguistic mediating factor in educational failure. It follows, then, that in the process of introducing this material I shall inevitably be making statements germane to educational success and failure. 24.2 Social group lexes Social group lexes have no status as linguistic subsystems. They simply refer to the range of vocabulary, either passive or active, that is available to the members of identified social groups. The more attributes that are combined to state the sociological boundaries of the social group, the more specific the lexes pertaining to the group will be. The studies reported here and elsewhere (Corson 1985) indicate that there are different kinds of lexes available to children from different sociological backgrounds and the instruments described below are designed to measure some of those differences. This fact is consistent with other studies of linguistic variation. Labov's (1968, 1973) influential studies show that people from different sociological backgrounds speak different kinds of English which in many respects deviate systematically and regularly from one another. 24.2.1

Two approaches to measuring social group lexes

The instruments described below were first used among Yorkshire and London adolescents to measure oral lexes. Later they were reapplied in the Australian context to measure the written and spoken lexes of certain groups of children from Sydney and the Illawarra region of New South Wales. 24.2.1.1 The measure of passive vocabulary The Measure of Passive Vocabulary examines access to certain semantic fields by presenting orally 49 words, each bracketed with a keyword, to

Social class differences in the lexicon

351

subjects who are invited to use each pair of words in one original sentence. Each keyword suggests the semantic field relevant to the target word. Seven fields are dealt with representing the seven educational knowledge categories identified by Hirst (1974), as representative of the map of specialist knowledge spanned by the contemporary school curriculum. Hirst's map of the curricular knowledge categories was adopted in these studies because of its wide (though not total) acceptance by educationalists. His scheme for categorising knowledge and experience is derived from the different ways that statements are formulated in each mode; each mode has its own logic for testing statements made within it. Hirst argues that while these seven modes are not constant or eternal, they do provide a mapping system for specialist knowledge as it is presently perceived. His headings for each of these seven modes, in slightly modified form, provide the themes for the seven semantic fields into which lexemes are grouped within the study. These fields are: formal logic and mathematics; the physical sciences; the human sciences; ethics; aesthetics; religion; and philosophy. Each sentence produced, if meaningful within the relevant semantic field, confirms the subject's grasp of the target word. Whether or not the target word is embedded only in the passive vocabulary of the subject, the context encourages its public articulation. The Measure itself performs the retrieval process. The words are delivered to subjects; they hear them uttered by another individual; their use by subjects is appropriate to the context of situation; the topic is apparently left to the subjects themselves to choose; and moderate encouragement to provide original sentences is resident in the relatively interesting experience of making up sentences from given component words. The measure relies on the postulations of the 'use theory' of meaning: knowing the meaning of a word is knowing how to use it in an original utterance. The meaning of a word is its use in a language. 24.2.1.2 The Graeco-Latin Instrument The Graeco-Latin (G-L) Instrument can be applied to either oral or written language. The researcher examines texts in order to measure the percentage of specialist words that entered English ultimately from Greek and Latin sources. Since these words are of a type that fill up the specialist knowledge categories of the secondary school curriculum, almost to the total exclusion of other types of word, the G-L Instrument is a complementary measure to the Measure of Passive Vocabulary. It measures active access to the specialist semantic fields. The application of the Instrument is a relatively simple matter, since the majority of G-L words in the language are unambiguous. Their origin is clear from their form, which usually betrays no change from the original apart from a modification of the suffix. Non-specialist words of G~L origin, whose antiquity and prevalence in English is such as to make them

352

DAVID CORSON

part of necessary, everyday language, are removed from the count by the operational rules detailed elsewhere. Those excluded words are mainly preRenaissance additions to the language that are chosen for equal use by language users in speech regardless of social class background. There are at least six features which attach to specialist words extracted by the Instrument which make them lexically 'difficult': (i) they possess a connotative meaning which is not extensional; (ii) they are often without suitable synonyms and may be defined only by the use of a number of other words and even then perhaps poorly; (iii) they give precision to texts and these may convey personal meaning more effectively than texts which lack them; (iv) they are not readily inserted in an abstraction ladder of superordinacy; (v) they allow their users to order thought where such an ordering of thought might not occur without the words themselves; (vi) they may be 'culturally determined' in that they represent meanings which are rarely readily translated word for word into the languages of other cultures. Plainly in the upper levels of education, as presently conducted, children cannot manage without a wide and ready access to specialist G-L words, although this tells us nothing about other words that might be distinctively common in the ordinary language lexes of those who are without a motivated specialist G-L vocabulary in the context of the school. In the Australian studies, both oral and written language were examined within two contexts: a descriptive task and an explanatory task. In the English studies only oral language was examined, again in the two contexts. Spoken or written texts, once collated, were transcribed and analysed using the G-L Instrument. The spoken texts were tape-recorded in response to an orally administered questionnaire. The written texts were gathered, after several weeks' interval in response to the same questionnaire in written form. 24.3 The research contexts

Children of 12 and 15 years were compared in each setting to measure development in adolescents who were on either side of the notional onset of Piaget's stage of formal operational thought, which tends to be first manifested in language use at adolescence, if at all. The children were matched in groups taken from much wider samples on the basis of their non-verbal reasoning scores, and cells consistent with accepted sociolinguistic cell size were drawn from the groups using the technique of 'systematic sampling'. The social class of children was determined by reference to the Registrar General's Classification of Occupations. There were no anomalous categories in applying this British register to the parents of children resident in Australia in this instance. Both parents' occupations were used wherever appropriate. All children

Social class differences in the lexicon

353

had experienced common curricula without a Classics component. Great care was taken in choosing subjects whose experiences in formal education had been relatively homogeneous so that any variations in scores between groups on the instruments would not be significantly due to educational inequalities produced by the institution of education itself. Four schools were used in the studies proper: School Y is located near a small country town which itself is between six and thirteen miles from three large south Yorkshire cities. Its student population of about 1,500 children is drawn one-third from the town itself and two-thirds from the many small villages within a five mile radius. The area includes light industrial works, some mining industry, small farming and agriculture. The school has a strong blend of children from middle class and working class backgrounds, but no children of non-British descent. School L is a neighbourhood school in north London. Its 1,100 children include a strong mixture of children from the various social classes. School W is on the fringes of the iron and steel foundries which provide direct or indirect employment for much of Wollongong's population. In the postwar period Wollongong became a boom town with the regional population passing 200,000, swollen by the highest immigrant inflow of any Australian city. In the immediate region of School W 32 per cent of the population were born in non-English speaking countries. The 1,000 students included a small group of upper-middle-class Anglo-Australian children, many children of Macedonian (Yugoslavian) parentage, many children of Italian or Portuguese parentage and about 40 per cent poorer working class Anglo-Australians. The children from non-English-speaking families were themselves poorer working class by family background. They used their parents' native tongue as the language of the home in early childhood and were all received into their primary school in the 1970s as English as a Second Language students. School S is a prestigious but 'non-selective' private school for boys, set among the wealthiest of Sydney's northern suburbs. Because of its high fees, its reputation, its long waiting-list for entrants, and its favouring of the sons and brothers of old boys, School S is very selective. It attracts students who are exclusively upper middle class by family background.

24.4 Social group lexes: England and Australia compared The range of social group lexes in Australia and England, whose differences might be of significance to researchers, is wide indeed. The limitations of this study, then, are clear. It looks at only two areas of England and only one part of Australia that includes two cities on the central to south coast of

354

DAVID CORSON

New South Wales. However, the findings are sufficiently clear-cut to allow certain inferences to be made which might be generalisable. 24.4.1

Passive vocabulary

Figure 24.1 indicates that there is an approximately monotonic development within the 12- and 15-year age period for all social groups in their command of passive vocabulary. Children from the same social backgrounds in Australia and England are very similar in their access to passive vocabulary on this measure regardless of age. Using the Pearson Test of Correlation and comparing nationalities on total scores on the .Measure, that is, English or Australian, a negative correlation coefficient ( — 0.13) is returned. This means that it would be very difficult to predict subjects' nationalities from their total scores recorded on the measure, or alternatively, to predict scores from nationalities. In access to individual semantic fields or knowledge categories there is also no discernible difference. The improvement by age on the measure is for each group an increase in passive vocabulary of about 7 to 9 words. This seems a good indication of the rate at which passive vocabulary increases for children over this three-year developmental span. The upper-middle-class children in both countries begin ahead of the rest and maintain that lead. At 15 years, the upper-middle-class Sydney group, for example, outscore the poorer working-class Wollongong group with very high significance in mean scores on the Measure as a whole and also in their use of vocabulary in the semantic fields of ethics and aesthetics and with high significance in the other f\NQ fields. The upper-middle-class Wollongong group, who have shared the same primary and secondary school experiences as their poorer working-class peers, outscore the latter at 15 years with high significance in mean scores on the Measure as a whole and also in the semantic fields of logic and maths, aesthetics and philosophy. No significant differences in mean scores occur when comparing the Macedonian and Romance groups with the upper-middle-class Wollongong group. Some explanation of the Wollongong figures may be useful in answering the question: Why do the children of non-English-speaking poorer working-class backgrounds outscore their Anglo-Australian working-class peers so conclusively at 15 years? These immigrant children are certainly manifesting the effects of a heavy concentration on spoken English in their primary education through English as a Second Language courses which centred on the standard variety of Australian English. They may also be reflecting in their results on this measure an observable school achievement rate above their Anglo-Australian working-class peers which is due, in part at least, to strong parental influence towards success in schooling, common

Social class differences in the lexicon

355

MC (Sydney) MC (London)

45

MC (Yorkshire) MC (Wollongong) WC (London) ROM (Wollongong) MAC (Wollongong)

40 WC (Yorkshire)

WC (Wollongong)

35

15

12 AGE

Figure 24.1 Line graph illustrating group means recorded on the measure by age and social group (N= 129). Key. Australian groups . English groups . MC (Sydney)=upper middle class Sydney (private school) boys; MC (London) = upper middle class London (comprehensive); MC (Yorkshire) = upper middle class Yorkshire (comprehensive); MC (Wollongong) = upper middle class Wollongong (comprehensive); WC (London) = poorer working class London (comprehensive); WC (Yorkshire) = poorer working class Yorkshire (comprehensive); WC (Wollongong) = poorer working class Wollongong (comprehensive); ROM (Wollongong) = Romance Language background (comprehensive); MAC (Wollongong) = Macedonian language background (comprehensive)

among closely knit immigrant communities and observed among Wollongong's immigrant iron and steel foundry workers. Furthermore, it is likely that the immigrant children are benefiting in their English from the transfer effects to be gained by true bilinguals who have their linguistic awareness and competence enhanced at an early age when the two conceptual frameworks represented by two distinct languages are integrated into the world view of one individual. It is clear that there are some important and educationally serious contrasts in the passive vocabularies of the Anglo-Australian and English

356

DAVID CORSON

poorer working-class children relative to others which may affect levels of comprehension and performance in the setting of the school. The fact that the 15-year-old poorer working-class Wollongong group mean is below that of the 12-year-old upper-middle-class Sydney group represents an important educational contrast at the level of passive vocabulary. This difference indicates that the poorer working-class children, in their final stages of schooling, have yet to develop even a passive command of certain aspects of a specialist lexis to which presumably they have had four years exposure in schooling. 24.4.2

Active vocabulary

Figures 24.2-5 present social group mean scores on the G-L Instrument in the four contexts (oral language, descriptive and explanatory tasks; and written language, descriptive and explanatory tasks). It will be seen that onlyfigures24.2 and 24.3 are directly relevant to the theme of this chapter, since the other figures do not include comparative information from England. Figures 24.4 and 24.5 are included, however, because they indirectly bear upon the theme of this chapter and also provide wider information about the Australian lexicon. Fromfigures24.2 and 24.3 it can be seen that active oral lexical selections in England and Australia are very similar at the two ages. Moreover, at 15 years there is a remarkable and statistically very strong parallel between the Australian and English social class groups. Plainly it can be concluded that at this age on this measure lexical selections occur according to horizontal sociological arrangements and not vertical ones; region is less important than social group background. This point is especially clear infigure24.3, where poorer working-class children, who are not immigrants from nonEnglish speaking backgrounds, are disinclined to use the high status lexes of the school even in an explanatory context which would seem to demand a use of vocabulary of this kind. Correlations of the subjects' scores in the G-L tasks with their nationalities (i.e. English or Australian) produced coefficients of only 0.22 for the oral descriptive (figure 24.2), of only 0.07 for the oral explanatory (figure 24.3), and of only 0.15 for total oral scores on the Instrument. An examination of figures 24.4 and 24.5 will also reveal that the Australian children perform with marked consistency across the written and oral tasks. This judgement is supported by correlation figures. Scores on the oral descriptive task with scores on the written descriptive task, and scores on the oral explanatory task with scores on the written explanatory task, returned coefficients of 0.49 (p=