Small-Language Fates and Prospects : Lessons of Persistence and Change from Endangered Languages: Collected Essays [1 ed.] 9789004261938, 9789004230514

Citation preview

Small-Language Fates and Prospects

Brill’s Studies in Language, Cognition and Culture Series Editors Alexandra Y. Aikhenvald (Cairns Institute, James Cook University) R.M.W. Dixon (Cairns Institute, James Cook University) N.J. Enfield (Max Planck Institute for Psycholinguistics, Nijmegen)

VOLUME 6

The titles published in this series are listed at brill.com/bslc

Small-Language Fates and Prospects Lessons of Persistence and Change from Endangered Languages Collected Essays

By

Nancy C. Dorian

LEIDEN | BOSTON

Cover illustration: Looking to Achiltibuie from Polbain pier, Coigach, Wester Ross, Scotland. Library of Congress Cataloging-in-Publication Data Dorian, Nancy C.  Small-language fates and prospects : lessons of persistence and change from endangered languages : collected essays / By Nancy C. Dorian.   pages cm. — (Brill’s Studies in Language, Cognition and Culture ; 6)  Includes index.  ISBN 978-90-04-23051-4 (hardback : alk. paper) — ISBN 978-90-04-26193-8 (e-book) 1. Endangered languages. 2. Scottish Gaelic language—Dialects—Scotland. I. Title.  P40.5.E532S36 2014  491.6’3—dc23 2014002474

This publication has been typeset in the multilingual ‘Brill’ typeface. With over 5,100 characters covering Latin, ipa, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see brill.com/brill-typeface. issn ����-���� isbn ��� �� �� ����� � (hardback) isbn ��� �� �� ����� � (e-book) Copyright 2014 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill nv incorporates the imprints Brill, Brill Nijhoff, Global Oriental and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change. Brill has made all reasonable efforts to trace all rights holders to any copyrighted material used in this work. In cases where these efforts have not been successful the publisher welcomes communications from copyright holders, so that the appropriate acknowledgements can be made in future editions, and to settle other permission matters. This book is printed on acid-free paper.

For the MacRae family of Achnahaird, Coigach, and Golspie, East Sutherland. A hundred thousand thanks!

∵

Contents Foreword ix Sources xi Introduction  1

part 1 Language Change in an Obsolescent Language  31 1 Grammatical Change in a Dying Dialect (1973)  33 2 The Fate of Morphological Complexity in Scottish Gaelic Language Death: Evidence from East Sutherland Gaelic (1978)  66 3 Making do with Less: Some Surprises along the Language Death Proficiency Continuum (1986)  93 4 Negative Borrowing in an Indigenous Language Shift to the Dominant National Language (2006)  115

part 2 Speaker Skills and the Speech Community in a Receding Language Context  135 5 The Problem of the Semi-Speaker in Language Death (1977)  137 6 Language Shift in Community and Individual: The Phenomenon of the Laggard Semi-Speaker (1980)  146 7 Defining the Speech Community to Include its Working Margins (1982)  156 8 Abrupt Transmission Failure in Obsolescing Languages: How Sudden the ‘Tip’ to the Dominant Language in Communities and Families? (1986)  167 9 Age and Speaker Skills in Receding Language Communities: How Far do Community Evaluations and Linguists’ Evaluations Agree? (2009)  178 10 Linguistic Lag as an Ethnic Marker (1980)  193

viii

contents

part 3 Language Shift and Language Maintenance  203 11 Language Loss and Maintenance in Language Contact Situations (1982)  205 12 The Value of Language-Maintenance Efforts which are Unlikely to Succeed (1987)  223 13 The Ambiguous Arithmetic of Language Maintenance and Revitalization (2011)  234 14 Purism vs. Compromise in Language Revitalization and Language Revival (1994)  247 15 Western Language Ideologies and Small-Language Prospects (1998)  264 16 Bi- and Multilingualism in Minority and Endangered Languages (2004)  284

part 4 Language Use  309 17 Stylistic Variation in a Language Restricted to Private-Sphere Use (1994)  311 18 Telling the Monolinguals from the Bilinguals: Unrealistic Code Choices in Direct Quotations within Scottish Gaelic Narratives (1997)  329 19 Celebrations: In Praise of the Particular Voices of Languages at Risk (1999)  347

part 5 Fieldwork: Methods, Problems, Insights  369 20 Gathering Language Data in Terminal Speech Communities (1986)  371 21 Surprises in Sutherland: Linguistic Variability amidst Social Uniformity (2001)  391 22 Documentation and Responsibility (2010)  409 23 The Private and the Public in Language Documentation and Revitalization (2010)  425 Author Index  445 General Index  449

Foreword Nancy Dorian is a star. She is the founder of at least one burgeoning field within linguistics. She has made a lasting impact on a language community, and has published much of lasting value on many topics to do with the fate of endangered languages and language shift, and on what can be done about this. Language endangerment, language obsolescence, shift and loss are among the most prominent concerns of today’s linguistics, especially sociolinguistics. Nancy Dorian put them on the map. Her book Language death: the life cycle of a Scottish Gaelic dialect (University of Pennsylvania Press, 1981) was the first systematic investigation of a language on the way out – including intergenerational variation between speakers, their insecurity, loss of proficiency, and the irreversible changes under the influence of the majority language (English). This book became an instant classic, and so did Nancy Dorian. Investigating obsolescence: studies in language contraction and death (Cambridge: Cambridge University Press, 1990), edited by Nancy Dorian, opened completely new perspectives on how languages may retreat, and contract, as they gradually fall into disuse. Nancy Dorian has had an outstanding career in linguistics. She has published many books, and several score articles – all of them highly influential, and much quoted. She has done an immense amount for the Scottish Gaelic community she has been working with – helping maintain the language, and understand the why and the how of its dynamics. In 2012, Nancy was the recipient of the prestigious Kenneth Hale award, by the Linguistic Society of America, for her research on Scots Gaelic ‘that spans a period of almost fifty years, perhaps the most sustained record of research on any endangered language, and for her effective advocacy for the cause of endangered language preservation and revitalization’. Nancy Dorian’s voice in support of minority and endangered languages was one of the earliest, and continues to be one of the most prominent. Nancy’s illustrious career has in many ways shaped the profile of modern linguistics in the true sense. As the editor of the section on ‘Small languages and small language communities’ of the International Journal of the Sociology of Language, Nancy Dorian helped so many scholars and fieldworkers to openly talk about the issues faced by the communities they work in. Through her inspiring work, and her warm and engaging personality, Nancy Dorian managed to bring together scholars from different parts of the world, getting them to make joint discoveries and work on similar topics, making their research richer and more interesting. Nancy Dorian has served as a source of inspiration, and as an informal mentor to many – including us. To have a special volume of Nancy Dorian’s papers – many of them classics, but not all of them easy to locate – has been something we have always wished for. This volume is it. Here we have twenty three papers, and an introduction, by Nancy, which

x

foreword

brings them all together, as the culmination of her life’s work. Each of these is a paramount achievement in the areas of endangered languages, language variation, language shift and maintenance, and the methodology and practice of linguistic fieldwork. It is plainly an honour for us to have this volume within our monograph series. Few if any linguists have displayed such insight, and brilliance as Nancy Dorian. She remains an admirable role model, and the source of motivation and encouragement for all real linguists. Colleagues like Nancy make our discipline worthwhile.

Alexandra Y. Aikhenvald R.M.W. Dixon N.J. Enfield

Sources Chapters 1–23 have been previously published. The current versions have undergone varying amounts of revision. Chapter 1, ‘Grammatical Change in a Dying Dialect’ was published in Language 49, 2: 413–438, 1973, used with permission. Chapter 2, ‘The Fate of Morphological Complexity in Scottish Gaelic Language Death: Evidence from East Sutherland Gaelic’ was published in Language 54, 3: 590–609, 1978, used with permission. Chapter 3, ‘Making do with Less: Some Surprises along the Language Death Proficiency Continuum’ was published in Applied Psycholinguistics 7: 257–276, 1986, used with permission. Chapter 4, ‘Negative Borrowing in an Indigenous Language Shift to the Dominant National Language’ was published in the International Journal of Bilingual Education and Bilingualism 9, 5: 557–577, 2006, used with permission. Chapter 5, ‘The Problem of the Semi-Speaker in Language Death’ was published in the International Journal of the Sociology of Language 12: 23–32, 1977, used with permission. Chapter 6, ‘Language Shift in Community and Individual: The Phenomenon of the Laggard Semi-Speaker’ was published in the International Journal of the Sociology of Language 25: 85–94, 1980, used with permission. Chapter 7, ‘Defining the Speech Community to Include its Working Margins’ was published in Susanne Romaine (ed.), Sociolinguistic Variation in Speech Communities, London: Edward Arnold, 1982, pp. 24–33, used with permission. Chapter 8, ‘Abrupt Transmission Failure in Obsolescing Languages: How Sudden the ‘Tip’ to the Dominant Language in Communities and Families?’ was published in The Proceedings of the Twelfth Annual Meeting of the Berkeley Linguistics Society, pp. 72–83, 1986, used with permission.

xii

sources

Chapter 9, ‘Age and Speaker Skills in Receding Language Communities: How Far do Community Evaluations and Linguists’ Evaluations Agree?’ was published in the International Journal of the Sociology of Language 200: 11–25, 2009, used with permission. Chapter 10, ‘Linguistic Lag as an Ethnic Marker’ was published in Language in Society 9: 33–41, 1980, used with permission. Chapter 11, ‘Language Loss and Maintenance in Language Contact Situations’ was published in Richard Lambert and Barbara F. Freed (eds.), The Loss of Language Skills, Rowley, Mass: Newbury House, 1982, pp. 44–59, used with permission. Chapter 12, ‘The Value of Language-Maintenance Efforts which are Unlikely to Succeed’ was published in the International Journal of the Sociology of Language 68: 57–67, 1987, used with permission. Chapter 13, ‘The Ambiguous Arithmetic of Language Maintenance and Revitalization’ was published in Handbook of Language and Ethnic Identity, edited by Joshua A. Fishman and Ofelia García, Oxford: Oxford University Press, 2011, pp. 461–471, used with permission. Chapter 14, ‘Purism vs. Compromise in Language Revitalization and Language Revival’ was published in Language and Society 23, 4: 479–94, 1994, used with permission. Chapter 15, ‘Western Language Ideologies and Small-Language Prospects’ was published in Endangered languages. Language Loss and Community Response, Lenore A. Grenoble and Lindsay J. Whaley (eds.). Cambridge: Cambridge University Press, pp. 3–20, 1998, used with permission. Chapter 16, ‘Bi- and Multilingualism in Minority and Endangered Languages’ was published in Handbook of Bilingualism, Tej K. Bhatia and Wm. C. Ritchie (eds.), Oxford: Blackwell Publishers, 2004, pp. 437–459, used with permission. Chapter 17, ‘Stylistic Variation in a Language Restricted to Private-Sphere Use’ was published in Sociolinguistic Perspectives on Register, Douglas Biber and Edward Finegan (eds.), Oxford: Oxford University Press, 1994, pp. 217–32, used with permission.

sources

xiii

Chapter 18, ‘Telling the Monolinguals from the Bilinguals: Unrealistic Code Choices in Direct Quotation Within Scottish Gaelic Narratives’ was published in the International Journal of Bilingualism 1: 441–54, 1997, used with permission. Chapter 19, ‘Celebrations: In Praise of the Particular Voices of Languages at Risk’ was published in Ogmios. Newsletter of the Foundation for Endangered Languages 12: 4–14, 1999, used with permission. Chapter 20, ‘Gathering Language Data in Terminal Speech Communities’ was published in The Fergusonian Impact. Volume 2. Sociolinguistics and the Sociology of Language, Joshua A. Fishman et al. (eds.). Berlin: Mouton de Gruyter, pp. 555–575, 1986, used with permission. Chapter 21, ‘Surprises in Sutherland: Linguistic Variability Amidst Social Uniformity’ was published in Linguistic Fieldwork, Paul Newman and Martha Ratliff (eds.), Cambridge: Cambridge University Press, pp. 133–151, 2001, used with permission. Chapter 22, ‘Documentation and Responsibility’ was published in Language & Communication 30: 179–185, 2010, used with permission. Chapter 23, ‘The Private and the Public in Language Documentation and Revitalization’ was published in New Perspectives on Endangered Languages, José Antonio Flores Farfán and Fernando Ramallo (eds.), Amsterdam: John Benjamins, 2010, 29–47.

Introduction 1

Early Days

In American descriptive linguistic practice during the first decades after World War II, it went almost without saying that any speech form being newly described would be presented as spoken by fully fluent adults whose first language it was. If this was not possible because no such speakers remained, the reader was alerted to possible shortcomings in the description (Haas 1941; cf. Evans 2001: 261). Interest in such matters as, say, the language acquisition of children or the effects of language contact was certainly not absent, but the development of distinct subfields devoted to child language, language contact, non-standard speech, pidgins and creoles, aphasic speech, language attrition, and language obsolescence was still only on the horizon. All the same, while the preeminence of fluent-speaker descriptive objectives went more or less unchallenged for several post-war decades, questions relating to the social context of language use and the effect of social context on language structure itself gradually attracted increased attention. Like many another researcher working late in that era with a small, recessive speech form in unequal competition with an expanding majority language, I considered the chief objective of my linguistic work to be describing the speech form in question and placing it officially on the record (chapter 23). But also like other researchers working with a recessive speech form, I eventually found that the social context in which this speech form was receding gave rise to interesting questions: who were the people who still persisted in speaking the minority language under these conditions, and what was the form of the language they continued to speak like, especially among the youngest of them? The article reprinted as chapter 21 in this collection sets out the circumstances in which I began in 1963 (and continued thereafter) to do linguistic fieldwork in eastern Sutherland, on the far northeast coast of Highland Scotland, in the three former fishing villages of Brora, Golspie, and Embo. Fisherfolk Gaelic, long sustained by the occupational and social separateness of its speakers, had been losing ground to English as the local fishing industry declined and then disappeared. The Gaelic spoken by the Gaelic-English bilinguals of the area (there were no remaining monolingual Gaelic speakers) showed marked regional characteristics common to all three villages, but it also showed notable differences village by village. This local Gaelic was unwritten, and the only previous record of it took the form of answers to a questionnaire administered

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�02

2

introduction

by peripatetic fieldworkers from the Gaelic division of the Linguistic Survey of Scotland in 1953 (Brora) and 1958 (Golspie and Embo). The Gaelic of this region was highly distinctive, as is often true of dialects spoken at geographic peripheries, but by the 1950s and ’60s its survival was in doubt. Both its distinctiveness and its fragility made its description important to the Linguistic Survey, whose fieldworkers had sampled the small population of former fisherfolk via several speakers in Brora, but via only one speaker each in Golspie and in Embo. This constituted a very restricted sample for Embo in particular, since something close to half of the adult Embo population (more than 100 people) still spoke Gaelic five years later, when I began working there. The focus of the Survey’s questionnaire was furthermore quite narrow: “to gather information on the synchronic reflexes of the Common Gaelic phonological system” (Ó Dochartaigh 1997: 54). While it was a lengthy document and took a good deal of time to administer, it dealt extensively only with phonology. Other aspects of East Sutherland Gaelic speech remained unexplored and unrecorded as the 1960s began. From the point of view of the Linguistic Survey, the speech varieties in use in Brora, Golspie, and Embo represented particular dialects of Gaelic, and while their loss might reduce the number of existing forms of Gaelic speech and remove valuable historical information, it did not represent a threat to the existence of the Gaelic language. Associating the word “death” with “Gaelic”, as I did in the title and subtitle of a book that appeared in 1981, may consequently have seemed provocative and extreme to the general run of Gaelic specialists. But for local Gaelic speakers, who by then constituted an isolated speech island in an area where English had been dominant for over half a century and was becoming more so with every passing year, the issue was quite different. The local Gaelic was Gaelic to them, and what they saw as threatened was one of their two languages. When they talked about the likely end of their Gaelic, they spoke of it not as a dying dialect but as a dying language, as in the local context it surely was. The terminology of death and dying was in fact already in use among local speakers (“That’s how I’m saying it’s a dying language”, said a Brora bilingual in 1972; “We thought it was dying”, said another Brora bilingual in a 1974 field interview, explaining how it was that she and her husband had not transmitted Gaelic to their children). For the most part I have adopted local bilinguals’ point of view, speaking here and elsewhere of “language” loss and “language” endangerment in the East Sutherland villages: one of two languages spoken in these villages was in sharp decline and was likely to pass out of use. The papers collected in this volume represent most particularly the development of my own interests, of course, as I continued to work with this small,

introduction

3

at-risk language for many years; yet in a more general sense they also reflect the development of what has meantime become known as the field of language endangerment. Linguists, sociologists, linguistic anthropologists, and missionary linguists were by the second half of the twentieth century increasingly engaged with the many speech forms in various parts of the world that had few or no young speakers and so could be viewed as potential candidates for language loss. As a result, a body of research dealing with the condition of those speech forms, the consequences of the likely loss of some (or even many) of them, the prospects of maintaining or revitalizing them, and the feasibility of documenting them had begun to emerge. The overall and individual volume introductions to Austin and McGill’s four-volume Endangered Languages collection offer a good account of this development, a development so recent that a preponderance of the papers included date only from the previous twenty years (Austin and McGill 2011). During this same period my own East Sutherland fieldwork interests expanded to encompass a number of perspectives well beyond those of my original Survey-related work. I continued to work with East Sutherland Gaelic periodically during the rest of the 1960s and all of the 1970s, and although there followed a health-related hiatus of a little more than a decade, fieldwork resumed in 1990 and continued until 2010. (It continues still, for that matter, but to an extent sadly limited by the loss of nearly all local speakers.) The resumed “fieldwork” was made possible by tape recordings and long-distance telephone calls and also in a number of happy instances through visits paid me in the U.S. by East Sutherland Gaelic speakers. The descriptive and dialectological focuses of the earliest work were soon joined by two more focuses that came to the fore in the 1970s. The first of these was a focus on oral history and traditional lore, a line of inquiry that arose from a need to understand the circumstances of fisherfolk life and the particular cultural environment of these communities. The second was a focus on the linguistic changes that could be detected as the role of Gaelic altered within the former fishing communities, passing from dominant language among the oldest of the bilinguals to only partially controlled language among the youngest. Evidence of age-related proficiency differences among speakers led in turn to interest in the ways in which the least proficient speakers made use of their language skills and continued to function as members of the bilingual community. In the 1990s another two focuses emerged. One was the documenting of distinctive forms of East Sutherland Gaelic language use, something that came to seem ever more important as fluent command of this speech form declined and the prospects for its future continued to dim. The other was recognizing

4

introduction

and accounting for the apparently long-standing prominence of individual variation in local fisherfolk Gaelic, distinct from geographical variation and, in a striking departure from the mainly urban variation documented in Labovian correlational sociolinguistics, socially neutral. Throughout all these years, furthermore, essentially as a natural outgrowth of minority-language research in this setting, a perennial focus was the force of certain language ideologies that affect minority language maintenance and transmission, both in Britain and elsewhere. Finally, the experience of fieldwork itself, together with the ethical, social, and theoretical complexities that research work in small minoritylanguage communities casts up, has over time become a focus in its own right, not only for me but for many other researchers (see for example the 2001 Newman and Ratliff collection and the 2010 Innes and Debenport journal issue from which ch. 21 and ch. 22 are drawn). 2

From Focus on Dialectological Features to Focus on Features of Obsolescence

Because of my original link to a dialectological survey, documenting geographical variation was a preoccupation from the outset. The proportions of that variation in a set of villages separated from one another by a maximum distance of 17 road miles surprised me, even though the very existence of a Gaelic Division of the Linguistic Survey of Scotland testified to the presence of many distinctive regional forms of Scottish Gaelic (ch. 21). Because villagerelated variation was so prevalent, I needed multiple sources per village in order to establish local validity for every form recorded. I searched out a pool of speakers that included a minimum of nine speakers, three per village, and this proved to be an important precaution when I found that well over two hundred lexical items and a number of grammatical features were in fact distinctive in the Gaelic speech of at least one of the three villages (Dorian 1978: 152–58). In the late nineteenth century and the early twentieth century, when the fully fluent Gaelic speakers of the 1960s and ’70s were children, families in the former fishing villages might have as many as a dozen children, sometimes more, and the sibling set extended in quite a few cases beyond the 20 years conventionally taken to be a generation. The strength of nuclear family ties was especially apparent in residence patterns. It was fairly common for a widowed parent to live with one of his or her children, for married couples to have unmarried siblings living with them, and for unmarried or widowed

introduction

5

siblings to share a house with one another. All of these arrangements were to be met with among the people who became my sources, and close consanguinial and affinal kin ties produced visiting patterns that often added drop-in Gaelic visitors to fieldwork sessions (ch. 21). As time went on I was exposed to the usage of a relatively large number of speakers, which in turn led to a growing awareness on my part of differences that emerged both in responses to elicitation tasks and in ordinary conversation and corresponded by and large to the ages of the speakers. Here, it seemed, was a ready-made environment for examining linguistic change in progress where it had not often been examined up to that time, namely in a speech form that was no longer being transmitted and so was likely soon to pass out of use. By the 1960s, there was nothing left of the fishing industry that had been a major presence on the east coast of Sutherland in the 19th and early 20th centuries. Between the two World Wars fishing had gone into decline in Brora, Golspie, and Embo, and after the second World War it disappeared. Brora and Golspie still had distinct residential areas that had previously housed only fisherfolk and still housed many fisherfolk descendants (Lower Brora in the former, the West End in the latter), and Embo – though with a much reduced population as compared with the early years of the twentieth century – still retained the fishing village layout it had previously had, with a series of straight parallel streets of two-family houses running down to the sea. Fishing had kept the bilinguals of the three villages poor, separate, and stigmatized. The circumstances that brought each of these characteristics into being are discussed in a number of the studies included here, so for present purposes it can suffice to say that Gaelic-speaking parents increasingly, and often quite consciously, declined to speak Gaelic with their children or grandchildren, even though most of them continued to speak Gaelic habitually with their contemporaries and their elders (Dorian 1981: 83, 104–06; ch. 16). That is, Gaelic was fading from the scene not by disuse among fluent speakers, but by fluent speakers’ failure to use Gaelic regularly with their children or require Gaelic from them. Remarkably, some of those children acquired a degree of Gaelic all the same, through their own interest and effort. Although they did not reach their parents’ level of proficiency, or in some cases their older siblings’ level of proficiency, this resistance to abandonment of a stigmatized language from within a stigmatized group was itself a matter of research interest (ch. 3, ch. 6; Dorian 1981: 109,112). In a language-shift setting such as this, it seemed that any number of questions might be raised about age-related proficiency differences in the local Gaelic. For example, how systematic were they? Could an age-based proficiency

6

introduction

continuum be established, with older speakers using relatively conservative structures more or less predictably and their juniors using less conservative structures with similar predictability? Did markedness or frequency affect the maintenance or loss of a structure? Did grammatical or phonological elements without a parallel in English, increasingly a dominant-language presence even within the former fishing communities, pass out of use in Gaelic before elements that had an English parallel? Did conservative features that decreased in the usage of younger speakers fall out of use altogether, or might they disappear in some environments while surviving in others? Did they merge with other structures, maintaining their grammatical function while changing their composition? Were contact effects from English largely lexical, or did phonological, morphological and syntactic effects appear as well? Did code-shifting increase among younger speakers? Questions such as these were not specific to East Sutherland fisherfolk Gaelic, of course, but would arise quite naturally in connection with any small language retreating under growing pressure from a dominant language. The opportunity to explore them was simply exceptionally good in the former fishing villages of eastern Sutherland. 3

Assembling a Database

For me as a fieldworker interested in these questions and others like them, the most immediate challenge lay in assembling a body of data that could offer answers to some of them. My original evidence that such questions might be raised and answered in these villages came both from elicitation (translation tasks posed to speakers of various ages) and from exposure to conversation in multi-speaker environments. What I most needed in order to gain a better perspective on these matters was a large body of directly comparable material, something that could not be achieved in this setting through freely spoken material alone. (There were for example no origin myths, traditional stories, disaster accounts, or the like that a variety of speakers might be expected to reproduce along very similar lines.) The best prospect for achieving comparability was to present the same set of sentences for translation to an agedifferentiated sample of speakers, and from 1970 onwards I set about doing just that (while of course continuing to gather freely spoken material as well). In 1974 and 1976 in particular, I tried to establish the parameters of discernible language change via translation tasks presented to four older fluent speakers from all three villages, four younger fluent speakers from Embo (where more subtle age-related differences were in evidence than in Brora and Golspie), one

introduction

7

formerly fluent speaker from Embo (1974 only), and from 6 to 8 demonstrably imperfect speakers from Brora and Embo. The 120 sentences of the 1974 translation battery and the 150 sentences of the 1976 translation battery were designed to elicit structures that would set elements of ongoing change off against elements of stability in the local Gaelic (chs. 1–4; chs. 5–9; Dorian 1981: 117–21). In this work I had two major advantages. One was that by 1974 I had been working with East Sutherland fisherfolk Gaelic for over a decade and spoke the local dialect myself, although with some mixture of Golspie and Embo forms (and also, of course, with some anglicisms and other mistakes). This made it possible for me to design my elicitation sentences with an eye to highlighting change. The other advantage was that the speakers I worked with were highly adept translators. This came about in good part because local practice in reported speech was to use the language expected by the interlocutor rather than the language in which a remark was originally made (ch. 18). With language shift underway, kin networks in these villages included both older people who were bilingual and younger people who were monolingual in English. When individual speakers passed along in conversation any remark they thought would interest someone else, they would frame the remark in whichever language they normally used with that particular conversation partner, even if the original remark had actually been made in the other language. Of course this very frequent conversational translation was more spontaneous and less self-conscious than formal translation tasks, but the general effect of these reported speech habits was an ease with translation that made elicitation unproblematic for a good many bilingual speakers. There was a great advantage in being aware in advance of certain characteristics of local speech when designing elicitation tasks. Knowing, for example, that diminutives were exceptionally freely used in the local Gaelic made it possible to add an important check for nominal gender to the elicitation batteries. A single test sentence such as “The wee glass fell and it broke” could provide a triple check for feminine gender: the initial consonant of the noun (which differed after the article according to gender in certain phonological classes), the replacement of the pronoun in the second clause, and the diminutive suffix added to the noun ‘glass’. For conservative older speakers, the same sentence with a traditionally masculine noun such as ‘bowl’ would show differences in all three respects. As in the earlier dialect-descriptive work, the availability of a relatively large speaker sample was essential, and I enlarged my pool of sources accordingly; but in this subsequent work it was the inclusion of lowproficiency speakers that was particularly important.

8 4

introduction

The Novel Inclusiveness of Language Obsolescence Work

Established field practice in descriptive linguistics favored a certain kind of speaker as source and disfavored others, as noted above. Especially when working with small and endangered languages, linguists sought out elderly representatives of what appeared to be the most fluent and traditional form of local speech (e.g. Thomason 2001: 237). These were often also socially prominent people within the local community, and linguists were in any case inclined to exclude socially marginalized individuals and imperfect speakers (Berge 2010: 59). Frequently, synchronic data was all that was available, but diachronic and comparative evidence was respected, in that grammatical distinctions present in related languages and in reconstructed protolanguages were diligently searched for in the contemporary form of local speech. In the East Sutherland Gaelic case, very nearly all of the language data available represents apparent time. Fisherfolk Gaelic in these villages is unwritten, and there are no surviving materials from earlier periods, apart from the lexical items recorded by the Survey workers who came through the area in the 1950s. It was unfortunately seldom possible to compare the usage of individual sources across longer periods, since all too many were lost to death over the years and in other cases our early work together was too narrowly restricted to phonological matters to bear on the language change topics that interested me later. But holding the elicitation frame constant while a relatively broad spectrum of speakers of different ages and proficiencies responded produced easily distinguishable sets of differences, and insofar as the correlation between age and certain features of speech held across the sample, the differences could be taken to reflect change in progress. Use of quantifiable apparent-time data emerged strongly in the 1960s within correlational sociolinguistics, where it developed as a key technique in the work of William Labov and his students. In the 1970s another departure from established descriptive techniques emerged, this one in fieldwork with receding languages. Researchers looking at communities where marked language shift was underway began deliberately to include the verbal output of imperfect speakers among their study materials. Wolfgang Dressler, a pioneer in language obsolescence studies, pointed out as early as 1972 that ignoring language data from imperfect speakers blocked linguists from obtaining material that could be important to solving theoretical linguistic problems (Dressler 1972). Whereas Bloomfield had heaped scorn on the Menomini produced by imperfect speaker White Thunder, whose Menomini he characterized as “atrocious”, with its “small” vocabulary, “barbarous” inflections, and sentences constructed on “a few threadbare models” (Bloomfield 1964: 395), Dressler purposely sought out 12 imperfect younger speakers whose second-language Breton he could

introduction

9

compare with that of 20 older native speakers. He described the weak Breton of “terminal” speakers as pidginized and the only slightly less weak Breton of “preterminal” speakers as creolized, but in studying their Breton output he was able to identify certain word-formation processes that appeared particularly likely to be lost by imperfect speakers. In a study of East Sutherland fisherfolk Gaelic, also including imperfect speakers and likewise undertaken in the early 1970s (ch. 2), I was able to compare the morphologically rich plural and gerund formations produced by fully fluent speakers with those produced by imperfect speakers and determine that while simplification appeared in imperfectspeaker forms, it was very much less extreme than might have been anticipated (and far short of the morphological simplification typical of pidgins). On the basis of responses to the elicitation batteries of 1974 and 1976, in fact, I concluded that East Sutherland fisherfolk Gaelic could be said to be “dying . . . with its morphological boots on” (ch. 2); Thomason (2001: 236) later identified a similar and still more striking case involving Montana Salish. Although an age-and-proficiency continuum proved identifiable in East Sutherland fisherfolk Gaelic, what factors other than age as such were involved, especially in the lower reaches of the continuum, was difficult to determine. What was clear was that there were some less than fully fluent speakers within these enclaved communities of Gaelic-English bilinguals, capable of speaking Gaelic and viewed by themselves and others as members of the fisherfolk population, who produced a Gaelic that was not considered the equal of the Gaelic spoken by fully fluent members of the speech community (see the articles in section II). Some of the imperfect speakers (those identifiable as “formerly fluent speakers”) had been fully fluent speakers into adolescence, but they had married English monolinguals and lived elsewhere for long periods, seldom having the opportunity to use Gaelic. At the time when I worked with them, several of them were in various stages of recovering their Gaelic. Other imperfect speakers had lived elsewhere for briefer periods and then returned to their home villages; some of these speakers were also married to monolinguals. A few had lived locally all along yet also spoke at less than fluent levels in the 1960s and ’70s; some of these were younger children in families where older children had introduced more and more school-learned English into the home. Fully fluent speakers in Embo explicitly identified the imperfect speakers in their midst (apart from the “formerly fluent” group, who seemed to be exempted as the lapsed fluent speakers they were) as people who made mistakes in their Gaelic, and I coined the metaphorical term “semi-speaker” to designate them (ch. 5). Whether contemporary semi-speakers had once been fully fluent child speakers of Gaelic and had lost proficiency (one Brora case was clearly of this sort – see Dorian 1981: 81 – and probably two others as well), or whether their acquisition had been partial and they had never achieved

10

introduction

full fluency (a history likely for at least one Brora semi-speaker and three of the Embo semi-speakers), was difficult to determine. But since the one Brora woman who had indisputably been a fluent child speaker at school entry age was a particularly low-proficiency semi-speaker as a woman in her late fifties, it seemed that in Sutherland a history of full-fluency acquisition in childhood did not bear directly on proficiency later in life. This outcome differs from Sasse’s very clear-cut findings in the Arvanitika-speaking village of Boeotia in Greece (Sasse 1992: 62). There he found that all speakers identifiable as semi-speakers had had parents who had not talked to them in Arvanitika; instead they had acquired their Arvanitika by listening to fluent older speakers and having occasional interchanges with them. Fluent Arvanitika speakers, in sharp contrast, were raised by adults who spoke to them in that language and had shown a positive attitude toward the language. In another departure from the expectations the Boeotian case might suggest, I found in work with multi-sibling bilingual families both in East Sutherland and in Berks County, Pennsylvania, where Pennsylvania German was still spoken by some members of local secular (non-Anabaptist) families, that parental policy on using the heritage language at home had less effect on acquisition than birth order did (ch. 6). The effect of various acquisition histories remains something that invites more study. The importance of imperfect speakers to following the patterns of maintenance, shift, and linguistic change in East Sutherland fisherfolk Gaelic would be hard to overstate. Their Gaelic, when set directly beside that of fully fluent speakers, highlighted (morpho)phonological, lexical, and grammatical features of East Sutherland Gaelic that were subject to change as age and proficiency dropped. Their continuing use of Gaelic, in spite of grammatical and lexical deficiencies evident to their elders (and criticized by those elders), highlighted factors of personal history and local social structure that supported language loyalty (ch. 3, ch. 6, ch. 11; Dorian 1981: 107–10). The fact that most of them were women reflected the generally better integration into home-village kin networks of younger women as compared with younger men. The striking integrity of their receptive skills contrasted – vividly in a number of cases – with the shortcomings of their productive skills, and their excellent knowledge of sociolinguistic norms contrasted at times all too obviously with the imperfect knowledge (and so the greater likelihood of committing social offenses) of the guest linguist who had learned to speak their language. Their general acceptance as members of the bilingual community, despite considerable failings in their Gaelic, indicated that the conservative linguistic completeness and invariance especially valued by descriptive linguists was not valued to a comparable degree among community members (ch. 7, ch. 9).

introduction

11

Once information about the language skills of imperfect minority-language speakers began to be available, the possibility of comparing them with classroom second-language learners arose: foreign-language instructors were interested in whether the factors that favored home-language retention among imperfect minority-language speakers also came into play (or could be brought into play) in better retention of classroom-acquired languages (ch. 11). In more recently elaborated areas of inquiry, too, imperfect speakers are now of major interest, namely in first-language attrition studies and in studies of heritage language speech production (see for example Schmid 2011 and Polinsky 2011). In these latter cases the subjects are generally observed in a university setting or some other environment where conditions can be experimentally controlled; they are tested, that is, in a setting removed from a community or a home in which the ancestral language is still spoken. By no means all community language settings lend themselves to testing, to be sure, but in East Sutherland I was unusually fortunate in having available a sizeable number of good-natured, tolerant sources who submitted cheerfully to testing in the form of the extensive translation tasks I administered in 1974 and 1976. This made it possible to establish in unusual detail the proficiency levels among East Sutherland Gaelic speakers, comparing the grammatical capacities of individuals relative to one another and of semi-speakers relative to fully fluent speakers. In addition to the translation task batteries, I experimented with read-back tests: the results of elicitation tasks demonstrating nominal gender assignment via morphophonological signals were read back to some fluentspeaker sources at a later time, but with the gender-related morphophonological feature altered, to see whether the speakers would accept the alteration or insist on the original version (ch. 1). All of this testing was extremely fruitful, permitting direct and relatively fine comparison of individuals’ grammatical control. But equally crucial to determining important aspects of the semi-speakers’ role in the bilingual community was a long-term observational component. All but one of the semi-speakers included in my studies were observed over the years in natural interactions within Gaelic-speaking kin networks in the home village (or in one case in an exile-community network in London). It was above all through this extended observation of the semi-speakers’ strong kin network integration, their comfortable acceptance as community members, their exceptional receptive language skills, and their willingness to use the weaker of their languages with more proficient speakers, that their considerable linguistic and social contribution to the survival of Gaelic into the twenty-first century in the smallest of the former fishing villages (Embo) was assessable.

12

introduction

While identifiably imperfect speakers have a special place in the study of receding languages, even the most fluent and habitual speakers in such speech communities can be presumed to have fewer opportunities to use the minority language than their predecessors did and are likely for that and other reasons to use the language for fewer functions. Hill, like Haas, reminds us that in that case even fluent speakers will show loss of some structural possibilities that would have been appropriate to contexts that the receding language is now seldom used for or no longer used for at all (Hill 1973; Hill 2001: 176–77). In the former fishing communities of East Sutherland, the disappearance of fishing itself removed the occupational activities to which Gaelic had once been central and, as population decline followed economic decline, reduced the number of fellow-speakers available as conversation partners. In the Gaelic of the former fishing villages, just as Hill found in the Nahuatl (Mexicano) of the Malinche Volcano region of Mexico, some syntactic structures of a certain degree of complexity showed weakness even among otherwise strong local speakers (#3). Very striking, for the fisherfolk communities, is a recurrent twentieth century pattern of being out of step in terms of local language behavior (ch. 10). That is, quite apart from less proficient individuals who showed some deviation from conservative grammatical, phonological, or lexical conventions for the local Gaelic, the fisherfolk as a group deviated from whatever the norm for linguistic behavior was in the region overall. At one point the deviation was positively evaluated, by other Gaels at any rate: at the beginning of the twentieth century fisherfolk Gaelic was lauded as exceptionally free of English influence. But as the century progressed, subsequent deviations, each in turn clearly perceived and commented on by the surrounding monolingual population, were very negatively evaluated: first a persistent Gaelic monolingualism, then a conspicuously imperfect English, and finally, in the last half of the century, persistent bilingualism when all other segments of the population had become monolingual in English. In the retreat of Gaelic and the advance of English, linguistic lag came to belong to the negative stereotypes about fisherfolk identity, adding one more facet to regional perceptions of fisherfolk backwardness and difference. 5

From “Language Death” and “Language Obsolescense” to “Language Endangerment”

Determining the differences between the Gaelic spoken by the oldest and most fully fluent of fisherfolk Gaelic speakers and the youngest and least fully fluent was key to learning what changes were taking place in this speech form

introduction

13

as transmission became less routine and conditions for acquisition less ideal. At the end of §2 above, a rich array of questions was raised to which it seemed possible that developments in East Sutherland fisherfolk Gaelic might offer some answers. Not one of those questions was likely to be asked by anyone but a linguist, however, and certainly none of them by members of the local speaker population. For the most part researchers working with small, receding languages have had a set of objectives that privilege Western scientific modes and topics of inquiry and their objectives and are often – indeed usually – not shared by the speaker community (Hill 2002; Dorian 2002). In the 1970s it was common to speak of “dying languages” and “language death” in connection with languages that were not being transmitted to children. The biological metaphor was controversial from the outset, however, given that language is primarily a social rather than a biological phenomenon; and in addition, many language communities found it offensive to have the label “dying language” attached to their ancestral speech form by outsiders who had no direct connection to the ethnic community and often no personal experience of the language at all (see especially Hill 2002). Both objections were valid enough, and there has been a decline in the use of “language death” terminology in the rhetoric of outsiders with professional linguistic expertise but without personal experience of the languages they are referring to, much less ethnic connection to them. But anthropologist Bernard Perley, a member of a Canadian First Nation, considers that all of the metaphorical labeling characteristic of expert rhetoric has been detrimental to small indigenous communities’ sense of vitality and restorative possibility, where language is concerned (Perley 2012: 135, 141). In addition, he asserts, simplistic notions of “saving” languages by documenting them have been counterproductive, producing more attention to the artifacts of documentation than to speakers (op. cit. 134). One problem inherent in working with receding languages is that such languages commonly show changes directly related to reduced use. Comparison between forms used by earlier populations, if such records are available, or by older speakers as opposed to younger speakers, frequently demonstrate shrinkage of one or many sorts: reduction of phonological inventory, loss of grammatical distinctions, decline in lexical options. These are not generally perceived as positive developments, either by ethnic community members or by descriptive linguists, although not necessarily for the same reasons. Perhaps inevitably, a process of pejoration has appeared where the description of these change processes is concerned. This has happened for example in connection with the term “semi-speaker”, to which negative overtones have become attached. I was at pains when I introduced the term to celebrate the remarkable way in which semi-speakers put limited skills to maximum use (ch. 3) and

14

introduction

to stress the exceptional resilience of these speakers in resisting the language shift going on all around them (ch. 6), but metaphors develop their own resonances and the term has acquired negative overtones through its association with “imperfect” speakers whose Gaelic is measurably different from that of their elders. For still less active speakers, Flores Farfán has suggested the term quasi-speaker (Flores Farfán 2001); but it is likely that this or any similar term would in time attract negative overtones for the same reason the term semispeakers did. The perspectives of community members who wish to protect, preserve, revitalize, or revive their heritage languages are more or less inevitably distinct from those of researchers, even in the case of researchers who would be extremely happy to see those results and to assist if possible in producing them. Tonya Stebbins’ realistic account of the difficulties posed for both parties, linguist and community members, in community directed research with Coast Tsimshian in British Columbia, is usefully revealing in this respect (Stebbins 2003). Community-member responses such as Perley’s also remind today’s researchers of the negative effects, potential or actual, of professional dispassion, a necessary but socially unnatural component of the research role (ch. 23). At the same time there is a potential for equally negative effects from insufficient ethnic community awareness of serious obstacles to successful maintenance, revitalization, or revival efforts, arising from such factors as the difficulty of adult language acquisition as compared with child language acquisition (Long 1993; Dorian 1995), insufficient acknowledgment of and provision for register differentiation (Will 2012, drawing on McEwan-Fujita 2011), lack of cross-dialect experience and hence absence of dialect tolerance (Peacock 2010; Chatsis, Miyashita, and Cole 2013), and the inadvertent introduction of new social control mechanisms that emerge along with writing systems (Woolard 1998: 23). Exceptionally useful in discussing these obstacles is Peacock’s account of the many pitfalls encountered in the course of revival efforts in his own Dakota nation, stemming for example from insufficient awareness of general features of second-language teaching and learning and from unfamiliarity with revitalization and revival efforts and outcomes in other heritage language communities (Peacock 2010). Whether for ethnic community members or for outside researchers, there are definitional problems with the notion of language maintenance and survival: who can be said to qualify as a “speaker”, and can a point be identified at which a language has become extinct for lack of “speakers”? Determining just who can appropriately be considered a speaker is both methodologically and theoretically problematic. To address this question first among the East Sutherland fisherfolk descendants, self-identification proved not to be

introduction

15

a reliable gauge of Gaelic speaking ability. One low-proficiency semi-speaker disclaimed speaker status but tested slightly better than a cousin who was a particularly eager speaker and enthusiastically claimed her speaker status. Speech performance was unsuitable as a measure of ability because motivations to produce Gaelic speech fluctuated unpredictably according to both external and internal circumstances. One formerly fluent speaker simply avoided using Gaelic, almost to the point of rudeness, leaving the question of his ability to do so uncertain until unusual circumstances (the presence of a Gaelic radio interviewer looking for an interviewee) flushed out his conversational capacity. Another formerly fluent speaker suddenly and somewhat mysteriously reversed her decades-long insistence that she really no longer spoke Gaelic and began to converse eagerly over the phone, both with me and with a brother-in-law in the home village. It was seldom possible to test the capacity of (near)-passive bilinguals to put together Gaelic sentences, but when two with demonstrably good comprehension provided some limited spoken material, they proved quite different in their ability to produce phonologically unremarkable lexicon (one could, one could not). There are undoubtedly others, at least in the smallest village where Gaelic continued longest, who know quite a few high-frequency lexical items and some phrases – the chunk learned “residue knowledge” that can serve as a “phatic symbol of identification” (Sasse 1992: 64; ibid., quoting Tsitsipis 1983). In all likelihood there will continue for some time to be fisherfolk descendants who use certain local Gaelic lexical items – possibly in some cases without realizing that they are Gaelic – long after there are no active speakers remaining in their kin circles. In other settings where languages have receded or are currently receding, the difficulties in pointing to a particular time as the termination date for a speech form may arise from quite different causes. Local language ideologies may prevent imperfect speakers from revealing that they have some capacity in the local language, as Aikhenvald found was originally the case among the Tariana in the Brazilian Amazon; it was only after Aikhenvald offered Tariana teaching sessions for the ethnic community, and then especially after the deaths of the oldest and most puristic of the small group of remaining speakers, that a number of younger and less fully proficient Tariana felt free to openly use the less conservative form of Tariana they actually knew (Aikhenvald 2003 and 2013). Early acquisition without membership in the community of birthright speakers may inhibit highly competent speakers from claiming skilledspeaker knowledge of one of their languages. This is a familiar phenomenon in Aboriginal Australia, where languages are associated with particular territories and are “owned” by the inhabitants of a given territory; fluent speakers from outside the ownership group are reluctant to claim knowledge, even

16

introduction

when they speak fluently, or to act as sources for such a language (Evans 2001: 250–261). For at least one part of Australia Eades points to special difficulties in establishing when an Aboriginal language ceases to be spoken: “. . . for many Aborigines the use of a language-group label to identify a person remains long after that language is no longer spoken. In fact it is often difficult to say when a language is no longer spoken, since some Aboriginal groups continue to mix in some words from Aboriginal languages when they are speaking what is essentially Standard English” (Eades 1981: 13). In southeastern Australia, Aboriginal languages believed extinct were found in common use “in at least rudimentary form” as secret languages designed to confound the police (Wurm 1991: 15). Reporting on South America, where languages believed to be extinct are rediscovered from time to time (he instances Cholón in the Peruvian Andes), Adelaar recommends that linguists refrain from statements that a native language is no longer spoken: “Categoric statements by linguists to the effect that there are no speakers left of a particular language often do more harm than good as they may keep other linguists from continuing the search” (Adelaar 1991: 51). In one recognized type of language retreat, “bottom-to-top death”, everyday use of the local language disappears, but some ritual use is still preserved for ceremonial purposes. Campbell and Muntzel (1989: 185) tell of a source who provided a religious text in Chiapanec (an Otomanguean language of Mexico) that he reserved for performance but spoke no Chiapanec otherwise except for a small number of isolated “remembered” vocabulary items. Fluent learners from outside the ethnic group would not normally be included in tallies of remaining speakers, perhaps, but how is a fluent outsider who becomes a transmitter of the language to be reckoned? This is the surprising role that linguist William Shipley took on late in life, transmitting the Mountain Maidu language of northern California to a grandson of the Maidu speaker who had contributed most to making Shipley himself a fluent speaker (Shipley 2000). Even languages that have no active speakers remaining may be considered potentially viable, especially by members of the ethnic group in question, if documentation is full enough to allow them to mount a wellsubstantiated reclamation effort, as has happened in the case of the “sleeping” language Miami (Leonard 2008). Indeterminacies about what constitutes a speaker and what level and frequency of use justifies the claim that a language is still spoken are encountered in discussions of language revitalization, too. Irish has long been a key case in this regard. If a substantial number of Irish citizens have a second-language knowledge of Irish (enough to take an Irish-language newspaper or listen to broadcasts in Irish, say) but do not actively use Irish on any regular basis, is their existence a sign of the success of Irish “revival” efforts or of their failure (ch. 13)?

introduction

17

Reviewing the complex history of Rama revitalization in Nicaragua, Grinevald points out that revitalization “has many facets, one of the most essential being the recreation of a link with an ancestor language, to develop a relationship and a certain familiarity with the language, for self image and identity purposes” (Grinevald 2005: 211). No new fluent speakers of Rama have been produced by the Rama Language Project, but a language previously despised by its own ethnic population has gained legitimacy among that population and within the region, achieved a written form, and been introduced to successive waves of primary school children who have valued and enjoyed the experience. Hill (2001) reminds us that communities differ not only as to what they consider good speech (e.g. laconic and terse versus highly elaborated) but even as to what they consider constitutes being a speaker at all (e.g. being able to count to 10, recite a prayer, carry on a conversation). Her conclusion is that “the identification of language decay cannot be accomplished exclusively on structural grounds but requires ethnographic investigation of local ideologies and preferences which are invariably complex and often disputed” (Hill 2001: 177). 6

The Two-Way Street of Ideology: Theirs and Ours

Ideological questions are indeed central in the study of endangered languages, but bidirectionally so rather than unidirectionally. That is, it is essential to know what local beliefs and conceptions affect the likelihood that a group will maintain or give up a threatened language, but it is equally essential to know what beliefs and conceptions on researchers’ parts may affect their perceptions of observed language behaviors and their interpretation of them (ch. 15, ch. 16). Of these two challenges, the former is the easier to deal with, since one’s own perspectives seem ordinary while the other fellow’s are likely to call attention to themselves by seeming unusual to the researcher. Tariana repugnance at any evidence of language mixing (Aikhenvald 2001), Eastern Pomo pragmatism about adhering to whatever languages appear to be particularly useful and letting others go (McLendon 1978), Western Mono embrace of individual variation in language (Kroskrity 2002), Tolowa skepticism about the possibility of a general linguistic description of their language (Collins 1998), Pueblo insistence on limiting outsiders’ access to cultural materials, including language (Debenport 2010), are all ideological positions that registered clearly with the linguists who worked with these languages, in large part because they were distinct from the linguists’ own ideological positions. More difficult to recognize are the assumptions we as researchers may bring into the field from our own experience – especially from our experience as

18

introduction

speakers of standardized written languages – without full awareness of their hold on our thinking. James Milroy has repeatedly emphasized that exclusive experience with highly standardized languages encourages an expectation of uniformity and invariance, even among linguists, and promotes an inappropriate uniform-state model of what is actually a highly dynamic phenomenon (Milroy 1999: 17; 2000: 11–12); although “everyone knows” that language is variable, many people, including scholars of language, nonetheless consider linguistic invariance desirable (Milroy 1992: 3). The “ideology of standardization” (Milroy 2000: 11) assumes an idealized uniformity and promotes for example the definitive “correct” dictionary entry, the grammaticality judgment, and the authority of the written language, none of which may exist or apply in certain language settings. This preoccupation with conservative and definitive grammatical completeness may be carried into the field by linguists with objectives that are formed not only by primary personal experience of standardized languages but also by professional training and its ideologies (ch. 9). For example, as descriptivists we aspire to gathering whatever complete paradigms might be available (especially if documentation for related speech forms suggest that they might be there), and when we find paradigmatic gaps in our data we are likely to be disconcerted. In attempting to fill these and other gaps (for example, “missing” cover terms in certain semantic fields), we aspire, in keeping with our preference for the complete and the invariant, to locating definitive uniform responses. Hence the ready assumption of a very knowledgeable Celticist colleague of mine that if conjunctions and pronominal paradigms were variable in East Sutherland fisherfolk Gaelic, as they were, the variability must reflect the obsolescent state of the speech form (ch. 21; see also Dorian 2010: 21–22). This overlooks the possibility that different paradigms may coexist in the usage of different individuals, or even of one individual, and that more than one grammar may be present among speakers of “the same” language in “the same” speech community. The evidence is very strong for East Sutherland fisherfolk Gaelic in particular that individual variation, even where grammatical forms are concerned, is a long-established feature of these unwritten village speech forms, and there is evidence of similar personally patterned variation in other small communities without extra-community norming and without community-internal social stratification (Dorian 2010: 271–87; see also the final page of this Introduction). Not surprisingly, our sources do not share our linguistic preoccupations or assumptions. Collins describes the good-humored patience with which his Tolowa language consultants endured his “focused paradigm elicitation” and “focused and narrow elicitations of contrastive alternates and distributional possibilities”, but he eventually realized that their orientations to language

introduction

19

were altogether different from his: “Simply put, they were interested in words not grammar” (Collins 1998: 260). Kroskrity, similarly, found that the linguistic focus of Arizona Tewa speakers’ very long-standing puristic ideology was strictly lexicographical, with no comparable attention paid to instances of grammatical convergence resulting from the very same history of language contact (Kroskrity 1998: 110). Yet another tacit ideological point of view, one that generally went unacknowledged among early researchers working with endangered languages, was identified by Susan Gal: writings about the receding languages of small and mostly rural communities tended to be perfused with a pastoralist perspective that looked back to a more intact past, linguistic as well as social; this was reflected in the use of the metaphor “language death” itself and in an emphasis on structural loss and current-speaker inadequacy (Gal 1989: 316). The focus was on the disappearance of conservative structures used by the oldest speakers, while innovative aspects of younger speakers’ speech, such as those she identified in Oberwart Hungarian, were less likely to be acknowledged (Gal 1989: 315–16). Particularly pervasive in the Western world has been an assumption that linguistic homogeneity is somehow fundamental, with monolingualism the normal condition and the ideal nation a bounded territory in which a single language is spoken (Irvine and Gal 2000: 63, 76). Most of our accounts of biand multilingualism appear in the writings of Western scholars who present these states as unusual or problematic conditions, settings for imbalance in language dominance and strength (e.g. “Bilingualism . . . is a natural setting for the unraveling of native language abilities”, Seliger and Vago 1991: 3) and sustainable only so long as practical necessity requires them (“People will not indefinitely maintain two languages when one will serve across all domains”, Edwards 1994: 110). Influenced by such widespread but mostly unacknowledged assumptions, Westerners, including linguists, can find it deeply surprising to encounter well-established multilingual populations who unproblematically maintain fluency in several languages, even when one of those languages is a regional lingua franca that potentially serves all practical needs (ch. 16). Remarkable, too, to researchers of Western European cultural background, are accounts of multilinguals who in old age take pleasure in adding to the repertory of languages they already speak (ch. 16; see also the Aboriginal Australian role of the “polylingual specialist” and “linguistic virtuoso”, described in the sources cited in Brandl and Walsh 1982). Intolerance of bilingualism, particularly pronounced in the U.S., can leave us surprised, too, to learn of larger populations who matter-of-factly accept the long-term presence of a minority population speaking an allophone language in their midst (ch. 16). Relatively

20

introduction

rarely encountered in the multilingualism literature produced by Westerners, likewise, is the positive self-image that some multilingual populations derive from their knowledge of more languages than neighboring peoples possess (ch. 16). Committed furthermore to an ideology that recognizes only one “native language” per speaker, linguists have been reluctant to consider the possibility that individuals in multilingual regions may have acquired from their earliest years more than one language in what we ordinarily think of as native speaker depth and can therefore reasonably serve as expert consultants for more than one language (ch. 16). How much skepticism about the value of bi- and multilingualism in particular leaks out, in our dealings with minority populations? Parental fear already prevails in many minority communities that acquisition of the home language will interfere with successful performance in dominant-language schooling (e.g. Tsitsipis 1984: 123; Schmidt 1985: 25; Kuter 1989: 82), undermining support for revitalization programs. Some of the questionable ideological assumptions we unconsciously adopt undermine the validity of our own conclusions, but it would be more unfortunate still to find that negative Western attitudes toward bilingualism contribute to undermining the will of small language communities to maintain their languages. Not the least of the responsibilities we bear as researchers is an obligation to scrutinize the ideologies that underlie our scholarly work and to recognize their likely impact, not only on our own research results but also on the community under study, since Western researchers are only one of the latest in a long series of outside forces to enter the local community in the service of values generally alien to that community (ch. 24). 7

Documentation: The Impossible Task?

Despite some urgent statements in the early 1990s (Dixon 1991, Hale et al. 1992), the need for a massive effort to document the many speech forms likely to pass out of active use in the foreseeable future was slow to gain any real traction among linguists. In the 1970s and ’80s, it was hard for linguists engaged with endangered languages to get their colleagues, let alone the media and the reading public, to pay attention to the oppressive fates and dim prospects of many small language communities scattered around the globe. This is scarcely the case now, however, and in an ironic turn of events the ubiquity and the dramatic character of warnings about the “endangerment crisis” are now themselves the subject of critical commentary (e.g. Moore 2006, Muehlmann 2011). Since the late 1990s (e.g. Himmelmann 1998), discussions about documentation have come to center on the advantages of (cross-disciplinary) team

introduction

21

fieldwork, on the interplay between documentation and description, on the amount and type of archival material that can be considered adequate, and on archiving practices that can ensure the survival and usability of documentary material. (It is almost comical to recall that in the 1970s and ’80s discussion was more likely to center on whether a dissertation describing an endangered language would lead directly to unemployability.) Current supporters of what might be termed the documentation imperative consider it essential that linguists prioritize documentation above other goals and equally essential that the resulting corpus be maximally large and diverse (Himmelmann 1998, Woodbury 2003). Documentation without simultaneously extensive analysis can be considered acceptable, although in practice some level of analysis is implicit in the transcription phase of documentation (Berge 2010: 53–55). In my own view, the documentation enterprise poses some unusually difficult ethical problems, first and foremost the simple impossibility of gaining the genuinely informed consent of one’s sources for scholarly use of the materials they provide (ch. 22, ch. 23; cf. Himmelmann 2008: 343). Scholarly purposes are incomprehensibly distant from the experience of nearly all of the small-language groups we might work with, as I found even in a first-world setting, and this difficulty is only magnified where cultures differ more sharply. In addition, some procedural decisions about how we will conduct fieldwork can have unintentionally damaging consequences for the people who speak the endangered languages we wish to document. If we use a language of wider communication instead of learning to speak the language being documented, for instance, we transmit a negative message about the value in the wider world’s eyes of the local language (ch. 23). The effect of what Bradley has called “parachute linguists” is also unlikely to be favorable: fieldworkers who arrive to document an endangered speech form, show apparently intense interest in it, expend what are obviously major resources on recording it, but then after a relatively brief period disappear as suddenly as they appeared (Bradley 2007: 143). Many other issues are also troublesome. For example, how do we as fieldworkers position ourselves vis-à-vis local sources: are we primarily experts who extract information according to our own ideas of what is important, or are we equally learners who ask for help in acquiring some knowledge of the language about which local people are the closest thing to experts? If, while establishing what the direction and scope of linguistic change is in a dwindling population of mixed-proficiency speakers, we call attention to the “innovative” (nontraditional) forms used by the younger speakers, do we thereby contribute to lower-proficiency speakers’ sense that their language skills are inadequate to promoting and transmitting their heritage language? What effect is it likely

22

introduction

to have on long-term dialect fortunes if each of a number of settlements pre­ sents somewhat different forms of a language but time and funding limit full documentation to only one of them? Issues of adequacy also arise. If we engage in anything short of almost indiscriminate audio and video recording (and even then), how confident can we be of the comprehensiveness of the record? How can our successors (or for that matter even we) evaluate the representativeness, the overall accuracy, or the significance of the materials we obtain (cf. the reinterpretation in ch. 9 as compared with ch. 5)? In areas where shift is at a very advanced stage when we arrive, how quickly and reliably can we gauge the suitability and even the basic competence of a few available sources recruited only with considerable difficulty (see Childs 2009 for an extreme case)? Documenting a language is a formidable challenge, and the number of languages still only minimally documented or wholly undocumented will set linguists a furious-paced professional agenda for the foreseeable future. The goals of this agenda will no longer be so single-minded as they were only a few decades ago, when the motivation for description and documentation was primarily (solely, for a good many linguists) to leave a record for contemporary colleagues and scientific posterity. Increasingly, if belatedly, linguists will be engaged in some kind of support effort – to devise or reform an orthography, provide a user-friendly dictionary, help produce primers or story collections, train ethnic group members to do linguistic work, create an archive, find or interpret earlier records, give historical evidence in land claim cases – for the indigenous groups whose languages they work with. Ultimately it will be for those groups and their ethnic inheritors to determine how much value outside “expertise” has had for them, and since responses have been very far from uniformly positive to date, we can anticipate that they will not be so in future, either. Elmer Miller’s thoughtful book about his experience in two successive but very different roles, missionary and anthropologist, among the Toba Indians of the Argentine Chaco, has lessons for linguists as well as anthropologists: the Toba found him considerably more useful to them as a missionary, with access to mission funds, than as an anthropologist, even though he produced ethnographic writings that would be of potential heritage value to their posterity. His retrospective account expresses concern, furthermore, lest his ethnographic writings on Toba religious and social life inadvertently serve to strengthen the nation-state’s efforts to dominate a minority that was still stubbornly resistant up to that point (Miller 1995: 172, 200–02; cf. the opposing viewpoints of Ladefoged 1992 and Dorian 1993 where the “political” role of linguistic scholars is concerned).

introduction

23

We can predict more securely that future linguists will value the efforts of today’s documenters than that indigenous groups will, but even there we should not be too confident of our success. For one thing, linguists are still drawn, almost irresistibly it seems, to the most unique features of any language they study and are inclined to overstudy these and neglect less glamorous features (ch. 19; Berge 2010: 57–58). For another, research priorities change, and so do descriptive and analytic directions. Our future successors are sure to fault us in our turn for gaps and skewings in our work. Suggestively, in this regard, Woolard reports that participants in the discussion group prefiguring an important 1998 volume on language ideologies “were struck by the apparent absurdity of nineteenth-century philology’s relentless reading of spiritual qualities from linguistic structures”, but then wondered whether “the singleminded reading of power into and out of communicative practices that has characterized our own late-twentieth-century sociolinguistics will look as ludicrously obsessive in another century’s retrospect” (Woolard 1998: 28). Already uncommon, multi-decade engagements with a single minority language are likely to become rarer still as concern grows over the number of languages still to be documented. An engagement of that extent offers time to review and expand or correct earlier observations and analyses, a luxury many endangered language researchers never have and even fewer will have in the future. In the case of my work with East Sutherland fisherfolk Gaelic, two experiences in particular have led me to greater appreciation of the advantages of multi-faceted and long-term work with the speech forms of a single area. As a result of my involuntary switch from on-site fieldwork to telephone fieldwork in the 1990s, the data that I recorded changed character considerably. Field linguists have traditionally been more inclined to record certain kinds of material than others (Foley 2003; Berge 2011: 56), and I was no exception: my early recordings ran heavily to stories, autobiographical narratives, interviews, and some folk tradition. Although I asked questions fairly frequently, these interchanges remained primarily informational. By contrast, the material recorded later over the phone was first and foremost conversational. My sources generously continued to do a good deal of translation-task work over the phone with me, focused at that stage on personal pattern variation (ch. 21), but most of the time we talked back and forth and the Gaelic on these tape recordings was very different in style from what I had recorded in the 1960s and ’70s. (For example, they were much less revealing of puristic English loanword avoidance and much more revealing of increasing use of English discourse particles, adverbs, and conjunctions.) Because what gets recorded is affected by so very many factors – e.g., how well source(s) and researchers know one

24

introduction

another or like one another, how many people are present and listening, how much factionalism or leadership competition exist within the population the researcher would like to record, what the recording medium is, how forthcoming the sources are (unknowns such as the behavior or even aptitude of previous researchers can be major influences; see Grinevald 2005), the race, gender, and/or age of the researchers and the sources, even the season of the year and the weather in some settings – it seems inevitable that any documentary record we produce will be skewed in ways and directions that we recognize poorly or fail to recognize at all. This does not of course mean that we should abandon the effort to document, only that we should be very cautious about supposing that a language for which we have copious audio and video recordings can be considered well or fully documented. The second notable experience concerned the documentation of that very personal pattern variation that by the 1990s I was especially interested in. Variation of various kinds was overwhelmingly abundant in the local fisherfolk Gaelic, and only once a very substantial corpus was available from a considerable number of speakers could anyone hope to determine what part of the variability represented geographical variation and what part personally patterned variation (or for that matter stylistic variation). In one case, for example, only the chance fact that a lexical item was both a geographical variant and a personally patterned variant made it ciear to me (belatedly at that) that two different kinds of variation were simultaneously involved. I had tracked the word airgiod ‘money, silver’ as a geographic variable (its Embo form differed from its Brora and Golspie form) long before I had acquired and sorted the larger number of instances from Brora and Golspie speakers that were needed to establish that it was also a personal pattern variant in those two villages, with different sets of speakers using a slightly different first-syllable vowel. It took years to untangle such complexities of variation, and since personally patterned variation turned out to be intimately linked to the social organization of these villages, despite the fact that the variables as such were not social group markers, no documentation effort short of a multiple-year engagement with the language would have made that untangling possible. The second of these fieldwork experiences is to me more disconcerting of the two. In unwritten languages without external norming, individual variation is a much more frequent phenomenon than has previously been recognized (see the literature cited in Dorian 2010, chapter 8; also Goddard 2010, Chatsis, Miyashita, and Cole 2013, and Aikhenvald forthcoming). In many endangered language settings it would be not just easy, but even natural, to take the profusion of variant forms for some sort of late-stage laxity or for the dissolution of

introduction

25

linguistic competence rather than the distinct linguistic phenomenon it is, and so to expunge it from a record that was unconsciously aimed at the invariance associated with standardized languages. This mistake would leave the linguistic record incomplete in much the same way that eliminating young speakers and imperfect speakers as sources once did. Long-term fieldwork does not of course guarantee that documentation will be ideally complete, let alone superior in some particular fashion. It will however be different, because there will be opportunity to review early-acquired data from a more broadly based late-fieldwork perspective and because the possibility of new late-stage information increases. In my own case personal circumstances set me on this path, rather than original intention, but in retrospect I take it to have been a useful path to follow. References Adelaar, Willem F.H. 1991. The endangered languages problem: South America. In R.H. Robins and E.M. Uhlenbeck, eds, Endangered languages, 45–91. Oxford: Berg. Aikhenvald, Alexandra Y. 2001. Language awareness and correct speech among the Tariana of Northwest Amazonia. Anthropological Linguistics 43(4):411–30. ———. 2003. Teaching Tariana, an endangered language from Northwest Amazonia. International Journal of the Sociology of Language 161:125–39. ———. 2013. Shifting attitudes in north-west Amazonia. International Journal of the Sociology of Language 222:195–216. ———. Forthcoming. Language contact and language blend: Kumandene Tariana of north-west Amazonia. Austin, Peter K., and Stuart McGill, eds. 2011. Endangered languages. New York: Routledge. Berge, Anna. 2010. Adequacy in documentation. In Lenore A. Grenoble and N. Louanna Furbee, eds, Language documentation: Practice and values, 51–66. Amsterdam: John Benjamins. Bloomfield, Leonard. 1964 [1927]. Literate and illiterate speech. In Dell Hymes, ed., Language in culture and society: A reader in linguistics and anthropology, 391–96. Cambridge: Cambridge University Press. Bradley, David. 2007. What elicitation misses: Dominant language, dominant semantics. In Peter K. Austin, ed., Language Documentation and Description 4:136–44. Brandl, M.M., and M. Walsh. 1982. Speakers of many tongues: Toward understanding multilingualism among Aboriginal Australians. International Journal of the Sociology of Language 36:71–81.

26

introduction

Campbell, Lyle, and Martha C. Muntzel. 1989. The structural consequences of language death. In Nancy C. Dorian, ed., Investigating obsolescence: Studies in language contraction and death, 181–96. Cambridge: Cambridge University Press. Chatsis, Annabelle, Mizuki Miyashita, and Debbie Cole. 2013. A documentary ethnography of a Blackfoot language course: Patterns of variationism and standard in the ‘organization of diversity’. In Shannon Bischoff, Deborah Cole, Amy Fountain, and Mizuki Miyashita, eds, The persistence of language: Constructing and confronting the past and present in the voices of Jane H. Hill. Amsterdam: John Benjamins. Childs, G. Tucker. 2009. How to pretend you speak a dying language when you don’t really know how to speak it: Methodological worries in documenting dying languages. In M. Brenzinger and A.M. Fehn, eds, Proceedings of the 6th World Congress of African Linguistics, 45–56. Cologne: Köppe. Collins, James. 1998. Our ideologies and theirs. In Bambi B. Schieffelin, Kathryn A. Woolard, and Paul V. Kroskrity, eds, Language ideologies: Practice and theory, 256– 70. New York: Oxford University Press. Debenport, Erin. 2002. Comparative accounts of linguistic fieldwork as ethical exercises. International Journal of the Sociology of Language 206: 227–44. Dixon, R.M.W. 1991. The endangered languages of Australia, Indonesia and Oceania. In Robins, R.H., and E.M. Uhlenbeck, eds, Endangered languages, 229–55. Oxford: Berg. Dorian, Nancy C. 1978. East Sutherland Gaelic: The dialect of the Brora, Golspie and Embo fishing communities. Dublin: Institute for Advanced Studies. ———. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. ———. 1993. A response to Ladefoged’s other view of endangered languages. Language 69(3):575–79. ———. 1995. Sharing expertise and experience in support of small languages. International Journal of the Sociology of Language 114:129–37. ———. 2002. Commentary: Broadening the rhetorical and descriptive horizons in endangered-language linguistics. Journal of Linguistic Anthropology 12(2):134–40. ———. 2010. Investigating variation: The effects of social organization and social setting. New York: Oxford University Press. Dressler, Wolfgang U. 1972. On the phonology of language death. Proceedings of the Chicago Linguistic Society 8:448–57. Eades, Diana. 1981. ‘That’s our way of talking’: Aborigines in southeast Queensland. Social Alternatives 2(2):11–14. Edwards, John. 1994. Multilingualism. London and New York: Routledge. Evans, Nicholas. 2001. The last speaker is dead. Long live the last speaker! In Paul Newman and Martha Ratliff, eds, Linguistic fieldwork, 250–81. Cambridge: Cambridge University Press.

introduction

27

Flores Farfán, José Antonio. 2001. Culture and language revitalization, maintenance, and development in Mexico: The Nahua Alto Balsas communities. International Journal of the Sociology of Language 152:185–97. Foley, William A. 2003. Genre, register and language documentation in literate and preliterate communities. In Peter K. Austin, ed., Language Documentation and Description 1:85–98. Gal, Susan. 1989. Lexical innovation and loss: The use and value of restricted Hungarian. In Nancy C. Dorian, ed., Investigating obsolescence: Studies in language contraction and death, 313–31. Cambridge: Cambridge University Press. Goddard, Ives. 2010. Linguistic variation in a small community: The personal dialects of Moraviantown Delaware. Anthropological Linguistics 52(1):1–48. Grinevald, Colette. 2005. Why the tiger language and not Rama Cay Creole? Language revitalization made harder. In Peter K. Austin, ed., Language Documentation and Description 3:196–224. Haas, Mary R. 1941. Tunica (Extract from Handbook of American Indian Languages, Vol. IV). New York: J.J. Augustin. Hale, Ken, Michael Krauss, Lucille J. Watahomigie, Akira Y. Yamamoto, Colette Craig, LaVerne Masayesva Jeanne, and Nora C. England. 1992. Endangered languages. Language 68(1):1–42. Hill, Jane H. 1973. Subordinate clause density and language function. In C. Corum, T.C. Smith-Stark, and A. Weiser, eds, You take the high node and I’ll take the low node: Papers from the comparative syntax festival, 33–52. Chicago: Chicago Linguistic Society. ———. 2001. Dimensions of attrition in language death. In Luisa Maffi, ed., On biocultural diversity, 175–89. Washington D.C.: Smithsonian Institution Press. ———. 2002. “Expert rhetorics” in advocacy for endangered languages: Who is listening, and what do they hear? Journal of Linguistic Anthropology 12(2): 119–33. Himmelmann, Nikolaus P. 1998. Documentary and descriptive linguistics. Linguistics 36:161–95. ———. 2008. Reproduction and preservation of linguistic knowledge: Linguistics’ response to language endangerment. Annual Review of Anthropology 37:337–50. Irvine, Judith T., and Susan Gal. 2000. Language ideology and linguistic differentiation. In Paul V. Kroskrity, ed., Regimes of language: Ideologies, polities, and identities, 35–83. Santa Fe: School of American Research Press. Kroskrity, Paul V. 1998. Arizona Tewa speech as a manifestation of a dominant language ideology. In Bambi B. Schieffelin, Kathryn A. Woolard, and Paul V. Kroskrity, eds, Language ideologies: Practice and theory, 103–22. New York: Oxford University Press. ———. 2002. Language renewal and the technologies of literacy and postliteracy: Reflections from Western Mono. In William Frawley, Kenneth C. Hill, and Pamela Munro, eds, Making dictionaries: Preserving indigenous languages of the Americas, 171–92. Berkeley: University of California Press.

28

introduction

Kuter, Lois. 1989. Breton vs. French: Language and the opposition of political, economic, social, and cultural values. In Nancy C. Dorian, ed., Investigating obsolescence: Studies in language contraction and death, 75–89. Cambridge: Cambridge University Press. Ladefoged, Peter. 1992. Another view of endangered languages. Language 68(4): 809–11. Leonard, Wesley Y. 2008. When is an “extinct language” not extinct? Miami, a formerly sleeping language. In Kendall A. King, Natalie Schilling-Estes, Lyn Fogle, Jia Jackie Lou, and Barbara Soukup, eds, Sustaining linguistic diversity: Endangered and minority languages and language varieties, 23–33. Georgetown: Georgetown University Press. Long, Michael H. 1993. Second language acquisition as a function of age: Research findings and methodological issues. In Kenneth Hyltenstam and Åke Viberg, eds, Progression and regression in language: Sociocultural, neuropsychological and linguistic perspectives, 196–221. Cambridge: Cambridge University Press. McEwan-Fujita, Emily. 2010. Ideologies and experiences of literacy in interaction between adult Gaelic learners and first-language Gaelic speakers in Scotland. Scottish Gaelic Studies 26:87–114. McLendon, Sally. 1978. How languages die: A social history of unstable bilingualism among the Eastern Pomo. In Margaret Langdon, Shirley Silver, and Kathryn Klar, eds, American Indian and Indo-European Studies, 137–50. The Hague: Mouton. Miller, Elmer S. 1995. Nurturing doubt: From Mennonite missionary to anthropologist in the Argentine Chaco. Urbana and Chicago: University of Illinois Press. Milroy, James. 1992. Linguistic variation and change: On the historical sociolinguistics of English. Oxford: Blackwell. ———. 1999. The consequences of standardization in descriptive linguistics. In Tony Bex and Richard J. Watts, eds, Standard English: The widening debate, 16–35. London: Routledge. Milroy, Jim. 2000. Historical description and the ideology of the standard language. In Laura Wright, ed., The development of Standard English, 1300–1800, 11–28. Cambridge: Cambridge University Press. Moore, Robert E. 2006. Disappearing, Inc.: Glimpsing the sublime in the politics of access to endangered languages. Language and Communication 26:296–315. Muehlmann, Shaylih. 2011. Von Humboldt’s parrot and the countdown of last speakers. Language and Communication 32:160–68. Newman, Paul, and Martha Ratliff, eds. 2001. Linguistic fieldwork. Cambridge: Cambridge University Press. Ó Dochartaigh, Cathair, ed. 1997. Survey of the Gaelic dialects of Scotland: Questionnaire materials collected for the Linguistic Survey of Scotland, Vol. 1. Dublin: Dublin Institute for Advanced Studies, School of Celtic Studies.

introduction

29

Peacock, John Hunt. 2010. The good news and the bad news about reviving my Spirit Lake Dakota nation’s endangered language. In Hywel Glyn Lewis and Nicholas Ostler, eds, Reversing language shift: How to re-awaken a language tradition, 132–39. (Proceedings of Conference FEL XIV) Bath: Foundation for Endangered Languages. Perley, Bernard C. 2012. Zombie linguistics: Experts, endangered languages, and the curse of the undead voices. Anthropological Forum 2(2):133–49. Polinsky, Maria. 2011. Reanalysis in adult heritage language: New evidence in support of attrition. Studies in Second Language Acquisition 33:305–28. Sasse, Hans-Jürgen. 1992. Language decay and contact-induced change. In Matthias Brenzinger, ed., Language death: Factual and theoretical explorations with special reference to East Africa, 59–80. Berlin: Mouton de Gruyter. Schmid, Monika S. 2011. Language attrition. Cambridge: Cambridge University Press. Schmidt, Annette. 1985. Young people’s Dyirbal: An example of language death from Australia. Cambridge: Cambridge University Press. Seliger, Herbert W., and Robert A. Vago. 1991. The study of first language attrition: An overview. In Herbert W. Seliger and Robert A. Vago, eds, First language attrition, 1–15. Cambridge: Cambridge University Press. Shipley, Bill. 2000. The Maidu story. Sandy Chung, Jim McCloskey, and Nathan Sanders, eds, The Jorge Hankamer Web Fest. Babel.ucsc.edu/jorge/, accessed 16 Mar. 2013. Stebbins, Tonya. 2003. Fighting language endangerment: Community directed research on Sm’algyax (Coast Tsimshian). Osaka: Endangered Languages of the Pacific Rim. Thomason, Sarah G. 2001. Language contact: An introduction. Washington D.C.: Georgetown University Press. Tsitsipis, Lukas. 1981. Language change and language death in Albanian speech communities in Greece: A sociolinguistic study (Unpublished Ph.D. dissertation, University of Wisconsin, Madison). ———. 1983. Language shift among the Albanian speakers of Greece. Anthropological Linguistics 25(3):288–308. ———. 1984. Functional restriction and grammatical reduction in Albanian language in Greece. Zeitschrift für Balkanologie XX(1): 122–31. Will, Vanessa Katharina Angela. 2012. Why Kenny Can’t Can: The language socialization experiences of Gaelic-medium educated children in Scotland. Unpublished University of Michigan dissertation. Woodbury, Tony. 2003. Defining documentary linguistics. Language Documentation and Description 1 :35–51. Woolard, Kathryn A. 1998. Introduction: Language ideology as a field of inquiry. In Bambi B. Schieffelin, Kathryn A. Woolard, and Paul V. Kroskrity, eds, Language ideologies: Practice and Theory, 3–47. New York: Oxford University Press. Wurm, Steven A. 1991. Language death and disappearance: Causes and circumstances. In R. H. Robins and E. M. Uhlenbeck, eds, Endangered languages, 1–18. Oxford: Berg.

Part one Language Change in an Obsolescent Language

∵

chapter 1

Grammatical Change in a Dying Dialect 1 Introduction Scottish Gaelic is a receding language throughout all of Highland Scotland, but most especially on the eastern side of the Highland mainland. In the county of Sutherland, in the extreme north of the Highland area, Gaelic has almost disappeared along the east coast during the past century and a half. The sole surviving pockets of Gaelic speech are located in the coastal villages of Brora, Golspie, and Embo, where the descendants of a former population of fisherfolk are bilingual in English and a unique local form of Gaelic. Fifty years ago fishing still existed as a livelihood. At that time the fisherfolk were occupationally isolated in a district of tradesmen and agriculturalists. They were not linguistically isolated, however, because the agricultural population of the surrounding countryside was still bilingual in English and a Gaelic very little different from that of the fisherfolk. By the 1960s and ’70s, when the fieldwork represented in this paper was done, the remaining Gaelic-English bilinguals had become a small and aging group, a true relic population. What was once occupational distinctiveness had become linguistic distinctiveness. This paper examines certain grammatical changes in progress in this isolated speech community.1 Gaelic is moving toward extinction in each of the three East Sutherland (ES) villages, but the process is much further advanced in the relatively large villages of Brora and Golspie (pop. ca. 1400 and 1300 respectively) than it is in Embo (pop. ca. 275). When fishing failed as an industry in Sutherland in the years after World War I, Brora and Golspie were growing villages, and the former fisherfolk were swamped in an expanding monolingual English-speaking population which has gradually absorbed them. Embo during the same period shrank drastically but preserved its Gaelic character, primarily because it existed in geographical isolation and solely as a fishing settlement, so that 1 This chapter is based on research begun in East Sutherland during a year’s residence in 1963–64, and continued at frequent intervals up to the present. A grant from the American Council of Learned Societies made possible a full summer’s work on the project in 1970, and the School of Scottish Studies was generous with space and facilities during that same period and others. I would like to thank Howard Hoffman for help with the graphs which appear in this paper.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�03

34

chapter 1

there was nothing to attract new population to the village. As a result, GaelicEnglish bilinguals continued to appear in Embo for about 25 years after active bilingualism ceased to be the norm for fisher-descendants in the two larger villages. The youngest fluent bilinguals in Embo are currently in their early 40’s. These are people who spoke Gaelic regularly at home and on the school playground in their childhood years, and who now explicitly claim Gaelic as their mother tongue, even when they acknowledge that they are more competent in English. In Brora and Golspie, a group with so consistently Gaelic a background can be found only among the 70- and 80-year-olds. The current bilinguals of Brora and Golspie had parents who were very nearly monolingual in Gaelic, for the most part. So did the older bilinguals in Embo. But the younger Embo bilinguals grew up in a community which lacked a monolingual Gaelic norm. These younger Embo bilinguals express doubt about the “correctness” of their Gaelic, and often remark that their Gaelic is inferior to that of their parents and grandparents. Explicit comment on the decline in the quality of their Gaelic focuses almost entirely on the lexicon, however: the younger speakers feel sure their elders had many more ‘words for things’ than they have themselves. There is a much lower awareness of only one instance of ongoing analogical leveling in the morphology, and some sporadic note is taken of certain phonological developments; but there is no awareness at all in the community of developments currently underway in the grammar of the so-called ‘initial mutations’. 2

Initial Mutations and Change

Initial mutations are phonological alternations in certain word- or root-initial consonants in a wide variety of syntactic environments. Syntactic environments requiring mutations are so numerous, in fact, that a Gaelic sentence is rarely spoken that does not contain at least one mutated morpheme, whether the variety spoken is East Sutherland Gaelic (ESG) or a western dialect. Since speakers are unaware of changes in the mutational grammar, mutational phenomena are not subject to conscious correction in attempts at ‘good’ linguistic behavior. Aside from the usual difficulties of eliciting very specific structures from linguistically unsophisticated speakers, then, these phenomena lend themselves particularly well to an investigation of ongoing change. Initially, it was as little apparent to me as to the ESG speakers themselves that anything like systematic change was underway in the grammar of the initial mutations. My notes seemed to show a dismal patchwork of inconsistencies and (from the point of view of the standard language) mistakes, haphazardly distributed over villages, speakers, and occasions. Since the dialect was so

grammatical change in a dying dialect

35

clearly in a terminal phase in the area, it was easy to suppose that this motley picture reflected a sporadically disrupted stage in the decay of the local form of the Gaelic language. It was only the gradual development of an ever wider age-range of informants that brought the situation into clearer focus. Some of my original and oldest informants died, and the speakers who replaced them in my sample were inevitably a little younger. In Embo, where a younger group of speakers is statistically preponderant, I originally aimed to work only with older speakers, i.e. people of 60 or more. But eventually I developed a younger group of informants as well, through sheer statistical availability and the friendly helpfulness of younger friends and acquaintances, speakers between 40 and 60. As the sample of speakers broadened, and younger speakers either replaced or supplemented older speakers, I gradually became aware that the apparent “mistakes” of the vanishing oldest speakers were somewhat commoner among the slightly younger speakers, and extremely common among the youngest Embo speakers. 2.1 Mutational Change and Terminal Linguistic Stages If generational patterning representing progressive stages of change is in fact discernible in the data collected, certain questions suddenly become not only possible but necessary. These questions deal both with the grammatical specifics of the Celtic dialect in hand, and also with the nature of grammatical change in dying languages in general. Among the former: Do all syntactic environments which require mutation show change in progress, or only some among the many? Are there limits to the extent of change possible – e.g., is one mutation merely substituted for another, or can a mutation be eliminated altogether? Is a change in the mutational system matched, or compensated for, by changes elsewhere in the grammatical system? And among the larger questions pertaining to the nature of linguistic “death”: Can grammatical change in the final stages of a language rushing toward extinction be shown to proceed in much the same orderly fashion as grammatical change in less drastic phases of linguistic evolution? How does a language compensate for or do without information lost by large-scale changes in its grammatical system? Where only a few last bilingual speakers survive as representatives of a vanishing speech community, can their grammatical system reliably represent the language in question for purposes of description or of historical reconstruction? Evidence from a single dying dialect can scarcely provide definitive answers to these questions for all moribund Celtic dialects or for dying languages in general. But the paucity of our information about the process of linguistic extinction makes all evidence about terminal linguistic stages welcome. As Lehmann (1964: 111) has noted:

36

chapter 1

Numerous instances are attested of languages going out of use: Cornish in the eighteenth century, Dalmatian in the nineteenth and today many indigenous languages throughout the world. Thorough documentation of the stages leading to their extinction would be of great interest to historical linguistics. For many languages of which we know are now extinct; the steps to their extinction may be understood more clearly if we have thorough descriptions of languages now on the way to extinction. In ES we have a well-defined linguistic microcosm in which to observe linguistic phenomena, an island of Gaelic speech isolated by geography, and by virtual Gaelic illiteracy, from other Gaelic dialects.2 In the village of Embo we have a speech community where Gaelic has a foreseeable and none-too-distant terminal date, and yet is still in vigorous use by a bilingual population which includes nearly 50% of the adults. In Embo, further, there is a more than four-decade age difference between the eldest fluent Gaelic speakers and the youngest, so that there is scope for age-correlated speech patterns to appear. Consequently, although the “conservative” mutational patterns discussed in this paper are established on the basis of usage common to the eldest speakers in all three ES villages, the focus on change throughout will necessarily be within Embo village, since only there do we find an age-span of fluent speakers which permits true generational contrast. 3

Change-Resistant Initial Mutations in Esg

In ESG the mutational system may be said to be the same for all age groups in the following senses: every speaker has retained an active use of initial consonant mutation as a syntactic device; and every speaker has a repertory of 2 Since the herring fishing began to die out more than 50 years ago, and the last of the GaelicEnglish bilinguals of the agricultural population died out during the same period, the East Sutherland fisherfolk and their descendants have been increasingly isolated linguistically. But even during the heyday of the herring industry, when both men and women travelled extensively around the coasts following the fishing, the aberrance of ESG made interaction with other Gaels difficult. This same aberrance makes contact with the standard language – in writing, on radio, or in church services – of relatively little reinforcement value to East Sutherlanders. No surviving East Sutherland Gaelic speaker is actively literate to the point where he or she could write a letter wholly in Gaelic, and this again reflects the distance between the standard language and the unwritten dialect. [Correction, 2014: G3 subsequently proved able to do so.] Passive literacy exists in a small number of cases, but it is almost exclu­sively confined to a limited ability to read the Bible, especially the metrical version of the Psalms.

grammatical change in a dying dialect

37

initial consonantal choices which includes (to use the traditional terminology) those of “lenition”,3 those of “nasalization”,4 and those of the unmutated radical initial. Given this basic system in common, and given a group of older, more conservative speakers who represent a mutational norm, the younger speakers’ departures from that norm might consist in (1) substituting the opposite mutation for the one their elders use; (2) using a mutation where their elders use none; (3) using either mutation interchangeably where their elders use only the one or the other; and (4) using no mutation at all where their elders obligatorily use one or the other. All these deviations do in fact occur, but not all of them in any one syntactic environment. Each group or sub-group of mutational environments shows deviant developments peculiar to itself, and consequently each must be examined separately with regard to the pattern of change involved. This paper will concentrate on two striking instances of change in progress, and contrast these with environments where little or no change appears. 3.1 Mutation in the Verbal System Of greatest interest, perhaps, are those structures which show no change at all in mutational pattern and remain resistant to change throughout the entire age-range of fluent Gaelic speakers. Such resistance is shown in ESG only by the verbal system: the mutational markings of tense, relative mode, and negation show no age-correlated variability whatsoever. (Such rare variants as occur are either idiosyncratic, peculiar to a given individual; or geographical, peculiar to a given village.)5 Mutational marking of dependent mode and interrogation is

3 In most cases lenition requires the replacement of an initial obstruent by a spirant. But in the case of initials which are themselves already spirants, and also the initial consonants of certain clusters, the replacement may be another spirant or zero. The lenition replacements in ESG are as follows (V = vowel, C = consonant, I = front vowel, and U = back vowel): Lph = [f]; LthV, LsV, LchI, LsI = [h]; LthC = zero; LčhU, LšU, L khy, LkI = [ç]; LkhU = [x] (Embo and Brora also have Lt, LkU = [x], and Golspie has Lt, LkU = [γ]); Lf = zero; Lmy, LmI, Lpl, Lpy, Lpr, LpV = [v]; Lpw = zero; Lč, LkI = [y]; Lky = zero; Lsn, Lsn´ = [r ~ n], Lstr = [r]. 4 In most cases, nasalization in ESG requires the replacement of an initial voiceless consonant by a voiced consonant – usually, although not always, the voiced equivalent of the original radical initial. The nasalization replacements in ESG are: Nph, Np = [b]; Nth, Nt, NsV = [d]; Nčh, Nč, Nš, Ny = [dž]; Nkh, Nk = [g]; Nstr, Nsn, Nsn´ = [dr]. 5 A considerable isogloss bundle separates Embo village from Brora and Golspie, with phono­ logical isoglosses predominant, but lexical, morphological, and mutational isoglosses repre­ sented as well. A very much smaller number of isoglosses runs between Brora and Golspie villages, and there are even rare isoglosses which link Brora and Embo as opposed to Golspie, or Golspie and Embo as opposed to Brora.

38

chapter 1

nearly as resistant: the only age-correlated feature is the optional extension of nasalization to the sibilants /s/ and /š/, where the majority of the older, conservative speakers nasalize these phonemes only in the nominal system.6 The counter-traditional development here is that younger speakers use a mutation, namely nasalization, where their elders use none. The model for the innovation is in this case perfectly clear: a morphophonological rule which formerly applied to only one word-class (noun) has been extended by analogy to apply to another word-class (verb) as well.7 Aside from this single analogical development, the verbal system of ESG is strikingly stable in its mutational patterns. Even the youngest fluent speakers never fail to lenite in the independent preterit or conditional, or to nasalize in the interrogative positive or the dependent mode after {kənN} ‘that (conj.)’, for example. So reliable are these mutations among fluent speakers, in fact, that if one encounters an ESG speaker who confounds them, that confusion can be taken as a sure indication that the speaker in question is not fluent and will make gross errors elsewhere in his Gaelic.8 The mutational conservatism of the ES verb system is the more striking because lexically the verb class has suffered more loss than any other; its lexicon is extremely weak and shows borrowing from English on a truly massive scale. Since there are plenty of verbs whose initial consonants are not susceptible to mutation, it is untenable to argue that the grammatical system can not do without the information supplied by the mutations. The mutational information, while prominent in the greater number of verbs, is redundant: morphology and word order supply the same information about tense, mode, negation, and interrogation. The mutational stability of the verb system remains unexplained, but it is perhaps significant that Gaelic has no words corresponding to English yes and no; all yes-no questions are answered by verb phrases with subjects and objects deleted. This means that even the very young Gaelic-speaking child is confronted almost immediately with the neces6 I.e., younger speakers may say /ən ǰo:ɫ ən ǰo:ɫədar/ [ən džo:ɫ ən džo:ɫədar] ‘Will the sailor sail?’, while older speakers are much more likely to say /ən šo:ɫ ən ǰo:ɫədar/ [ən šo:ɫ ən džo:ɫədar]. 7 This represents the development which Kiparsky (1968: 200) predicts in the movement of rules, so that the rule for the nasalization of the groove sibilant now affects more items than it did previously. 8 Imperfect bilinguals who misuse the verbal mutations exist; they include Brora and Golspie ‘semi-speakers’ under 60 and Embo ‘semi-speakers’ under 40. They are excluded from discus­ sion in this chapter by definition, since they are not fluent speakers. Their mutational usages reflect not genuine changes in the mutational system, but imperfect control of Gaelic in general, as revealed by many other irregularities in their speech.

grammatical change in a dying dialect

39

sity of handling the full verbal apparatus. Most of the highest frequency verbs are extremely irregular and do have mutatable initials. If we can assume early mastery of this system, and inordinately heavy reinforcement of its mutational phenomena because of the full mutational systems of the highest frequency verbs, then we may after all have at least a clue to the conservatism of the verbal grammatical system, including mutations. 3.2 Obligatory-Lenition Adjuncts One other group of mutational structures is highly resistant to change. These are lenitions which occur obligatorily after a small group of high-frequency morphemes: the adverbs {kle:L} ‘very’ and {rɔ L} ‘too’, the numeral {ta:L} ‘two’, and the adjective {a:L} ‘next’. Despite the fact that none of these mutating elements is ever deletable, and that the subsequent mutation consequently provides no grammatical or semantic information at all, it is only the very youngest of fluent speakers in Embo who ever fail to produce lenition after these elements. Such failures are extremely rare, and they always consist of the absence of any mutation rather than the substitution of nasalization. The occasional exceptional failure of lenition to occur in these paradigmatically isolated environments represents, not so much a change currently underway, but a point at which change might be expected to appear on a serious scale if still more generations of fluent bilinguals were on hand in Embo: this obligatory mutation might in time become optional, or perhaps even disappear. Such developments would seem, on the basis of comparative evidence from other dialects, to have appeared before in ESG. E.g., obligatory lenition after the adjectives sean ‘old’ and droch ‘bad’ is reported for two western dialects,9 but it is entirely optional in ESG. The total lack of grammatical information supplied by the obligatory mutations of the kind discussed in this paragraph makes it easy to understand how variability might have come into being with sean and droch, but difficult to understand why the variability is not further advanced with {kle:L}, {rɔL}, {ta:L}, and {a:L}, unless the sheer statistical frequency of these morphemes has a braking effect on the process of change. The structures discussed so far have shown themselves resistant to change. To find structures which are not resistant, but rather show well-developed patterns of ongoing change, we must turn to the pronominal and nominal systems. Here we find two grammatical sub-systems in flux, in each of which there is traditionally a set of paradigmatically related initial mutations: the pronominal system within the passive construction, and the case system. Since these are complex cases of change in progress, I will take them up in detail in 9 See Oftedal (1956: 200) and Holmer (1938: 99).

40

chapter 1

the following sections as examples of the evolution of mutational grammar in terminal bilingualism, hoping to gain thereby at least partial answers to some of the questions raised in §2.1 above.

The Passive Construction

4

The Possessive Pronouns

Although the term ‘passive’ suggests a verbal structure to the English speaker, the passive in ESG is essentially a nominal construction,10 centered on a possessive pronoun modifying a gerund (traditionally called the verbal noun in Gaelic grammar). It is the possessive pronouns in this construction which cause the mutation, and it is within the mutational system of the possessives that change is taking place. The choice of the passive as the construction in which to study the mutational changes involving the possessives is dictated by the fact that the passive is the only construction in which the possessive pronouns survive as a productive system in ESG. Even the full paradigm of possessive pronouns is difficult to elicit in ES, although the paradigm, once obtained, proves to parallel that of the standard grammar very closely. The possessive pronouns which cause mutation in conservative ESG are precisely the ones which produce mutation in the standard language, and the mutations produced are also the same:

1st person 2nd person 3rd person masc. fem.

10

11

sg

pl.

mə L tə L əL ə

nə11 nə ənN

The truly verbal passive, in which the category is expressed by the suffixal morphology of the verb itself, survives only as a fossil in the expression rugadh mi /rugu mi/ ‘I was born’ and possibly in one or two other less common expressions. The phonological shape of this possessive (1st and 2nd plural) is peculiar to ESG and other northern dialects, but the lack of consonantal mutation after it is characteristic of the standard language as well.

grammatical change in a dying dialect

41

Except for their appearance in the passive, all but two of these possessives are little-used in ESG, with the plurals rarest of all. The full range of possessives, with appropriate mutations or lack of mutation, can be elicited from older speakers, and even occasionally from some younger speakers, in only one environment other than the passive: fossilized fixed expressions involving kinship terms (‘when our father was alive’), body parts (‘at her back’, i.e. ‘behind her’), or other inalienables (‘they lost their lives’). Outside these fixed expressions, the pronouns of the 1st and 2nd singular occur freely;12 but the other singulars typically, and the plurals invariably, are replaced by an alternative possessive structure involving the definite article and the preposition aig /ig/ ‘at’, which conjugates for person. Thus with a clothing term, e.g., we would typically get a syntactically suppletive paradigm: {khačh ənN bel məL phεčan}13 {khačh ənN bel təL phεčan} {khačh ənN bel ənN phεčan ig} {khačh ənN bel ənN phεčan εkh}

‘Where is my sweater?’ ‘Where is your-sg. sweater?’ ‘Where is the sweater at-him?’ ‘Where is the sweater at-her?’

and so forth through a plural formed with the aig-construction exclusively. 4.1 The Passive Fortunately there is no syntactic alternative to using the possessive pronouns in the passive, so that the passive can be counted on to produce a possessive. There are actually two different passive constructions in use in ESG, each of which calls for the possessive.14 The chief differences between them are that they use different finite verbs and that one requires a subject pronoun while the other (in the usage of conservative speakers) does not. In the passive without subject pronoun, the finite verb is supplied by the irregular verb dol 12

13

14

I.e., they occur freely within the semantic range open to possessives in ESG, chiefly inalienables and things that can be worn or carried on the person. If the noun modified does not fall into this semantic range, the alternative aig-construction will be used. /bel/ ‘is’ (interrogative) shows nasalization in Embo, lenition in Golspie and Brora. The phonological form given for paitean ‘sweater’ is also the Embo version; Golspie and Brora have a close [e] in the same word. Since most of the informants whose speech is cited in this study are from Embo village, Embo forms will be used throughout unless otherwise noted. Both are to be found in use in western dialects and in the written standard, but grammars of Gaelic, whether prescriptive or descriptive, often ignore them. Neither is mentioned in, e.g., Maclaren’s Gaelic self-taught (n.d.), or in Oftedal’s relatively full treatment of a Hebridean dialect (1956), although examples of both appear in the sample texts Oftedal provides.

42

chapter 1

‘go’; this construction will be referred to as the dol-passive. The full construction is made up of a finite verb followed by a subject which consists of a verbal noun modified by a possessive pronoun: {Lga məL khumal ə stε} /xa mə xumal ə stε/15 Chaidh mo chumail a staigh, lit. ‘went my keeping in’, i.e., ‘I was kept in.’ The possessive pronouns of the 3rd singular and plural usually appear homophonously as /ə/ or, after a preceding word-final vowel, as zero; hence only the mutations resulting from the possessive can specify the subject of this construction for the 3rd person, if the subject is a pronoun rather than a noun: {Lga əL khumal ə stε} /xa xumal ə stε/ ‘He was kept in’. {Lga ə khumal ə stε} /xa khumal ə stε/ ‘She was kept in.’ {Lga əN khumal ə stε} /xa gumal ə stε/ ‘They were kept in.’ The second passive construction takes its finite verb from the irregular verb bith ‘be’, and will be referred to hereafter as the bith-passive. The bith-passive consists of finite verb plus subject pronoun, followed by a prepositional phrase in which the possessive once again modifies a verbal noun: {Lba mi er məL khumal ə stε} /va mi er mə xumal ə stε/ Bha mi air mo chumail a staigh, lit. ‘I was on my keeping in’, i.e., ‘I was kept in.’ In this construction, too, the possessive pronouns of the 3rd persons usually appear homophonously as /ə/ or zero, but here the presence of a preceding subject pronoun makes the information provided by the mutation redundant. 4.2 Change in the bith-passive In view of the redundancy of the mutations in the bith-passive, it is not surprising to find that younger speakers tend to generalize lenition, the commonest mutation required by three of the possessives, including the two most

15

Standard Gaelic chaidh ‘went’ is not the immediate source of East Sutherland /γa/ (Golspie) ~ /xa/ (Brora and Embo). In order to account for the voiced Golspie version, the underlying base form would have to be either {Lta} or {Lka}; I have arbitrarily chosen the latter to represent the base here. Note that the symbol a, in phonetic as well as phonemic and morphophonemic transcriptions, is used for typographical convenience to represent a back unrounded vowel.

grammatical change in a dying dialect

43

frequent, throughout all persons. Embo speakers in their 40’s and 50’s,16 for example, produced the utterances {Lba aǰ er əL khumal} /va aǰ er ə xumal/ ‘They were kept’ {Lba i er əL khumal} /va i er ə xumal/ ‘She was kept’ with inappropriate lenitions; yet no confusion results, thanks to the presence of the subject pronouns. In these examples, younger speakers deviate from the conservative mutational pattern for possessive pronouns either by producing a mutation where older speakers have none (3sg. fem.) or by producing the opposite mutation where older speakers have nasalization (3pl.). Younger speakers also show a rather rare alternative deviation in the 3pl., namely no mutation where older speakers have nasalization, as in E11’s {Lba aǰ er ə throgal l´eš ənN šεnuar} /va aǰ er ə throgal l´eš ə ǰεnuar/ ‘They were raised by their grannie.’ This occasional absence of mutation in the 3pl., though rare, has an obvious model in the normal absence of mutation in the other persons of the plural and in the 3sg. feminine. 4.3 Change in the dol-passive Because the initial mutation or lack of mutation with the possessives follows an explicit subject pronoun in the bith-passive, it is informationally unimportant that many younger speakers use initials which are inappropriate from the point of view of both traditional Gaelic grammar and conservative ESG grammar. In the dol-passives, on the other hand, any such use of inappropriate initials would seem to be disastrous, since there is normally no subject pronoun to specify person and number. Younger speakers obviate the apparent need to follow conservative mutational patterns in the dol-passive by introducing an intrusive subject pronoun, borrowed by analogy to the bith-passives. Some younger speakers occasionally introduce this subject pronoun and still preserve the conservative mutation, as E10 does in the sentence {Lga aǰ ənN pa:u s ə ɫɔx} /xa aǰ ə ba:u s ə ɫɔx/ ‘They got drowned in the loch.’ 16

E17 and E11 respectively in a chronological numbering of informants which makes E1 the oldest Embo informant, E2 the second-oldest, and so forth through E19.

44

chapter 1

But once the subject pronoun is present, the younger speakers are much more likely to extend lenition to all persons, exactly as they do for the most part with the bith-passives. Thus E16 produced a 3pl. with lenition in the sentence {Lga aǰ əL khur a max} /xa aǰ ə xur ə max/ ‘They were put out’ where traditional grammar would call for a nasalization. Analogy to the bith-passives goes even beyond the introduction of a subject pronoun for most younger speakers. The preposition /er/ ‘on’ is most commonly introduced along with the subject pronoun, and this reduces the difference between the two passive constructions to one of finite verbs only. This was true, for example, of the younger speakers E17 and E18, who could respond to instructions to begin a passive with either finite verb, but who used only the one or the other, usually with no difference in the rest of the construction, when left to their own devices. E17 is a bith-passive user and E18 a dol-passive user, but their usages differ only with respect to those finite verbs; witness E17’s {Lba aǰ er əL khrɔxu} /va aǰ er ə xrɔxu/ ‘They were hung’ and E18’s corresponding {Lga aǰ er əL khrɔxu} /xa aǰ er ə xrɔxu/ ‘They were hung’ each with subject pronoun, preposition /er/, and lenition analogically extended to the 3pl. If we order the dol-passives obtained from Embo speakers by age of speaker, as in Table 1 below, we find the analogical development of the resemblance between the dol-passive and the bith-passive plainly set forth. There is a total absence of analogical constructions for the two oldest speakers; they use the dol-passives in the traditional form. In the second decade-group, however, analogical forms appear: most prominently with the subject pronoun only, but once with both subject pronoun and preposition. In the next decadegroups, the most fully remodeled analogical constructions become dominant, and finally E19, the youngest fluent speaker, produces only forms identical with bith-passives except for the finite verb. The same results are presented in graph form in Figure 1, where the percentage of conservative dol-passive forms without any analogical intrusions is plotted in terms of the speakers’ age-groups. The function moves dramatically from 100% in the case of the oldest speakers to zero in the case of the youngest speakers.

45

grammatical change in a dying dialect Table 1.1

Embo speakers’ dol-passives in the 3sg. f. and 3pl.

Decade group

Speaker

No Subj. Pron., No Preposition

Subj.Pron. only

80’s

E1

9

–

70’s

E2 E3 E4 E5

6 3 2 4

60’s

E6 E7 E8 E9

– 3 7 1

50’s

E10 E11 E12 E13 E14

1 – 7 1 2

40’s

E16 E17 E18 E19

– – – –

9 (100%)

15 (68%)

11 (52%)

11 (27%)

0

– 2 1 3 1 – – 1 3 – – 2 3 4 1 3 –

Subj.Pron. And Prep.

0

6 (27%)

2 (10%)

8 (19%)

8 (22%)

–

0

Totals

1 (5%)

Totals

8 (38%) 2 5 5 3 7 22 (53.5%) 12 3 9 5 29 (78%)

Totals

– – 1 – 2 1 1 4

Totals

Totals

We have already seen that the analogical remodeling of the dol-passive after the pattern of the bith-passive allows variability in initial consonants to appear in the verbal noun of the dol-passive as in that of the bith-passive. Table 2 shows the number of actual instances of passives produced with appropriate vs. inappropriate17 initial consonant for both passives in the interview

17

‘Inappropriate’ here means either lenition or no mutation at all, since the environments are the 3sg. feminine and the 3rd plural.

46

chapter 1

120 100 80 60 40 20 0

80

70

60

50

40

Figure 1.1. Percentage of conservative (non-analogical) dol-passives plotted according to the age-group of the speakers.

situation.18 The three eldest speakers show no instances of inappropriate initials at all, and instances remain rare through the third decade-group. The significant rise in constructions with inappropriate initial consonants appears among speakers in their 50’s – i.e., among the first decade-group of what we will label “younger speakers”. Here, suddenly, more than 50% of the structures produced have an inappropriate initial. And in the youngest decade-group of all, the statistics abruptly reverse, so that inappropriate initial consonants become more than twice as common as appropriate ones. It is of interest that the speakers represented in this table include two mother-and-son pairs, in which the mother is at least 25 years older than her son: E4 is the mother of E19, and E6 is the mother of E18. The few examples available for E6 show some foreshadowing of her son’s usage, since she too 18

The interview called for Gaelic translations of English originals. Past experience had shown that the bith-passives were normally produced as translation equivalents of the English be-passive, whereas I was especially interested in developments in the dol-passive; hence the English models given were all get-passives. Among older speakers this proved to produce a bias against the bith-passive, which I allowed to go uncorrected because I already had ample information from early fieldwork on the conservative bith-passive. Younger speakers tended to use one of the two finite verbs to the exclusion of the other, more often bith than dol, and in such cases I instructed them to “begin that sentence with /xa/” or “begin that sentence with /va/” in order to test for the structure of both passive constructions. These procedures account for the over-all preponderance of dol-passives in the data.

47

grammatical change in a dying dialect Table 1.2

Incidence of appropriate and inappropriate initials in the 3 sg.f. and 3 pl. of both passives 3 sg.f. and 3 pl. bith – passive Appropr. Inappropr. Initial Initial

3 sg.f. and 3 pl. dol – passive Appropr. Inappropr. Initial Initial

9 9 6 3 3 5 17 1 4 8 5 18 4 2 10 4 6 26 9 1 1 – 11

Decade group

Speaker

80’s

E1

–

–

70’s

E2 E3 E4 E5

60’s

E6 E7 E8 E9

50’s

E10 E11 E12 E13 E14

40’s

E16 E17 E18 E19

2 – 5 3 10 – – – 1 1 8 3 – 3 – 14 1 5 – – 6

– – 1 1 2 – – – – 0 3 2 – 2 1 8 1 10 1 2 14

– 0 – – 1 2 3 2 – – 1 3 2 3 2 2 6 15 7 4 11 5 27

Totals

Totals

Totals

Totals

Totals

produces more inappropriate than appropriate initial consonants. But E4, in the next-older decade-group, differs sharply from her son. Her use of appropriate vs. inappropriate initials runs four to one in favor of appropriate initials, whereas her son’s usage runs zero to seven the other way. Since the son is unmarried and lives in his mother’s household, the contrast highlights rather dramatically the differing usage of the eldest and youngest speakers.

48

chapter 1

The results of Table 2 are presented in graph form in Figure 2, where the percentage of forms with appropriate initial consonant (taking both passives together) is plotted in terms of the speakers’ age-groups. Except for a slight rise between the 70 to 60 decade-groups, the function exhibits a clear downward movement from 100% for the octogenarian E1 to 29% for the youngest decade-group. 4.4 Summary of Changes in the Mutational System with the Passive The mutating elements in the passive construction are the possessive pronouns, three of which are homophonous in the presence of a preceding consonant and even entirely deletable in the presence of a preceding vowel. In such an environment the subsequent mutation carries a heavy informational load if it is the sole indicator of person and number, as in dol-passives where the subject is not a noun. This might seem to be the least likely environment in which to find change underway in the mutational grammar. But because the grammar provides another passive structure in which a subject pronoun makes the mutational information redundant even in environments with a homophonous possessive, the younger speakers are able to analogize a subject pronoun into the dol-passive and dispense with the conservative highly differentiated mutational grammar of the possessives.19 Given the development of 120 100 80 60 40 20 0

80

70

60

50

40

Figure 1.2 Percentage of inappropriate initial consonants in both passives plotted according to the age-group of the speakers 19

It should be noted that no variation occurs in dol-passives unless the subject pronoun is intruded. Subjectless dol-passives are mutationally perfectly orthodox, even in ESG.

grammatical change in a dying dialect

49

mutational redundancy in the dol-passive and the built-in mutational redundancy of the bith-passive, one might imagine that the use of mutations would disappear altogether in these environments, but this does not happen. Instead, far the commonest development is that the obligatory nature of lenition after the two common (and non-deletable) possessives {məL} and {təL} is extended to possessives of all persons and numbers. This reduces the whole range of possessives to the status of mutationally meaningless obligatory-lenition adjuncts on the order of the adverb {kle:L}, the numeral {ta:L}, and so forth. The fact that some of them can be deleted, unlike {kle:L} and {ta:L}, is not important, because the remodeled analytic grammar of the passive requires that they occur in conjunction with subject pronouns. There is no serious loss to the grammar in this whole process, since the grammatical category ‘passive’ survives and continues to contrast with the category ‘active’. 5

The Case System

Turning now to the nominal grammatical system, it is within the co-ordinate case systems of the article and the masculine noun that we find mutational variability which correlates once again with the age of the speaker.20 5.1 The Genitive The most weakly age-correlated variability occurs in the remnants of an attributive noun structure; in this structure the genitive case of standard Gaelic 20

By considering only the noun accompanied by the article, we deliberately exclude the vocative from consideration. Aside from the vocative, the noun without article shows no initial mutations, and the final mutations of the case system do not survive for either the definite or the indefinite noun. As for the vocative, lenition of the initial consonant is the sole sign of its occurrence, in nearly all ESG nouns. (The preceding vocative particle a of the standard language is as good as non-existent in the dialect, and the word-terminal morphology of the vocative exists as a relic in at most five or six words, optionally at that.) The only mutational “change” apparent in ESG vocatives is occasional unmutated initials produced by the very youngest Embo speakers. These speakers are the same ones who sometimes fail to mutate after {kle:L} or {ta:L}, and the failures are equally rare. These two cases of mutational failure might seem to be functionally very dissimilar, since mutation after {kle:L} and {ta:L} is totally redundant informationally, whereas mutation with the vocative is normally the sole marker of the grammatical category. But in fact the voca­ tive is so strongly marked by phonological signals (e.g. non-final contour with a sharp preceding rise in pitch and a drop in pitch for the vocative noun itself) that loss of the mutation produces no loss of information: phonological markers simply take over for the morphophonological marker.

50

chapter 1

grammar is partially preserved. However, the structure is poorly represented in ESG in general, since prepositional phrases are strongly preferred to attributive noun constructions, and my materials are much poorer for this structure than for the other mutational environments discussed so far. Variability occurs only in the case of attributive structures which have a masculine definite noun in second position.21 Here the sole sign of the older genitive structure is a lenited initial in that masculine noun, and the competing structure is a nasalized (i.e. nominative) noun. The relic genitive of ESG has none of the word-final morphonological apparatus of the standard-language genitive, e.g. brogach / prɔgax/ ‘lad’, gen. brogaich /prɔgiç/, or ceann /khyãũ:n/ ‘head’, gen. /khĩ:n´/.22 The contrast between a mutationally conservative genitive and an innovative nominative can be seen in the renditions of two older-generation Golspie speakers and two younger Embo speakers for the phrase ‘the lad’s father’. All the speakers, old and young, would have preferred to use the prepositional phrase an athair aig a’ bhrogach ‘the father at (i.e., of) the lad’, but when persuaded to drop the aig, the two older speakers gave a lenited genitive version, athair a’ bhrogach /a:r ə vrɔgax/, whereas the two younger speakers substituted a nasalized nominative, athair am brogach /a:r ə(m) brɔgax/. Neither the older nor the younger speakers concerned used the lenition-genitive or the nasalization-nominative exclusively, however. One of the same Golspie 70-year-olds produced a nasalized mathair am boiroinnach /ma:r ə(m) bɔrn:ax/ ‘the woman’s mother’ in the course of a lengthy tape-recorded story, and the same two young Embo speakers produced a lenited cu a’ chiobair /khu: ə çi:bar/ in the test-phrase ‘the shepherd’s dog’. But on the whole it is true that older speakers are more likely to use the lenited genitive in noun-attributive constructions than are younger speakers. Of the genitive in general it can probably be said that, with the loss of virtually the entire word-final morphology and morphophonology of that case, the sense of the genitive as a distinct grammatical category is greatly weakened. Whether as cause or effect, the syntactic environments which call for a genitive in the standard language fail to do so in ESG: there are no prepositions which require the genitive, and verbal nouns take nominatives rather than genitives as their objects.23 Even the definite 21 22

23

Feminine definite nouns in second position show no trace of the unlenited forms of the standard grammar; instead, a lenited nominative always occurs in ESG. It is possible to supply an ESG pronunciation for the missing genitives in these cases because the genitive singular coincides with the plural, and ESG does preserve the pala­ talized plurals. Nominative and accusative are not distinguished. The term nominative will be used to cover the general nominative/accusative form.

grammatical change in a dying dialect

51

noun attributive structures are replaced by prepositional phrases in the vast majority of cases. With this over-all weakening of the category ‘genitive’, the mutational signals of the genitive, instead of taking up the slack left by the loss of word-final genitive markers, are losing ground themselves. Nonetheless, the ground lost is yielded not to an unmutated form, but to nasalized nominative forms. That is, it is still unthinkable that no mutation at all should occur after the definite article. Mutation as a phenomenon survives even where a particular mutationally signaled case does not.24 5.2 The Nominative On the surface, there is relatively little mutational change underway in the nominative case in ESG. Two classes of nouns, one defined by phonology and one by grammatical gender, show no mutational change at all. Any noun of either gender with initial /th t čh č š/ nasalizes in the nominative after the definite article; and feminine nouns with any other initial consonant lenite after the definite article in the nominative.25 Among masculine nouns, mutational variability may appear wherever the initial consonant is other than /th t čh č š/, but only masculines with initial labials and velars (/ph p kh k/) will be considered here and in §5.3. Masculines with initial labial and velar stops offer the most profitable opportunity for investigation – because (1) there are many of them, (2) the group forms a coherent phonological class which shows comparable phonetic behavior under mutation, (3) they are affected by both mutations, (4) the mutational replacement is never zero, and (5) there is strong consensus among older speakers on what the mutational forms of these masculines should be in the nominative and dative.26 Where the term ‘masculine’ is used throughout the remainder of this paper, it is masculines of this restricted phonological class with initial labial and velar consonants that are intended. 24 25

26

An unmutated feminine genitive singular in the definite noun does not survive in ESG, even in fossilized phrases. This rule is invariable, but the class “feminine” is itself variable. Gender-class membership may vary from village to village, so that a noun which is masculine in one village (e.g. Golspie /pu:/ ‘shop’) is feminine in the others (Brora and Embo /pu:/); but such geographicallybased differences are rare. A much commoner source of fluctuating gender-class mem­ bership is the gradual transfer of part of the feminine noun-class into the masculine class. Even older speakers show an occasional idiosyncratic masculine which is feminine for their peers, but among younger speakers the transfer can occur on a fairly large scale. Reasons 1 and 5 do not hold for s-initial masculines; 3 and 5 do not hold for m-initial masculines; 4 and 5 do not hold for f-initial masculines. There is no other phonologicallyconstituted and phonetically coherent sub-group of type 2.

52 Table 1.3

chapter 1 Nominative mutations after the definite article

/khe:li/ ‘ceilidh’ /khrãũ:n/ ‘mast’ /khu:/ ‘dog’ /khath/ ‘cat’ /kho:th/ ‘coat’ /khyo:/ ‘fog’ /khyɔ:rd/ ‘tinker’ /khɔr/ ‘kettle’ /khũãn/ ‘ocean’ /khɫɔxar/ ‘mason’ /khɫu:th/ ‘cloth, rag’ /khogu/ ‘war’ /khɔrkh/ ‘oats’ /khrã:ĩ/ ‘bone’ /khalax/ ‘rooster’

N after Art.

L after Art.

7 speakers 7 6 6 7 7 7 7 6 7 7 6 7 6 7

0 speakers 0 0 0 0 0 0 0 0 0 0 0 1 0 0

My early attempt (1968) in the tracking down of the scale and limits of mutational variability involved the use of a vast battery of nominative and dative test-sentences for nouns of both genders and all mutatable initial consonants. Subsequent work on mutational variability drew heavily on the results of this earlier testing, and went on to test further the especially rich class of /kh/initial masculines, which includes more extremely common nouns than does any other. Table 3 shows the results of the earliest testing for the nominative case of definite nouns in 15 /kh/-initial masculines. The respondents were four Golspie speakers, all around 70 years of age; one Brora speaker near 80; and four Embo speakers, one over 60, one in his 50’s, and two in their 40’s. However, there was never a full complement of nine responses because married couples insisted on working together, and only the first response from such a pair was taken as valid – i.e. unbiased – whichever spouse it came from. The table shows a near-monolithic solidarity in favor of nasalization in the nominative after the definite article. The sole instance of lenition after the article came from an Embo speaker in his 40’s. Largely because I was convinced that in daily speech I had met with more variability than these tests showed, I decided to check the apparent solidarity

grammatical change in a dying dialect

53

of this response from another angle. After an interval of from two weeks to two years, the same sentences were read back to five of the original respondents, under the pretense of checking to see whether they were recorded correctly; but in the read-back, the mutation other than the one originally given was deliberately substituted. On this test the results became less unanimous. In 27 cases with /kh/-initial masculines, older speakers insisted on their own original nasalization; younger speakers insisted on their original nasalization in only 13 cases. Older speakers accepted a contrary lenited version in just 6 cases, whereas younger speakers accepted a lenited version in 16 cases. That is, older speakers rejected lenition of these masculines in the nominative more than 80% of the time; younger speakers accepted lenition for the same masculines better than 50% of the time. This evidence of some weakening of nasalization in the nominative among younger speakers was confirmed by a later (1970) test which called for a Gaelic translation of the English sentence ‘Who threw the stone that killed the stonemason?’ This sentence was simultaneously a test for the preservation of gender and for the nominative mutation of the masculine noun. The Gaelic /khɫɔx/ ‘stone’ is a reliable feminine for older speakers, while /khɫɔxar/ ‘stonemason’ is a reliably masculine agent-noun derived from it. Three Golspie and two Brora speakers, all over 70, lenited the feminine /khɫɔx/ according to the invariable rule requiring lenition of definite feminines in the nominative (unless in initial /th t čh č š/), and also nasalized the masculine /khɫɔxar/, again properly from the point of view of conservative usage. The results for Embo speakers are presented in Table 4. Older Embo speakers did exactly as their Golspie and Brora counterparts had done; but younger Embo speakers showed a weak tendency to treat the feminine /khɫɔx/ as a masculine, and the youngest group of Embo speakers showed a deviation toward a lenited masculine nominative after the definite article. Once again the mother-and-son pairs E4 and E19 and E6 and E18 were represented, and the difference in their usages highlighted the departure of the younger generation from the conservative norm, namely nasalization after the article in the nominative.27 5.3 The Dative In standard Gaelic, the dative case is required in the article and the noun after all simple prepositions, while the genitive case is required after compound

27

The symbol < represents an eventual self-correction in the direction the symbol indicates. No such symbol appears between E16’s lenited and nasalized versions of /khɫɔxar/; she gave them both as equally valid alternatives.

54

chapter 1

Table 1.4

Embo speakers’ masculine and feminine nominatives

Decade group L

80’s 70’s

60’s

50’s

40’s

E1 E2 E4 E5 E6 E7 E8 E9 E10 E11 E14 E16 E17 E18 E19

/khɫɔx/, fem. N

L

< E10 E12

< E17

E16 E18 E19

/khɫɔxar/, masc. N

E1 E2 E4 E5 E6 E7 E8 E9 E10 E11 E12 E14 E16 E17

prepositions.28 As noted above, no use of the genitive with prepositions survives in ESG. But in spite of the loss of the prepositional genitive, it is still possible in the prepositional environment, as in the genitive and nominative environments, to have either of two variant forms of the definite noun, namely a lenited form (the survival of the historical dative) or a nasalized form.

28

The prepositions themselves may or may not be mutating elements which cause leni­ tion in a following indefinite noun; in ESG, for example, the prepositions {tə(L)} ‘to, for’ and {kən(L)} ‘without’ lenite an immediately following noun optionally. But the mutat­ing properties of the preposition have nothing to do with the mutation of the definite noun after the preposition, which is a product of the case system alone. Thus the prepo­sition /ri/ ‘to’ does not itself lenite; witness the popular Embo proverb {LkaL ses phokh fɔɫi ri pɔɫ} /xa hes phokh fɔɫi ri pɔɫ/ ‘An empty purse won’t stand against (to) a wall.’ But a prepositional phrase with the definite noun, ‘against (to) the wall’, would be regularly lenited by conservative speakers: {riš əL pɔɫ} /riš ə vɔɫ/.

55

grammatical change in a dying dialect

Table 5 shows the results of the original (1968) prepositional tests on the same masculine /kh/-initial nouns used for the nominative tests.29 For speakers of all ages a lenited dative was the commonest response, but from the Table 1.5

Responses to preposition plus article plus /kh/-initial noun (1968)

Noun

Preposition

Nasalization Older Younger Speakers

/khe:li/ ‘ceilidh’

/ɔrn/ ‘for’ /ig/ ‘at’ /ɔrn/ ‘for’ /yε/ ‘off of’ /er/ ‘on’ /tə/ ‘for’ /ɔrn/ ‘for’ /yε/ ‘off of’ /fo/ ‘under’ /thrε̃/’through’ /as/ ‘in’ /tə/ ‘to’ /ɔrn/ ‘for’ /as/ ‘in’ /vɔ/ ‘from’ /as/ ‘in’ /ig/ ‘at’ /l´eš/ ‘with’ /kus/ ‘to’ /ɔrn/ ‘for’ /as/ ‘in’ /yε/ ‘out of’ /er/ ‘on’ /vɔ/ ‘from’ Totals

1 – 3 – 2 – 4 – 1 2 – 1 5 – 1 – – – 1 3 2 – 1 – 27

/khrãũ:n/ ‘mast’ /khu:/ ‘dog’ /kho:th/ ‘coat’

/khyo:/ ‘fog’ /khyɔ:rd/ ‘tinker’ /khɔr/ ‘kettle’ /khũãn/ ‘sea’ /khɫɔxar/ ‘mason’ /khɫu:th/ ‘cloth’ /khogu/ ‘war’ /khɔrkh/ ‘oats’ /khrã:ĩ/ ‘bone’ /khalax/ ‘rooster’

29

– 2 2 1 1 1 3 2 1 2 1 1 – – – – 3 – – 1 – – 1 22

Lenition Older Younger Speakers

4 6 2 5 3 3 – 4 5 3 5 4 – 4 3 4 4 5 5 2 4 4 3 4 86

3 2 2 2 4 1 1 1 3 2 2 1 1 4 – 2 3 – 1 1 1 1 1 2 41

There are only 14 nouns in the prepositional tests; /khath/ ‘cat’ was inadvertently omitted.

56

chapter 1

point of view of standard grammar the number of nasalized nominative forms was striking. As between older and younger speakers, the absolute number of nasalized nouns was not greatly different, but there was a striking difference in the prepositions which produced the nasalizations for each group of speakers. Just one preposition, /ɔrn/ ‘for’,30 accounted for 15 of the 27 nasalizations in the older speakers’ sentences, whereas all the other prepositions taken together accounted for only 12 additional older-speaker nasalizations. For the younger speakers /ɔrn/ was only one of a great variety of prepositions producing nasalizations; /l´eš/ ‘with’ was also so strong a nasalizer with younger speakers, for instance, that one might expect it to have produced just as many nasalizations if it had cropped up as often in the test-sentences.31 The prominent role played by /ɔrn/ in producing nominatives among older speakers suggested the possibility that, in a conservative stage antedating the usage of the younger speakers represented in the tests, prepositions in ESG might have differed from each other in their tendency to ‘govern’ the dative, and that /ɔrn/ in particular might characteristically have been followed by the nominative rather than the dative – possibly as a replacement for a still earlier genitive such as the standard language requires. To test for regular case differences after prepositions, three /kh/-initial nouns were elicited in sentences with one preposition which had produced mostly nasalizations in the original prepositional tests overall, namely /ɔrn/, and in sentences with one preposition which had produced lenitions equally strongly throughout the same original battery of test sentences, namely /as/ ‘in’. Table 6 shows the results of this test over the full age range of Embo speakers.32 The oldest Embo speakers did in fact show a regular pattern of nasalization after /ɔrn/ and lenition after /as/; and taking these results together with the suggestive older-speaker results of Table 5, we can establish a conservative ESG 30

31

32

Historically /ɔrn/ is a compound preposition which governs the genitive in standard Gaelic. It contains as components the preposition air ‘on’ and noun son ‘sake’. The very oldest ESG speakers occasionally use an uncontracted form /ɔrsɔn/; except in this rare form, the preposition air has the front vowel /er/ rather than a back vowel. There were several more informants in the prepositional tests than in the nominative /kh/-initial tests; thus the total number of responses is sometimes higher. On the other hand it is often possible to use more than one preposition in a given sentence, so that the total number of responses for any one preposition is sometimes rather low. E5 here is unique in using (probably as an affectation) the standard pronunciation /ersɔn/ instead of the local /ɔrn/. His mutational behavior after this idiosyncratic prepositional form is unpredictable; he sometimes lenites, as would (in masculines) the west-coast speakers among whom he long worked and from whom he presumably borrowed the form, and he sometimes nasalizes, as his Embo peers would after the local form /ɔrn/

57

grammatical change in a dying dialect Table 1.6 Noun Prep. Speaker

E1 E2 E3 E4 E5 E6 E9

Embo speakers’ mutations with /orn/ and /as/ /khɔr/ /ɔrn/ N L

N N N N

N N

L

/as/ N L



L L L L L



L

N

/khrãũ:n/ /ɔrn/ N L

N N N N

N N

/as/ N L

L L N1 L2 N1 L2 L L L

L



E10 E11 E12 E13 E14 E15 E16 E17 E18 E19

N N N N

N N N N N



L



L L L L L



L

N N L

N

N N N N N N

N N

N N

L L



L L



L L

N N

N N

/kho:th/ /ɔrn/ N L

/as/ N L

– – L N L N L N L N L N N N N older speakers younger speakers N N N N N N N N N N



L



L L

N N

N N N N N

N1 = N was given first. L2 = L was given second.

norm in which /ɔrn/ governs the nominative rather than the dative case. It is interesting to note that, in the original 1968 battery of prepositional tests, only /er/ ‘on’, one of the component elements of /ɔrn/, also showed a weak nominative-governing pattern (more speakers nasalized than lenited in three of nine test cases). Otherwise all the other prepositions tested followed /as/ in favoring the dative. A reflection of this can be seen in the overall favoring of lenition, by younger speakers as well as older, in the responses of Table 5. Still, older speakers produced less than a third as many nasalizations as lenitions in Table 5, despite the high number of /ɔrn/-sentences, whereas younger speakers in the same sentences produced nasalizations in better than half the

58

chapter 1

cases. And the age-grouped responses of Table 6 show clearly that the actual change in progress in this part of the case system is a movement away from the dative and toward the nominative. With the exception of the idiosyncratic E5, who uniquely used the standard Gaelic form of the preposition /ɔrn/ and twice lenited after it, no older speakers substituted lenition for nasalization after /ɔrn/. On the other hand, older speakers gave a total of five nasalizations after /as/, two of them as first-named variants but the other three as sole versions. Younger speakers, for their part, showed three lenitions after /ɔrn/, but the trend ran more than five times as strongly in the other direction: there were 16 instances of nasalization after /as/. The existence of a preposition which ‘governed the nominative’ in conservative ESG usage must have provided a ready model for nasalization of definite masculines in the prepositional environment, and current younger-speaker usage is increasingly adopting that model. Tape-recorded materials confirm the picture provided by Table 6,33 at the same time offering evidence that a lenited dative is in fact the conservative norm away from which current usage is moving. Older and younger speakers alike show absolutely invariant lenition of the very highest-frequency masculines in the dialect after every preposition but /ɔrn/. The masculines /pal/ ‘village’, /khɫɔdax/ ‘shore’, /phɔrsth/ ‘harbor, ferry’, and /khũãn/ ‘sea’ never nasalize after any preposition but /ɔrn/, regardless of the age of the speaker, and here again we see a conservative norm reflected by the strong resistance to prepositional nasalization with the highest-frequency masculines available. On the other hand, leaving aside this special group of high-frequency nouns, the instance of prepositional nasalization is much higher on tape for younger speakers than for older. To contrast three Embo speakers well-represented speaking freely on tape and widely separated in age: the octogenarian El uses the nominative after the preposition /l´eš/ ‘with’ in addition to /ɔrn/, but otherwise uses the dative; 33

Lengthy tape-recorded reminiscences and stories were collected from as many informants as were capable of speaking comfortably with a recorder going. Most of the tapes were made in the form of interviews between the investigator and the informants, in which questions were asked and answered. The exceptionally interesting tape from which the data on E14 and E16 prepositional usage is taken was made in the absence of the investiga­ tor (but for her benefit) and represents a husband reminding a wife eight years his junior of how things used to be when they were young and Embo was still very much a fish­ ing village. The tape from which the scanty material on E5 and E13 is taken was a lettertape made by those two speakers and the investigator for an emigrant sister in Australia who had been home to visit the year before. The tape was therefore in no way made for research purposes, but by permission of the two Embo speakers a copy was made to that end before the original was mailed.

grammatical change in a dying dialect

59

the husband-and-wife team E14 and E16 not only nasalize four-to-one after / l´eš/, but they also nasalize all but the highest-frequency masculines after /ig/ ‘at’, and they show sporadic nasalization after /kus ~ kəs/ ‘to’, /as/ ‘in’, /er/ ‘on’, and /thrε̃/ ‘through’. To take another case, the brother-sister pair E5 and E13 are poorly represented on tape; yet even in the meager material available, their age difference is reflected in the fact that E5 lenites two of three masculines, while E13 nasalizes four of five. 5.4 Summary of changes within the case system The most conservative mutational case-system pattern available from older ESG speakers can be presented in the following paradigm:

Masc Fem.

Nom./Acc. N L Gen. L L Dat. L L Since lenition is by far the commonest mutation in the paradigm overall, one seemingly natural line of development, if change is to occur, would be the introduction of lenition into the masculine nominative. This line of development seems at first glance to be weakly represented. Table 3 shows virtually no lenited masculine nominatives with the definite article; but the read-back tests show a strong tendency among younger speakers to tolerate the use of lenition with the masculine nominative, and the /khɫɔx : khɫɔxar/ test shown in Table 4 indicates that such lenition does actually occur among the youngest speakers. However, this apparent line of development is ruled out as a serious contender in the pattern of change in the ESG case system, both by the actual statistics of mutational use in the masculine nominative and by mutational developments in the oblique cases. Nasalizations still predominate in the nominative, not only in the results of such tests as those of Table 3 and (less strongly) Table 4, but also in the tape-recorded evidence. In the lengthy tape recording of E14 and E16, for example, there is not a single case of lenited masculine nominative. The briefer recordings of two other younger speakers, E10 and E13, show the same pattern: nasalization without exception in the masculine nominative. Further, even for such young speakers as E18 and E19, who lenite the masculine nominative /khɫɔxar/, nasalization is the strongly preferred mutation for masculines elsewhere, as Table 6 indicates. This preference

60

chapter 1

is shared by the younger group of speakers as a whole, and in the genitive as well as the dative. For some speakers, namely those who still retain a nasalized masculine nominative, there is thus a movement toward a caseless system for both masculines and feminines,

Masc Fem.

def. art.

N

L

while a mutational distinction based on gender-class survives. There are probably no speakers who use this system exclusively, since at least the highestfrequency masculines faithfully preserve a lenited dative; yet a strong tendency in this direction can readily be detected among younger speakers. The situation with regard to the case system of ESG is still unresolved, as Table 3 and Table 4 show. There are competing tendencies within the dialect; and since Gaelic is terminal in the area, there will be no clear resolution of the situation such as we might expect if there were upcoming generations. The actual transitional case-system mutational grammar which exists for most younger speakers, however, could probably best be charted in terms of favored variants: Masc Fem. Nom. Gen. Dat.

N ~ L N ~ L N ~ L

L L L

The boldface mutational symbols represent current younger-speaker preferences, and only the prepositional environment which traditional Gaelic grammar recognizes as the dative remains unresolved in terms of a mutational preference. The highest-frequency masculines show great resistance; but over the rest of the masculine lexicon, the nasalization option can be shown to be increasing. The reasons for the movement toward a caseless nasalized definite masculine noun are twofold, I think. First, the oblique cases are syntactically and morphologically weak in ESG as compared with the standard language. The genitive is rarely required syntactically and has lost both its characteristic inflection (gen. sg. fem. -e) and its characteristic word-final morphophonology (masculine and feminine palatalization with internal vowel change). The dative, too, is called for only in one syntactic environment, the prepositional

grammatical change in a dying dialect

61

phrase; and it, too, has lost the (rare) word-final palatalizations with internal vowel change which one declension shows in the standard language. Case as a surface-structure phenomenon has virtually disappeared in ESG. Second, under these circumstances, the definite article ceases to take on varying morphological shapes within a case system, and moves toward the status of so many other morphemes in the language: it becomes a paradigmatically isolated morpheme, much like {kle:L}, {ta:L} etc. Grammatically it is not redundant, however, since it signals the gender of the following noun. Once again we find that the loss or threatened loss of certain mutational interrelationships causes no informational loss to the language. In a suggestive article on Gaelic grammar, Borgström notes (1968: 13), speaking of the fuller system of western dialects: “The intricate morphology of [Gaelic] cases . . . is only an intellectual play without much informative value”. As in the verbal system, nouns with unmutatable initials occur plentifully, and there is sufficient grammatical apparatus available outside the mutational system to make the syntactic role of any given noun abundantly clear. ESG replaces the genitive of definite-noun attributives by prepositional phrases, and the prepositions themselves make the “dative” lenition of masculine nouns informationally unnecessary. The striking thing about the mutations of the ‘case system’ in ESG is not that syntactic expansion of the nominative is giving rise to a single mutated definite-noun form dependent on gender, but that the expanding nominative preserves a gender-distinctive mutation in a dialect in which gender is itself a rapidly weakening category (see fn. 25). 6

Language Change and Terminal Bilingualism

In this final section we will return briefly to a consideration of the questions raised originally in §2.1. 6.1 Preservation of Mutation as a Phenomenon One of the most impressive features of mutational change in ESG is the extreme rarity of total disappearance of a mutation. Even where a mutation is paradigmatically isolated and informationally functionless, as with the obligatory lenitions after {kle:L}, {ta:L}, etc., mutational loss is a rare occurrence. Evidently, despite Borgström’s accurate observation on the low informational value of even the paradigmatically embedded mutations, the sheer statistical frequency of initial mutation as a morphophonological phenomenon is an overwhelming factor in favor of the preservation of mutations. By reason of

62

chapter 1

his silence on the subject, it seems safe to assume that Breatnach 1964 also found no loss of mutation in his search for terminal characteristics in areas of relic Irish Gaelic speech. Perhaps the relative independence of initial mutation from specific grammatical categories has even in the end contributed to its longevity. At any event, it is undeniable that it has survived far better in ESG than the word-final consonant mutations (chiefly palatalizations) that were so intimately linked with the genitive and dative cases; even the coincidence of some of these final mutations with the final mutations of irregular plural formation did not operate to preserve the phenomenon outside the plural itself. 6.2 Processes of Change Changes in the initial mutations, if not actually caused by changes elsewhere in the grammar, are certainly facilitated by them. This we saw most clearly in the genitive, where the increasing syntactic rarity of the case itself at the very least encouraged the encroachment of the nominative in the definite-noun attributive structure; and in the dol-passive, where the introduction of a subject pronoun led to the loss of possessive mutations distinctive for person. The processes by which these mutational changes occur are familiar to all students of language change. We find analogical leveling occurring on a fairly large scale, with the extension of the nominative with the article into all the environments which formerly called for other cases; and we note the adoption of analytic grammar in preference to synthetic grammar in the substitution of prepositional phrases for the attributive-noun genitive, and the substitution of subject pronouns (with regular lenition) for the grammatically marked mutations of the possessive pronouns. 6.3 Loss to the Grammatical System Not one of the mutational changes we have observed has resulted in a serious loss of information to the grammar of the dialect. The ultimate redundancy of the mutations is such, in most cases, that word order (immediate post-nominal position in attributive-noun structures) or morphology (the subject pronouns in the bith-passive) or even phonology (suprasegmentals in the vocative) can provide whatever information the mutations previously provided, if any. In the one case where there was no redundancy and the mutations carried a truly significant grammatical load, the dol-passives in the 3rd person of both numbers, analogical introduction of subject pronouns provided the margin of redundancy needed to allow for a leveling through of lenition to all persons.

grammatical change in a dying dialect

63

6.4 The Rate of Change If the kinds and processes of change which we find within the mutational grammar of ESG are familiar, what of the rate of change? Is it in any way unusual to find, over a four-decade range of speakers, clear evidence of age-graded speech usage, with the youngest speakers a considerable grammatical distance from the oldest speakers? Systematic studies of grammatical change within a single population are not known to me, but we have the evidence of Gauchat’s work in Charmey (1905) and Labov’s work in New York City (1966) that the processes of sound change at any event work rapidly enough to make the study of phonological change in progress possible for the single linguist working within a currently available population. The Embo situation is, of course, essentially a truncated one; the optimal time for a study of grammatical change in progress there would have been 20 years ago, when the youngest fluent speakers of today were 20 years old and there was over 60 years of age difference between the eldest and the youngest speakers, with the eldest speakers more nearly approximating the old monolingual norm than today’s eldest speakers do. (It may be said, however, in defence of the norm represented by the upper-limit age group in the present study, that the octogenarian E1 is distinctly more comfortable and more proficient in Gaelic than in English.) But even the 40-yearplus span available in Embo today proves adequate to establish the kinds and directions of change in progress in the mutational system. Three special features in the Embo situation may have facilitated change in the dialect and aided in producing a rate of change perceptible even within a truncated 40-year span. One is the feature of virtual illiteracy. Zengel’s study (1962) of the effect of literacy on vocabulary change suggests that literacy may be expected to retard the rate of change. By the same token, illiteracy might speed the rate of vocabulary change, and perhaps other forms of language change as well. The extremely small size of the Embo population (ca. 100 Gaelic speakers) and the relatively reduced form of Gaelic which survives in the village may also be factors in the rate of linguistic change (cf. Stevick 1963). The lexicon and grammar of ESG are sharply constricted in comparison with relatively healthy western dialects. The resultant reduction of grammatical and lexical alternatives is perhaps typical of a language with restricted use among people who share a great deal of common knowledge and experience. In these circumstances the variable features tend to be of relatively high frequency within a very dense communication network. The third feature which may have affected the rate of language change in Embo is the relatively long drawn-out period of isolative bilingualism. In terms of profiles of linguistic death, this may be a far more unusual feature than illiteracy or small population size. Neither the Golspie

64

chapter 1

nor the Brora Gaelic speakers enjoyed a period of social and linguistic isolation comparable to Embo’s, and, perhaps as a consequence, the full-scale changes traceable in Embo are only embryonically present in Golspie and Brora. In fact, precisely because of the lack of a younger generation of fluent bilinguals, a description of the ESG dialect in Golspie and Brora is conspicuously more difficult, because less obviously and clearly patterned, than is a description of ESG based primarily on Embo materials. Studies of western Gaelic dialects have generally aimed at codifying a phonological and grammatical norm for the communities concerned, which has meant that they were by definition unlikely to deal extensively with variability in usage. Detailed studies which focus on variation and change are needed for western dialects before the seeming rapidity of change in ESG can be properly assessed. 6.5 Terminal Speakers and Historical Linguistics One fact that the study of terminal bilingualism in East Sutherland makes abundantly clear is that the last speakers of a dying language can be a very misleading source of information about the grammar (and presumably also phonology and semantics) of the language they represent. It would not be possible to reconstruct the mutations of the case systems or the possessive pronoun system and their role in an earlier stage of ESG grammar from the usage of today’s 40-year-olds. Hymes (1962: 115) speaks of the importance of the individual speaker’s personality for the survival of a language, and relates that the Siouan language Ofo was preserved because a linguist accidentally discovered a last, assimilation-resistant Ofo speaker who had ‘practiced the language frequently to herself in the years since all other speakers had died’. The present Embo study includes several such resistant personalities, among them the young bachelor E19, who has a unique reputation in Embo for answering in Gaelic even when spoken to in English (a highly unusual, almost impolite departure from local custom). The conservative linguistic attitudes of E19, however, have done nothing to retard the rate of change in his language – he is in fact a prime exponent of all the changes discussed in this paper. It should probably be assumed, wherever a grammar is written or a proto-language reconstructed on the basis of materials gathered from a few last remaining speakers, that the stage of the language represented by those speakers is markedly deviant even in terms of the recent history of the language.

grammatical change in a dying dialect

65

References Borgström, Carl Hj. 1968. Notes on Gaelic grammar. Celtic Studies, ed. by James Carney and David Greene, 12–21. London: Routledge & Kegan Paul. Breatnach, R. B. 1964. Characteristics of Irish dialects in process of extinction. Communications et rapports du Premier Congrès de Dialectologie Générale, 141–5. Louvain: Centre International de Dialectologie Générale. Fishman, Joshua A. 1968. Readings in the sociology of language. The Hague: Mouton. Gauchat, Louis. 1905. L’unité phonétique dans le patois d’une commune. Aus romanischen Sprachen und Literaturen: Festschrift Heinrich Morf, 175–232. Halle: Max Niemeyer. Holmer, Nils. 1938. Studies on Argyllshire Gaelic. Uppsala: Almqvist & Wiksell. Hymes, Dell H. 1962. The ethnography of speaking. Anthropology and human behavior, ed. by T. Gladwin and William C. Sturtevant, 13–53. Washington, D.C.: Anthropological Society of Washington. Reprinted in Fishman, 99–138. Kirparsky, Paul. 1968. Linguistic universals and linguistic change. Universals in linguistic theory, ed. by Emmon Bach and Robert T. Harms, 171–204. New York: Holt, Rinehart & Winston. Labov, William. 1966. The social stratification of English in New York City. Washington, D.C.: Center for Applied Linguistics. Lehmann, Winfred P. 1964. Historical linguistics. New York: Holt, Rinehart & Winston. Maclaren, James. n.d. Gaelic self-taught. Third edition. Glasgow: Alex MacLaren. Oftedal, Magne. 1956. The Gaelic of Leurbost, Isle of Lewis. (A linguistic survey of the Gaelic dialects of Scotland, 3; Norsk Tidsskrift for Sprogvidenskap, supplementary vol. 4.) Oslo: Aschehoug. Stevick, Robert D. 1963. The biological model and historical linguistics. Language 39. 159–69. Zengel, Marjorie S. 1962. Literacy as a factor in language change. American Anthropologist 64.132–9. Reprinted in Fishman, 296–304.

chapter 2

The Fate of Morphological Complexity in Scottish Gaelic Language Death: Evidence from East Sutherland Gaelic 1 Introduction In recent years, increasing attention has been focused on pidgin and creole languages, in part as a kind of proving ground for both linguistic and sociolinguistic theory. Not least among the interests of recent writers have been issues of simplification and convergence as linguistic processes; witness the very substantial third section of Hymes 1971. In its preface, Hymes writes of four “ ‘moments’ which a theory of pidgin and creole languages must integrate”. Of these, the first two are: “(1) the universal tendencies to adapt speech, and varieties of a language, by simplification in some circumstances, expansion in others; (2) the occurrence of these tendencies in situations of language contact, so as to give rise to partial confluence of linguistic traditions”. While fully acknowledging the great value of pidginization and creolization studies in the investigation of simplification (and/or elaboration) and confluence in language use, I submit that the study of language death has much to offer in these same areas of investigation, and that so far it has been much too little tapped as a source of information in these matters. This is not to say that simplification and confluence appear in language death in the same degree, at the same points – or for the same reasons – that they do, say, in pidginization. Indeed, I hope to show in this paper that they need not. But this, I think, only makes it the more important that we include the special case of language death when we venture on the topics of simplification and confluence. It has, of course, long been recognized that dying languages characteristically show reduction of one kind or another – or, most often, of many kinds at once. Early reports tended to be quite general and to be impressionistically rendered. Thus Bloomfield 1927 characterized White-Thunder’s Menomini as “atrocious”: “His vocabulary is small; his inflections are often barbarous; he constructs sentences of a few threadbare models.” Krauss 1963–70 offers many comments on failings in the Eyak texts he collected from a last few speakers, but most of them are general rather than specific (“inappropriate here”, “distorted”, “confused towards end”, all from p. 44); and they are not systematized to show in what ways the language is suffering changes in its patterns. Miller

The fate of morphological complexity in language death

67

(1971: 119) offers a generalization about the terminal Shoshoni language: “Younger speakers do not always have a complete control of the grammar and phonology, but the area which shows the greatest impoverishment is vocabulary.” Quite recently, studies have begun to appear which treat simplification and confluence in dying languages with something more nearly approaching the detail and scope with which they have been treated in pidginization studies, although there is still no full-length volume comparable to Mühlhäusler 1974. Thus Dressler 1972 discusses rule loss in the phonology of a dying variety of Breton; Hill 1973 traces the loss of stylistic options in the syntax of two dying Californian languages; Dorian 1973 demonstrates grammatical change, in the direction of analogical simplicity and analytical restructuring, for a terminal Scottish Gaelic dialect; and Dorian 1976 notes the survival of grammatical gender in the same dialect, but with reduction in the number and coherence of the signals of that category, as well as increasing confluence between Gaelic and English use of gender-signaling pronouns. Dorian 1977a reports a hierarchy of morphophonemic decay in terminal East Sutherland Scottish Gaelic, related apparently to avoidance of grammatical syncretism and to the presence or absence of parallel grammatical categories in English. Hill & Hill 1977 detail the impact of Spanish lexicon on beleaguered Nahuatl in Central Mexico, in terms both of those areas of the vocabulary most affected and of the consequences of the massive lexical importation for language loyalty to Nahuatl. In the present study, I wish to pursue the issues of simplification and confluence in language death by examining closely the fate of morphological complexity in a terminal Scottish Gaelic dialect. The structures chosen for investigation represent the extreme in morphological complexity for this dialect and for Scottish Gaelic in general. The noun plural and the gerund in Scottish Gaelic are particularly high-frequency structures, and they are formed in a rich variety of ways.1 This richness is essentially gratuitous. Some of the devices for the formation of plurals and gerunds are phonotactically or morphophonemically capable of operation with only certain groups of nouns or verbs; but others are potentially capable of extension to ALL nouns or verbs. That is, if a simplification process appeared in Gaelic and continued to its ­logical extreme, there is no inherent reason why it should stop short of complete uniformity in the morphological formation of all noun plurals; and

1 The gerund is normally known as the ‘verbal noun’ in the study of Gaelic; but ‘gerund’ is adopted here for brevity, and for the sake of maximum terminological differentiation of the two structures under discussion.

68

chapter 2

s­ imilarly with gerunds. How far this is from the actual case in the dying dialect at hand will appear in due course. The terminal speech form which provides the data for this study is the East Sutherland dialect of Scottish Gaelic. Because it exists in considerable isolation from all other Gaelic dialects, and because it is succumbing not in competition with some more prestigious or normalized variety of Gaelic, but rather in competition with an entirely different language (namely English), East Sutherland Gaelic (ESG) can, for our purposes, be termed a dying language. That is, the struggle for dominance, and currently for survival, involves two languages, Gaelic and English. Each is in fact represented by a regional variety of a language of greater national currency. While East Sutherland English will be called simply English throughout, since its structure is essentially that of standard English in the formation of noun plurals and gerunds, East Sutherland Gaelic will be identified by means of the local label because it differs in some of the particulars discussed here from the standard Gaelic of textbooks. ESG is spoken by fewer than 150 people, all of them bilingual in English, on the east coast of the county of Sutherland in the extreme north of mainland Scotland. In the two largest East Sutherland villages, Brora and Golspie (pop. 1200–1300 in each case), less than 5% of the population speaks Gaelic. Fluent speakers in these villages are between 70 and 80 years of age. Between the ages of about 45 and 65, a small number of speakers can make themselves understood in Gaelic; but their Gaelic, in terms of the norms of the older group, is imperfect in many ways. These I have called semi-speakers (Dorian 1973: 417). In the little village of Embo (pop. ca. 275), the same two groups exist, but the semi-speakers here are between about 35 and 45; transitional between the older fluent speakers and the semi-speakers is a group of younger fluent speakers whose Gaelic is very good, yet shows certain departures from the conservative norms of the older fluent speakers (Dorian 1973). That Embo should have developed a wider spectrum of Gaelic proficiency probably reflects its greater isolation, smaller size, and lower number of resident English speakers. Perhaps half the adult population of Embo speaks Gaelic, and the death of the language will be slightly more protracted there. But Gaelic is dying as surely in Embo as in the other two villages – since, as in Brora and Golspie, no young people are acquiring the language. In order to determine whether changes were appearing in the complex morphology of noun plurals and gerunds in ESG, a large number of test s­entences designed to elicit those structures were presented in English, for translation into Gaelic, to members of the three groups:2 older fluent speakers (o.f.s.), 2 A distinction between o.f.s. and y.f.s. is not made by the communities themselves; it is established on the basis of differences in grammatical usage between age-groups, as reported in

The fate of morphological complexity in language death

69

younger fluent speakers (y.f.s.), and semi-speakers (s.s.) The sample consisted of 4 o.f.s., 4 y.f.s., and from 5 to 8 s.s. The two fluent-speaker groups remained constant throughout the study; a core of 5 s.s. also remained constant, while three others were or were not represented according to their availability when the tests were administered – and, in the most extreme case, according to the speaker’s ability to produce the material. (That is, one “s.s.” participated in only the briefest and simplest subset of test sentences.) For fluent speakers, testing which requires translation of English stimulus sentences into Gaelic poses no problem. All are accomplished bilinguals, and the position of Gaelic in local life is such that translation between Gaelic and English is a commonplace activity, since even kin-linked interaction groups are likely to include both monolinguals and bilinguals. S.s., on the whole, find the translation of relatively simple sentences from English to Gaelic – and only relatively simple sentences can be used, if responses are to be expected across all three groups – to be the most congenial of the three tasks I have thus far asked of them (these were production of isolated Gaelic words from the lexicostatistical list of core vocabulary, translation of simple English sentences into Gaelic, and free Gaelic conversation in interviewing). Because of the restricted role of Gaelic in the lives of the s.s., no elicitation of Gaelic from them could be without stress. They are not regular speakers of Gaelic, and even the best of them use Gaelic only in brief bursts, interrupted by longer stretches of English in their interactions outside test situations. From my considerable interaction with some of the s.s., I believe that their test results in these translation tests are closer to the upper end of their proficiency spectrum in Gaelic than to the lower end; e.g., one s.s. steadfastly professes not to be a speaker of Gaelic at all, and expresses amazement that she can regularly produce the sentences I cajole her into attempting. Free conversation in Gaelic would be less likely to produce good results for such an informant than the translation of single, relatively simple, set sentences. 2

The Syntactic Environment of the Noun Plural and the Gerund

It will be useful first to consider the position of the noun plural and the gerund in the Scottish Gaelic sentence. The ESG noun must be expressed in the plural in all places where a noun would be plural in English, except that the singular Dorian 1973. The communities do show some explicit recognition of a group corresponding to the s.s. here, and these community judgments of less-than-fluent proficiency prove to have demonstrable linguistic correlates. Very fine discriminations can be made by the community and confirmed by linguistic testing (Dorian 1977b).

70

chapter 2

appears after certain numerals (2, 20, and any multiple of 20; the numbers 3 through 5 for certain high-frequency nouns). Syntactically there is no agreement between number in the noun and number in the verb, because the verb does not show number. The definite article shows number, but adjectives and demonstratives do not; if the noun phrase includes the definite article, then number will be expressed twice, once in the article and once in the noun itself: /vriš əm brɔgax peg šən ən ĩn´ag/ broke the lad wee that the window ‘That wee lad broke the window.’ /vriš nə prɔgiç peg šən nəh ĩn´agən/ broke the lads wee that the windows ‘Those wee lads broke the windows.’ Pronoun replacements for the noun also show number, e.g. ‘the window’ – ‘she, it’;3 ‘the windows’ – ‘they’. Far and away the commonest use of the Scottish Gaelic gerund, and the only use tested for this study, is formation of the progressive verb phrase. Progressive aspect is much more prominent in Gaelic than in English. It applies to verbs which would not normally permit progressive aspect in English (‘see’, ‘hear’, ‘understand’, ‘believe’); and it is often used with the inflected verb ‘to be’ in periphrasis, in lieu of inflected forms of the lexical verb itself: /priši will-break

mi/ becomes /pi I will-be

mi prišu/ I breaking ‘I will break’

/vrišu mi/ becomes /viu mi prišu/ would-break I would-be I breaking ‘I would break’ These are all derived from the root /priš/ ‘break’ (with gerund /prišu/). The non-Celtic reader will have no trouble in accepting the prominence of the noun plural in any normal stretch of Gaelic speech. Because of the extreme preference for progressive verb phrases, the gerund achieves a prominence very nearly as great. Consequently there is no difficulty whatever in devising 3 ‘Window’ is grammatically feminine in ESG, but feminine pronoun replacements are giving way to masculine in the singular (the plural pronoun does not express gender). This seems in fact to be a case of confluence; ESG /a/ ‘he, it’ is associated with English it and extended to all inanimates.

The fate of morphological complexity in language death

71

simple stimulus sentences which call for noun plurals and gerunds; and where either plural or gerund is difficult for the informant to produce, it is strictly the token, and not the type, which causes the difficulty. 3 Morphological Devices in the Formation of the Noun Plural and the Gerund Table 2.1 below presents the morphological devices which operate in the formation of the noun plural and the gerund, as elicited from o.f.s. (taken as representing a conservative norm of sorts) in the test sentences. One device which may require a word of explanation is final mutation. In Scottish Gaelic, this is the substitution of a different final consonant for the one in the root of the word. ESG often substitutes a palatal equivalent of the original, sometimes with a concurrent change (from back to front) in the immediately preceding (and in that case always lightly stressed) vowel. Or the reverse may occur: a non-palatal equivalent is substituted for a palatal phoneme. The set of substitutions is limited and fixed, and occurs with a restricted set of words; i.e., it is not productive. The consonant substitutions which are typically word-final may occasionally become internal through the addition of a suffix; since the same pattern of substitutions appears, the term final mutation will be stretched to cover such cases. Table 2.1 shows some overlap in morphological devices for the formation of the noun plural and gerund: four of the first six devices are common to both. In noun plural formation, only syncope (in type X) is wholly unique; in gerund formation, only subtraction (in types VI–X) and the zero formation (in type XI). But vowel alternation and quantity change, which appear independently as devices in the formation of noun plurals, appear only in combination with other devices in the formation of gerunds. With regard to suffixation, the number of different suffixes which are used as the sole device in the two structures also proved to be remarkably similar (though the forms are almost entirely distinct): 9 different suffixes appeared independently in forming the noun plural, and 8 in forming the gerund (though neither figure represents a maximum number of suffixes for the structure). Among the exemplars of Table 2.1, two of the gerund formations elicited are idiosyncratic, supplied by only one o.f.s. in the entire sample; moreover, they are idiosyncratic even in the much larger sample which I have used for other work on ESG. Thus, for most speakers, subtraction actually combines only with vowel alternation plus suffixation (gerund type IX) and with suffixation (gerund type X).

72

chapter 2

Table 2.1 Noun Plural

I. suffixation II. final mutation III. suppletion IV. quantity change  + suffixation V. final mutation  + suffixation VI. vowel alternation VII. vowel alternation  + final mutation VIII. vowel alternation  + suffixation IX. vowel alternation  + final mutation  + suffixation X. vowel alternation  + syncope  + suffixation XI. quantity change

Singular

Plural

/pre:g/ /phũ:nth/ /thε/

/pre:gən/ /phũ:nčh/ /thro:r/

‘lies’ ‘pounds’ ‘houses’

/phyu:r/

/phyuriçεn/

‘sisters’

/se:x/ /mãkh/

/se:çεn/ /mĩkh/

‘dishes’ ‘sons’

/thəu:ɫ/

/thwi:l´/

‘holes’

/khu:/

/khɔ̃ n´/

‘dogs’

/yax/

/yəiçu/

‘horses’

/tarəs/ /ĩn´an/

/tɔrsĩn/ /ĩn´an:/

‘doors’ ‘onions’

Root

Gerund

/ĩ:š/

/ĩ:šu/

‘telling’

/furiç/ /rax/

/furax/ /tuɫ/

‘waiting’ ‘going’

/kyɔ:r/

/kyɔru/

‘cutting’

/čhirəmiç/ /ibəriç/

/chirəməxu/ or /chirəmixu/ /ibər/

Gerund

I. suffixation II. final mutation  vowel alternation III. suppletion IV. quantity change  + suffixation V. final mutation  + suffixation, with  or without vowel  alternationa VI. subtraction

‘drying’ ‘working’ (Continued)

The fate of morphological complexity in language death

73

Table 2.1 (Continued) Gerund

VII. subtraction  + quantity change VIII. subtraction  + final mutation  + suffixation IX. subtraction  + vowel alternation  + suffixation X. subtraction  + suffixation XI. zeroc

Root

Gerund

/e:riç/

/eri/b

‘rising’

/kɫu:əškh/b

/kɫu:əsth/

‘moving’

/khrəǰ/ or /khrεǰ/

/khresčən/

‘believing

/e:šiç/ /khrεkh/

/e:šnax/ /khrεkh/

‘listening’ ‘selling’

a Final mutation in gerund formation was normally accompanied by vowel alternation in the most conservative ESG usage of recent years (that of speakers now dead), in the presence of the suffix /-u/. Final mutation without accompanying vowel alternation is now commoner, even among o.f.s., before /-u/. b This is an idiosyncratic form supplied by only one o.f.s. in the sample. c ‘Zero’ means no morphological change; the gerund is identical with the root.

Although both noun plural formation and gerund formation are morphologically rich, the number and prominence of nouns or verbs which can use one or another of these formations differs markedly. Furthermore, the productivity of the formations is anything but uniform, even within the conservative norm provided by o.f.s. Many noun plurals and gerunds are formed by suffixation, but of the 8 or 9 suffixes available in each case, only one or two are productive for fluent speakers in the sense that they attach readily to loanwords or serve as analogical alternants for rarer forms. No formations other than suffixation are productive in these senses, either for nouns or verbs. Some non-productive formations nonetheless apply to large numbers of nouns or verbs. Quantity change is a very common noun plural formation (type XI) for fluent ESG speakers, and the zero model is very common in conservative gerund formation (type XI). On the other hand, the various plurals with vowel alternation (types VI–X) involve relatively few nouns; but like their analogs in English (the foot/feet group), they include some of the highest-frequency nouns in the language, which makes

74

chapter 2

the device of vowel alternation feel more important than it is statistically – and similarly with suppletion among the gerunds (type III). Consequently, if we find statistical advances in the use of the productive suffixes, it will scarcely be surprising; and it is at least conceivable that we might find some statistical advance in the use of quantity change in noun plurals and the zero formation in gerunds. However, a reduction in vowel alternation in plural formation or suppletion in gerund formation could represent a striking change in the dialect without looming very large statistically, because of the prominence of the words concerned. If we now look in detail at the noun plural formations produced by the three groups of speakers, and then in similar detail at the gerund formations of the three groups, we will be able to follow certain developments in this rich area of terminal ESG morphology. 4

The Noun Plural

Table 2.2 gives the noun plural formations of all three groups of speakers. The percentage of all noun plural formations within the test sentences is given for the formation type, and beside it the number of actual occurrences. The table is arranged in descending order of frequency according to the usage of the o.f.s., who represent a sort of conservative norm. Table 2.2

Formation Type

/-(ə)xən/ (I)a final mutation (II) /-ən/ (I) vowel alternation + final  mutation (VII) vowel alternation + /-(V)n(´)/b (VIII) /-čεn/ (I) lengthening of final consonant (XI) /-n(:)/ (I) /-içεn/ (I)

o.f.s. # %

y.f.s. # %

s.s. # %

48 18 28 10 27 10

34 12.5 24 9 24 9

 87 19  23 5 124 27

24 9 21 8 20 7.5 19 7 16 6 14 5

28 10.5 33 12 25 9 15 5.5 16 6 15 5.5

 11 2.5  26 5.5  18 4   5 1  11 2.5  24 5 (Continued)

75

The fate of morphological complexity in language death Table 2.2 (Continued)

Formation Type

vowel alternation (VI) vowel shortening + /-içεn/ (IV) vowel alternation + syncope +  /-in/ (X) suppletion (III) /-in(:)/ (1) vowel alternation + final mutation  + suffixc (IX) /-čax/ (1) final mutation + /-εn/ (V) vowel shortening + /-in(:)/ (IV) suffixes unique to a single group of  speakersd zero

o.f.s. # %

#

y.f.s. %

s.s. # %

14 5 10 4

12 4.5 12 4.5

18 4 14 3

 8 3  6 2  4 1.5

 8 3  8 3  2 1

 6 1  9 2  7 1.5

 3 1  3 1  2 1  –

 4 1.5  –  3 1  3 1

 7 1.5  –  6 1  –

 5 2  –

 2 1  1 .5

21 4.5 42 9

a References in parentheses are to the general types of Table 2.1. b V stands for ‘vowel’; the parentheses around the palatalization symbol indicate that either the dental nasal or the palatal nasal may occur. c Elsewhere, when suffixation combines with a non-suffixal device, the suffix is constant; but in this one combination, despite the small number of cases, more than one of the plural-forming suffixes is involved. d Unique o.f.s. suffixes are /-čax/, three instances; /-ç/, /-ču/, one instance each. Unique y.f.s. suffixes are /-u/, /-rən/, one instance each. Unique s.s. suffixes are /-əx/, /-čxən/, /-iç/, /-ig/, /-əns/, one instance each; /-s/, 6 instances; and /-an/, 10 instances.

A number of observations may be made on developments in noun plural formation in the light of the data presented in Table 2.2. 4.1. If all instances of plurals formed by simple suffixation (i.e. suffixation alone, type I) are tabulated as in Table 2.3, this device proves to have gained significantly in the usage of s.s. as compared with fluent speakers. Virtually all this gain results from the increasing use of the suffix /-ən/ among s.s.; /-ən/ shows a 17% gain over fluent-speaker use of the same suffix, while a few other suffixes (/-čεn/, /-n( :)/) drop off slightly in s.s. use.

76

chapter 2

Table 2.3 SPEAKER GROUP

#

o.f.s. y.f.s. s.s.

%

134 118 292

50 44 63.5

4.2. Among the devices of relative prominence (those constituting 5% or better) in the plural formations of fluent speakers, the two which show the greatest decline among s.s. are vowel alternation combined with final mutation, as seen in Table 2.4, and lengthening of the final consonant, in Table 2.5. Table 2.4 SPEAKER GROUP

TypeVII Type VIII Type VI Type X Type IX

o.f.s.

y.f.s.

s.s.

#

%

#

%

#

%

24 21 14  8  3

9 8 5 3 1

28 33 12  8  4

10.5 12 4.5  3 1.5

11 26 18  6  7

2.5 5.5 4 1 1.5

TABLE 2.5 o.f.s.

Type II Type VII Type IX Type V

y.f.s.

s.s.

#

%

#

%

#

%

28 24  3  2

10  9  1  1

24 28  4  3

 9 10.5 1.5  1

23 11  7  6

5 2.5 1.5 1

Vowel alternation and final mutation can in fact be seen to have declined in popularity as pluralization devices in general among s.s. – separately and in

77

The fate of morphological complexity in language death

various combinations, including combination with each other. Because of the very high frequency of many nouns which traditionally had vowel alternation as an element in their plurals, the 10% drop in the use of vowel alternation in pluralization among s.s. is a rather salient feature of their speech. The decline in the use of final mutation, also at the 10% level when compared with the o.f.s. norm, is perhaps less salient in terms of the frequency of some of the words affected; but it does represent a conspicuous shrinkage in the use of a peculiarly Celtic morphological device. The situation where quantity is an element in pluralization shares with with final-mutation pluralization the fact that the device in question (quantity change) plays no role whatever in English pluralization. The one plural formation which depends solely on a difference in quantity, lengthening of the final consonant (type XI), falls off across the three groups of speakers (7%, 5.5%, 1%). Furthermore, as Table 2.6 shows, length is reduced or lost (largely the latter) among s.s. in plurals which o.f.s. form solely by the suffixation of a long nasal (after stem-final vowel or consonant) or of a vowel plus long nasal (after stem-final consonant; types I and XI). Table 2.6 SPEAKER GROUP

long half-long short

o.f.s

y.f.s.

s.s.

#

%

#

%

#

%

34  4  1

87 10 2.5

29  4  –

88 12  –

 8  5 10

35 22 43

The fate of this group of plurals is in one sense different from that of all the others: its decline rests on a phonological development among s.s. – a very heavy loss of the phenomenon of consonant length in general, and a considerable weakening of the phenomenon of vowel length. Phonologically, the s.s. show a marked tendency to lose phonemes and prosodemes of Gaelic which are not shared by English; thus they generally fail to use the ‘dark’ velarized /ɫ/ or the palatal fricative /ç/, and vowel nasalization as well as vowel and consonant length become sporadic or disappear. Where quantity change is concerned, this means that, despite the frequency of the device in noun pluralization among fluent speakers, there can be no question of any increase in the phenomenon among s.s. However, there are at least two ways in which

78

chapter 2

this group of nouns might be treated, for pluralization purposes, by the s.s.: addition of a suffix compatible with their phonotactic shape, or loss of all overt plural markings. The latter is the usual development, and brings us to the next morphological device. 4.3. Zero plurals, which are unknown among o.f.s. and highly exceptional among y.f.s., constitute 9% of all s.s. plurals – a striking development. Of the 42 zero plurals supplied by the s.s., 18 (or 43%) derive from the loss of final consonant length in plurals where such length would be the sole pluralization marker among conservative speakers (type XI). Back-formation of a singular identical to the /-(V)n(:)/ plural, with loss of final consonant length, accounts for three more instances.4 The rest of the s.s. zero plurals (50%) are independent of developments in consonant length. 4.4. While certain rarer o.f.s. suffixes are totally unrepresented among y.f.s. and s.s., none of the pluralization devices as such (or combinations) in use among o.f.s. is completely absent in the other two groups – not even those constituting less than 5% of the o.f.s. plurals, like suppletion or final mutation plus /-εn/. 4.5. Each of the three groups of speakers has some idiosyncratic plural formations (all cases are of idiosyncratic suffixes, except the three y.f.s. instances of vowel shortening plus /-in(:)/); but the s.s. have proportionately slightly more. The percentages for the three groups respectively, in the usual order, are 2, 2, 4.5. 5

The Gerund

The gerund formations of ESG present a picture of morphological variety almost as great as that of the noun plurals. Table 2.7 presents the full range of gerund formations elicited by the test sentences from the three groups of speakers, with the number of occurrences of each type and the percentage of the total which that number constitutes. Again the table is arranged in descending order of frequency, according to the usage of the o.f.s. as representing a conservative norm. As with the noun plurals, certain developments are evident in the data presented in Table 2.7.

4 Fluent speakers, e.g., have sg. /khu:rsti/, pl. /khu:rstin:/; two s.s. have sg. and p1. /khu:rstin/.

79

The fate of morphological complexity in language death Table 2.7 SPEAKER GROUP

o.f.s.

y.f.s.

s.s.

#

%

#

%

#

%

41 25 16

26 16 10

38 27 20

24 17 12.5

62 42  4

27 18.5  2

14  9

 9  6

16  4

10 2.5

58  3

25.5  1

 8

 5

 9

5.5

11

 5

 8  7  4  4  4  3

 5  4.5  2.5  2.5  2.5  2

11  9  1  4  4  4

 7 5.5 .5 2.5 2.5 2.5

 6 11  –  7 10  –

2.5  5  –  3 4.5  –

 3

 2

 4

2.5

 –

 –

 2  2  2  1

 1  1  1  .5

 3  –  –  1

 2  –  –

 3  2  –  –

 1  1  –  –

 1

 .5

 3

 2

 1

 1

 .5

 –

 –

 –

 –

 1

 .5

 –

 –

 –

 –

 –  –

 –  –

 3  1

 1

FORMATION TYPE

/-u/ (I)a zero (XI) final mutation + vowel  alternation (II) /-al/ (I) vowel shortening +  /-u/ (IV) vowel shortening +  /-al/ (IV) /-tən/ (I) subtraction (VI) /-šən/ (I) suppletion (III) /-əm/ (I) final mutation +  suffixation (V) vowel alternation +  subtraction +  /-sčən/ (IX) /-i/ (I) /-ən/ (I) /-d/ (I) vowel shortening +  /-tən/ (IV) subtraction +  /-nax/ (X) vowel shortening +  subtraction (VII) subtraction +  final mutation  + /-th/ (VIII) vowel alternation /-kən/ (I)

 –  –

 –  –

.5

.5

.5 (Continued)

80

chapter 2

Table 2.7 (Continued) SPEAKER GROUP

o.f.s.

y.f.s.

s.s.

#

%

#

%

#

%

 –  –  –  –

 –  –  –  –

 –  –  –  –

 –  –  –  –

1 1 1 1

.5 .5 .5 .5

FORMATION TYPE

/-iç/ (I) /-in/ (I) /-x/ (I) infixation

a References in parentheses are to the general types of Table 1.

5.1. The percentage of all gerunds formed by simple suffixation (suffixation alone, type I) has risen among s.s. as compared with the other two groups of speakers, as shown in Table 2.8. Virtually all this gain is caused by the increasing use of the suffix /-al/ among s.s., for whom it has become exceptionally productive. Table 2.8 SPEAKER GROUP

o.f.s. y.f.s. s.s.

#

77 73 145

%

49 46 63.5

5.2. Among the devices of relative prominence in the formation of the o.f.s.’ gerunds, the one which shows the greatest decline in use is final mutation plus vowel alternation (type II), which drops off sharply in s.s. use: 10%, 12.5%, 2%. This same decline appears if we tabulate all occurrences of final mutation, in combination not only with vowel alternation but also with other devices (types II, V, VIII), as in Table 2.9. Table 2.9 SPEAKER GROUP

o.f.s. y.f.s. s.s.

#

%

20 24 4

13 15 2

The fate of morphological complexity in language death

81

5.3. Change in vowel quantity in combination with suffixation (type IV) has declined in gerund formation, as shown in Table 2.10, even when we omit from consideration the single idiosyncratic o.f.s. case of vowel shortening plus subtraction (type VII; cf. Table I, fn. b). Table 2.10 SPEAKER GROUP

#

%

o.f.s. y.f.s. s.s

18 14 14

11.5 9 6

5.4. Although subtraction as a gerund-forming device shows little difference across the three groups when it appears as the sole device (type VI: 4.5%, 5.5%, 5%), its popularity declines somewhat among s.s. when combined with other devices (types VIII, IX, X): 3%, 4.5%, and .5% for the sole device (type VI: 4.5%, 5.5%, 5%), its popularity declines somewhat among s.s. when combined with other devices (types VIII, IX, X): 3%, 4.5%, and .5% for the three groups respectively. 5.5. Gerund formation, unlike plural formation, shows a loss of certain nonsuffixal formation types among s.s. as compared to o.f.s.; this is not true among y.f.s. (if the two highly idiosyncratic o.f.s. types noted in Table 2.1 are omitted from consideration as models). One such loss involves final mutation – which, as we have noted, is weakening generally among s.s. (final mutation plus suffixation, type V); the other is the complex formation made up of vowel alternation plus subtraction plus suffix (type IX). 5.6. As in plural formation, though less markedly, the s.s. prove to be most given to idiosyncratic formations of the gerund; this is especially true when “idiosyncratic” is taken to mean not simply “unique to that (group of) speaker(s)”, but “representative of a new development in morphological formation”. I make this distinction in connection with the gerund because two “unique” o.f.s. gerunds, those in suffix /-d/, represent a conservative norm which is being lost in the dialect and fails to appear even among the y.f.s. group in this sample. Thus the number of truly idiosyncratic, or innovative, gerunds produced by o.f.s. is only two, and the percentage of such gerunds for the three groups respectively, taken in the usual order, is 1, 0, 3.5. 5.7. Zero formation of gerunds increases in a small, steady way across the three groups of speakers (16%, 17%, 18.5%).

82

chapter 2

6 Comparison of Developments in the Formation of Noun Plurals and of Gerunds It is striking that, in both cases we have examined, a sharp gain is registered in the use of simple suffixation in the formation of the item in question. Even more striking is the fact that the statistics in the two cases almost match one another, as seen in Table 2.11. Table 2.11 SPEAKER GROUP

PLURAL

GERUND

o.f.s. y.f.s. s.s.

50% 44 63.5

49% 46 63.5

It is important to note that the marked increase in the use of simple suffixation by s.s. is not foreshadowed by any comparable increase among y.f.s. as compared with o.f.s.; on the contrary, y.f.s. use simple suffixation slightly less in each case. That is, the rise in simple suffixation is not part of a gradual continuum of change within the language, but a phenomenon characteristic of, and peculiar to, s.s. use of the language. Indeed, the only plural formation which shows much increase among y.f.s. is type VIII, vowel alternation plus suffixation. Such a shift in favor of a doubly-marked plural is altogether in line with Kury�owicz’s first ‘law’ of analogical change (1966: 162) – whereas the changes evinced by s.s., changes thus perhaps characteristic of language extinction in general, are never in the direction of Kury�owicz’s ‘morphème bipartite’, either in noun plurals or in gerunds. Also of interest is the fact that the rise in simple suffixation, relative to all other formation types, results very largely (indeed almost exclusively) from the increased use of a single ‘favored’ suffix: /-ən/ in the noun plural, and /-al/ in the gerund. That is, the sense of plurality in nouns, and gerund formation in verbs, resides for s.s. increasingly in one strongly productive allomorph. In the case of /-ən/, the s.s. are ‘over-using’ a suffix which is by all odds the commonest noun plural suffix in ESG (though not, as it happened, the commonest one in the test sentences for any but s.s.).5 The case of /-al/ is the more interesting, because here the s.s. are favoring a suffix which is not normally the commonest 5 The statistical preponderance of the /-ən/ plural allomorph was underplayed in the test sentences, so as to avoid the planting of a built-in model for analogical change in the test itself.

The fate of morphological complexity in language death

83

gerund-forming suffix in the language at large (as indeed it is not for any group of speakers in the test sentences). In this regard, the s.s. again demonstrate their relative independence of the y.f.s. position, since the y.f.s. group shows only very small differences from o.f.s. norms for the /-u/ and /-al/ gerund allomorphs. Whereas simple suffixation shows a marked gain in use among s.s., the peculiarly Gaelic device of final mutation shows a decline in use, as shown in Table 2.12. This is despite the fact that, like simple suffixation but unlike vowel alternation or subtraction, it plays a modest role, both alone and in combination with other devices, in forming BOTH noun plurals and gerunds in the usage of conservative speakers. Table 2.12 PLURAL

GERUND

SPEAKER GROUP

#

%

#

%

o.f.s y.f.s s.s.

57 59 47

21 22 10

20 24  4

13 15  2

Again it is conspicuous that the drop-off in the use of final mutation as a morphological device is both characteristic of and peculiar to s.s.; it is not consonant with the behavior of y.f.s., who actually show slightly more use of final mutation than the o.f.s. in each case. The reasons for the lesser use of final mutation as a gerund-forming device than as a plural-forming device in all three groups of speakers, when in theory ESG offers more opportunities for gerund formation by final mutation,6 are complex and interesting, but outside the scope of this paper. The dramatic falling-off of gerund formation by final mutation among s.s. vis-à-vis fluent speakers reflects the transfer of many verbs originally of this class into the suffixation class. Vowel alternation in the noun plural (and on a much smaller scale in the gerund) and subtraction in the gerund have also fallen off somewhat in the usage of s.s., but the drop in the use of subtraction is found only where it is combined with other devices. The drop in the use of vowel alternation is rather more marked, however, and again shows the pattern of s.s. decline as a breakaway development, unrelated to y.f.s. usage; see Table 2.13. 6 ESG has more verb roots in final /-iç/, to give potential gerunds in /-ax(u)/, than noun roots in final /-ax/, to give potential plurals in /-iç/.

84

chapter 2

Table 2.13a PLURAL

GERUND

SPEAKER GROUP

#

%

#

%

o.f.s. y.f.s. s.s.

70 85 68

26 31.5 15.5

19 24  7

12 15  3

a Note that there is an overlap here with Table 12, since some vowel alternations occur in conjunction with final mutation in noun plurals (types VII, IX) and in gerunds (type II).

An interesting development in this connection is that s.s. have introduced (albeit on a very small scale) the use of vowel alternation as sole device into gerund formation – i.e. vowel alternation not combined with any other device – but fail to use vowel alternation plus subtraction plus suffixation (type IX) at all, and make very little use of final mutation plus vowel alternation (type II). Here, then, s.s. are even more deviant in terms of types than in terms of number of tokens. Morphological formations involving quantity changes occur in both noun plurals and gerunds, and in each case they show a decline in use among s.s. Here, for the first time, is a case of continuity over the three speaker groups, in that we see a continuum of decline. The s.s., with a relatively low use of quantity in the formation of the gerund, represent an extreme of a decline which begins with the y.f.s. This is not the case in noun plurals, however. Table 2.14 offers an overview of developments in all plural and gerund formations in which quantity change plays a part for o.f.s. In this tabulation, the half-long nasals of Table 2.6 are assigned in the dichotomy to plural formations which show a quantity distinction – i.e., they “have length” as a feature of the formation. Table 2.14 PLURAL

GERUND

SPEAKER GROUP

#

%

#

%

o.f.s. y.f.s. s.s.

49 48 27

18 18  6

19 14 14

12  9  6

The fate of morphological complexity in language death

85

Innovatively idiosyncratic formations not shared by any other group of speakers are commonest among s.s., although the differences are not large in either case (innovatively idiosyncratic plurals: 2%, 2%, 4.5% across the three groups; innovatively idiosyncratic gerunds: 1%, 0%, 3.5%). There is some question whether even “innovatively idiosyncratic” means the same thing in connection with s.s. that it does for fluent speakers. As is expected in any speech community, I have found idiosyncrasy in many areas of ESG usage at all levels of proficiency (e.g. in the phonological shape of words, in gender assignment of nouns, in semantics – all among fluent speakers as well as among s.s.). Nonetheless, fluent-speaker idiosyncrasy tends to be stable; i.e. the same idiosyncrasy will be elicited from the same speaker repeatedly, for the most part (although with extremely low-frequency words this generalization does not always hold). S.s. idiosyncrasy, however, shows rather less stability across a number of occasions, or simply occurs variably on a single occasion. In a close study of 79 plural formations from a particularly weak s.s., I.F., I found two instances of an idiosyncratic plural suffix (/-an/) used for two nouns on one occasion, while well-known plural suffixes (though in one case not the usual suffix) were used on other occasions for the same nouns. For two further nouns on a single occasion each, the same person also variably offered a plural formation unique to s.s. (the zero plural) and a suffixal plural formation of well-known type (though not appropriate for the noun in question by fluentspeaker standards). Note that, by the definition of “innovatively idiosyncratic” used in this study, I.F.’s analogically-formed plurals in well-known suffixes are not included. They represent only an analogical extension of a morphological feature well-established in the dialect. Her /-an/ plurals and her zero plurals, however, are idiosyncratic in the sense used here: they represent a wholly new development in the morphology of the dialect: in the one case, a novel plural suffix; in the other case, a novel plural formation, namely the absence of any overt pluralization marker. Zero formation, which is well represented for all groups of speakers in gerunds and shows a slight increase across the three groups taken in the usual order, also shows a rather sizeable intrusion (9%,) into noun plurals in the speech of s.s. Two or three reasons for this development can be suggested. One is that, since ‘plural’ is often marked in the noun phrase by an unambiguously plural definite article, the plural marker in the noun may be omitted as redundant. The locus of marked plurality may thus change without loss of the grammatical category. This was in fact the case in 21 instances (out of 42) where s.s. failed to give the noun itself a plural marker; but there was a residue of 21 other instances where either the syntactic construction did not permit the use of a

86

chapter 2

definite article (13 instances) or where the definite article was called for but omitted (8 instances). Thus failure to mark the noun plural morphologically is not wholly explained by a built-in redundancy.7 A second possible explanation is intralinguistic influence from one class of words to another. A goodly number of morphological devices are common to both noun plurals and gerunds; perhaps the s.s. are simply extending this overlap to yet another device (as with the introduction of vowel alternation alone into gerunds, noted above). Or, as a third explanation, there is the possibility of interlinguistic influence. English, of course, has a small class of nouns, largely the names of edible animals (deer, sheep, fish etc.), which require no plural marker. The fact that English is the dominant language of all the s.s. may make it so much the easier to develop a parallel zero plural in Gaelic. As it happened, four s.s. instances of zero plural involved edible animals (rabbit, lamb); two others involved a domestic animal (horse). Most probably all three of the factors suggested operate conjointly to produce the s.s.’ predilection for zero plurals. Finally, it is perhaps worthy of comment that suppletion – a device which is itself used on a very small scale in the formation of both noun plurals and gerunds by o.f.s., and does not combine with any other device in either type of construction – shows no appreciable change across the three groups of speakers in either case. Only two nouns with suppletive plurals in conservative usage appeared in the test sentences, one a very high-frequency noun, the other much less so. Most s.s, retained the high-frequency suppletive plural; most lost the low-frequency suppletive plural. It should be noted, however, that high frequency of the noun is not a guarantee of the preservation of a non-productive plural formation. Some very high-frequency nouns which have vowel-alternation plurals for fluent speakers developed suffixal plurals for most s.s., or joined the zero-plural group. Suppletion is undoubtedly reinforced in gerund formation by two factors: the verbs which form their gerunds by suppletion are truly among the very highest in frequency in the language; and their other forms

7 The nouns which appeared as zero plurals for the s.s. ran rather heavily to ones which would have been consonant-length plurals (type XI) for fluent speakers (16 instances out of 42). Almost exactly as many were of the vowel alternation type, with or without suffixation (types VI and VIII, 15 instances). Scarcely any were of the suppletive type (type III; two instances of the same noun, from the same speaker); and, as one would expect, relatively few were of the simple suffixation type (type I; 7 instances). The final mutation type, with or without suffixation and vowel alternation (types II, V, IX), supplied only two instances (separate instances of the same noun, for two different speakers).

The fate of morphological complexity in language death

87

(“principal parts”, so to speak) are also suppletive, even more so than the gerunds. 7 Conclusions It is clear from §6 that the s.s. group emerges as very much the innovators in terms of changes in the part of the ESG morphological system which we have examined. (This is in fact the richest part of that system; no other part of the inflectional or derivational morphology of ESG shows such complexity.) Although I have been at pains elsewhere (Dorian 1973) to demonstrate that the y.f.s. do not always coincide in their usage with the o.f.s., in the complex morphology of noun plurals and gerunds they are seldom far from the o.f.s. norm. The s.s., however, show many departures from that norm. One might suppose that the linguistic discontinuity of the s.s. would be a linguistic parallel to a similar generational discontinuity in social life, i.e. that s.s. would interact relatively little with fluent speakers, forming a group apart, both socially and linguistically. This is not the case. The s.s. do have peer-group ties to one another in several cases, but their ties to the older generation, especially to the generation of their grandparents, are so conspicuous, in the linguistic autobiographies that I have collected from a number of them, as to lead me to propose, only half-jocularly, a ‘grandmother factor’ in the genesis of the s.s. as a linguistic phenomenon, at least in ESG. Some social explanation is, after all, required to account for the fact that these younger members of the community, whose command of Gaelic is weak and imperfect by their own ready admission, continue to be willing (eager, in a number of cases) to speak a language which is clearly dying and has extremely negative prestige on the local scene. Repeatedly, they report a favorite female figure (other than the mother) in the first or second ascending generation – usually a grandmother, but sometimes a greataunt or a cousin at one or two removes – with whom Gaelic was the sole or favored language of communication in their early years.8 Most of them maintain close ties with older Gaelic-speaking kin. Social discontinuity is, then, not the explanation here for linguistic discontinuity. The formal discontinuity which we note does correlate positively with a functional discontinuity, however, as the very labels ‘fluent speaker’ and ‘semispeaker’ imply. Fluent speakers are habitual users of Gaelic. The settings in which ESG can be used appropriately have been drastically reduced in the past 8 Michael Silverstein (p.c.) reports that this is ‘exactly the case over and over in [his] WascoWishram Chinookan work as well’.

88

chapter 2

fifty years (from work, religion, community, and home to only the last in Brora and Golspie, and only the last two in Embo); but the fluent speakers, even the youngest, do use Gaelic almost exclusively as a home language. This is true for none of the s.s., although three of them live in settings where they could achieve regular home use of Gaelic if they wished it and felt capable of it. The s.s. are all speakers of English first and foremost, who reserve their use of Gaelic for interaction with a few highly specific individuals (mothers or other older-generation kin; in one case a spouse) and a few equally restricted settings (most notably, settings where Gaelic serves a concealment purpose, such as joint shopping trips or joint journeys on public transportation). The functional discontinuity observed for ESG is perhaps reflected here in the fact that the rich, even over-rich, morphological categories studied prove resistant to marked change until the lower reaches of the proficiency continuum of speakers. This is particularly striking because the morphological complexity of the two structures concerned would seem almost to invite simplification via leveling, and certainly to provide the means: simple suffixation, which is clearly attractive to the s.s. as a device, could easily be extended to account for all noun plurals and gerunds. Y.f.s., who do show departures from the o.f.s. norm in other aspects of ESG structure, seem almost entirely impervious to the possibilities for simplification in these complex morphological structures. These y.f.s. not only share, both actively and passively, the morphological norms of the o.f.s., they also share their norms for the use of Gaelic. Fluent speakers, both older and younger, adhere to the norm of habitual home use, as noted above, while s.s. are deviant in this regard. Evidently the functional factor of habitual home use is sufficient to prevent the onset of morphological simplification among y.f.s., while the lesser use of Gaelic among s.s. facilitates it. S.s. are not altogether unaware of fluent-speaker morphological norms. They often react strongly to a conservative usage heard after completing test sentences with me, saying ‘That’s what I should have said!’ or ‘That’s the right thing, isn’t it?’ But they share at most receptive morphological norms with the fluent speakers, not productive ones. Other observations can be made about simplification in ESG morphology as examined here. Simplification certainly exists in the morphological performance of the s.s.; the marked rise in the use of simple suffixation (especially a rise produced by a single “favored” suffix) is a clear case of a movement toward simplicity in a highly complex morphology. Nonetheless, a great deal of that complexity remains. Final mutation, vowel alternation, suppletion, quantity change, subtraction, zero formation – all are still in use among s.s., although to a lesser extent than among fluent speakers. If one accepts Samarin’s concept (1971: 119, 123) of “substantive pidginization” and its wide distribution, then one

89

The fate of morphological complexity in language death

might speak of a kind of pidginization process in this case of terminal ESG morphology. But it should be clear that radical morphological simplification, as found in many pidgins,9 is not characteristic of ESG, even among its most halting speakers, and even very near the point of its extinction. The tables throughout this study show s.s. performance in terms of group behavior, of course; but there is actually a fairly wide range of proficiency within the s.s. group itself.10 For this reason I undertook a separate analysis of the two youngest and weakest s.s.’ responses in the area of most plentiful data, namely noun plurals. For I.F. I had a total of 79 noun plurals, for her cousin I.H. a total of 64. I.F. formed simple-suffix plurals for 53 of her 79 nouns. The most conservative o.f. [speakers] formed simple-suffix plurals in 38 of the same 79 cases. Comparable analysis for I.H. indicated that where I.H. formed simplesuffix plurals for 42 of her 64 nouns, the most conservative o.f. [speakers] pluralized by simple suffix in 27 of the same cases. Table 2.15 presents a full analysis of the data. Table 2.15 I.F.

total plurals plurals by simple suffixation zero plurals plurals by neither simple  suffixation nor zero

o.s.f. %

#

%

#

79 53  9

100  67  11

79 38  0

17

 21.5 41

100  48   0

#

I.H. %

o.s.f. #

64 100 64 42  65.5 27  7  11  0

 51.5 15  23.5 37

%

100  42   0  57.5

There is, in fact, a sizeable discrepancy between the pluralization practices of these s.s. and those of o.f.s. The latter use plurals formed by devices other than simple suffixation (and zero, which is non-occurrent for the o.f.s. group) 30% more than I.F., and 34% more than I.H.

9

10

Cf. Reinecke on Pidgin French in Vietnam (1971: 51–2): “Except for a few isolated set forms, S[tandard] F[rench] inflection has been dropped”; “the P[idgin] F[rench] verb is invariable.” Three of the group are relatively strong speakers, as s.s. go; two are of intermediate proficiency; two are quite weak speakers; one is so weak as to have been barely usable and should not really be termed a “speaker”.

90

chapter 2

If we return to the comparison of simplification in language death with simplification in pidginization, it seems to me that we have here a bottom line. I.F. and I.H. are at so low a level of proficiency that below them we reach passive bilinguals, individuals who have such difficulty in constructing a coherent sentence that they cannot properly be termed “speakers” of any kind. One such was included in these tests in a minimal way, and he attempted to translate the simplest of the test sentences; but he could barely be said to manage that, and produced only very few data. Taking I.F. and I.H. as an extreme for the language death situation, then, we still find better than 20% retention of pluralization by devices other than simple suffixation and zero. The devices represented are vowel alternation (with and without suffixation), final mutation (with and without suffixation), lengthening of final consonant (one instance only, however), and vowel alternation plus syncope plus suffixation. In summary: not only is the quantity of morphological complexity much greater than one would expect to find in “classical” pidginization, but the variety of allomorphs providing the quantity is fairly astonishing. If simplification is very much less evident in this language death situation than in pidginization as we normally recognize it, how do matters stand with regard to confluence? Overt influence from English in the complex ESG morphology that we have examined is relatively small, despite the bilingualism of all speakers and the dominance of English for s.s. The most direct influence from English is the appearance of a few /-(ən)s/ noun plurals (1.5%, 7 out of 459) in the usage of the s.s. There is possible English influence in two other areas: the appearance of zero plurals among s.s., mentioned above; and perhaps the growing dominance of a single suffix in each case, where the overwhelming dominance of the sibilant plurals and the -ing gerund in English may play some part. But while some of the devices most alien to English show a marked weakening (final mutation and quantity change), there is no wholesale dropping of devices not found in English, and no wholesale importation of high-frequency English elements in either case, not even in the usage of the very weakest s.s. Here, too, then, we are far removed from an extreme case of the “confluence of linguistic traditions” which Hymes exhorts students of pidginization to take into account in their theoretical approaches to the subject. Dressler (MS), in an investigation of decay in the derivational morphology of a dying Breton dialect, reports that weaker speakers (comparable to the s.s. of this study) “were incapable of evaluating new derivations”. He terms his weakest informants’ usage “pidginized”, and says that the best of the weak group is “comparable to a creole language”. This is far stronger terminology than 1 would be prepared to apply to the ESG of s.s., given the results reported

The fate of morphological complexity in language death

91

in this chapter. Indeed, ESG might be said to be dying, at least with regard to noun plurals and gerunds, with its morphological boots on. In the absence of detailed comparable studies of other dying languages, it is hard to interpret the resistance of ESG to large-scale morphological simplification. A number of factors seem to be of possible significance: the relatively protracted obsolescence of the language, with the number of speakers dwindling only gradually; the strong inter-generational ties characteristic of this community, with women of the second ascending generation playing a key role in the linguistic socialization of the s.s.; and the “integrative” rather than “instrumental” role (see Lambert 1967: 102) of Gaelic in East Sutherland, where the language is emblematic of membership in a particular subgroup in the population – the fisherfolk and their descendants – but no longer serves any economic purposes. Pidgins, by contrast, tend to spring into being rather quickly and to have a function considerably more “instrumental” than “integrative”; one may suppose that they figure less often in wide cross-generational linguistic socialization. The fact that ESG, in its terminal state, behaves so differently from the “typical” pidgin in terms of morphology bears out my contention that language death needs to be added to – but not equated with – pidginization as a source of data on simplification and confluence in language contact. But it is also clear that we need studies of language extinction as rich and broad as those on pidginization before we can venture generalizations about the circumstances under which dying languages will in fact resist wholesale morphological simplification and confluence. Whether all the factors noted above are necessary to the preservation of morphological complexity, or some one or two of them, or whether still other factors not evident here are often operative, remains to be determined. References Bloomfield, Leonard. 1927. Literate and illiterate speech. American Speech 2.432–39. Dorian, Nancy C. 1973. Grammatical change in a dying dialect. Lg. 49.414–38. ———. 1976. Gender in a terminal Gaelic dialect. Scottish Gaelic Studies 12.279–82. ———. 1977a. A hierarchy of morphophonemic decay in Scottish Gaelic language death: the differential failure of lenition. Celtic Linguistics – 1976 (Word, vol. 28), 96–109. ———. 1977b. The problem of the semi-speaker in language death. International Journal of the Sociology of Language 12.23–32.

92

chapter 2

Dressler, Wolfgang. 1972. On the phonology of language death. Papers from the Eighth Regional Meeting, Chicago Linguistic Society, 448–57. ———. MS. Wortbildung bei Sprachverfall. Hamp, Eric P.; Fred W. Householder; and Robert Austerlitz (eds.) 1966. Readings in linguistics II. Chicago: University of Chicago Press. Hill, Jane H. 1973. Subordinate clause density and language function. You take the high node and I’ll take the low node: Papers from the Comparative Syntax Festival, 33–52. Chicago: CLS. ———, and Kenneth C. Hill. 1977. Language death and relexification in Tlaxcalan Nahuatl. International Journal of the Sociology of Language 12.55–70. Hymes, Dell (ed.) 1971. Pidginization and creolization of languages. Cambridge: Cambridge University Press. Krauss, Michael E. 1963–70. Eyak texts. Photocopy. Kurylowicz, Jerzy. 1949. La nature des procès dit ‘analogiques’. Acta Linguistica 5.121–38. [Reprinted in Hamp et al., 158–74.] Lambert, Wallace E. 1967. A social psychology of bilingualism. Journal of Social Issues 23.91–109. Miller, Wick. 1971. The death of language, or Serendipity among the Shoshoni. Anthropological Linguistics 13.114–20. Mühlhäusler, Peter. 1974. Pidginization and simplification of language. (Pacific Linguistics, B26.) Canberra: Australian National University. Reinecke, John E. 1971. Tay Boi: notes on the pidgin French of Vietnam. In Hymes, 47–56. Samarin, William J. 1971. Salient and substantive pidginization. In Hymes, 117–40.

chapter 3

Making do with Less: Some Surprises along the Language Death Proficiency Continuum 1 Introduction The kinds of loss, both general and particular, reported by now for obsolescent speech forms are legion. Phonological, morphophonological, morphological, syntactic, semantic, and lexical complexity or richness have all been shown to undergo reduction as languages are used less and/or used by a dwindling number of speakers. The reductions involved can range from outright disappearance of forms, constructions, categories, and stylistically or semantically related options, to a drop-off in the types or overall number available (or at least in use). Both of these outcomes can result, it seems, from a variety of circumstances. The number of conversation partners with whom use would be appropriate may decline to a prohibitive level; audiences for certain types of verbal performance may vanish. The structure of the language itself may make the merger of two previously parallel sets of forms or syntactic constructions likely. The relative difficulty or rarity of certain forms, constructions, lexical differentiations, and so forth, can lead to partial or total disuse (with or without some residual fossilization). The existence of similar structures or items in both superordinate and subordinate language can lead to the retention of just one of several original options, via convergence, in the subordinate language. In general it appears difficult to maintain marked complexity such as elaborate oratorical style, rich allomorphic variety, or minutely differentiated lexical fields under circumstances of prolonged and intense language contact, particularly if one of the languages enjoys significantly greater prestige and official support. It may be difficult even in relative geographical or social isolation if the economic vitality of a community falls off too sharply or if the number of speakers falls too low (the two developments perhaps occurring in tandem). 2

Scottish Gaelic

Over a number of years and in a considerable number of papers I have been concerned to document the processes of reduction in a dialect of Scottish

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�05

94

chapter 3

Gaelic approaching extinction in the extreme northeast of the Highland mainland, East Sutherland Gaelic (ESG). For the most part the processes dealt with have appeared, and been presented, as changes in the structure of ESG apparent in the speech of different age groups in a population divisible into a proficiency continuum. This continuum ranges from fully fluent speakers (themselves subdivisible into older fluent speakers (OFSs) and younger fluent speakers (YFSs) in the smallest and most isolated village, where the number of speakers was highest and densest and the disappearance of the dialect correspondingly slow as compared to the situation in the two other villages where ESG was spoken) to imperfect “semi-speakers” (SSs) dominant in English and more proficient in English, but for a variety of reasons (see Dorian, 1980) still making some use of Gaelic. In the present paper I would like to take the proficiency continuum and the reductions and losses across its range as well-established and discuss instead the two extremes of that continuum in rather different terms: the fully fluent speakers in terms of the occasional gaps and failings which appeared in their Gaelic, and the semi-speakers in terms of the sometimes surprising strengths or compensatory mechanisms which made their Gaelic more serviceable than its structural integrity (or lack thereof) might have seemed to warrant. This accounting seems a useful corrective to any overly simple notion of straightforward decline, with “superb” speakers at the top and “terrible” speakers at the bottom. Even though there is such a pattern (as the very term proficiency continuum implies), the decline of a language is too complex a matter for things to be quite so linear. It seems time to acknowledge and set out the deviations from any imagined straight downward path of decline, lest the phenomena associated with language death and contraction be taken, quite erroneously, to be more obvious and less interesting than they actually are. 3

The Speakers of East Sutherland Gaelic

It should be noted at the outset that the fully fluent speakers in this study, no matter what their numbers in a particular village and no matter how frequent their opportunities actually to speak Gaelic on a regular daily basis in later life, had all been raised in households where Gaelic was the first language and in a social setting where segregation of ESG speakers (residentially, occupationally, in social intercourse, and in marriage) was still largely the norm. They had learned Gaelic first, as children, spoke it either exclusively or much better than English when they began school, continued to use it habitually in their parents’ homes, and in many cases could still be considered Gaelic-dominant

making do with less

95

when included in my study if the criterion for dominance was taken to be preference for Gaelic under conditions of emotional or physical stress. Use of Gaelic is above all interlocutor-governed in East Sutherland; certain individuals trigger the use of Gaelic in certain other individuals on a habitual basis, whereas matters such as topic or setting, though sometimes relevant, are unimportant by comparison. One fairly effective measure of Gaelic dominance was the degree to which a fluent speaker used Gaelic with me or in my presence under conditions of excitement, fatigue, anger, worry, and the like. I was a speaker, but not ideally fluent, nor could I be counted on to be a perfect decoder, most especially under emotional or physical circumstances such as when the native speaker’s tempo picked up, the noise level rose, or the number of disfluencies sharply increased. If a fluent speaker persisted in Gaelic, or broke into Gaelic, under such circumstances when I was the interlocutor or among just a very few interlocutors, the likelihood was very great that that speaker was simply more comfortable in Gaelic to a degree where a less-thanideal conversation partner could not inhibit the preference for use of Gaelic. It is on the basis of participant-observation over a fifteen-year period (1963–1978) during which there were a good many such occasions that I venture the opinion that many of the fluent speakers with whom I worked should be considered Gaelic-dominant, including – very conspicuously – the youngest of them. A smaller number of individuals could best be considered truly balanced bilinguals, almost entirely responsive to the interlocutor factor (or that factor combined with topic and setting factors) regardless of any stress factor. And a few (including one of the oldest – probably, judging by his remarks, by reason of emotional reaction to the local bias against the Gaelic-speaking fisherfolk) had become English-dominant. By contrast, two or three elderly people were so obviously Gaelic-dominant (more fluent and better able to express themselves in Gaelic) that no close observation of their behavior under stress conditions was necessary for assessment of preference or relative abilities. Since I will be pointing to weaknesses in the Gaelic of the fully fluent here, it seems important to establish that despite their ever diminishing numbers and steady movement into the upper reaches of the population pyramid in East Sutherland, these were speakers of high verbal skill who used their native language for a quite normal range of purposes. They could narrate, argue, joke, gossip, tease, discuss health, community, national affairs, and business matters; they were fully able to exchange news, advice, plans, and opinions on matters grave and trivial. All could quote Biblical passages and psalms; some – with very few exceptions, males – controlled the relatively formal language of prayer. All knew some proverbial lore and children’s rhymes. A relatively small proportion (again mostly males) could read familiar Biblical Gaelic passages;

96

chapter 3

a few could read nonreligious material in Gaelic. Very few could write Gaelic with any accuracy or freedom as to subject matter – probably fewer than half a dozen. Because of the ease with which English loanwords are taken into ESG and the remarkable degree to which they are adapted to Gaelic grammar and morphophonology, even the most up-to-date and technical subjects could be discussed in Gaelic, and were. Extremes of code-switching were very rare for this reason, except that certain particular people were notorious for it as part of their personal style and were strongly condemned for it; such individuals were excluded from the sample of the fully fluent in this study, since the community did not in general regard them as fully fluent in speech behavior, whether or not they might have been capable of full Gaelic fluency by dint of greater effort and self-discipline. One peculiarity of the fully fluent sample merits special mention: On the whole, the younger half of the fluent sample had more opportunity to use Gaelic regularly, at the time of the study, than did the older half. This reflects the simple fact that the younger individuals were more likely to have living spouses, siblings, and schoolmates available as regular conversation partners, whereas the older individuals were likely to have lost some of their regular Gaelic conversation partners to the higher mortality natural for their age group. There were certain noticeable consequences of this phenomenon, most particularly in lexical recall. The more elderly sometimes had more difficulty retrieving vocabulary items, especially in isolation and on demand, than their younger counterparts, even when the words in question were in fact well known to them and could be heard in their spontaneous speech. No such age or isolation factor appeared in grammar, however. The older fully fluent speakers were for the most part better able to supply conservative structures, although in this connection personality factors came into play: some individuals among the ranks of the fully fluent, older or younger, could be pressed for the most conservative or elaborate structures in their repertoire more successfully than others (that is, without becoming confused or hesitant and with relish for the challenge to their abilities). 4

Troublesome Structures: Morphological

Most of the structures which proved difficult for fully fluent speakers were syntactically somewhat complex and had probably fallen out of common use as the language receded. One exception was a deficiency in morphological structure, namely uncertainty about the formation of ordinal numbers as opposed to cardinal numbers. Gaelic numbers are notoriously difficult compared with

making do with less

97

English numbers, since Gaelic numerals are mostly based on multiples of 20 and call for special rules of morphophonology and for slightly irregular choices between singular and plural in any accompanying nouns. Gaels all over the Highlands will tend to resort to English even in the middle of a Gaelic conversation in order to name a year (e.g., “1962”) or to mention a high number (e.g., “I lost 76 lambs during the bad weather”). The difficulties are apparent if one translates the Gaelic for “76 lambs” element-for-element into English; the most common ESG formulation would give “six-lambs-teen on three twenty” in English.1 Among fluent East Sutherlanders this line of least resistance (use of English for high numbers) is generally taken, too, but when asked to produce a higher Gaelic cardinal number, they prove able to do so – flawlessly, in fact, though they may hesitate for a bit before coming up with a number of this kind. Ordinals caused absolutely no difficulty until the first multiple of 20 came into play at “40th,” whereupon people who were able to form “20th” by (correctly) adding /-u/ to the numeral “20” and “30th” by (correctly) adding /-u/ to the numeral “10” in the construction which translates literally as “10th and 20” in ESG became slightly uncertain about what to do with the /-u/ they knew to be required for ordinal formation. Although at this point in ordinal formation no one appended it (incorrectly) to the numeral “2,” some (incorrectly) left it off the numeral “20,” producing in effect the cardinal instead of the ordinal. This problem continued to appear in the rest of the round decade ordinals up to but not including “100th” (which is not based on a multiple of 20; neither is “50(th),” however, and rare substitution of “50” for “50th” did occur among fluent speakers). In between the round decade numbers a more severe problem cropped up, most notably in the numerals where figures in the teens must be added to the multiples of 20: that is, in the 30s, 50s, 70s, and 90s (the pattern being literally “third (noun) teen on 20” for “33rd,” for example, and so forth). Asked to produce the phrase “78 men,” fluent speakers had no real trouble. Asked to produce “the 78th man,” they showed uncertainty about where the obligatory /-u/ should be attached: to the “-teen” element of “18,” or to the “3” element of the “3 twenty” to which the “18” must be added? Neither is correct, though the former was a more common attempt. The /-u/ belongs on the “20” element to give (literally) “the 3 20th man and 18.” The problem arises because 1 There are some extra complications not readily discernible even in an element-for-element translation. For just one example, 2, 20, and all multiples of 20 are normally followed by the singular of any noun, while all other numbers take the plural of any accompanying noun. But nouns which occur with very high frequency after numbers sometimes appear in the singular after 3, 4, 5, and 10 in ESG: “year” and “time, occurrence” are two such examples.

98

chapter 3

there are too many elements in such a numeral (“3,” “20,” “8,” and “-teen”), all of which present themselves at least theoretically as candidates for the obligatory ordinal-forming /-u/ to individuals who simply do not make use of such high ordinals – and perhaps never did. The Subjective Element in the “troublesomeness” of Troublesome Structures In the case of these ordinals, speakers were well aware that they were producing structures which felt or sounded wrong, but they were at a loss as to how to improve them. Other morphological formations which were incorrect from the point of view of orthodox Gaelic produced no disquiet among fluent ESG speakers and had apparently become acceptable even when obviously out of line with high frequency fossils on the orthodox model still remaining in active use in the dialect. An example of this sort is adverb formation, which traditionally preposes the unstressed element /kə ~ ku/ to an otherwise unchanged adjective (/kəh ~ kuh/ before vowel-initial adjectives). The traditional model survives well in such frequent expressions as /kə tɔ̃ n/ “badly,” /kəh εçar/ “soon, promptly,” /kə mã/ “well,” /ku ɫua/ “quickly,” /kə na:dərax/ “naturally,” and sporadically in some other relatively common adverbial phrases. Asked to form adverbial phrases not in familiar use, most speakers used an unadorned adjective form by itself or, seemingly for emphasis, added the heavily stressed adverb /kle:/ “very” before the adjective form. There was no unease about such phrases and seemingly no awareness of the discrepancy between the fossils with /kə ~ kəh/ and the productive formations without the unstressed adverbial particle. The very occasional individual formation which did include /kə/ (a rare example was /va khĩãd ku thaiax/ “He was looking carefully” from an Embo OFS) occasioned no comment one way or another if other speakers were present. (The same Embo OFS on the same occasion actually dropped the /ku/ from the phrase /ku ɫua/ “quickly”, where it is usual, also without provoking any comment from another fluent speaker who was present when adverbial phrases were being elicited and attention was consequently focused on the structure.) The anomalies of adverb formation are within the range of acceptable variability for ESG, then, whereas the anomalies of ordinal formation represent a loss of control in productive morphological skill and are felt as such. There is clearly a subjective or conventional element in these matters. Both anomalies are “mistakes” from the point of view of traditional grammar. Even in ESG both structures follow traditional models in the more commonly used range of expression (high frequency adverbial phrases, ordinals up to “40th”). But the anomaly of adverbial phrases identical with adjectives does not perturb fluent speakers, whereas the anomaly of ordinal modifiers identical with 4.1

making do with less

99

cardinals does. And this despite the fact that context makes the intended meaning quite clear in both cases. 5

Troublesome Structures: Syntactic

Most of the structures which proved problematic for fully fluent speakers involved syntax rather than morphology. Although it seems very likely that the constraints of a translation task increased the difficulty of the structures, it has to be kept in mind that all of these speakers were being asked to produce many other complex structures via translation tasks as well; yet only the ones discussed here were notable for the rather general and self-conscious difficulty they caused speakers of quite considerable skill. Two out of the four problematic syntactic structures shared the feature that they used essentially the same construction twice in immediate succession. There seemed to be something inherently confusing about such structures, since one of the two which proved troublesome is relatively short and simple, while the other is relatively long and involved. I never heard either one used in spontaneous speech. Rather, I realized that the structures would exist (in one case because it was so logically obvious, in the other because a colleague asked me how something would be expressed in Gaelic and it dawned on me what the answer would have to be, at least in a conservative form of ESG), and they would come out a certain way in ESG, given the grammar of the dialect. I then set out to see whether they could be elicited. I should note, again, that I went after a few other structures which did not crop up in spontaneous speech in the same fashion and got responses without hesitation and without faulty grammar or distorted meaning.2 The simpler of these two problematic structures called for two verb complements in a row, each a gerund. Given that people could build sentences of the type “I’m trying to hear him” and also of the type “I’m making him go,” it stood to reason that they should also be able to construct sentences which combined the two types, to produce sentences such as “I’m trying to make him go.” Yet when I asked for (among others) the sentence “I’m trying to make him hear me,” entirely fluent speakers produced defective translations with various missing elements. The defective productions typically dropped one 2 An example would be “You ought to have a dog,” a nonobvious structure in ESG because there is no lexical verb “to have” to serve as the complement of “ought.” The translation turned out to be, roughly, “It’s incumbent on the dog to be at you,” and no one had the slightest trouble producing it on demand.

100

chapter 3

complement or another, to give, for example, “I’m making him hear me,” or, more oddly, “I’m trying him to hear me.” One resourceful fluent speaker made an end-run around the difficulty by phrasing the sentence altogether differently: “I’m trying to make that he’ll hear me”; but the structure of the expression “to make, cause to happen” was itself slightly defective in this version (it lacked a prepositional element). Another sentence of the same general type (“Are you trying to make me leave?”) gave rise to similar difficulties; exceptional was the similarly structured sentence “He’s trying to make her put a fire on,” which four out of four speakers asked got right on the first try. By contrast, four out of five speakers gave defective versions of “I’m trying to make him hear me,” and four out of eight gave defective versions of “Are you trying to make me leave?” Two of the four who gave defective versions of the former were tested on two different occasions separated by at least two years, with the same sentence presented on each occasion; one got it right on one occasion and right on the second try on the other occasion, while the other speaker never arrived at a correct version despite many tries on one occasion but produced a correct version – barring a missing initial consonant change required – on the third try, on the other occasion. In my subsequent efforts to fathom the curious difficulty caused by this double verb-complement structure, I found that passive ability to handle it was much better than active ability. On a different field trip from the one during which I did most of the investigation of active control of this structure, I presented the sentence “I’m trying to make him hear me” in ESG to eight fluent speakers and asked for an English translation. Six of the eight produced accurate English translations instantly and easily.3 In view of the difficulties with active control of the double verb-complement structure, it was not surprising to find that fluent speakers also had trouble with prepositional possessive relatives of the type “The woman in whose house it happened died.” Such sentences would require, in conservative ESG, two identically structured prepositional relative clauses in immediate succession: literally, “Died the woman at (-whom) was the house in (-which) happened it” (with the italicized portions representing exactly parallel structures). Because the verbs in this sort of sentence have to be inflected for tense, involving not only the infamous initial consonant mutations of Celtic

3 One YFS chose an alternative meaning, namely “show,” for the verb which also translates as “try,” and then ignored the expression “to make” in favor of a continuation which made sense with “show”: “I’m showing him how to hear me.” One OFS used “try” in her translation, but weakened the causative sense of the remainder: “I’m trying for him to hear me.”

making do with less

101

(as was also true of the double verb-complement structure) but also particles and dependent as opposed to independent verb forms, they can be considered difficult as compared to the double verb-complement sentences. In the latter, the verb complement is always the gerund, changeless except for the initial consonant mutation if that consonant is one susceptible to the mutation (always lenition in this construction, a mutation which generally replaces stops with spirants and spirants with other spirants or zero: /khɫĩ:nčən/ “hearing,” /fiaxən . . . xɫĩ:nčən/ “making . . . hear,” /fa:gal/ “leaving,” /fiaxən . . . a:gal/ “making . . . leave”). Fluent speakers asked to produce a Gaelic translation of the English sentence “The woman in whose house it happened died” vastly preferred to avoid the double prepositional structure. Two anomalously began the sentence with nouns (Gaelic is a Verb-Subject-Object language); they then tried to repair, producing respectively “The house in which it happened, the woman died” and “The woman in the house it happened, she died.” With each of these speakers (both very cooperative and resilient sources, luckily), I asked them to try again, beginning the sentence with “died,” as would in fact be normal for Gaelic. The elder of the two then got the structure right; the younger produced “The woman in the house it happened died.” After several more tries, each close but slightly off, I said the conservative Gaelic structure myself. He responded, “That’s right enough.” Mindful of Labov’s experience with nonstandard English speakers who proved unable to repeat a formulation outside the patterns of their own dialect even when well motivated (Labov, 1979, pp. 332–334), I then asked him to repeat the version I had said. He did so with no difficulty whatever. Another YFS who thoroughly liked challenges was posed the problem; he rephrased, but produced a good ESG sentence, namely “The woman who belonged to the house where it happened died.” Pressed for another version, he gave “The woman at (-whom) was the house where it happened died.” Asked to keep this general structure but work in the preposition “in” after the noun “house,” this man then produced the double prepositional (possessive) relative on the third try. A quite different sort of construction which struck me as possibly troublesome for the fully fluent even before I began to investigate it closely seemed predictably difficult both because the conjunction involved was of low frequency even in the positive and also because the negative clauses introduced by that conjunction became semantically very close to a much higher frequency construction. Thus, a shift to the higher frequency alternative seemed likely. This was in fact what happened. Most fluent speakers were familiar enough with the conjunctival phrase /nas ɫə: na/ “unless,” but even in the positive they were inclined to use “if . . . not”

102

chapter 3

in its stead.4 Asked to render “He’ll do it unless he’s too tired,” three out of four speakers initially produced “‘He’ll do it if he isn’t too tired’”; one shifted to /nas ɫə: na/ when pressed to translate the word “unless” more literally. Asked for the Gaelic of “I won’t go unless he goes, too,” four out of seven fluent speakers produced the /nas ɫə: na/ construction, but the other three, all OFSs, used “if . . . not.” One of the three who used “if . . . not” was pressed for “unless” but was unable to produce it on this occasion, although she did use it in other scattered translation-test sentences on a variety of occasions. Presented with a Gaelic sentence which used /nas ɫə: na/, eight fluent speakers (out of eight asked) translated it unhesitatingly, uniformly rendering the conjunction as ‘unless’ in English. In the negative, when Gaelic-to-English translations were requested, the switch-over to a form of “if” was more prevalent than in the positive, as predicted. For example, eight out of eight fluent speakers asked rendered /nas ɫə: na/ as “unless” in their English translation of the Gaelic sentence “He won’t sell this house unless he gets another house,” but only four out of seven fluent speakers rendered /nas ɫə: nax/ as “unless” in the sentence “He won’t get [= inherit] the shop unless he doesn’t marry her.” In this latter sentence, three fluent speakers used “if” in their translations, all incorrectly, to give “if he doesn’t marry her.” Among the four who preserved “unless” in their translations, three also made the “unless” clause positive, reversing the sense of the original. One caught his own mistake and then supplied a (correct) negative version of the English “unless” clause, but three others carried on with the positive. One of them, a YFS, spontaneously translated her English version back into Gaelic, and in doing so she also made the Gaelic version of the “unless” clause positive, that is, the reverse of what had been given her to translate. Semantically either “. . . unless he marries her” or “. . . unless he doesn’t marry her” is a possible sequel to “He’ll inherit the shop . . .”; this probably made the translation task more difficult. In the Gaelic-to-English translation of another sentence, “I’ll see him tomorrow unless he doesn’t come,” shifting to the positive in the “unless” clause would make no sense, and indeed no one did that in this instance. Three fluent speakers offered English translations; all used “unless” without hesitation, and all gave negative versions of the second clause. (One of the three was a YFS who had made the negative “unless” clause

4 The form of this conjunctival phrase varies slightly across the three ESG-speaking villages. It is given here unvaryingly in the Embo form simply for consistency’s sake. But when Gaelic sentences were presented for translation into English I always altered the form, both positive and negative, to suit the native village of the person to whom the sentences were presented.

making do with less

103

positive in the sentence discussed just above, translating “. . . unless he doesn’t marry her” as “. . . unless he marries her”; she showed no such tendency here.) It is quite clear from the cumulative evidence scattered through my field notebooks from over a good many years, as well as from more concentrated and focused testing, that most fluent speakers of ESG were able to form sentences in all tenses using /nas ɫə: na/ and /nas ɫə: nax/ to express “unless” and “unless . . . not” if they were specifically asked to do so. It is equally clear that they rarely chose to do so. Not one of the more than 70 Gaelic “unless” and “unless . . . not” sentences recorded in my notebooks came from a freely spoken narrative, and I have a marginal notation of one YFS’s comment that she wouldn’t use the “unless” form that she produced for me if she were speaking completely naturally. This evidence – plus the marked tendency to mistranslate by reversing the negation, where either a negative or a positive subordinate clause introduced by “unless” is semantically sensible – indicates that a conjunctival use of “unless” and “unless . . . not” lingers on in the dialect but has become somewhat unnatural and consequently somewhat confusing and prone to inaccurate use. The final construction which proved difficult for fluent speakers involved what would appear to be a pluperfect, from the English point of view, in condition contrary-to-fact clauses introduced by “if.” That is, a pluperfect verb form distinguishes this construction from a more common one with which ESG speakers frequently confuse it, when the two are contrasted in English. But ESG cannot really be said to have a pluperfect verb form, in any tense or aspect, and clauses which contain the pluperfect in standard English are typically expressed in the simple past not only in the Gaelic of these bilinguals but also in their English. Standard English “I didn’t know who he was – I’d never seen him before” corresponds to “I didn’t know who he was – I never saw him before” in both languages for the bilingual speakers of ESG. One method of expressing in ESG the notion of time still earlier than that of an accompanying past-tense clause is to use a passive. Although an ESG speaker would normally say simply “We ate our dinner before he came” (rather than “had eaten”), it is possible to say, as one fluent speaker pressed for a “pluperfect” did, “Our dinner was eaten before he came.” In the case of “pluperfect” condition-contrary-to-fact sentences, it is significant that the condition in question is past and whatever might have followed on it did or did not do so; this is opposed to ordinary conditions, where the situation still holds and whatever might follow on it may or may not yet happen. The translation task I posed most frequently and systematically to force this issue was chosen to make this crystal clear. I asked for Gaelic translations of the following two sentences in immediate succession: “If the rope had broken,

104

chapter 3

we would have been killed,” and “If the rope broke, we would be killed.” In the first English sentence it is now certain that the rope did not in fact break, and we were consequently not killed, though we might have been. In the second English sentence our fate is still undetermined – we may yet be killed, though it is only speculation rather than prediction (whereas “If the rope breaks, we will be killed” is predictive). Earlier translation-task experience had indicated that the direct contrast between the two English sentences was necessary if I was to get any Gaelic responses at all which reflected the fact that the condition was past and no longer pertained in one instance but not in the other. The distinction which appears pluperfect in form in English can be made in ESG, but it involves the use of a passive state-of-being (the “passive” being more nominal than verbal in ESG in any case; see Dorian, 1973, for discussion); literally “If the rope were at/to its breaking . . .,” (that is, “broken”) as opposed to “If the rope would break. . . .” Despite the direct contrast in the later testing there were some failures. One YFS, asked first for the “pluperfect” contrary-to-fact and then for the regular conditional, gave the conditional in both cases. Realizing that his responses were identical, he pondered a moment and then said, “It’s exactly the same. Funny, that, isn’t it?” He proved unable to produce a “pluperfect” sort of construction. Another YFS gave a “pluperfect” for the regular conditional, an extremely unusual mistake, then gave the regular conditional in the “if”-clause of what should have been the contrary-to-fact with “pluperfect” in her effort to differentiate. In confusion at that point, she got the result clause wrong – she gave it as a simple past tense rather than any sort of conditional, although she was an excellent speaker who normally had no trouble whatever with conditionals as such. A Golspie OFS gave identical renderings for both English sentences until I objected and pressed strenuously for a “pluperfect,” at which point she did produce one. One Embo OFS and one Golspie OFS got the contrary-tofact completed condition right on the first try, given the direct contrast in the English stimulus sentences. On a second immediate-succession-contrast test, the same YFS who had remarked that his first two sentences were “exactly the same” also gave identical uncompleted-condition sentences in this case: He produced “If he came, we would leave” in Gaelic as a translation both for that sentence in English and also for “If he had come, we would have left.” The other two speakers who made mistakes in the first test (reversal in the one case, semantic collapse in the second) now got the completed condition right in the second pair of test sentences. Evidently one round of direct-contrast testing was enough to allow them to straighten the confusion out and handle the distinction. But once again, both the presence of failed attempts scattered throughout my field note-

making do with less

105

books over the years and also marginal notations in those notebooks about the difficulty of eliciting the completed contrary-to-fact condition indicate that the contrast between a possible condition which still may apply and a possible condition which was not fulfilled is maintained poorly and is recognized as confusing and seen as difficult, even by fully fluent speakers. Apart from these four syntactic environments and the one morphological one described first, other notable grammatical difficulties manifested by fully fluent speakers were either idiosyncratic (peculiar to a given speaker or to one family) or clearly associated with a lexical item or a grammatical category which had passed out of common use across the board. There were certainly other deviations from the most conservative grammatical norm (the merger of the two passive constructions is a case in point; see Dorian, 1973), but the speakers were still able to express their meaning clearly and to differentiate the structure concerned from other structures. They showed no recurrent confusion or hesitation about the utterances they offered in such cases. 6

The Rarity of “Troublesomeness” and Possible Reasons for the Instances Found

To a researcher who spent a good deal of time over more than a dozen years deliberately thinking up posers for a group of speakers of a language clearly fading from existence, with opportunities for use diminishing along with the number of available habitual conversation partners, it seems remarkable that there are so few constructions as this which cause self-conscious difficulty. The facts that the population was largely illiterate or at best literate only in a very limited fashion in their mother tongue, and that they got little or no positive reinforcement for their skills in Gaelic until I appeared on the scene and began to express a linguist’s admiration and appreciation for their abilities,5 likewise make the relatively small number of “troublesome” constructions remarkable. As to the nature of the constructions which caused the self-conscious difficulties and the reasons for its being these constructions rather than others, a number of possible factors seem to be involved. Only in one case does it seem likely that English has in any direct sense driven out Gaelic (although increased use of English and decreasing use of Gaelic is surely a general factor in all decline and decay of Gaelic in East Sutherland): As noted earlier, the formation of higher numerals is notoriously complex in Gaelic, and all over the Highlands native speakers of Gaelic show a tendency to substitute English 5

My opinions were contrary to the norm: my informants were looked down upon within the local setting for being Gaelic speakers.

106

chapter 3

for high Gaelic numerals, both cardinals and ordinals. Also as noted earlier, the general absence of a “pluperfect” in both Gaelic and East Sutherland English makes difficulties in contrasting anterior and present time in conditional sentences understandable, since real need for that sort of distinction would be felt relatively rarely; and the semantic substitutability of “if” and “if . . . not” for “unless . . . not” and “unless” makes use of the higher frequency conjunction “if” an attractive alternative in that case. If anything, the retention of “unless” by any speakers at all becomes a bit hard to explain. It is a low-frequency conjunctival phrase and not strictly necessary. The most likely explanation for retention is that /nas ɫə:/ is the comparative of “little” and occurs fairly frequently in the meaning “less,” thus keeping the general expression /nas ɫə:/ in common use. The connection of less with “unless” in English translation may be relatively salient for some individuals. As for the prepositional possessive relative, its complexity contrasts sharply with the extreme simplicity of subject and object relatives in Gaelic. These latter use a relative particle undifferentiated as to case, number, or gender followed by a verb phrase identical with that of a main clause; even a good number of SSs succeed in forming them. By contrast just the prepositional relative (let alone the prepositional possessive relative) is a complicated structure. In its most conservative form it calls for a prepositional structure with a verb form as object, and the verb (or its accompanying particle) must show the initial consonant mutation known as nasalization, which typically voices voiceless consonants and replaces spirants with stops. Furthermore, the verb or verb phrase must appear in the dependent rather than the independent form. Since the prepositional relative phrase requires two of these relatively complex structures in a row, it is not unexpected to find that fluent speakers have trouble with them, even speakers who can form the single prepositional relative satisfactorily. It is also the case that there are two possible forms of the purely prepositional relative (e.g., “That’s the broken chair that I was sitting on”) in ESG, one of which postpones the preposition until the end of the clause, probably under the influence of English, and permits the substitution of a simple relative without mutation and without dependent verb form for the more complicated structure called for by a nonpostponed preposition. This alternative is not available in the prepositional possessive relative, with its doubled-up structure; but some speakers may nonetheless be struggling mentally to locate such an alternative, since it’s common enough (preferred, even, by some speakers) for the prepositional relative as such. The sole construction troublesome for fully fluent speakers which seems particularly unexpected is the double verb-complement structure. It is not really very difficult; it is both logical and relatively short. Only the end-to-end linking of almost identical structures makes it confusing, and it was strictly on

making do with less

107

that basis that I foresaw the difficulties fluent speakers might have with it and began to ask for translations of sentences which would require it. 7

Unexpected Strengths of Semi-Speakers

What about the possibly surprising strengths of the Gaelic used by semispeakers (SSs)? In a sense, of course, all of their abilities are surprising. I will take most of their capacity for granted, all the same, and discuss here only a few of the more unusual aspects of their use of Gaelic. For the SSs, such unusual characteristics are usually individual rather than group-wide, since their Gaelic is much less stable and much more idiosyncratic than that of fluent speakers. Their acquisition histories are not uniform: some began as fluent child speakers and lost capacity; some were always imperfect speakers; one or two were passive but not active childhood bilinguals who somewhat later made an effort to develop some active skills. Furthermore, some have had much more sustained interaction with fluent speakers than others: a few have lived all their lives in households where fluent speakers used Gaelic regularly within their hearing, while most others got the bulk of their exposure and active experience from elderly kinfolk who survived for varying periods of the SSs’ lives. Individuals within the latter (and larger) group have had quite variable opportunities to hear and use Gaelic since the death of the crucial older person(s) in their kinship networks. All of the SSs have extraordinary receptive skills in Gaelic – an ability to decode messages which is dramatically out of line with their ability to encode messages. However defective their phonology, morphophonology, morphology, and syntax, they have the stunning ability of the native speaker to understand virtually everything said in their presence, regardless of noise level, speed, or faulty articulation. This is the more striking because it contrasts with my own skilled-learner shortcomings in these circumstances, despite the fact that my Gaelic is more fluent and more grammatical than theirs. Their decoding skills are not merely contextual, furthermore, since they can provide actual translations if asked or if they see that someone is not following.6 When it comes to their active skills, SSs have to make do with very much less than the resources fluent speakers have available if they want to join in Gaelic interactions. They can achieve a surprising amount of success by sticking largely to fixed phrases and very high-frequency vocabulary and grammatical 6 Confirmation of the exceptional decoding skills of very imperfect speakers (and even one completely passive bilingual) is provided by comprehension tests carried out by Schmidt (1985) among the youngest users of Dyirbal in North Queensland, Australia.

108

chapter 3

constructions (see Dorian, 1982, for a particularly extreme case of reasonably successful interaction by a “speaker” of drastically limited resources). They were probably rarely pressed to their verbal limits so sharply as they were by the English-to-Gaelic translation tests which I asked them to undertake so that I could match their results with those from a sample of fluent speakers. I hoped by this device to explore the precise areas of difficulty and the points at which their active skills broke down. The seven SSs who did the bulk of the translation tests (a number of others did various subsets) included two relatively good SSs, two of intermediate skills, and three very weak SSs. Five were very willing and for the most part also eager to try their hand at the sentences. Two of the very weak “speakers” were less enthusiastic: One never used Gaelic spontaneously except on rare occasions to keep a secret from her son or some other nonspeaker, but because of my ties to other members of her extended family she was unable to refuse to help me and did the tests fairly graciously and unexpectedly successfully, considering how little she used her Gaelic. The other was a nervous, rather high-strung woman who found the task stressful, even though she had known me for many years and was essentially generous spirited and quite willing to be helpful if she could. I had of course no opportunity to compare the ordinary conversational skills of the nonuser with her test results; the abilities of the nervous test-respondent corresponded very well to what I came to know of her Gaelic eventually from listening carefully to her interactions with kinfolk who were much better speakers. 8 Conjunctions Despite the fact that conjunctions are a closed class, and a relatively small one, they represent a trouble spot for SSs. There are a number of reasons for this. Most conjunctions require choices about mutation or nonmutation of the initial consonant in the main verb of the clause which they introduce or in a particle which precedes it; this is sometimes the same as and sometimes in addition to a choice between a dependent and an independent form of the verb. Some of the conjunctions are of low frequency and simply do not make their way into the SSs’ repertoires; a few conjunctions have several different forms depending on what tense they appear with, or, especially, on whether the clause is negated or not.7 7 Negation is a special problem because a few conjunctions (and also some particles) have a negative counterpart which subsumes the negation, so that no other negating element appears in the clause; the negative form may or may not resemble the positive form in sound.

making do with less

109

The test sentences were not actually designed to explore the control of conjunctions, since conjunctions overall seemed a more complex matter than most SSs could be expected to handle. Nonetheless a number of the test sentences did contain conjunctions in the natural course of things. Although the coverage is far less systematic than would have been the case if conjunctions had been a real focus of the tests, the number and range of such sentences offer an interesting glimpse of SS ability to deal with relatively challenging material. Where a SS simply did not know the conjunction in question, several strategies appeared. Rarest by far was borrowing from English. One of the two strongest SSs adopted /khəs/ (from English because) in the sentence “He didn’t come because he’s too tired,” but the rest of those asked (none of whom were familiar with fluent-speaker /wəl/ “because”) generally omitted the conjunction and left the causal connection to be inferred from juxtaposition of “He didn’t come” and “He’s too tired.” It should be noted that “because” is not the commonly used conjunction in Scottish Gaelic that it is in English; although ESG use of /wəl/ is moderately high, the form is a local one and many western dialects seem to prefer causal inference of the type demonstrated by the ESG SSs or else substitution of “and” (“He went home early and him tired with working in the field”). Substitution of a more common conjunction, where the general meaning permitted, was a useful strategy among SSs for making the complex more expressable. A strong SS, who knew and used the conjunction “when” on occasion, nonetheless circumvented it in two cases, once via a shift to “and”: “She lost her life when she was only five” was rendered as “She lost her life and she wasn’t but five.” A very weak SS made a similar substitution of “but” for a conjunctival construction she didn’t know at all: “He understands everything, but can’t speak,” for “Although he can’t speak, he understands everything.” One of the intermediate-level SSs made the same sort of substitution of “but” for “although . . . not” in the same sentence. Omission, with juxtaposition carrying the burden of semantic/syntactic relatedness, was the strategy employed by the weakest of all the SSs in this same sentence; she rendered it as “He has no speaking; he’s understanding everything.” She employed the same juxtapositional strategy for “if” in several cases, for example, “I won’t see him again; he’s going away tomorrow” for “I won’t see him again if he goes away tomorrow,” and “Don’t give (him) £5; you’ll lose your money” for “You’ll lose your money if you give him £5.” The second of these juxtapositional substitutions for “if” preserves the sense of the original; the first one does not. One of the two strongest SSs and one on the intermediate level allowed “if” to stand in for “unless,” which would have worked if they had used the negative form of the conjunction. But since they either did not control that form or

110

chapter 3

failed to employ it, their substitution resulted in reversal of meaning: “I won’t go if he goes, too” instead of “I won’t go if he doesn’t go” = “. . . unless he goes.” 9 Merger The most common problem which appeared in SS efforts to handle conjunctions was merger (syncretism): the falling together of conjunctival forms which are distinct in fluent-speaker ESG. In one case, the conjunction “if,” the collapse was of no consequence at all as far as meaning was concerned, and none of the SSs made any effort to avoid it or showed any awareness of its occurrence. For the most part they simply eliminated the special form of the conjunction “if” which traditionally appears with the conditional and used the dominant form /mə/ in all tenses and aspects. Two of the weakest, however, and one of the intermediate-level SSs, used both forms variably; the /na ~ nə/ form traditionally reserved for the conditional occasionally appeared with other tenses, and /mə/ appeared freely with the conditional. Most mergers were more problematic. One of the two strong SSs merged “unless” and “until,” with fatal consequences for meaning; that was her only merger apart from the various “if”-forms. The weakest SSs had the severest potential problem with syncretism among their conjunctival forms. The least skilled of all was largely unable to prevent loss of meaning; she seemed to have developed a sort of all-purpose conjunction /tə/, which she used to mean “that,” “when,” “if,” and “until,” although she occasionally used other forms for all of these except “when” and “until.” Of the 18 instances in which she managed to supply a “conjunction” at all, 10 were /tə/. She deleted almost all potential instances of “that” (not a permissible deletion in traditional ESG); on the one occasion when she did use an explicit form, however, it was (incorrectly) /tə/. Of the four meanings for which she used /tə/, only “when” was suitable; she used /tə/ acceptably for three out of four instances of “when,” in fact, deleting in the fourth instance (i.e., no conjunction appeared, but juxtaposition allowed the sentence to remain interpretable). Unfortunately, since she used /tə/ for so many other purposes as well, her “when”-sentences were not interpretable except by context. I scribbled into the margin of a field notebook that this woman remarked plaintively at one point in the testing, “It’s the little words that get me, it’s not the big words,” and on the whole that was true of her Gaelic: Her open-class lexicon was better than her closed-class repertoire of prepositions, conjunctions, and particles. More interesting were the efforts of the other two extremely weak SSs. Each of them made use of forms which were altogether incorrect but were

making do with less

111

also unique, so that their use avoided the ambiguity of merger and allowed more possibility of interpretation by context and repetition. Unhappily, these unique forms were also nonce-forms for the corpus, and so the question of whether they would be used consistently by the SSs in question cannot be answered. Each of these women showed less merger than might have been expected: Partial merger was more common, but the overlap was relatively small, considering how minimal their overall skills were. One of them made consistent use of /kə/ (normally “that”) to mean “when”; as if to compensate, she adopted a strategy of deletion (7 out of 8 instances) for “that,” the only exception being an occurrence with the very high frequency verb “to be.” Three of the 8 instances of “that” in the test corpus occurred in clauses with the verb “to be”; but if inconsistencies and exceptions are to appear, this is certainly likely to be the verb to produce them, since it is highly irregular and has more than the standard number of distinct forms. So it is not entirely unexpected that this speaker produced her one syncretistic use of /kə/ “that” in that particular environment. She used both /mə/ and /na/ variably for “if”; she also extended /mə/ to mean “unless,” and, like the other SSs who used this strategy, she failed to negate and produced a meaning opposite to that of the stimulus sentence. This speaker made one further use of /mə/: She tried substituting it for “until,” a conjunction she clearly did not control. She also clearly did not like the result, since she immediately tried twice again, discarding /mə/, and produced a unique (and incorrect) form /khən ə/ on each of the subsequent attempts. Though wrong, this new effort is much closer to the correct /khəs ə/; she preferred it, and in sticking with it she eliminated the partial merger of “until” with “if” and “unless.” Her cousin, the third of the very weak group of SSs (and the one who does not spontaneously make active use of Gaelic), had some striking successes with conjunctions. She got “when,” “until,” and – most surprisingly – the negative “that . . . not” form correct. She missed on “unless,” but did not merge it with “if”; instead she produced a unique (incorrect) form /kən ə/. Her only syncretistic use of conjunctions was a very unusual one. She never deleted “that,” but she used /ə/ (or a variant of it) three times instead of her usual /kə/ (5 instances). This produced partial merger, because she also used /ə/ (or one of its variants), as well as /mə/ and /na/, for “if.” Furthermore, /ə/ is the normal positive form of the relative particle, and this speaker does make quite regular and correct use of that particle; so the syncretism becomes marked when the relative is taken into consideration as well. (The relative does not need to be considered for other speakers, since no conjunction takes the form /ə/ for them.) All the same, there is no conjunction for which this SS used /ə/ (or one of its variants) as her dominant form, much less her sole form; she does use it

112

chapter 3

consistently and exclusively wherever the positive relative particle is required, so once again the potential for confusion is held down to some extent. Among the seven SSs most closely studied, then, only one has a severe problem arising from syncretistic use of conjunctival forms. The problem of missing forms (conjunctions they do not know) is evident, but it is dealt with partly by rephrasing so that the context, and especially juxtaposition, compensate, or so that substitution of a more common conjunction becomes possible; also partly by inventing unique forms; and partly, indeed, by the potentially dangerous practice of allowing one and the same form to stand for two (or more) meanings – that is, syncretism. Overall, it can be said that most of the SSs dealt reasonably well with the problem posed by conjunctions, especially when one considers that in a less constrained situation they would no doubt resort more freely to rephrasing strategies. They were working on a translation task and had long since become accustomed to that task. They knew me and my methods well, and they understood in a general sense that I was looking for direct comparison. (Many had more fluent kinfolk who also worked with me; though I always presented this particular set of stimulus sentences to single individuals, they knew that I was presenting the same set to many other people as well.) They tried to meet my challenge by actually translating when they could. 10

Subtle Compensations

Just how subtle SSs’ compensatory mechanisms could be became clear to me when I discovered that the most eager among my three weakest well-studied SSs had contrived – surely well below the conscious level – to preserve a semblance of gender and case distinction in a phonologically defined class of nouns where these categories are traditionally marked in ESG, despite the fact that she has lost one entire mutation (nasalization) normally required to make the distinctions.8 She did this by contrasting absence of any initial mutation with lenition, where SSs whose morphophonology was more intact contrasted nasalization with lenition, as did fluent speakers (Dorian, 1981, p. 133). 8 This SS lives in London, as she has most of her life. She spent her first six years in a grand­ mother’s household in Embo, however, and Embo Gaelic was the habitual language of her parents, to whose home in London she moved after age 6. The parents spoke only Gaelic to each other and to nearly all adult visitors to the house, although not regularly to this elder child until she came to insist on it. Thereafter, her mother in particular became a fairly regu­ lar Gaelic conversation partner for her in the London environment.

making do with less

113

This same weak but eager SS took a notion one summer to correspond with me in Gaelic – an undertaking which only becomes extraordinary when the fact that she is utterly illiterate in Gaelic emerges. Despite her lack of any notion of how to write Gaelic, despite the facts that Gaelic has quite a lot of phonemes not present in English and that the English orthography therefore doesn’t provide any obvious method of rendering them, and despite the fact that this woman’s Gaelic is defective grammatically and morphophonologically, there were only two words (in a total of 358 words of Gaelic across four letters) which I was unable to decipher. She managed to devise writings good enough to make her meanings clear, for the most part, with a bit of ingenuity also on my end as decoder. In fact it was considerably easier for me to decode her messages than to adopt her system of encoding so as to be able to write back; again to her enormous credit, however, she was consistent and clever enough in her renderings so that I could use her writing system to respond, even though I had to suppress my own knowledge of Gaelic orthography (incomplete to be sure – ESG is an unwritten dialect and I have never formally studied Gaelic at any time) in order to do so. It is hard to think of more dramatic proof of how much a very weak “speaker” of a dying dialect can do with very little than this woman’s success in communicating by writing through a language which she controlled poorly and in which she was illiterate. 11

Concluding Remarks

Nothing written in this paper is intended to deny the phenomena of loss in a dying speech form. It is as undeniable after all these pages as it was at the outset that ESG is a dying dialect and that as such it is undergoing various sorts of reduction and contraction in what are sure to be its last five or six decades (assuming that the youngest of the SSs live to a ripe old age). The point is not to pretend that loss in form does not accompany severe decline in use (decline either by number of speakers or by function, or both), but rather to try to achieve a bit of balance in the overall picture. It is possible, it seems, to be a very skillful and fluent speaker of a dying speech form and still show manifestations of hesitation, confusion, and even complete inability at certain points of weakness in the grammatical control of that speech form. Equally, it seems possible to have minimal active control of the grammar of a dying speech form and yet to make the most of what control remains, so that with imagination and persistence (and some goodwill on the part of the receivers in the sender/receiver interactions, certainly) quite a bit

114

chapter 3

can be conveyed successfully. Linguists and psychologists may want to reexamine the parameters of the native speaker and speech community concepts in the light of both these extremes.9 References Dorian, N. C. 1973. Grammatical change in a dying dialect. Language, 49, 413–438. ———. 1980. Language shift in community and individual: The phenomenon of the laggard semi-speaker. International Journal of the Sociology of Language, 25, 85–94. ———. 1981. Language death. Philadelphia: University of Pennsylvania Press. ———. 1982. Defining the speech community to include its working margins. In S. Romaine (Ed.), Sociolinguistic variation in speech communities, 25–33. London: Edward Arnold. Labov, W. 1979. Locating the frontier between social and psychological factors in linguistic variation. In C. J. Fillmore, D. Kempler, & W. S.-Y. Wang (Eds.), Individual differences in language ability and language behavior, 327–40. New York: Academic. Schmidt, A. 1985. Young people’s Dyirbal: An example of language death from Australia. Cambridge: Cambridge University Press.

9 I have argued elsewhere (Dorian, 1982) that the abilities of SSs and their ready inclusion in verbal interactions among fluent speakers warrant a redefinition of the speech community.

chapter 4

Negative Borrowing in an Indigenous Language Shift to the Dominant National Language 1 Introduction One change process widely recognized in language-contact situations appears under various names, among them negative borrowing: over the course of time linguistic features not shared by both languages are susceptible to disappearance from use among bilingual speakers (Sasse 1992: 65, Thomason 2001: 231). In language-shift settings, where use of one language is expanding while use of the other is receding, receding-language features that have no counterpart in the expanding language may be particularly susceptible to gradual reduction and eventual loss. Thomason cites a phonological instance from Fenyvesi’s study of a variety of American Hungarian spoken in McKeesport, Pennsylvania: the loss of voicing assimilation in clusters of obstruent consonants (Thomason 2001: 231). Mougeon and Beniak, investigating contact effects in Ontario French, discuss this same phenomenon under the rubric “covert interference”, noting that it can be difficult to establish that the primary causal factor is absence of an equivalent form in the expanding (majority) language, since internal system pressure may also be involved in the reduction or loss of a feature or variant in the receding (minority) language (1991: 10–11, 159–60). Silva Corvalan, in a study of Los Angeles Spanish, recognizes four kinds of linguistic transfer in language-contact settings, the fourth of them being loss of a category or a form in one language that does not have a parallel category or form in the system of the other. This she terms indirect transfer, offering as an example the loss of adjective gender marking in some varieties of Los Angeles Spanish in contact with English (1994: 4). She notes, in agreement with Mougeon and Beniak’s position, that transfer “leads to, but is not the single cause of, convergence, defined as the achievement of greater structural similarity in a given aspect of the grammar of two or more languages, assumed to be different at the onset of contact”; as a second factor she recognizes “pre-existing internally motivated changes”, with convergence “most likely accelerated by contact” (1994: 4–5, italics in original). As a test of the hypothesis that features without a parallel in the expanding language will, under conditions of language contact leading to shift, be particularly susceptible to reduction and loss, this paper looks at four grammatical

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�06

116

chapter 4

features of the Scottish Gaelic spoken by bilingual fisherfolk and their descendants in the village of Embo, on the east coast of the county of Sutherland in Highland Scotland. All four grammatical features were showing some change in progress across a 55-year age-and-proficiency continuum at the time when I first began to work with this receding dialect in the early 1960s, and I have tracked their ongoing use among a dwindling number of bilinguals ever since.1 While two of the four features had at least rough parallels in English grammar, two did not, and the question raised here is whether the features with no parallel in English grammar (unmatched structures hereafter) will show swifter and more extensive loss in receding Gaelic than the features which have an English parallel (matched structures hereafter). 2

The Fisherfolk Gaelic of East Sutherland

The fishing villages of east-coast Sutherland came into being around the beginning of the 19th century. At that time great landlords all over the Highlands, pursuing a new source of cash income, summarily removed their Gaelicspeaking tenants in order to create sheep farms. In East Sutherland the inland evictees were resettled at the coast under conditions that required them to take up fishing as a livelihood: too little land was allotted to make agricultural self-support possible, hooks and lines were distributed, and in at least one case Scots-speaking fisherfolk from the Moray coast well to the south were settled among the evictees to model fishing skills (Dorian 1981: 29–37). During the next century and a half the evictees’ descendants became skilled fisherfolk, but their livelihood was a dangerous and precarious one offering only intermittent financial reward. They had arrived at the coast as desperately poor evictees, and their low-income livelihood guaranteed continuing poverty; as a result the fisherfolk population suffered a considerable degree of social stigmatization. Living in a few densely populated streets beside the sea, all following a single occupation and marrying among themselves, the East Sutherland fisherfolk retained their Gaelic speech even after most other occupational groups in the region had made a gradual transition to English. In 1963 there were about 105 adults in Embo village who spoke Gaelic, but there were no 1 I worked with Embo speakers on site between 1963 and 1978; after a decade-long hiatus caused by health problems, I resumed work with surviving sources in 1991 by tape-recording and long-distance telephone. (Phone conversations were recorded with the permission of the other party.) In several cases during the late 1980s and the 1990s direct in-person work was also made possible by visits paid me in the U.S.

negative borrowing in a shift to the dominant language

117

longer any monolingual Gaelic speakers and there were also no children who spoke the language. The population of Embo was roughly 250 at the time, and active bilinguals constituted at least 40–50% of the adult population;2 among the older people Gaelic was universal. For most of the bilinguals Gaelic had been their first language, used in the family on a regular basis while they were growing up and for the most part also in everyday village life; for at least a few of the oldest remaining speakers Gaelic was their dominant language as well. All Gaelic speakers were fluent in English. The youngest speakers who still made use of Gaelic in daily life did so chiefly vertically, with older members of their kin networks, rather than horizontally among themselves, and English was the dominant language for these speakers. East Sutherland Gaelic was a dialect of the extreme northeastern periphery, sharply different in many respects from more mainstream dialects of the western Highlands. The local dialect, not taught in the schools or used in writing, was unaffected by any standardization processes. Passive literacy in a standardized Biblical form of Gaelic was to be met with among a good many men who had been trained to precent (line out for congregational singing) the metrical version of the psalms, but the differences between this archaic written language and the actual spoken Gaelic of the village were too great to produce any significant transfer of literacy or to provide any grammatical modeling for local speakers. 3

Unmatched East Sutherland Gaelic Structures

Among the grammatical structures that were very clearly showing change in progress in Embo Gaelic were locational adverbs and the vocative case, two prominent elements of Gaelic grammar without a parallel in English. A brief sketch of each structure follows. Locational adverbs were originally paired with a set of directional adverbs. Locational ‘out’ (‘He’s out’) was /(ə) mwĩ(ç)/, directional ‘out’ (‘He went out’) was /(ə) max/, locational ‘up’ was /hurəd/, directional ‘up’ was /(ə)n ɔ:rd/, and so forth. But in Embo Gaelic the distinction between locational and directional ‘in’ had been lost at some earlier point, and the distinction between the

2 Because census figures are entered by parish, no official population figures exist for Embo village as such. In response to a request for a population estimate for Embo, the General Register Office for Scotland estimated that the village had a population of 260 at the time of the national census in 1971.

118

chapter 4

l­ocational and directional members of other adverbial pairs was weakening, with the directional forms increasingly being used in both environments.3 Vocative case was traditionally marked in Embo Gaelic by lenition of the initial consonant of the noun,4 lenition being a morphophonological process that replaces initial stops and affricates with fricatives and initial fricatives with other fricatives or zero. (A vocative particle, seldom present in the surface structure, was the trigger for the lenition.) Conservative Gaelic speakers used initial-consonant vocative marking routinely with such high-frequency forms of address as personal names and endearments, but they also applied the vocative easily and freely to inanimate objects (e.g. ‘Chair, get out of my way!’). In each of these cases elimination of the traditional Gaelic structure would render Embo Gaelic more like English, since adverbs are not distinguished by form according to whether they indicate location or direction in English and direct address is marked in English only by means of suprasegmental features.5 In the tables presented here, Embo Gaelic speakers are assigned numbers by age, with the oldest speaker whom I worked with designated E1 and the youngest E42; only speakers who provided data relevant to the structure under consideration are entered in the tables. Ages are normed to 1970, roughly the mid-point of my on-site fieldwork; the ages of speakers who had died by 1970 are also normed to 1970, but the age is entered within square brackets. Four categories of speakers are recognized in terms of general proficiency levels. .

a.

Those of my Embo sources aged 64 and above were regarded in the community as skilled senior speakers, notable both for the lexical richness of their Gaelic and for their command of local knowledge (e.g. mastery

3 Expansion of the directional adverbs into additional environments was probably encouraged by the fact that verb phrases involving any of the adverbs in question (e.g. ‘die out’, ‘tie up’, ‘shut down’) invariably used the directional form. 4 In a few cases there was in addition a change in the final consonant of a noun in the vocative case, and in still fewer a vocative suffix appeared; these word-final features were very seldom met with among younger speakers and were sporadic even among older speakers. 5 The fact that initial consonant mutation is itself unmatched as a grammatical device in English does not appear to be a significant factor in the weakening of vocative marking in Embo Gaelic. Lenition alone marks the independent past tense for most verbs as well as the vocative case for most nouns, and no comparable weakening of independent past-tense marking occurs across the age-and-proficiency continuum. Even the obligatory but noninformational lenitions that traditionally appear after the numeral /ta:/ ‘two’ and the adverbs /kle:/ ‘very’ and /rɔ/ ‘too’ are retained to a surprising extent among the youngest fluent speakers and the less than fully fluent speakers (semi-speakers).

negative borrowing in a shift to the dominant language

b.

c.

d.

119

of complex kinship ties); they are designated older fluent speakers (OFSs) here. A somewhat younger group, the individuals aged 41 through 60, were (or had once been) fully fluent and were not distinguished within the community from their elders; those of them who had remained fully fluent will be referred to here as younger fluent speakers (YFSs) in recognition of their relative chronological status within the age-and-proficiency continuum. Those speakers next to whose coding an asterisk appears constitute a special proficiency group, the formerly fluent speakers (FFSs). Purely by age they fall into the younger-fluent-speaker group, and like others of the same age they had grown up in wholly Gaelic-speaking families. (The vigorous use of Gaelic in their families of origin was clear from the fact that all of them had fully fluent younger siblings in the village.) But the FFSs, all married to English speakers and in most cases domiciled in Englishspeaking environments for long periods, had gone many years without using the language regularly before they became part of this study’s speaker sample. Their Gaelic was distinctly rusty compared to the Gaelic of their agemates and their younger siblings, and they are therefore recognized as formerly, rather than currently, fluent. The youngest sources in the sample were aged 40 or below and were considered by fully fluent members of the community to make “mistakes” in their Gaelic. They will be referred to here as semi-speakers (SSs).

One striking difference in the degree to which Embo speakers used Gaelic regularly can be recognized across these groupings. Apart from the formerly fluent individuals, all OFSs and YFSs had some conversation partners with whom they used Gaelic as their routine medium of daily communication. This was not true of any of the SSs, nor of any of the FFSs, partly because of a less complete command of Gaelic (both groups, but especially the SSs), partly because of their life circumstances (FFSs), and in the case of the SSs also because of the degree of shift which the community had reached by the time of their birth and childhood. Table 1 presents the freely spoken and elicited data from speakers across the age-and-proficiency continuum for traditional use of locational adverbial forms and for non-traditional use of directional adverb forms to indicate location. Table 2 presents the number (freely spoken and elicited data combined) and percentage of traditional vs. non-traditional forms used by individual speakers in environments traditionally calling for locational adverbs.

120 Table 4.1

chapter 4 Locational-adverb forms used traditionally in environments expressing location vs. directionaladverb forms used innovatively in environments expressing location. No shading = OFSs, light shading = YFSs, dark shading = SSs. Fr = freely spoken, el = elicited. The number entered in brackets represents forms used solely in direct repetition of a non-local speaker and is not counted in reckoning percentages in Table 2. locational adverbs used traditionally to express location

speaker

mwĩ (ç) ‘out’ fr el

E3 [85] m. E4 82 m. 9 E6 75 m. 3 E7 ?74 f. E8 ?71 f. E9 70 f. 2 E10 70 m. 1 E13 [67] f. 3 1 E14 65 f. 2 E15 65 m. 2 E17 64 f. E20 58 f. 2 E22 58 f. 9 1 E23 57 m. E24 57 f. 5 E26* 54 f. E27 54 m. 10 10 E28 51 m. E29* 50 f. 1 2 E30 49 m. 2 E32* 47 m. E33 46 m. E34 45 f. E37 41 m. E38 40 f. E39 38 f. 2 2 E40 36 f. [3] E41 31 f. E42 30 f.

hurəd ‘up’ fr el

2 15

stã:n ‘down’ fr el

1 28 1 1

1

directional adverbs used innovatively to express location

hau:ɫ ‘over’ fr el

4

1

1 1

18

1

4

8

2

1

3

1

2

1

1

3 1 5

1 6

16

3

1 1

2

1 2

vã:n ‘down’ fr el

nũ:ɫ ‘over’ fr el

1

2

1 1 1

1

19 2 1

3

nɔ:rd ‘up’ fr el

2

1

2

max ‘out’ fr el

2

1

16 3 23

1

5

3 1 2 17 16 4 6 1 1 3 43 11 77 11 19

2

3 4 1 2 26

1

10

4

5 2 6 6 4 3 6 30 1 4 8

2 10 3 1 12

2

1 1

39

5 2

4

121

negative borrowing in a shift to the dominant language

Table 4.2 Number and percentage of traditional locational-adverb forms used in environments expressing location ( freely spoken and elicited instances summed), versus number and percentage of instances of directional-adverb forms used traditionally in environments expressing location ( freely spoken and elicited instances summed). Percentages 90 and above boldfaced. No shading = OFSs, light shading = YFSs, dark shading = SSs. traditional locational forms # %

speaker

E3 E4 E6 E7 E8 E9 E10 E13 E14 E15 E17 E20 E22 E23 E24 E26* E27 E28 E29* E30 E32* E33 E34 E37 E38 E39 E40 E41 E42

[85] m. 82 m. 75 m. ?74 f. ?71 f. 70 f. 70 m. [67] f. 65 f. 65 m. 64 f. 58 f. 58 f. 57 m. 57 f. 54 f. 54 m. 51 m. 50 f. 49 m. 47 m. 46 m. 45 f. 41 m. 40 f. 38 f. 36 f. 31 f. 30 f.

3 57 4 1 2 2 3 7 2 2 31 2 21

100 96.6 100

6

85.7

48

92.3

7 4 6 1 2 3 8 4

3.6 100 66.7

1

3.2

100 100 100 87.5 100 100 47 100 95.5

13.3 14.3 33.3 44.4

non-traditional directional forms # %

2

3.4

1

12.5

35

53

1 3 1 61 4 1 189

4.5 100 14.3 100 7.7

3

33.3

13 18 16 5 111 1 30

86.2 85.7 66.7 55.6 100

96.4

96.8

122

chapter 4

Table 3 presents the freely spoken and elicited data from speakers across the age-and-proficiency continuum for traditional vocative-case marking and for absence of traditional vocative-case marking, giving the total number and percentage of such forms for each speaker. Table 4.3 Number and percentage of instances of traditional vocative forms with initialmutation marking vs. number and percentage of instances of unmarked forms used in direct address. No shading = OFSs, light shading = YFSs, dark shading = SSs. Fr = free, el = elicited, tot = total. Percentages 90 and above boldfaced.

speakers

E4 82 m. E6 75 m. E7 ?74 f. E9 70 f. E10 70 m. E13 [67] f. E16 64 m. E17 64 f. E21 58 f. E22 58 f. E23 57 m. E24 57 f. E26* 54 f. E27 54 m. E29* 50 f. E30 49 m. E31 49 m. E33 46 m. E34 45 f. E35 43 f. E37 41 m. E38 40 f. E40 36 f. E41 31 f. E42 30 f.

trad. vocative marking fr el tot %

1 4 16 1 14 2 13 4 40

2 1 7 1 9

1 6 4 11 12 2 3 14 7 7 20 1 43 49

3 9 2 2 7 12 3

1 6 5 11 16 18 4 28 9 20 20 5 83 49 2 3 1 16 2 3 7 21 3

85.7 100 100 100 100 100 100 90 95.2 100 100 92.2 81.7 100 100 66.7 66.7 23.1 35 19.8 30

no vocative marking fr el tot %

1

7

5 4 2 57

1

1

14.3

1

1 1

10 4.8

7 11 5

7 11 12

7.8 18.3 100

3 1 6 11 28 7 11

8 1 10 13 85 7 11

33.3 33.3 76.9 65 80.2 70 100

negative borrowing in a shift to the dominant language

123

For both the locational adverbs and the vocative case, it is clear that the traditional structures are best preserved by the older speakers and less well preserved by the youngest speakers (which of course is what marks them as cases of change in progress), but on the whole vocative-case marking persists better than does use of the traditional locational adverbs. OFS E17, fewer than half of whose locational adverbial forms were traditional, produced traditional vocative marking at 100%; FFS E26*, who showed no retention of locational adverbial forms at all, produced more than 90% traditionally marked vocatives. Speakers as young as E34 and E35 retained 66.7% vocative marking, whereas in the same age-range E34 and E37 used locational adverb forms in only 33.3% of the environments that traditionally called for them. Ten of 23 speakers (43.5%) used 100% traditionally marked vocative structures, while only nine of 25 (36%) used 100% traditional adverb forms in locational-adverb environments. 4

Matched East Sutherland Gaelic Structures

Turning now to Embo Gaelic grammatical structures showing change in progress but matched by a generally parallel structure in English, we find the negative imperative and the negated past tense of the (irregular) verb ‘to be’ expressed both by traditional forms and by non-traditional forms. The means used to express the structures in question differ considerably in the two languages, naturally enough, but each language offers constructions which express the same sort of distinction. The negative imperative in Embo Gaelic took one of two forms in the usage of the oldest speakers: the negative particle /(n)a/ followed by an unlenited gerund, or the negative particle /(x)a/ followed by a lenited gerund. Since both /(n)a/ and /(x)a/ frequently appeared as /a/, the presence or absence of lenition was often the only indication of which particle was being used. Gaining ground across the age-and-proficiency continuum was an alternative negative imperative structure in which /a čhe: ǰ/ ‘don’t go’ was used as an invariant “dummy” negative imperative, followed by the lenited gerund (or, in the usage of some of the youngest speakers, by the lenited verb root). These had become the sole form of the negative imperative for a few young speakers.6

6 This development does not appear to reflect English influence. Negative imperative expressions with ‘don’t go’ + gerund (e.g., ‘don’t go making trouble’), moderately common in American English, are not in use in any form of English to be heard in eastern Highlands (or anywhere else in Scotland, so far as I’m aware).

124

chapter 4

The traditional negated past tense of the irregular verb ‘to be’ differs in Embo Gaelic from the negated past tense of all other verbs in not showing the preterite particle /tə/ in the surface structure. In the negated past tense most verbs have the structure /(x)a/ + /tə/ + Lverb root (negative particle + preterite particle + lenited verb root); apart from ‘to be’, even the few that have suppletive past-tense forms still show the preterite particle. For ‘wasn’t/weren’t’ the equivalent structure is /(x)a/ + /rɔ/ (negative particle + suppletive past-tense form of ‘to be’); no preterite particle appears. Although this is the traditional structure, the prevalence of the preterite particle /tə/ with all other verbs, and even in certain other grammatical structures involving /rɔ/, has led for some speakers to the appearance of /tə/ in the negated past tense of ‘to be’ as well. Tables 4 and 5 present the data for traditional and non-traditional forms of the negative imperative and of the negated past tense of the irregular verb ‘to be’.7 As before, better preservation of traditional structures by older speakers than by younger speakers marks these as instances of change in progress. And again, as in the case of the two unmatched structures, one of the grammatical structures is retained in traditional form somewhat better than the other: 13 of 28 speakers, or 46.4%, used 100% traditional forms of the negated past tense of ‘to be’, as compared with eight of 21 (38.1%) who used 100% traditional Embo forms of the negative imperative. Speakers as young as E37 and E38 used 100% traditional negated past-tense forms of ‘to be’, whereas E31 was the youngest speaker to use 100% traditional forms of the negative imperative. Table 6 presents percentages for traditional versus non-traditional forms of all four structures, unmatched and matched. 5

Embo Age-and-Proficiency Continuum Results

It must be stressed that the use of traditional structures is by no means to be equated with such capacities as fluency or ease of self-expression on a speaker’s part. Speakers E3 through E37 – excepting only the four FFSs – all spoke with full fluency and expressed themselves with normal ease in Embo Gaelic. Among the FFSs E30* was ultimately the most fluent (her fluency increasing after her return to Embo residence in the 1990s), but she was conspicuously non-traditional in the forms she produced for these structures. E6, the least traditional among the OFSs who provided all four of these structures, was an 7 Negative imperatives using /a/ followed by an unlenitable consonant were discounted for Table 4 purposes, since they could not be assigned either to /(n)a/ or to /(x)a/.

125

negative borrowing in a shift to the dominant language

Table 4.4 Number and percentage of traditional negative imperative constructions with /(n)a/ + unlenited gerund or /(x)a/ + lenited gerund vs. number and percentage of non-traditional negative imperative constructions with /a čhe:ǰ/ + lenited gerund or /a čhe:ǰ/ + lenited root. No shading = OFSs, light shading = YFSs, dark shading = SSs. Fr = free, el = elicited, tot = total. Percentages 90 and above boldfaced. trad. negative imperative with /(n)a/ or /(x)a/ /(n)a/ + gerund Speakers

E4 82 m. E6 75 m. E9 70 f. E10 70 m. E13 [67] f. E17 64 f. E21 E22 E23 E24 E26* E27 E29* E30 E31 E33 E34 E35 E37 E38 E39 E40 E41 E42

58 f. 58 f. 57 m. 57 f. 54 f. 54 m. 50 f. 49 m. 49 m. 46 m. 45 f. 43 f. 41 m. 40 f. 38 f. 36 f. 31 f. 30 f.

fr

4 2

3 3 1

el

2 2 1 8 6 9 6 1 1 2 9 5

1

1

2

1

/(x)a/ + lenited gerund fr el

%

2 5 5 23 6 18

100 83.3 100 100 100 94.7 100 25 100 78.6 100 37.8

2

17 2 2 22 17 14 1 2

9 1

11 1

78.6 50

2

4 2

10.3 50

4

66.7

2 2

5

6

1

1

+ lenited + lenited gerund root tot

3 4 9

11 1 1 11 8 6

3

non-trad. negative imperative with /a čhe:ǰ/

fr

el

fr

el

1

1

tot

%

1

16.7

1

5.3

1

1

6

6

75

1

5

6

21.4

5

18

23

62.2

1 3

1 3 1 5 35 2 45 6 2

21.4 50 100 89.7 50 100 100 33.3

100

5 1

4 22 2 38 4 1

1 1 13 2 2

126

chapter 4

Table 4.5 Number and percentage of conservative (irregular) negative past tense instances of the verb ‘to be’ without preterite particle vs. number and percentage of innovative (regular) negative past tense instances with preterite particle. No shading = OFSs, light shading = YFSs, dark shading = SSs. Fr = freely spoken, el = elicited, tot = total. Percentages 90 and above boldfaced.

speakers E3 [85] m. E4 82 m. E5 77 f. E6 75 m. E7 ?74 f. E9 70 f. E10 70 m. E13 [67] f. E14 65 f. E15 65 m. E17 64 f. E18* 60 f. E20 E21 E22 E23 E24 E25 E26* E27 E29* E30 E31 E32* E33 E34 E36 E37 E38 E39 E40 E41 E42

58 f. 58 f. 58 f. 57 m. 57 f. 56 m. 54 f. 54 m. 50 f. 50 m. 48 m. 47 m. 46 m. 45 f. 41 m. 41 m. 40 f. 38 f. 36 f. 31 f. 30 f.

traditional absence of preterite part.: /(x)a rɔ/ fr el tot % 3 49

3 55

100 75.3

2 10 5 3 2 23

2 4 2 15 18 3 2 464

25 100 100 100 94.7 100 22.2 99.1

8 4 14

4 8 4 26

6 1 2

4 5 13

441 4

12

9

10

76.9

88 88 17 29

36 28 3 2 1

124 116 20 31 1 12 1 12 1 12 87 22

93.9 83.6 7.3 100

8 1 22 8

1 4 1 11 65 14

18 2

18 4

1

6

24.7 75

1

5.3

7

7 4

77.8 .9

7 18 213

1 3 1 1 5 40

1 3 1 8 23 253

24

4 14

4 38

80 76

4 79 3 17

4 167 3 144

12 100 100 100

4

100 100 100 100

1

12

innovative introduction of preterite part.: /(x)a t rɔ/ fr el tot %

100 20 24 100 100 88

88 127

23.1 6.1 16.5 92.7

127

negative borrowing in a shift to the dominant language

Table 4.6 Percentages of traditional structures (left) and non-traditional structures (right) used by Embo speakers across the age-and-proficiency continuum for two Gaelic grammatical constructions without parallels in English grammar (columns 1, 2, 5, 6) and two Gaelic grammatical constructions with English parallels (columns 3, 4, 7, 8). No shading = OFSs, light shading = YFSs, dark shading = SSs. Percentages 90 and above boldfaced.

speaker

1 locatnl adv

2 voc

3 4 neg imper neg past ‘be’ w/o part

5 locatnl adv

6 voc

7 8 neg imper neg past ‘be’ w. part

trad.

trad.

trad.

trad.

nontrad.

nontrad.

nontrad.

nontrad.

%

%

%

%

%

%

%

%

E3 [85] m. E4 82 m. E6 75 m. E7 ?74 f. E8 ?71 f. E9 70 f. E10 70 m. E13 [67] f. E14 65 f. E15 65 m. E16 64 m. E17 64 f. E18* 60 f.

100 96.6 100

E20 E21 E22 E23 E24 E26* E27 E29* E30 E31 E32* E33 E34 E35

100

58 f. 58 f. 58 f. 57 m. 57 f. 54 f. 54 m. 50 f. 49 m. 49 m. 47 m. 46 m. 45 f. 43 f.

100 100 100 87.5 100 100 47

95.5 85.7 92.3 3.6 100

85.7 100 100 100 100

100 100

90 95.2 100 100 92.2 81.7 100 100

100 83.3

100 100 100

94.7

100 25 100 78.6 100 37.8

66.7 66.7

100 100 94.7 100 22.2 99.1

100

100 100 100 76.9 93.9 83.6 7.3 100

3.4 14.3

16.7

12.5

24.7 75

5.3 77.8

53

4.5 100 14.3 100 7.7 96.4

5.3

.9

10 4.8 75 7.8 18.3 100

21.4 62.2

23.1 6.1 16.5 92.7

100

66.7 13.3

100 75.3 25 100

78.6 50

100 20 24

33.3 86.2

33.3 33.3

21.4 50

80 76

(Continued)

128

chapter 4

Table 4.6 (Continued) 1 locatnl adv

2 voc

3 4 neg imper neg past ‘be’ w/o part

5 locatnl adv

6 voc

7 8 neg imper neg past ‘be’ w. part

trad.

trad.

trad.

trad.

nontrad.

nontrad.

nontrad.

nontrad.

speaker

%

%

%

%

%

%

%

%

E37 E38 E39 E40 E41 E42

14.3

23.1

100

85.7

76.9

100

41 m. 40 f. 38 f. 36 f. 31 f. 30 f.

33.3 44.4

35

10.3 50

19.8 30 3.2

66.7

100 88

66.7 55.6 100 96.8

65 80.2 70 100

89.7 50 100 100 33.3

12 100 100 100

accomplished Gaelic speaker and a ready conversationalist. Among the SSs the two who spoke most comfortably in Gaelic were E39 and E40, but the degree to which they used traditional forms for these structures was sharply different: E39 was the most traditional and E40 the least so. The community itself, furthermore, had no experience of Gaelic standardization and was not in the least puristic about these structures. With just one exception, neither endorsement of the traditional forms nor disapproval of the non-traditional forms was ever expressed by speech community members over the 40-year period of this study. The single exception was a remarkable 1995 conversation with E29* during which she commented extensively on the use of personal names in direct address. But even though she was explicit (by way of examples of what one would and wouldn’t say) about the need for lenited initial consonants in direct address, she expressed this wonderingly, as a curious property on the part of her native language, rather than as a “rule” which some people “violated” or were “wrong” about. She did not so much as mention having heard other speakers fail to produce the lenition she was endorsing. All four of the grammatical structures under consideration here are commonplace, occurring at high frequency in ordinary conversational interaction. However two of them, the negative imperative and the vocative, did not occur frequently in my freely spoken corpus, even though that corpus accu-

negative borrowing in a shift to the dominant language

129

mulated (especially for the speakers who survived longest) to a considerable size over the more than 40-year period of work with this speech variety.8 The greater part of the vocatives that I recorded in spoken material, like most of the negative imperatives, appeared as quotations within reported speech; elicitation was relied upon for additional examples. Environments calling for locational adverbs and for the negated past tense of ‘to be’ are by contrast so extremely frequent that most sources who provided any freely spoken material at all provided instances of these structures, though additional instances were also elicited. The high frequency of occurrence of all four of these structures was broadly useful, since it led either to ample freely spoken instances or to unproblematical translation-task results, all speakers being thoroughly familiar with the grammatical environments concerned. There are various ways of looking at the data presented in Table 6 so as to compare the expression in obsolescent Embo Gaelic of the unmatched and matched structures. One approach might consist simply of counting the number of cases in which speakers’ retention of the traditional grammatical structure is above 50%. For the two unmatched structures, percentages could be reckoned in 48 instances; of these, 32, or 66.7%, were scores of better than 50%. For the matched structures the equivalent figures are 33 better-than-50% scores out of 49 instances, or 67%. By this reckoning, retention is higher by only the barest margin in the two matched cases as compared with the two unmatched cases. Another approach is comparison of the number of speakers who no longer make any use at all of the various traditional structures. Three speakers showed no use of locational adverbs and two no use of the vocative (unmatched structures); three showed no use of the traditional negative imperative structures and three showed no negated past tense of the verb ‘to be’ without the preterite particle (matched structures). In nearly all cases the speakers who produced no traditional forms were FFSs or SSs, but one fully fluent YFS (E23) does appear among those who used no traditional locational-adverb forms, suggesting a slightly greater susceptibility to loss for that unmatched structure. It must be noted, however, that semi-speaker data taken separately do not support a notion of greater susceptibility to loss for unmatched structures. In the eight instances in which SSs provided data for unmatched structures, there were two instances of zero use of traditional forms (25%), whereas in the ten instances in which they provided data for matched structures there were five instances of zero use of traditional forms (50%). The negative borrowing principle would have predicted contrariwise a higher incidence of completely 8 There are a great many instances of my own name used in direct address in the corpus, but because it begins with an unlenitable consonant this yielded no marked vocatives.

130

chapter 4

absent traditional forms in the structures without a match in their dominant language. Yet another approach might be to compare, speaker by speaker in the case of those who provided examples of both sorts of structures, the averaged percentage of traditional expression of unmatched structures with that of the averaged percentage of traditional expression of matched structures. E6, for example, produced two unmatched structures at a total of 185.7% and two matched structures at a total of 108.3%, which gives an unmatched-structure average of 92.9% and a matched-structure average of 54.2%. Allowing any single-instance percentage to stand as the average for that kind of structure, the results of this comparison are as follows (number of structures totaled given in parentheses; unm. = unmatched, m. = matched, higher averages for matched structures boldfaced, higher averages for unmatched structures boldfaced and italicized): OFSs unm.

m.

YFSs

unm.

m.

FFSs

E3

(1) 100 (2) 87.7 (2) 54.2 (1) 100 (2) 100 (2) 100 (2) 97.4 (1) 100 (1) 22.2 (2) 96.9

E20

(1) 100 (1) 90 (2) 95.4 (1) 100 (2) 92.9 (2) 87 (2) 100 (1) 100 (2) 40 (1) 66.7 (2) 18.7

(1) 100 (1) 100 (2) 100 (1) 25 (2) 88.5 (2) 91.8 (1) 100 (1) 100 (2) 51.3 (1) 50 (2) 50

E26* (2) 46.1 E29* (2) 1.8 E32* (1) 66.7

E4 E6 E7 E9 E10 E13 E14 E15 E17

(1) 100 (1) 96.6 (2) 92.9 (1) 100 (2) 100 (2) 100 (2) 93.8 (1) 100 (1) 100 (2) 73.5

E21 E22 E23 E24 E27 E30 E31 E34 E35 E37

unm. m.

(2) 86.3 (2) 22.6 (1) 100

SSs

unm.

m.

E38

(2) 34.2 (1) 44.4 (2) 9.9 (1) 30 (2) 1.6

(2) 55.2 (2) 69 (2) 0 (2) 0 (2) 33.4

E39 E40 E41 E42

negative borrowing in a shift to the dominant language

131

The ultimate interest in this comparison lies in how often the average percentage of traditional expression of matched structures is higher than the average percentage of traditional expression of unmatched structures, since some such imbalance is what the notion of negative borrowing would predict. For five OFSs and three YFSs there is no difference, because they preserved 100% traditional expression of both kinds of structure, either in one or in both of the instances representing the category. For two OFSs (E13 and E17), for five YFSs (E21, E22, E27, E34, and E37), for all three FFSs, and for three of the five SSs (E38, E39, and E42), the difference was in the predicted direction, with higher average percentage of traditional expression for the matched structures. Contrary to the prediction, however, the reverse was true for three OFSs (E4, E6, and E15), for three YFSs (E23, E24, and E35), and for two SSs (E40 and E41): they showed higher average percentages for traditional unmatched-structure expression. For the two OFSs and one of the YFSs (E24), introduction of the preterite particle into the negated past tense of ‘to be’ was the non-traditional feature that caused the discrepancy in favor of unmatched-structure expression, since they preserved higher rates of all the other traditional structures. For the other YFS (E35) better retention of traditional vocative marking than of traditional negative imperative marking was the source of the discrepancy. For both of the SSs use of some traditional vocative marking also produced the discrepancy, since none of their other structures showed any traditional forms at all. In terms of the overall picture it appears that the formerly fluent speakers are particularly susceptible to negative borrowing, since all three of them showed considerably greater use of non-traditional forms in the unmatched structures. This is a reasonable enough outcome, in view of the fact that they spoke mainly English during most of their adult lives. OFSs appear more resistant to negative borrowing than YFSs, with only two out of 10 OFSs showing the predicted imbalance in favor of matched structures, whereas five out of 11 YFSs conformed to the prediction. This again is an unsurprising outcome, given the greater role of Gaelic in the lives of OFSs, during whose childhoods Gaelic was unrivaled as the dominant language of Embo village. The chief surprise comes with the SSs, for two of whom the residual tenacity of traditional vocative marking produces results contrary to the prediction entailed by the negative borrowing concept. 6

General Discussion

The Embo village data offer only rather modest support for negative borrowing as a language-contact phenomenon in obsolescence. This is noteworthy

132

chapter 4

in view of the fact that three of the changes in progress produce greater regularity and therefore one sort of greater simplicity. Eliminating the distinctive locational adverb forms would result in a single set of adverbs used both as independent adverbial forms and as constituents in verb phrases (see note 3). Introducing exclusive use of /a čhe:ǰ/ as an invariant dummy negative imperative would eliminate a choice between two negative particles with different effects, one producing no mutation and one requiring lenition. Introducing the preterite particle into the negated past tense of ‘to be’ would eliminate the distinction between that verb and all others in one very high-frequency environment.9 As for the fourth case, eliminating vocative marking would reduce by one the number of grammatical categories requiring special segmental marking in the local Gaelic. Direct address would then, as in spoken English, be signaled only by suprasegmental features. The cases of the negative imperative and the vocative pose an interesting challenge to the negative borrowing principle. In the (matched) negative imperative case, nine of 21 speakers (43%) used conservative forms at a 90–100% level, while in the (unmatched) vocative case, a higher proportion (13 of 23 speakers, or 56.5%) used conservative structures at that level. At the other end of the retention scale, six of 21 speakers (28.6%) used conservative forms at a level below 50% in the (matched) negative imperative case, while a smaller proportion (five of 23 speakers, or 21.7%) used conservative forms of the (unmatched) vocative at that low level. That is, not only did a greater proportion of the speakers perform at a very high conservative level in the unmatched case, but a greater proportion performed at a very low conservative level in the matched case. In each case, particles are central to the construction, and yet it is the unmatched structure – in which furthermore the particle is not usually present in the surface structure – that survives better at the low end of the age-and-proficiency continuum. Part of the explanation may lie in the fact that only a single particle requiring lenition is involved in the conservative vocative structure, while a choice between two particles, one producing no mutation and the other requiring lenition, is involved in the conservative negative-imperative structure. Non-structural factors may also be at work in the unusually strong persistence of the vocative in this speech variety. Women in East Sutherland use both endearments and their interlocutor’s given name 9 An alternative route to regularization is exemplified in this speaker sample by E38: she regularized all the forms of ‘to be’ that involved /rɔ/ by eliminating the preterite particle /tə/ from every one of them, including from structures that routinely showed /tə/ for nearly all Embo speakers, e.g. after conjunctions or particles that require the initial mutation known as nasalization.

negative borrowing in a shift to the dominant language

133

far more frequently during conversational exchanges than men do, and all of the semi-speakers in the sample here are female. The relatively high frequency of marked vocatives in the speech of older women with whom they frequently interacted may have affected the retention rate among the youngest women in the sample. As Mougeon and Beniak (1991), Silva-Corvalán (1994), and Jones (2005) discuss, negative borrowing is not the only effect at work in cases of contact and attrition, since internal system pressures are likely to play a role. The better preservation of the vocative structure among the youngest female speakers of Embo Gaelic suggests that gender-related speech style may have an unexpected effect. While generalizations about the likely loss of unmatched features (e.g. Andersen 1972: 97) have an inherent plausibility, based on the potentially greater efficiency for the bilingual brain of working with matching structures, it may be that to be effective a match must be structurally closer than is the case in the Gaelic/English parallels looked at here. Perhaps, too, negative borrowing is less characteristic of structural-change processes shared by the entire community at the very end of an indigenous speech form’s existence than of more individual-centered attrition processes such as first-language attrition among immigrants and exiles or second-language attrition among successful learners who later use the learned language infrequently. In whole-community language-contact settings, perhaps it is long periods of regional co-existence that are more likely to eliminate unshared grammatical structures, as in the case of Sprachbund areas. In these cases a good deal of unmatched-feature loss may already have taken place before any late-obsolescence stage. Some of the differences between East Sutherland fisherfolk Gaelic and more mainstream Gaelic dialects suggest that negative borrowing may have operated to this effect over the relatively extended co-existence period prior to the more dramatically obsolescent phase which this study documents, so that by the 1960s East Sutherland Gaelic was already structurally less different from English than many other Gaelic dialects are, even though most local Gaelic speakers were highly proficient at the time and some still Gaelic-dominant. References Andersen, Roger W. 1972. Determining the linguistic attributes of language attrition. In Richard D. Lambert and Barbara F. Freed (eds). The loss of language skills (pp. 83–118). Rowley, MA: Newbury House Publishers.

134

chapter 4

Dorian, Nancy C. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. Fenyvesi, Anna. 1995. Language contact and language death in an immigrant language: The case of Hungarian. University of Pittsburgh Working Papers in Linguistics 3, 1–117. Jones, Mari C. 2005. Transfer and changing linguistic norms in Jersey Norman French. Bilingualism: Language and Cognition 8(2), 159–175. Mougeon, Raymond, and Beniak, Édouard. 1991. Linguistic consequences of language contact and restriction: The case of French in Ontario, Canada. Oxford: Clarendon Press. Sasse, Hans-Jürgen. 1992. Language decay and contact-induced change: Similarities and differences. In Matthias Brenzinger (ed.) Language death: Factual and theoretical explorations with special reference to East Africa (pp. 59–80). Berlin/New York: Mouton de Gruyter. Silva Corvalán, Carmen. 1994. Language contact and change: Spanish in Los Angeles. Oxford: Clarendon Press. Thomason, Sarah G. 2001. Language contact: An introduction. Washington D.C.: Georgetown University Press.

part two Speaker Skills and the Speech Community in a Receding Language Context

∵

chapter 5

The Problem of the Semi-Speaker in Language Death The fieldworker who is investigating a dying language has by definition a limited pool of potential informants. This pool may in fact consist of only one person, or it may number a few hundred. Always it is in the process of contraction, and often the fieldworker has a sense of great urgency in his struggle to record and analyze the language. He must find the best possible informants as quickly as possible and try to exhaust them as sources. In the language death situation, however, there may be cause to question the intactness of the material gathered. In my own fieldwork in a terminally Gaelic-speaking part of Scotland, I discovered considerable differences in the Gaelic of the oldest available fluent speakers and the youngest, the Gaelic of the latter showing reduction and loss in certain areas in comparison with the former (Dorian 1973). I would like to consider here the evaluation problem which faces the investigator in the terminal language community: how is he to gauge the completeness and intactness of the version of the language which he receives from his informants? Oftentimes evaluation is possible by external or internal clues. If there is more than one speaker, the investigator may be able to compare one version with another (Swadesh 1948: 230–31). An isolated last speaker may betray the uncertainty of his productions by the manner of delivery (“pathetically halting”, Krauss 1963–70: 7). These hints are not always available, however. The most difficult case is surely a lone last speaker of some fluency, where there is neither a comparison available nor a markedly deficient manner of delivery. Faced with this situation, Haas (n.d.: 10) made a judgment based on the sociolinguistic probability that the language as represented by a last, isolated speaker was, as she put it, “a mere remnant of what the language must have been when many speakers used it as their only means of communication”. Haas’ assumption here is a common one, namely, that any language which continues to be spoken by only a very few people will exhibit a much reduced form as compared with the same language in vigorous use by a rich linguistic community. Exceptions will certainly be found in those cases where a language dies with extraordinary rapidity and without replacement by some other language; this was the case with Tasmanian and with the Yahi language in California, where the last speakers were monolinguals or near-monolinguals for most or all of

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�07

138

chapter 5

their lives. But on the whole the assumption that the reduced use of a language will lead also to a reduced form of that language seems realistic. Haas’ assumption can best be tested in a terminal language community where a continuum of proficiency is available, from full fluency to the barest skills necessary for communication in the dying language. Such a continuum is available in the coastal East Sutherland area of mainland Scotland where I have worked for 11 years. In a total pool of Gaelic speakers which numbered about 140 in 1972, there were at the upper end of the spectrum a few individuals who were more comfortable and proficient in Gaelic than in English,1 in the middle range many who were skilled bilinguals, fluent in both languages, and at the lower end some who could make themselves understood in imperfect Gaelic but were very much more at home in English. These last I have called “semi-speakers” (Dorian 1973: 417). It is the identification of these semispeakers which constitutes a major problem for the fieldworker dealing with a dying language, since he needs to know how representative and how reliable his data are. If even the youngest fluent speakers showed notable grammatical change in their Gaelic as compared with the oldest fluent speakers, then the semi-speakers would presumably show still more radical departures from the conservative norm, and data recovered from semi-speakers would need to be handled with caution in the writing of grammars or in the reconstruction techniques of historical linguistics. One of the questions that is of interest to the investigator in the field is whether the community’s own judgments of proficiency have any basis in fact: is there any significant difference between the usage of the youngest of those who have a reputation for fluency and that of the eldest of those whose skills are little thought of? Can the community itself evaluate linguistically to the extent of identifying the semi-speaker whose language is “reduced”? Bloomfield’s experience with the Menomini (Bloomfield 1927) shows that such judgments are made even in illiterate societies, but the question still remains as to how finely and accurately they may be made. The smallest and most strongly Gaelic speaking of the East Sutherland villages, Embo (population ca. 275), provides an excellent test case. The oldest children of the skilled septuagenarian speaker B.R. are all considered fluent speakers, down to and including the fourth, the 45-year-old son A.R. But the two daughters who are next in line, J.R. and W.R., are considered less than fluent, even though the elder of the two, J.R., is only one year younger than A.R.2 Since A.R. habitually speaks Gaelic by preference whenever circumstances permit (that is, when he has a 1 There were no Gaelic monolinguals, however, and had been none for perhaps 40 or 50 years. 2 All three of these children are unmarried and live in the mother’s household.

the problem of the semi-speaker in language death

139

Gaelic-speaking conversation partner), while J.R. much more often chooses to speak English, it would be quite possible that the community was responding more to A.R.’s greater language loyalty than to his greater proficiency when judging him fluent in comparison to his sister. To test the Gaelic skills of A.R. and J.R. and other speakers in the proficiency continuum, the same 115 English sentences were presented for translation into Gaelic to 16 speakers of varying ages. The sentences were chosen to contain a fair sampling of the grammatical signals of East Sutherland Gaelic and of the obligatory morphophonemic phenomena of the dialect. The speakers questioned included one octogenarian, 3 septuagenarians (including B.R.), 3 speakers in their 50’s, and the 45-year-old A.R., among those generally ajudged fluent speakers. The 7 supposedly less-than-fluent speakers were all women.3 They ranged in age from their early 60’s to their early 30’s, according to the stage the language has reached in their village on its passage to extinction: in Embo an individual in his 40’s can still be a fluent speaker, whereas in Brora, the other village surveyed, the fluent speakers are all in their 70’s and 80’s. One speaker who was an unknown quantity was also included. On the theory that originally fluent speakers who have been away from the home community for many years without opportunity to practice may lose considerable proficiency, a 58-year-old Embo exile was questioned in her home in the Lowlands;4 she left Embo 40 years ago and has returned only for brief holidays since. Her husband was an English monolingual, and she has had very little chance to use Gaelic during her 40-year exile. The results of immediate interest here are those of B.R. and her son and daughter, A.R. and J.R. There are 30 years separating the children from the mother, only one year separating the children from each other. If the community judgment of A.R. as a fluent speaker is accurate, then in some significant respects his performance must be either more like his mother’s than like his sister’s, or at least more like his mother’s than his sister’s is. This is certainly not the case in every test. Neither of the children has retained the vocative plural inflection of the mother, for example; and where the mother showed 100% of the obligatory morphophonemic changes in the vocative, A.R. showed only 17% as compared with J.R.’s 57%.

3 This is fortuitous. I heard about possible male counterparts, but they were either unwilling or unable to serve as informants during the limited time available for the study. 4 One of the fluent speakers in his 50’s is also an exile, but his wife is a Gaelic speaker, too, and Gaelic is the normal language of the home. Their daughter, likewise living in exile, served as one of the imperfect-speaker informants.

140

chapter 5

But the overall results do indicate a pronounced difference between the Gaelic of A.R. and that of his sister J.R. Although neither A.R. nor J.R. used the most conservative form of the passive, A.R.’s passives were superior to his sister’s in that he knew and used both of the available finite verbs with which the passive can be formed and in that there were no constituent elements of the passive missing. J.R. used only one of the finite verbs and was twice missing a necessary preposition in the formation of the passive. A.R.’s choice of word order for pronoun objects agreed with the conservative choice of older speakers like his mother in 6 out of 7 instances; J.R.’s in only 3 out of 7. A.R. controlled two less common conjunctions either not supplied or incorrectly supplied by his sister. A.R. used two of the three available forms of the negative imperative, J.R. only one. In one test A.R. outperformed both his mother and his sister: each of the women twice (out of 8 opportunities) substituted analytically-formed phrases for the usual synthetic forms of the conjugated preposition do “to”,5 whereas A.R. used synthetic forms in all 8 instances. J.R. failed in 3 instances out of 13 to produce an obligatory morphophonemic change in the initial consonant of an adjective, A.R. in only 1 instance (and B.R. in none). Where there was a choice of prepositions, A.R.’s choice coincided with the choice of conservative speakers like his mother 6 out of 7 times, J.R.’s only 3 out of 7 times. The greatest difference between the Gaelic of A.R. and that of J.R., however, comes when the retention of irregularities is considered. Of the 17 irregular noun plurals tested, B.R. retained them all (although she offered one regularized alternative in addition), A.R. 15, and J.R. only 9. Of the 16 irregular verb stems tested, B.R. and A.R. retained them all, J.R. 13. A highly irregular first person singular conditional inflection is missing in the speech of J.R. (and replaced by an analytic construction), but it appears without fail in that of B.R. and A.R. Similarly J.R. regularizes the future by carrying the predominant inflection into the first person singular; neither her mother nor her brother ever does so. If we attempt to analyze the respects in which J.R.’s Gaelic differs from her brother’s, we find the following phenomena each represented more than once: 1) 2) 3)

absence of a stylistic option (negative imperative; passive) substitution of an analytic construction for a synthetic one (conjugating preposition; 1st person conditional) analogical leveling (noun plurals; verb stems; conditional; future)

5 B.R. was the only fluent speaker to do this, but four of the semi-speakers other than J.R. also did so at least once.

the problem of the semi-speaker in language death

141

While I suspect that all of these phenomena are in fact characteristic of languages in decline, only one of them here seems to have significance for the community judgment of proficiency. The first two were displayed not only by the putative semi-speaker J.R., but also by her mother. If we plot the performances of B.R., A.R., and J.R., we find the following: 1) Option 1 – – –

Neg. Imper. Passive Option 2 Option 3 Finite verb 1 Finite verb 2 B.R. 3 – – B.R. 6 A.R. 1 A.R. 2 A.R. 3 A.R. 3 – J.R. 3 – J.R. 6

2) Conjugating Prep. do 1st Pers. Conditional synth. constr. anal. constr. synth. constr. anal. constr. B.R. 6 2 B.R. 5 – A.R. 8 – A.R. 5 – J.R. 6 2 J.R. – 5 It is only in the last case, the case of the analytic treatment of the first person conditional, that we find B.R. and A.R. clearly lined up on the one side of the statistics versus J.R. on the other. As it happens, the use of an analytical construction for the first person singular conditional also constitutes a case of analogical leveling, because all of the other persons of the conditional, both singular and plural, are likewise formed analytically in East Sutherland Gaelic. If we now look at the difference in performance among the three speakers on the measure of analogical leveling as such, the pattern which emerges is striking: 3) Analogically leveled noun plurals (opportunities: 17) B.R. 1 (offered in addition to the irregular form) A.R. 2 J.R. 8 Analogically leveled verb stems (opportunities: 16) B.R. – A.R. – J.R. 3 Analogically leveled 1 sg. conditional (opportunities: 5) B.R. – A.R. – J.R. 5

142

chapter 5

Analogically leveled 1 sg. future (opportunities: 5) B.R. – A.R. – J.R. 5 It should be noted that analogical leveling is not confined in East Sutherland Gaelic to those whom the community designates as less-than-fluent speakers. Analogical leveling in the passive can in fact be shown to be a change in progress throughout the entire Embo community (Dorian 1973). But the sudden dramatic upsurge of analogical leveling in J.R.’s speech, as compared with that of her brother a year older, seems actually to be a defining characteristic of her status as a semi-speaker. This supposition receives confirmation from the fact that it was her excessive analogical leveling in gerund formation (a measure not included in the tests) which was first commented on by the fluent speakers who called my attention to her as an imperfect speaker. A high incidence of analogical leveling is most useful as a criterion for semi-speaker status at the upper range of semi-speaker proficiency; it clearly emerges as the single most prominent difference between the Gaelic of A.R. and that of J.R., who is actually quite a good imperfect speaker in comparison to some of the others. But while some degree of analogical levelling is characteristic of all the semi-speakers in my sample (and also of the exile speaker J.F.), there is simply less left to level in the speech of the more extreme semispeakers. The Brora semispeaker J.M., for example, levels 7 irregular noun plurals, but in addition she has no marked plural form at all for two others; she levels 4 irregular verb stems, but was unable to provide any form for 5 others. And there is no use looking for levelling in the first person of her conditional, because she has lost the grammatical category “conditional” altogether. The future is vestigial in her speech. This is clearly a much more drastic kind of reduction. If J.M. were the last surviving “speaker” of East Sutherland Gaelic, she would, I think pose little problem in identification as a semi-speaker for the fieldworker. That is, there would be no difficulty in evaluating the intactness of her Gaelic. There is simply too much missing in her speech, and the investigator would be inclined to be suspicious about a speaker who could not distinguish between “I sell”, “I will sell”, and “I would sell”.6 Other losses, however, representing phenomena less common in the world’s languages, might be harder to 6 This sort of inference is risky, but combined with other evidence (halting manner of delivery; surprising lexical gaps; and especially the presence of the missing categories in related dialects or languages) may be helpful.

the problem of the semi-speaker in language death

143

spot. The young Embo semi-speaker I.F., for example, has all but lost one entire initial mutation, that is, one of the omnipresent morphophonemic changes common to the languages of the Celtic language family. Were there no other speakers, no other extant Celtic languages, and no written records, the only evidence in I.F.’s speech that the morphophonological system of “nasalization” had ever existed would be the occurrence of one seemingly irregular (but consistent) alternation between the citation form [thε] “house” and its counterpart [dε] after the definite article. Considering the tendency toward analogical leveling which we have already noted in the performance of semi-speakers, the utterly consistent appearance of such an apparent irregularity ought probably to be given great weight in the analysis of a language where the only informants are probable semi-speakers. Certainly the performance of all the semi-speakers in my sample indicates the accuracy of Haas’ assumption that reduction in the use of a language will be matched by reduction in its structure. Some of the kinds of reduction noted for East Sutherland Gaelic may prove to be universally characteristic of dying languages when more evidence is in; I am thinking particularly of loss of entire grammatical categories and of reduction of stylistic options. The latter has already been documented by Hill for Luiseño and Cupeño (Hill 1973). Substitution of analytic for synthetic structures, on the other hand, can of course only occur in languages with polymorphemic word structure. Analogical leveling is again potentially universal, since it may be either morphological or syntactic. Conclusion Evaluation procedures for terminal speakers are commonly possible. Often there will be earlier accounts of the same language to compare with the productions of the terminal speaker (and probable semi-speaker), as there were for Biloxi (Haas 1968: 77) and Luiseño (Hill 1973: 35); or there will be several speakers to compare with each other (Krauss 1963–70; Swadesh 1948: 230–31). Failing that, there is the possibility of the tell-tale irregularity, as with I.F.’s fossilized nasalization; or, on the other hand, the suspicious absence of any irregularity, which suggests vast analogical leveling; or the puzzling absence of an expected grammatical category, such as some provision for expressing the sense of the conditional in contrast with the future. On the basis of my work with terminal East Sutherland Gaelic, I would give a positive answer to the following three questions:

144 1. 2. 3.

chapter 5

Is reduced use of a language accompanied by a reduction in form? (i.e., are there semi-speakers?) Can the more proficient members of the language community realistically pinpoint the onset of that reduction? (i.e., can the community accurately identify the semi-speakers?) Can the investigator hope to evaluate the intactness of the version of a language which he derives from a last few speakers? (i.e., can the investigator spot semi-speaker performance?)

Data from other terminal language communities will be required before we can judge the general validity of these answers. I have already mentioned above that rapidity of extinction can produce a negative answer to the first question.7 This suggests that “reduced use” applies essentially to the individual speaker, not to the community. If the language is used less and less by the community as a whole, yet certain speakers continue to use it almost exclusively, those loyalists will not become semi-speakers. The semi-speakers among whom the language will appear in reduced form are the individuals who themselves use the language less, whether because they have moved out of the community (exiles like J.F.) or because they are the pivotal figures in a local language shift. Answers to the second question may vary in response to the precise kinds of change which are involved in the reduction process. Some may be more “visible” than others; that is, they become linguistic stereotypes (Labov 1970: 73). Analogical leveling proved to be a stereotypical form of reduction in East Sutherland Gaelic, whereas morphophonemic confusions, which are rife in the speech of semi-speakers, seem to produce no comment. In general here we need much more detailed accounts of the reductive features of language death and their salience to the native community.8 Fuller accounts from fieldworkers of their experiences in elicitation from terminal speakers can throw light on the third question. Can we realistically 7 Hill, for example, reports that there seem to be no semi-speakers among the Luiseño and Cupeño: “You either speak fairly well or not at all” (personal communication). In the Cupeño case this may again reflect the speed of the total extinction process; the Luiseño approach to extinction has been somewhat slower, but is still rapid compared to that of East Sutherland Gaelic (Hill 1973: 34). 8 Krauss (1963–70), for example, commonly merely notes that the form is suspect (“inappropriate here”, “distorted”, “confused towards end”, “inconsistent”, “confused”, all from p. 44) and only rarely specifies the feature which provokes his comment (“ti.1-class-mark . . . missing”, p. 116); very occasionally he provides the judgment of a second informant (“Lena rejects this morpheme, or more probably, sporadic allomorphic variant . . .”, p. 45).

the problem of the semi-speaker in language death

145

look for clues from the informants’ manner or from the coherence of the data to provide the means for evaluating a corpus? In sum, much more work needs to be done on the incidence of the semispeaker phenomenon, on the social and linguistic circumstances which give rise to it, and on the linguistic features typical of semi-speaker performance. References Bloomfield, Leonard. 1927. Literate and illiterate speech, American Speech, Vol. 2: 432–39. Dorian, Nancy C. 1973. Grammatical change in a dying dialect, Language, Vol. 49: 413–38. Haas, Mary R. n.d. Tunica (Extract from Handbook of American Indian Languages, Vol. IV). New York, J. J. Augustin. ———. 1968. The last words of Biloxi, International Journal of American Linguistics 34:77–84. Hill, Jane H. 1973. Subordinate clause density and language function, You Take the High Node and I’ll Take the Low Node (Papers from the Comparative Syntax Festival, ed. by C. Corum, T.C. Smith-Stark, and A. Weiser). Chicago, Chicago Linguistics Society. Krauss, Michael E. 1963–70. Eyak Texts. Photocopy. Labov, William. 1970. The study of language in its social context, Studium Generale 23:30–87. Swadesh, Morris. 1948. Sociologic notes on obsolescent languages, International Journal of American Linguistics 14:226–35.

chapter 6

Language Shift in Community and Individual: The Phenomenon of the Laggard Semi-Speaker It is possible, looking at a community as a whole, to speak of language shift even where not a single speaker has changed his linguistic habits. If a high emigration rate, a high in-migration rate, or a differential birth or death rate resulted in a statistically marked change in the ratio of speakers for two languages in a community, a shift would have taken place despite stable patterns of attitude and use (language loyalty). Linguistic groups can be ‘swamped’, going from majority to minority position in a short period; this happened in Glamorgan County, Wales, in the process of industrialization, when English immigrants flooded in during the period 1861 to 1911 (Lewis 1978: 277). Such sudden ‘swamping’ is probably less common, however, than a slow attrition in which, during each successive generation, some community members belonging by birthright to one linguistic group change their linguistic affiliation and move wholly or in part into another linguistic group. The one group ultimately has fewer speakers, using the language in fewer domains and often in a form which is structurally not quite intact (reduced phonology, loss of allomorphic or syntactic options, floods of loanwords); the other group gains speakers and sometimes even additional functions. The reasons for such affiliation changes are usually fairly obvious. One language has more prestige and wider currency than the other and consequently attracts the lion’s share of government, media, and school use; one language has a larger pool of speakers with wider dispersion than the other, so that it is advantageous to many who interact with the large, widely dispersed group to learn its language; and so forth. Mackey (1973) speaks of language power, language attraction and language pressure as three key concepts in what he calls “geolinguistics”; these forces, for which he offers assessment formulae, are in his view “ultimately responsible for the life and death of languages” (1973: 3). With Mackey’s paper a rich explanatory approach to language shift is achieved. * The research reported in this paper was supported by grants from Bryn Mawr College (1974), the American Philosophical Society (1976), and the National Science Foundation (1978) (BNS 77-26295). I am indebted to Michael Silverstein for helpful comments on the first version of this paper. The conclusions reached in this paper were presented in part at the 24th Annual Conference of the International Linguistic Association, March 1977.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�08

language shift in community and individual

147

Predictively the situation is less advanced. Although it is usually easy to see why a language has been or is being given up, it can be very difficult to understand how or why a language persists in the face of what would seem to be very heavy odds. Fishman (1971: 312–324) instances a number of cases where languages have been maintained counter to certain common generalizations about the conditions required to keep a language alive: where nationalistic sentiments are absent, for example; where rural isolation does not apply; where high linguistic or cultural prestige is not a factor. Work which is currently underway in Nova Scotia (Mertz 1978) may shed some light on the causes of differential language retention by whole communities. Two Cape Breton communities are being studied in order to determine the reasons why one of the two gave up its Gaelic speech whereas the other retained it. In the present chapter I would like to look at language retention on the level of the individual rather than on that of the community, in order to shed some light on sources of language loyalty. It is commonplace, in communities where a language shift is underway, for some individuals to be in the vanguard and some to lag behind. Where the laggards are individuals who are more proficient in the language which is giving way, their position is easily understood. They learned one language first and better, it remains both emotionally and linguistically dominant for them, and they choose to go on speaking it even though they may be well aware that it is rapidly becoming the less-favored language on the local scene. There is little to cause surprise or require much explanation here. Another group of laggards may exist whose resistance to shift is far more difficult to fathom. These are individuals whose mastery of the language which is gradually being given up is incomplete, so that they are imperfect speakers whose performances are riddled with what an older, more competent generation could only consider mistakes. In investigations of two communities where a local currency language is being given up in favor of English, I have found such imperfect speakers continuing to make some use of the increasingly disfavored language. These “semi-speakers”, as I have called them (Dorian 1973, 1977), persist in speaking a language which has low prestige and limited currency despite the fact that they speak it imperfectly and in some cases haltingly. This would seem to be a perverse stance, since all are fully proficient in English, have no contact with purely monolingual speakers of the disfavoured language, and thus have no compelling communicative need for the language they control less well. In eastern coastal Sutherland, in the extreme north of mainland Scotland, the local Scottish Gaelic dialect is being given up in the face of heavy pressure from English. The reasons for the shift are readily grasped. Gaelic has

148

chapter 6

been discouraged, even suppressed, in Scottish education and public life for some centuries (Campbell 1950). The economic base which supported the East Sutherland Gaelic (ESG) linguistic community disappeared with the end of the local fishing industry in the years between the two World Wars, and the patterns of residential segregation and endogamy which had kept the Gaelicspeaking fisherfolk apart began to weaken as a result. English, the language of the social elite locally as well as nationally, has the support of virtually all national institutions as they affect local life: law, education, government, the military and (preponderantly) the media. Standard Scottish Gaelic has minimal institutional support from one religious sect and from the media, the local Gaelic dialect none at all. The contest between English, as a language of wider currency with powerful governmental support, and ESG, as a language of restricted currency with no institutional support of any kind, has been unequal for generations. So long as the fisherfolk lived apart with an adequate economic base, they were able to retain their language even while English schooling and frequent commercial contacts with the English-speaking groups allowed them to become fluent in English as well. With the loss of the fishing and the separate marriage and friendship sub-population networks that an independent economic base made possible, shift to English set in rapidly in this last bastion of east coast Gaelic. In the vicinity of Hamburg, in Berks County, Pennsylvania, the German dialect which has been in use for over 200 years is likewise fading rapidly from the scene. Negative prestige attaches to Pennsylvania “Dutch” in the area, as to ESG in Sutherland, and the acceleration of urbanization, with the concomitant weakening of a rural farming economy and its special social network, has led to a shift away from the ancestral German dialect. The Hamburg area “Dutch” are not religious separatists (that is, not Anabaptists of any persuasion, whether Mennonite/Amish or Hutterite), and the move away from the farms has exposed them broadly to the values of the larger American society with its total commitment to English. There is minimal support for German (of the standard variety, however) from the church, and latterly a few schools have offered a little Pennsylvania Dutch in addition to regular courses in standard German;1 off and on some area newspapers print material, usually humorous, in the dialect. Beyond these slender efforts there is no institutional support for the dialect. Neither in eastern Sutherland nor in Berks County are the bilinguals unaware of the language shift in progress. People comment readily on the 1 This is very recent. Policy for many years in this century was the active discouragement of Pennsylvania Dutch in and by the schools (see Dorian 1978).

language shift in community and individual

149

change within living memory in the ratio of speakers of the local language to speakers of English. They are also acutely aware of the negative social prestige of the local language and have many anecdotes to relate of social discrimination against local language speakers; the English of the bilinguals is colored, in both regions, by the phonology of the home language, and this alone is sufficient to make them conspicuous in a disadvantageous fashion. In each case the focus of the social disfavor is probably primarily the charge of “backwardness”, being behind the times socially and intellectually, a kind of bumpkin status; in Sutherland but not in Berks County this stigma includes the reputation of excessive inbreeding in the bilingual community. Low income relative to other groups within the population, in days gone by, adds the notion of material backwardness to that of social and intellectual backwardness, in both cases. Although a great many speakers of ESG and Pennsylvania Dutch have moved out of the relatively isolated settings which spawned these stereotypes, the stigma tends to follow them. Given the position of ESG and Pennsylvania Dutch in these two areas, it seems extraordinary that anyone who had greater skills in English than in the limited-currency language would nonetheless choose to maintain his role as a speaker of that limited-currency language – especially when he was a pretty imperfect speaker at that. Still, this is just what I have found to be the case. The semi-speaker phenomenon is apparently not universally characteristic of dying languages and dialects. Hill (1973), who worked with two dying American Indian languages in California, found no semi-speakers: ‘You either speak fairly well or not at all’ (personal communication). Semi-speakers are common to ESG and Pennsylvania Dutch, however, and one can recognize them in a number of accounts of other fading language communities (among recent American Indian studies alone, for example, Krauss 1963–1970; Miller 1971; Salzmann 1969). Based on extensive interviewing of semi-speakers in eastern Sutherland and a lesser amount in Berks County, I have found three factors which operate to produce the social anomaly of the speaker who chooses to use a low prestige language which he controls imperfectly. These are: cross generational linguistic socialization outside the nuclear family; a highly valued sense of community identity, fostered especially by temporary or permanent exile; and a personality characterized, especially in the childhood years, by marked inquisitiveness and gregariousness.2 While all three factors need not 2 There is a fourth factor which can also produce semi-speakers: late birth order in a large family. Both in Sutherland and in Berks County I have worked with large families in which the youngest two or three children were semi-speakers; the presence of a considerable group of older siblings who have been to English-language schools seems to work against full

150

chapter 6

operate at once to produce a semi-speaker, I find that any combination of two is most likely to have that result. In a community where language shift is proceeding strongly, it is often the second ascending generation rather than the parental generation which transmits the local-currency language to the children; or it may be older siblings or cousins of the parents who play that role. Aware of the social penalties attached to use of the local-currency language and of the advantages of English, parents often deliberately choose to use English with their children. The mother of an ESG semi-speaker said: We never made any effort to ask them to speak the Gaelic. . . . Course we thought it was dying. That’s the trouble. We would never think of the Gaelic. . . . They couldn’t get through the world with Gaelic. That’s what we thou[ght] – took for granted. The Gaelic’s no use to you through the world. This woman’s daughter learned her Gaelic chiefly from a cousin of her father’s, but the commonest figure in the linguistic socialization of the semi-speaker is the grandmother. One woman, both of whose grandmothers were living during her childhood, was sent to reside with the one who had very little English while some of her siblings were sent to reside with the other. As a result she learned Gaelic much better than those brothers and sisters, despite the fact that her parents chose not to use it with their children. Some excerpts from her account: Although they spoke the Gaelic to each other, my mother and father, they didn’t speak it to us. . . . My father and mother used to go to the fishings, . . . and I stayed with this granny. And I heard nothing else but Gaelic. . . . She wouldn’t let me off with – eh – anything that wasn’t right. You know? She corrected me all the time. Because they couldn’t speak the English, at all. It’s – it’s Gaelic all the time. a­ cquisition of the parental language by the youngest members of the family. These semispeakers differ, potentially, from the others I shall discuss, in that the conscious or unconscious choosing to acquire the home language may not have been present; some of them seem to be more accidental semi-speakers than those who form the chief subject of this paper. However, where a late birth-order child also experiences one of the three factors mentioned above, a strong language loyalty may in fact result. For example, two late birth order semi-speakers in east Sutherland were also temporary exiles from the area, and now, living in the home village again, they are strong language loyalists.

language shift in community and individual

151

Another semi-speaker likewise claimed that she spoke better Gaelic than a sibling because of greater contact with her grandmother: I used to stay with my granny a lot, you see, this is the thing. And I – I suppose my granny and grandfather, they spoke Gaelic all the time. More so, I suppose – I suppose I heard my granny and grandfather more than my mother and father, really. . . . My sister’s older, but funny enough, I think I – I probably knew more than she did. And I think it was because I lived with my granny so much. I was never out of my granny’s. This is a theme repeated with striking frequency by semi-speakers: the old granny, the old auntie, the old cousin once or twice removed, who spoke the ancestral language to them when they were very young. Strong cross-generational ties outside the nuclear family thus play a significant role in maintaining the local-currency language among individuals whose own immediate households may have used it relatively little. Where this feature combines with that of strong community identity fostered by exile, the result may be an exceptionally language-loyal semi-speaker. One woman in whose fatherless childhood an aunt played the crucial linguistic role spent some years away from the community as a young woman, working in the city, in the neighborhood of two other girls from the fishing community: I remember when we were working away, when I was in Edinburgh and there were girls there from Brora, and we always went out and we spoke [Gaelic] together. You know, the three of us. Because, you know, we just liked speaking. . . . I like the Gaelic, I’m only sorry I never kept it up. Now, you know. I really like it very much. And I would just speak it all the time if I could. If my husband spoke it I – I would like to – to keep on with it. Perhaps the most improbable of the ESG semi-speakers is a woman in her late thirties who has lived in London ever since she was six years old and has still managed to maintain some hold on the language. Despite weak grammar and a limited vocabulary, she can speak fairly freely in an unconventional but intelligible fashion. The grandmother in whose household she and other family members lived (during the war) until she was six is the figure to whom she attributes the key linguistic influence in her life; her own parents followed during her childhood the familiar pattern of speaking Gaelic in her presence but not to her, though once she was old enough she began to insist on some Gaelic interaction with them, especially with her mother. Gaelic for this exile

152

chapter 6

has strong emblematic value; it signals her Highland Scottishness, which she greatly prizes, and the fishing-village heritage she identifies with: I think it’s a privilege, really, to speak [Gaelic]. [I]t’s a connection with the Highlands. And I just – well, Scotland to me is the place – I just enjoy talking it. And I – the older I get, the more I want to keep it. You know, I don’t want to lose it. I think the older I get, the more I speak it. This same exile semi-speaker exemplifies also the third feature typical of the genesis of the semi-speaker: the exceptionally inquisitive, gregarious personality. Even her grandmother showed a strong tendency to speak English rather than Gaelic to her, but eventually she herself demanded Gaelic from her parents and her grandmother: I don’t think it would have worried Mum and Dad if I didn’t speak it, ‘cause they n – I mean, they never bothered to tell me all the answers, every time I asked. But they never sort of said to me “Come on, sit down and we’ll teach you some Gaelic.” No. Even Granny. I used to ask her to speak Gaelic to me, and she’d still revert to English. As a child living in east Sutherland, and even later, living in a London household where every adult resident and visitor was a Gaelic speaker,3 this girl had an active desire to understand what was said around her: And anybody that came in, they were all Gaelic speakers. You know. Perhaps I was frightened of missing out on something if I didn’t know [Gaelic]. That this is a characteristic of one individual’s personality and not inherent in the situation is borne out by the fact that this semi-speaker’s cousin who was of the same age and who lived in her grandmother’s household in east Sutherland exactly as long as she did, then in London under exactly the same circumstances as she (Gaelic consistently spoken between the parents, adult visitors to the home likewise Gaelic speakers), neither speaks nor understands Gaelic

3 There was a large and rather cohesive group of related East Sutherland Gaelic speakers living in exile in London during her childhood. Adults in this group spent most of their leisure time in each other’s company.

language shift in community and individual

153

today.4 The genesis of the semi-speaker is thus far from inevitable. As with community language shift, where giving up a language is sometimes easier to understand than maintaining it, it is often easier to see why one individual did not become a semi-speaker than why another did. The London environment can scarcely be called favorable to Gaelic, and in neither home did the parents make any real effort to pass on a knowledge of Gaelic to the children. One cousin insisted on access to Gaelic; the other, less surprisingly, did not. The notion that there is a personality type which gives rise to the semispeaker is supported by a highly anomalous Pennsylvania Dutch case, where the speaker is a boy of 15. A dialect speaker of that age in the Hamburg area is almost unheard of; older speakers have sometimes been incredulous when I told them of this boy’s ability. Like the London speaker of ESG, he speaks an unconventional version of the dialect with aberrant grammar and a great many loanwords from English, but, also like her, he speaks it quite readily. It is characteristic of these two gregarious, out-going semi-speakers, in fact, that they speak more freely than many other semi-speakers whose linguistic control of the dialect in question is noticeably better. The mother of this young Pennsylvania Dutch semi-speaker gave an account of his acquisition of the dialect which virtually duplicates the London-based ESG semi-speaker’s account of her own dialect acquisition: We never – pushed him, you know. “Now say this”, or “Say that”, or “This is how you say it”. I mean, he was constantly asking us, you know. If there was something he didn’t understand. . . . He always wanted to know. He was constantly questioning you. . . . He just picked it up from being curious. Asking questions. And, of course, hearing it. Like so many other semi-speakers, this young boy was responsive to older people as a child: [I]f we had friends coming in, . . . he wouldn’t – he’d rather – ’stead of going playing with the – the children, he’d s – sit with the grown-ups. He’d talk, you know. 4 The two cousins are of identical social-class background and are both of good intelligence and similar educational background; differences of socioeconomic status or intelligence do not seem to play a role, therefore. The semi-speaker cousin has a strong interest in language in general which may or may not be shared by the monoglot cousin; it is impossible to say whether this interest develops out of her bilingualism or, perhaps, helped to produce it.

154

chapter 6

A great-grandmother is still alive, in this case, and is the matriarch of a fourgenerational farm family where two generations consistently use the dialect. The boy often visits the farm and works with his great-uncle, a proficient speaker of Pennsylvania Dutch. Conclusions Where the circumstances which favor one language over another in a languagecontact situation are rather overwhelming, it seems to me that explaining resistance to shift demands more of our attention than explaining language shift. It is usually all too clear why a local-currency, low-prestige language gives way to some other language of wider currency and higher prestige. Consequently, in investigating language death5 in eastern Sutherland and in Berks County, I have concentrated on the anomalous cases where an imperfect speaker of the local-currency dialect has persisted in making some use of that dialect despite his awareness of its demographically and socially weakening position and despite his imperfect control of its structure. In the cases presented here it can be seen that positive exposure to a language-loyal kinsperson, especially an older person, outside the nuclear family can counterbalance even the conscious decision of the parents not to transmit the local-currency language. Exile, temporary or even permanent, can foster a sense of community identity that favors maintenance of the threatened language. And, finally, there would seem to be a shift-resistant personality, characterized by curiosity and an outgoing nature; people of this personality type apparently need very little external encouragement beyond mere exposure to the language in the home, despite the well-known fact that exposure alone often produces the purely passive bilingual with no productive skills at all. These observations are explanatory rather than predictive, however. Where any two of the conditions prevail, the likelihood of a semi-speaker resulting is increased, but it is certainly not inevitable. For the individual as for the 5 I regard both these cases of dying dialects as legitimate cases of ‘language’ death, since in each instance a local speech form, isolated from any other dialect of the same language, is giving way not to a more standard or prestigious dialect of the same language but to a totally different language. Where two quite distinct languages were spoken in eastern Sutherland and in Berks County, only one language will survive; hence the term language death, although what is dying is a regional dialect whose extinction will not mean the end of either Gaelic or German.

language shift in community and individual

155

c­ ommunity, abandonment of a disfavored, limited-currency, low prestige language is more common than espousal. References Campbell, John Lorne. 1950. Gaelic in Scottish Education and Life. Edinburgh, W. and A. K. Johnston for the Saltire Society. Dorian, Nancy C. 1973. Grammatical change in a dying dialect, Language 49: 413–438. ———. 1977. The problem of the semi-speaker in language death, International Journal of the Sociology of Language 12: 23–32. ———. 1978. The dying dialect and the role of the schools: East Sutherland Gaelic and Pennsylvania Dutch in International Dimensions of Bilingual Education, ed. by James E. Alatis, 646–656. Georgetown University Round Table on Languages and Linguistics. Fishman, Joshua A. 1971. The sociology of language, Advances in the Sociology of Language I, 217–404. The Hague, Mouton. Hill, Jane H. 1973. Subordinate clause density and language function, You Take the High Node and I’ll Take the Low Node, ed. by C. Corum, T. C. Smith-Stark and A. Weiser. Papers from the Comparative Syntax Festival, Chicago, Chicago Linguistics Society. Krauss, Michael. 1963–1970. Eyak Texts. Photocopy. Lewis, Glyn. 1978. Migration and the decline of the Welsh language, in Advances in the Study of Societal Multilingualism, ed. by Joshua A. Fishman. The Hague, Mouton. Mackey, William F. 1973. Three Concepts for Geolinguistics. Publication B-42, International Center for Research on Bilingualism, Quebec. Mertz, Elizabeth. 1978. The Gaelic-speakers of Nova Scotia: Language maintenance and language death in comparative perspective, mimeo. Miller, Wick R. 1971. The death of language or serendipity among the Shoshoni, Anthropological Linguistics 13: 114–120. Salzmann, Zdenĕk. 1969. Salvage phonology of Gros Ventre (Atsina), International Journal of American Linguistics 35: 307–314.

chapter 7

Defining the Speech Community to Include its Working Margins For a Highland district, eastern Sutherlandshire has a relatively long history of use of English. Early in the twelfth century a prominent Moray family of proven loyalty to the Scottish crown was granted lands there to challenge Norse power in the northeast and establish a significant Scottish political presence (Crawford 1976–77). This family, which took the name de Moravia, was probably of Flemish origin (White 1953; Pine 1959); in any event they certainly did not stem from any Celtic or Pictish line native to the northern Highlands. Ennobled as early as the second Sutherland-based generation, they constituted a point of entry for English in a wholly Gaelic area. Their two principal seats of power, Dunrobin (site of the House of Sutherland’s castle) and the royal burgh of Dornoch, can be shown to have fostered the use of the English language before that language was in use elsewhere in the district (Dorian 1981: 14–15, 52). Although the shift to English in East Sutherland has been slow, it is now almost complete. Most present-day natives of the district are monolingual in English, the sole exception being the descendants of a distinctive ethnic group, the East Sutherland fisherfolk. Fisherfolk descendants are bilingual in English and Scottish Gaelic. They constitute a speech island, in that they are surrounded by English monolinguals and are not in contact with any other dialect of Gaelic. Furthermore, their Gaelic is of a distinctive East Sutherland variety which is unlike other Gaelic dialects.1 Since all Gaelic-speaking fisherfolk descendants are bilingual, they belong simultaneously to two speech communities, whereas their monolingual fellowvillagers belong only to one. This is the picture one arrives at by adopting Gumperz’s definition of the speech community: “any human aggregate characterized by regular and frequent interaction by means of a shared body of verbal signs and set off from similar aggregates by significant differences in language usage” (1971: 114). But while this definition allows clear-cut recognition of two readily identifiable groups – monolingual members of an English 1 The Gaelic of the fishing villages of Easter Ross showed the greatest resemblance to East Sutherland Gaelic, in all likelihood (see Watson 1974 and Dorian 1978: 143–4). The Gaelic of the Easter Ross fishing villages is all but extinct, however (Watson, personal communication).

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�09

defining the speech community to include its working margins

157

speech community, and bilingual members of both an English and a Gaelic speech community – it does not so clearly accommodate a third group which can be shown to exist in the region: low-proficiency ‘semi-speakers’ and nearpassive bilinguals in Gaelic und English. Semi-speakers are individuals who have failed to develop full fluency and normal adult proficiency in East Sutherland Gaelic, as measured by their deviations from the fluent-speaker norms within the community. At the lower end of the proficiency scale they are distinguishable from near-passive bilinguals by their ability to manipulate words in sentences: reminded of a forgotten Gaelic noun or verb, for example, they can nearly always build it into an intelligible Gaelic sentence, whereas near-passive bilinguals can rarely do so (although near-passive bilinguals know a good many lexical items and short phrases). At the upper end of the proficiency scale, semi-speakers are distinguishable from even the youngest fully fluent speakers of East Sutherland Gaelic by the presence in their speech of deviations from the local grammatical norms (recognized as “mistakes” by fluent speakers), and by the frequency of such deviations, as well as by the presence of a marked degree of analogical levelling and a tendency to eliminate syntactic redundancies (Dorian 1977 and 1980). Semispeakers differ among themselves in their grammatical and phonological abilities in Gaelic, however, and also in their manner of delivery. Some speak quite readily, though usually in short bursts; despite their phonological and grammatical deviations, they are generally known and accepted as Gaelic speakers of a sort. That is, they are considered part of the local pool of Gaelic English bilinguals by the fully-fluent speakers (who did in fact name most of them when asked to identify local Gaelic speakers). Others of the semi-speakers use Gaelic relatively little (and in one case, scarcely at all). They speak in a halting manner, and often leave sentences incomplete. Such semi-speakers are usually of low proficiency; but so is the occasional short-burst semi-speaker who uses Gaelic much more freely. That is, the amount of Gaelic actually spoken and the manner of delivery are not perfectly correlated with levels of grammatical proficiency. It is the low-proficiency semi-speakers who speak very little Gaelic, and also the near-passive bilinguals (whose verbal output is mainly short phrases and single-word utterances), that are of interest here, because they challenge the definitions of the speech community which have prevailed in recent years. In terms of their active use of Gaelic, they cannot easily be included in the East Sutherland Gaelic speech community. They speak only English with any readiness, and they speak mostly English in their day-to-day living. Some of them rarely make any active use of Gaelic. One young woman in this group claimed not to speak Gaelic at all, in fact, although when persuaded to undergo

158

chapter 7

a ­battery of translation tests she proved to control Gaelic grammar slightly better than a relative very near her in age who is an enthusiastic, even eager, short-burst semi-speaker. The reluctant semi-speaker, like some of the nearpassive bilinguals, cannot really be said to be “set off from [other] aggregates by significant differences in language usage”, if language usage is taken to mean active use of a speech variety by the individuals in question. Because some properties commonly taken to be important to membership in a speech community are absent among low-proficiency semi-speakers of East Sutherland Gaelic and among near-passive bilinguals in the same community, their claim to inclusion in an East Sutherland Gaelic speech community needs to be carefully considered. They can usefully be compared both with English monolinguals, who are readily excluded, and with fluent speakers of East Sutherland Gaelic and with high-proficiency semi-speakers, who are readily included. There is first and foremost the issue of language use and the norms that govern it in a given group. Fishman’s definition of the speech community, like Gumperz’s, includes this notion: “A speech community is one, all of whose members share at least a single speech variety and the norms for its appropriate use” (1971: 232). Labov’s concept of the speech community abandons any notion of uniformity in usage, but rests on a shared evaluation of patterns of usage: “The speech community is not defined by any marked agreement in the use of language elements, so much as by participation in a set of shared norms; these norms may be observed in overt types of evaluative behaviour, and by the uniformity of abstract patterns of variation which are invariant in respect to particular levels of usage” (1972b: 120–1). Thus Labov can perceive New York City’s complex native population as a single speech community because it shares regular patterns of subjective reaction to phonological variation: “it seems plausible to define a speech community as a group of speakers who share a set of social attitudes towards language” (1972b: 248). Since low-proficiency East Sutherland Gaelic semi-speakers and n ­ ear-passive bilinguals do not conform at all well to the prevailing fluent speaker norms for use of East Sutherland Gaelic, and are quite insensitive to many breaches of grammatical and phonological norms produced either by themselves or by others (e.g. by foreign learners of Gaelic), they would not seem to qualify for membership in the local Gaelic speech community by these criteria. Yet in certain important respects they are entirely unlike the English monolinguals who represent the clear-cut excludable group. The first is their outstanding receptive control of East Sutherland Gaelic, and the second is their knowledge of the sociolinguistic norms which operate within the Gaelic-speaking community. The fisherfolk descendants of East Sutherland number just under 100 at

defining the speech community to include its working margins

159

present2 and are distributed over three villages. The Gaelic speakers in any one village are complexly interrelated to one another, and there are also kin ties across villages; this is the result of about a century of forced endogamy among the fisherfolk. Kinship networks are also often friendship networks, so that relatives representing a variety of ages interact a good deal. Families also traditionally ran rather large, so that siblings sometimes span more than a decade in age ranges. As a result of these two facts, interaction networks which include both fully-fluent, Gaelic-dominant bilinguals and low-proficiency s­ emi-speakers or near-passive bilinguals, and even young English monolinguals, are fairly common. In one household, for example, four unmarried children, all adult, lived in the home of their Gaelic-dominant mother, while two somewhat older (and married) children lived nearby in the same village (village A). The three oldest siblings were fully-fluent bilinguals, the next two were high-proficiency semi-speakers, and the youngest of the siblings was a near-passive bilingual. Relatives who visited regularly in this household, often spending entire evenings there, included both fully-fluent bilinguals and English monolinguals. In another village (village B), various relatives gathered irregularly in the home of their eldest kinswoman, a fully-fluent woman who is now a nonagenarian. Most of them were likewise fully fluent, but the nonagenarian’s high proficiency semi-speaker daughter was often present, and also a low proficiency semi-speaker kinswoman who was a next-door neighbour. Most of my remarks about the claims of low-proficiency semi-speakers and near-passive bilinguals to membership in the East Sutherland Gaelic speech community are based on long-term participant observation of these two networks and another one composed of East Sutherland exiles from village A residing in and around London, plus less intimate knowledge of other similar networks. The networks I have observed and participated in – these three as well as others – have been altered by the deaths of one or more members over the 17-year period that I have known them; therefore I use the past tense to describe them even though they continue to exist in reduced form. The most striking feature of the cross-generational interaction networks among fisherfolk descendants was the ability of the low-proficiency members to participate in Gaelic interactions. Despite their very limited productive skills, they were able to understand everything said, no matter how rapidly or uproariously. They never missed the point of a joke or failed to grasp a significant tidbit of gossip. They occasionally supplied a translation of something difficult to hear or something poorly enunciated for the linguist-guest who spoke 2 The number was about 200 when I began fieldwork in East Sutherland in 1963–64.

160

chapter 7

a far more grammatical and likewise more fluent East Sutherland Gaelic than they did. The second notable feature of the participation of low-proficiency network members was its sociolinguistic ‘fit’. What they actually said might be very little, and some of their utterances were always grammatically deviant. But since their verbal output was semantically well integrated with what preceded in the conversation, and since it conformed to all the sociolinguistic norms of the dialect, the deviance could usually be overlooked. Often the semi-speaker or near-passive bilingual did not even have to finish the sentence; some fullyfluent member of the group could step in, if there was a marked hesitation, and supply the anticipated conclusion. Low-proficiency members of these networks, unlike the linguist-guest, were never unintentionally rude. They knew when it was appropriate to speak and when not; when a question would show interest and when it would constitute an interruption; when an offer of food or drink was mere verbal routine and was meant to be refused, and when it was meant in earnest and should be accepted; how much verbal response was appropriate to express sympathy in response to a narrative of ill health or ill luck; and so forth. Two approaches to the speech community which seem more adequate, in the sense that they do not define out of membership those who have low productive capacity but high receptive capacity and who conform to the sociolinguistic norms, are provided by Hymes (1974) and Corder (1973). Hymes proposes that the social group, rather than the language, be taken as the starting point, and that we then consider ‘the entire organization of linguistic means within it’ (1974: 47). This would enable us to start with the participants in Gaelic verbal interactions, in East Sutherland, including the low-proficiency semi-speakers and near-passive bilinguals, and define the speech community so as to include them. Their inclusion would be appropriate not simply because they are participants (so, after all, was the linguist-guest), but because they are highly successful participants whose receptive skills and knowledge of the sociolinguistic norms allow them to use their limited productive skills in ways which are unremarkable (that is, provoke no comment).3

3 I have noted that semi-speakers’ East Sutherland Gaelic is grammatically deviant in ways that are labelled “mistakes” by fluent speakers, which is true enough. But when left to their own devices, so that they can speak when they wish to, briefly and in the structures they are most comfortable with, semi-speakers are often able to reduce the deviance to the point where it can be overlooked, especially in the flow of a general conversation. Semi-speakers are sometimes also “rescued” by a fluent speaker from the necessity of finishing more adventurous sentences which they may have begun and in the middle of which they then hesitate.

defining the speech community to include its working margins

161

Corder takes the self-perceived group as the basis for the speech community: “A speech community is made up of people who regard themselves as speaking the same language; it need have no other defining attributes” (1973: 53; italics in original). This approach has the advantage of according well with the wellintegrated position in the Gaelic interaction networks of some individuals with extremely poor active skills. The very low-proficiency semi-speaker who lived next door to the (now) nonagenarian fluent speaker in village B expressed the complete ease she felt when her visits to her neighbour coincided with those of another relative, this one a verbally gifted, notably articulate fluent speaker: Semi-speaker: J., she’s fluent – Gaelic speaker. But any – any time she’s in, if she does [speak Gaelic], you know, I don’t – I just take it in my stride, as – just as if it’s English, you know? It doesn’t worry me in any way. Or I don’t get mixed up, and I know what they’re – I can join in the conversation, because I know everything they’re saying, you know. I haven’t to stop and think or anything. Investigator: Yes. Uh-huh. And if you joined in, would you join in in English or Gaelic? Semi-speaker: I would join in in Gaelic, you know. As best I could, y’ know.4 When this woman’s active skills were first tested, in the same year (1974) the above interview was taped, the testing was done in her fluent next-door neighbour’s home where she spent so much time. It proved a distressing experience for all participants: neither of the women, though neighbours for years, had realized how little active control of the dialect the younger woman had; nor for that matter had I, or I would not have exposed her to the embarrassment of “public” testing. She proved to be one of the very weakest speakers in my sample, yet none of us had noticed her failings as an active speaker, thanks to her skillful use of what proficiency she had and to her outstanding receptive skills and sociolinguistic knowledge. Although it’s clear that she acknowledges some weakness in her speaking abilities (“as best I could”), it’s also clear that she feels included in the interaction (“I know everything they’re saying”).5 4 This interview was conducted in English and is quoted here verbatim. It proved extremely difficult, almost impossible even, to interview this semi-speaker in Gaelic, since the presence of the tape-recorder produced a nervousness which compounded her difficulties with Gaelic. 5 A young man in Berks County, Pennsylvania, a low-proficiency semi-speaker of Pennsylvania Dutch, expressed exactly the same sense of inclusion in interactions with fluent Pennsylvania Dutch speakers, for the same reasons.

162

chapter 7

At least until the testing took place, her neighbour certainly considered her an adequate member of the Gaelic speech community. That was precisely why the testing proved so distressing to all of us – it showed plainly that she was actually less than adequate in productive East Sutherland Gaelic skills. But she regarded herself, and was in turn regarded by fluent speakers, as a speaker of East Sutherland Gaelic. What is interesting about this case (and others in which low proficiency semi-speakers and near-passive bilinguals participate successfully in Gaelic interactions) is that it highlights the minimum requirements for membership in a speech community. Fluency is not required, nor grammatical and/or phonological control of the speech variety common to the participants. The foreign learner can achieve those things and still be only a participant in a speech community and not a member if he or she does not also fully master receptive skills and sociolinguistic norms. As Hymes has so often insisted, communicative competence depends not only on knowing how to say something, but also on knowing how to say it appropriately (1964a, 1964b, 1967, 1971, 1974, for example). In fact it seems that knowing how to say relatively few things appropriately is more important than knowing how to say very many things without sure knowledge of their appropriateness. Low-proficiency semi-speakers, not to mention near-passive bilinguals, meet none of Fillmore’s criteria for fluency (in the sense of speaking one’s language well; Fillmore 1979: 93). In my experience their only productive skill which is even close to normal is control of what Fillmore calls formulaic expressions (1979: 91–2,94). There are a great many formulaic expressions which can be trotted out on suitable occasions; knowledge of their forms and their suitability enables the user to participate actively in the verbal interaction and helps to keep the interaction going forward smoothly, and thus earns the user a measure of social approval.6 Observation of semi-speaker success with these items offers support for Fillmore’s belief that “a very large portion of a person’s ability to get along in a language consists in the mastery of formulaic utterances” (1979: 92). I noted at the outset of the discussion of low-proficiency semi-speakers and near-passive bilinguals that any definition of the speech community which implied productive control of the language in question would eliminate these 6 In the absence of strong skills in the use (as opposed to the form) of formulaic expressions, my own strategy has been to master a good many East Sutherland Gaelic proverbs. It is easier on the whole [to know] when they are appropriate, and they also earn the user strong social approval for the same set of reasons noted in semi-speaker use of formulaic expressions, as well as for control of highly valued traditional material.

defining the speech community to include its working margins

163

apparent members of the East Sutherland Gaelic speech community, and likewise any definition which required sensitivity to the social evaluation of usage patterns. In connection with the latter criterion, I stated that these marginal East Sutherland Gaelic speech community members are insensitive to even fairly gross breaches of the local grammatical and phonological norms. Even if they were not insensitive in this respect, however, Labov’s notion of ‘a set of social attitudes toward language’ as a defining feature of the speech community would not apply in East Sutherland in anything like the way he found it to apply in New York City, where people who used certain forms at very different levels could assign social values to them quite uniformly (cf. also L. Milroy, this volume). The reason is that East Sutherland Gaelic is singularly lacking in patterns of social (as opposed to grammatical) evaluation of linguistic structures. A great deal of variation is characteristic of the dialect, morphophonemically, morphologically, and syntactically. Comment on this variation is confined almost wholly to patterns that correlate with a particular village, not with social groups within or across villages. That is, regional variation is the obsessive interest of East Sutherland Gaelic speakers, not social variation. Every East Sutherland Gaelic speaker is on the alert at all times for the intrusion of a variant characteristic of one of the other villages, and I myself provided endless material for this preoccupation because of the fact that I travelled regularly from village to village and often carried “alien” forms along with me through inability to switch cleanly enough from one village’s forms to the other’s as I went.7 A very large amount of morphonemic, morphological and syntactic variation not correlated with geography passed without notice among East Sutherland Guelic speakers. I pointed much of it out to the more thoughtful of my informants over the years, and invariably they said that they had never noticed it. In fact, I had a hard time getting them to notice it even when I produced two variants of the same structure in a row for them, so as to highlight the difference. Often I had to repeat the variants several times before they could spot the variation in question.8 When they did become aware of the variation, they had no strong feelings about the alternatives: they made no social judgements in connection with them and generally had no sense that one was more correct or suitable than another. 7 It was not uncommon for me to be in all three villages in the course of one day, and I was simply unable to monitor my speech carefully enough to guarantee only the correct forms for whatever village I was in at a given time. 8 Control of the phonology is one of the stronger points in my own East Sutherland Gaelic so that foreign accent [at least in the production of single words at a time] is not an explanation for this outcome.

164

chapter 7

The chief reason for the absence of social evaluation of linguistic variation must be the fact that the East Sutherland fisherfolk constituted until very recently an undifferentiated social group: all followed the same occupation, all were poor, all were members of a stigmatized ethnic group, none who remained in East Sutherland had more than the legal minimum of education. Intra-group status distinctions rested on skills or moral character, not on differences in occupation, education, or wealth. All present-day bilinguals grew up in active fisherfolk households, and all share the same status in the local social hierarchy as a result. How deep low-proficiency semi-speakers’ and near-passive bilinguals’ knowledge of regional variation is, I cannot be sure. Everyone in the Gaelicspeaking group, regardless of level of proficiency, can produce on request a short list of regional variants which are local stereotypes, much discussed and frequently imitated for the purpose of poking fun. But among fluent speakers, despite the fact that all produce the same small list of stereotypes when asked about regional variation, there is awareness of many more variants than they typically offer. If asked about words not among the stereotypes, they can often come up with the “alien” forms used in other villages, and when they listen to tape recordings made in another village they spot the regional variants readily, without any prompting. They are eager to discuss them, in fact, and will concentrate on them to the exclusion of content, oftentimes. I have neglected to press the low-proficiency semi-speakers and near-passive bilinguals to the limits of their skill in these matters, so that I am unable to say whether they can match the fluent speakers’ knowledge. Another issue on which I cannot at present shed any light is the position of true passive bilinguals in the speech community. I know that there are such people – individuals who understand what is said, but cannot produce Gaelic speech – since I once heard a young woman whose inability to pronounce East Sutherland Gaelic words was both evident and self-admitted translate a Gaelic conversation for the benefit of a foreign visitor. Unfortunately I have not myself worked with any true passive bilinguals, and so I have little notion of the actual extent of their passive abilities. Their existence is acknowledged by the bilinguals, who complain that one can’t count on keeping a secret through use of Gaelic if these people are about. Although they are sometimes “participants” in Gaelic interactions by dint of injecting English comments and responses into a Gaelic interchange which they have understood, I have never heard anyone deliberately address a Gaelic remark to them (except as a direct challenge to them to reproduce it – a test of their abilities), nor have I ever heard anyone refer to them as Gaelic speakers. In both these respects they differ from semispeakers of any proficiency, and even from the near-passive bilinguals, who are

defining the speech community to include its working margins

165

frequently spoken to in Gaelic despite their severely limited productive skills. Provisionally, then, I would exclude them from the East Sutherland Gaelic speech community. The behaviour of fluent speakers towards low-proficiency semi-speakers and near-passive bilinguals, on the other hand, indicates that the East Sutherland Gaelic speech community needs to be defined so as to include these marginal speakers. References Corder, S. Pit. 1973. Introducing applied linguistics. Harmondsworth: Penguin. Crawford, Barbara E. 1976/7. The earldom of Caithness and the kingdom of Scotland, 1150–1266. Northern Scotland 2:97–117. Dorian, Nancy C. 1977. The problem of the semi-speaker in language death. Inter­ national Journal of the Sociology of Language 12:23–32. ———. 1978. East Sutherland Gaelic: The dialect of the East Sutherland fisherfolk. Dublin: Dublin Institute for Advanced Studies. ———. 1980. The maintenance and loss of same-meaning structures in language death. Word 31:39–45. ———. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. Fillmore, Charles J. 1979. On fluency. In Charles J. Fillmore, Daniel Kempler and William S.-Y. Wang, eds, Individual differences in language ability and language behavior, 85–101. New York: Academic Press. Fishman, Joshua A. 1971. The sociology of language: An interdisciplinary social science approach to language in society. In Joshua A. Fishman, ed., Advances in the sociology of language, vol. 1, 217–404. The Hague: Mouton. Gumperz, John. 1971 [1968]. The speech community. Reprinted in Anwar S. Dil, ed., Language in social groups: Essays by John J. Gumperz, 114–28. Stanford: Stanford University Press. Hymes, Dell. 1964a. Introduction: Towards ethnographies of communication. In John J. Gumperz and Dell Hymes, eds, The ethnography of communication, American Anthropologist Special Publication 66, 6, part 2, 1–34. Menasha, Wisconsin: American Anthropological Association. ———. 1964b. Directions in (ethno-)linguistic theory. In A. Kimball Romney and Roy Goodwin D’Andrade, eds, Transcultural Studies in Cognition, American Anthro­ pologist Special Publication 66, 3, part 2, 6–56. Menasha, Wisconsin: American Anthropological Association. ———. 1967. Models of interaction of language and social setting. Journal of Social Issues, 23:8–28.

166

chapter 7

———. 1974. Foundations in sociolinguistics. Philadelphia: University of Pennsylvania Press. Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. Pine, L. G. Burke’s peerage. London: Burke’s Peerage Ltd. Watson, Joseph. 1974. A Gaelic dialect of N. E. Ross-shire. Lochlann 6: 9–90. (Norsk Tidsskrift for Sprogvidenskap, supplementary vol. vi.) Oslo: Aschehoug. White, Geoffrey H., ed. 1953. The complete peerage. London: St Catherine Press.

chapter 8

Abrupt Transmission Failure in Obsolescing Languages: How Sudden the “Tip” to the Dominant Language in Communities and Families? One of the regrettable but interesting things about language death is its long history. It’s anything but a new phenomenon, and we have a lot of extinct languages littering the shores of linguistic history to prove it. On the other hand, our own time seems a little curious in one special respect, namely in respect to the number of languages which have persisted with pretty fair strength for what seems like a long period, only to weaken in what seems like a rather short time and suddenly wind up in a downslide toward extinction. In this country and Canada, for example, some long-established populations with very distinctive customs and languages which have been secure for centuries are suddenly in trouble. The geographical region doesn’t seem to matter – it’s the same story regardless of location. Cajun French in Louisiana is in the same trouble as French Canadian in Maine. Pennsylvania Dutch (that is, German) among the secular (non-Anabaptist) Pennsylvania Dutchmen is threatened in the same fashion as Scottish Gaelic in Cape Breton. None of these is a particularly johnny-come-lately immigrant language – the oldest of them have been in place for several centuries, and their speaker populations have been relatively loyal and stable, sometimes also reinforced by continuing immigration (this is the case with Canadian reinforcement of the Frenchspeaking population in Maine and Highland Scottish reinforcement of the Gaelic-speaking population in Nova Scotia, whereas the Cajun and Penn Dutch populations seem to have recruited more by absorbing incomers or non-native locals than by major inflows of new immigrants). In general the twentieth century seems to be notable for the large number of languages which are either obviously dying out or showing marked signs of contraction such as simplifying structure, functional restriction, and loss of speakers at the margins of the community. Whether this century is actually any more characterized by these phenomena, or whether we’re only better informed about the number of cases and their wide geographical distribution, is unclear. Some people are inclined to argue that this is a particularly pernicious time for languages which are isolated, or enclaved, or represented by rather thin populations, or heavily outbalanced by languages of wider currency. People of © koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�10

168

chapter 8

this persuasion usually point to ease of modern travel, the “global village” phenomenon, the power of the modern nation-state to affect the lives of even its most outlying citizens, the savage thoroughness of the more modern instances of genocide or attempted genocide, the spread of literacy, the penetration of radio and television, and so forth. I think there is no denying any of these factors. They are all very real and very potent. Anyone who has worked with even a single threatened language can attest to the force of negative policies (or even only negative attitudes) spreading out from a central government and discouraging or perhaps penalizing speakers of languages or dialects other than the officially state-promoted language. Similarly the ouster of traditional activities which fostered minority languages – social gatherings like the ceilidh in Scotland and Ireland, pedagogically-oriented verbal routines such as Aesopian tales, fairy tales and rhyming genres (all directed toward children) in Albanian-speaking Greek communities (Tsitsipis 1983: 27), the most formal styles of public speaking in the Cupeño and Luiseño communities in California (usurped by English; Hill 1973: 45) – by passive or active verbal events which involve only or mainly the statepromoted language has a pronounced, unmistakably deleterious effect on the strength of the minority languages in most cases. This is the usual outcome, more or less the predictable outcome, and it surprises no one. It’s not the inevitable outcome, however, since people seem to be capable of quite remarkable segmentation of their lives, including linguistic segmentation. It’s hardly encouraging for a language to be excluded from the schools, ignored in broadcasting, discouraged in public life, and unprovided for in any officially sponsored activities whatever. But in some societies it seems to be possible for people to accept a very restricted role for their native speech form, such that they assume it will be used only in the hearthand-home sphere; they may even welcome the specialization of their mother tongue as an in-group marker. Where there is a deep gulf between the minority-language group and the dominant-language group, as with certain Native American tribes, the home language may be jealously guarded from members of the majority language group, treated along with things like religious ceremonials as a privileged form of in-group knowledge, not to be casually exposed to outsiders or shared with them. There are entire societies in which the home language has good standing but has been traditionally restricted in use without any threat to its ultimate viability (German Switzerland, where Schwyzertütsch is seldom written and almost never used in circumstances of any formality is a case in point), and of course quite a lot of societies exist in which the language of highest prestige is not the local language – most often where religion is involved, as in many Islamic but non-Arab societies.

abrupt transmission failure

169

Since there are recognized instances of all these exceptions to any general tendency to succumb to centralizing dominant-language pressures, the question may be why there aren’t more such exceptions rather than why there are any. In connection with the relatively long-standing ethnic communities now experiencing survival difficulties in the U.S. and Canada, it seems to be the temper of the times which works most against compromises which would allow continuance. Despite the “melting pot” myth, special provision for certain mother-tongue rights of long-established non-English populations was made in several cases into at least the early twentieth century: French in Louisiana (Kloss 1977: 112–113), German in Pennsylvania (ibid.: 146–147), Spanish in New Mexico (ibid.: 130–1:31), for example. Assimilative pressures have nonetheless been strong, of course, and the great nineteenth-century waves of European immigration undoubtedly created tensions for longer-established populations as concerns were increasingly voiced over the effect of home-school bilingualism on intelligence and on loyalty to the national state. Although functional segregation in language use is a perfectly feasible way of managing and maintaining two or more languages, unless the wider community is one in which this is the norm (as in German Switzerland and in Somalia, for example; see Pride 1971 for the latter case), there seems generally to be little support for this course and little understanding of its frequency of occurrence in a good many parts of the world. In most of western Europe and the areas colonized by western European nations, the prevailing attitudes have most definitely not been favorable to full-fledged linguistic dualism of any sustained kind. On the basis of my own work with two minority languages, one in Great Britain and one in the eastern U.S., and also on the basis of reports from other researchers working in similar settings, I would propose a rather widespread phenomenon which I have dubbed “tip” in describing the British case (Dorian 1981: 51). This phenomenon can be conceived metaphorically as a gradual accretion of negative feeling toward the subordinate group and its language, often accompanied by legal as well as social pressure, until a critical moment arrives and the subordinate group appears abruptly to abandon its original mother tongue and switch over to exclusive use of the dominant language. Because of the seeming suddenness of the switch-over, it’s rather like watching a structure slowly eaten invisibly away at the bottom topple over almost without warning. Yet when the tip has occurred and one begins to examine the period which led up to it, the tip is seldom if ever so sudden as it initially appeared. The most striking level at which tip occurs is, to my own perception, that of the family. I would like to introduce two cases, one among the Gaelic-speaking fisherfolk of East Sutherland in the extreme northeast of the Highland Scottish

170

chapter 8

mainland and the other among the secular Pennsylvania Dutch of the Hamburg area in Berks County, Pennsylvania. In each of these family cases there were a good many children: seven in the Gaelic-speaking family, and twelve in the Penn Dutch speaking family. The parents in each family were skilled bilinguals but spoke Gaelic and Penn Dutch (respectively) by habit and preference with each other and within the home generally at the outset of their family life. In each of these families the elder children – the first four of the seven in the Gaelic speaking family, the first nine of the twelve in the Dutch-speaking family – were raised as, and became, fully fluent speakers of the parents’ original mother tongue. In the Gaelic-speaking family no conscious change in the parents’ linguistic behavior toward the three youngest children seems to have taken place, whereas in the Dutch-speaking family there was an acknowledged though unexplained change of that type. In each family the three youngest children emerged as imperfect speakers (or, in the case of the youngest child in the Gaelic-speaking family, as a near passive bilingual with very little ability to generate utterances in the parental mother tongue). Several aspects of the two cases are especially interesting. One is that the parents’ intentions probably mattered relatively little, since the results were the same in a case where the parents deliberately changed their behavior and in a case where they didn’t. The behavior of the peer group outside the family and also the sheer number of older siblings who had attended English-languageonly schools and were using a good deal of English among themselves in or around the home most likely had more impact on the language-acquisition patterns of the youngest children than the parents’ own linguistic behavior or transmission plans, since in both homes the parents continued to use the original mother tongue with each other and with the older children (and with all the children in the Gaelic household). This means that the youngest children received at least a good deal of exposure to that language. Another interesting facet of the two cases is the clarity of the fully-fluent as opposed to the less than-fully-fluent demarcation line among the children, and the unimportance of the size of the age-gap where the demarcation line falls. No one in either family is in any doubt about which child is the last of the fully fluent and which is the first of the imperfect speakers. The three youngest children in each family are just as aware of their less-than-fully-fluent status as the older children are, although in the Gaelic speaking family the three youngest are not particularly sensitive about it whereas in the Penn Dutch family the three youngest mind very much that they are not as competent in Dutch as their older siblings. In the Gaelic family the last of the fully-fluent children is two years younger than the next oldest fully fluent sibling and only one year older than the first of his imperfect-speaker siblings. In the Dutch family the

abrupt transmission failure

171

last of the fully-fluent children is three years younger than the next oldest fully-fluent sibling and two years older than the first of the imperfect-speaker siblings. My data from tests among the Penn Dutch-speaking siblings are not yet fully analyzed. But I can present here two sets of results from testing of the Gaelicspeaking family, offering clear indication of how sharply the fluency line can be drawn between siblings only a year apart in age. In Table 1, Speaker 1 is the mother of the family. Speaker 2 her fourth child, and Speaker 3 her fifth child. Neither child was married; both lived in the mother’s household. There is only a year’s difference in age between the two siblings. The test was for analogical levelling in four structures, the frequent use of analogically levelled forms being a notable marker of the imperfect Gaelic of the less-than-fully-fluent. Identical sentences were presented to each of the three speakers, in individual elicitation sessions, for translation from English into Gaelic. (It should be noted that translation is a relatively natural, high-frequency occurrence in a community where kin networks include both bilinguals and monolinguals, since remarks or conversations in one language will often be recounted in translation at a later time to a kinsperson with whom the language of the original interchange is not the normal language of social interaction). All three of these speakers knew me well, were comfortable with me, and had done this kind of work with me before; I had been around the district over a period of a good many years and the test sentences were couched in a form of English which was reasonably normal for the local English dialect. Speakers 2 and 3 may be only a year apart in age, but Speaker 2 is much closer to his mother, 29 years older than he, than to his sister one year younger, in his linguistic usage on this measure. In Table 2, Speaker 2 is compared with Speaker 3 again and also with Speaker 4, the latter being the next younger sibling, another sister four years younger than Speaker 3 (and so five years younger than Speaker 2). The structures tested (by the same type of elicited translation procedure) were control of three tenses and control of three embedded structures (for discussion of the theoretical difference in the difficulty of the structures concerned, see Dorian 1982). Although Speakers 3 and 4 show some marked differences in their control of the particular structures tested (Speaker 3 being distinctly better than her sister at using the conjunction ‘that’ and distinctly worse at forming relative clauses and at constructing the conditional), the really striking difference is between their performances taken together as compared with their brother’s. He in fact misproduced only one form in the entire set of sentences, whereas his sisters misproduced 15 each. Overall, then, they performed much like each

172 Table 1.

Speaker 1 Speaker 2 Speaker 3

chapter 8 Analogically-levelled forms supplied by three members of a single Gaelic-speaking household in which Speakers 2 and 3 differ in age by only one year. Analogicallylevelled noun plural (opportunities: 17)

Analogicallylevelled verb stems (opportunities: 16)

1a 2 8

0 0 3

Analogicallylevelled 1st pers. sing. future verb (opportunities: 5)

Speaker 1 Speaker 2 Speaker 3

0 0 5

Analogically-levelled 1st pers. sing. conditional verb (opportunities: 5)

0 0 5

Total of # of forms potentially produced that analogically-levelled were analogical: forms produced

43 43 43

1 2 21

% of forms produced that were analogical

2 4.5 49

a: offered in addition to an irregular, non-analogical form

other, despite the four years between them; the brother, though only a year older than Speaker 3, performed quite differently – namely like the fully-fluent speaker he is. For the Berks County Pennsylvania Dutch I can’t be certain that intrafamily tip of the dramatic abruptness I found in the 12-sibling group is a frequent occurrence, since I worked extensively with two kin networks only. But among the East Sutherland fisherfolk I know of several similar cases where groups of siblings were sharply and abruptly divisible according to full fluency versus imperfect-speaker control of the local Gaelic, even though I undertook close testing of only this one highly available and highly cooperative family. The community at large was well aware of the phenomenon, in fact, and readily identified cases in their own kin networks or others. When it comes to tip on the community-wide level, the cessation of homelanguage transmission can seem equally sudden and surprisingly datable. In the smallest of the East Sutherland fishing villages, for example, I found that people were able to identify the last primary-school class whose members

173

abrupt transmission failure Table 2.

Comparison of control of three tenses and three embedded structures by three siblings in a single Gaelic-speaking household

Correctly formed

Speaker 2 Speaker 3 Speaker 4

Correctly formed

Speaker 2 Speaker 3 Speaker 4

Speaker 2 Speaker 3 Speaker 4

Past #/%

Future #/%

Relative #/%

23 of 23 / 100 24 of 24a / 100 23 of 23 / 100

17 of 18 / 94.5 15 of 18 / 83 16 of 18b / 89

8 of 8 / 100 3 of 8 / 37.5 6 of 8 / 75

‘that’ #/%

‘if’ #/%

Conditional #/%

9 of 9 / 100 8 of 9 / 89 1 of 9 / 11

5 of 5 / 100 4 of 5 / 80 4 of 5 / 80

10 of 10 / 100 6 of 11 / 54.5 8 of 10 / 80

# of errors

opportunities

% incorrect

1 15 15

73 75 75

1 20 20.5

a The number of instances of a given structure sometimes differs across speakers because a particular speaker offered two variants for a particular structure, each of which was recorded, evaluated, and counted in arriving at the tabulation. b The figures here differ very slightly from those recorded for Speaker 3 in Dorian 1982: 39, where she appears as WR; results of testing of the future were retabulated and recounted subsequently, with one additional instance recognized for WR.

regularly used Gaelic on the playground whenever they were let out to play during the schoolday. The class only one year younger, everyone agreed, might occasionally use Gaelic on the playground, but did so seldom; and they did not typically become, or remain, fully-fluent speakers, whereas their immediate predecessors did. No one could give a particular reason why this change in language behavior should have come exactly when it did, but they agreed on its timing. It was as if a consensus had tacitly been reached among the children – and that was that. Not merely coincidentally, the brother identified as Speaker 2

174

chapter 8

in the tables above was a member of the last primary-school class to use Gaelic regularly on the playground, and the sister identified as Speaker 3 was a member of the immediately following class, which did not regularly use Gaelic on the playground. This again suggests that even had the parents in that family made a concerted effort to keep their last three children in the fluent-speaker fold, the climate among the youngsters themselves would have made it an extremely uphill battle. Very strict and very determined parents are certainly known to succeed in producing fluent bilingual children, and then to succeed in maintaining that bilingual fluency in their children, within communities unfavorable to the phenomenon; but in my own experience most such cases involve either middle-class (often intellectual) parents, or, alternatively, an only child. (One exception which comes to mind did involve a Scottish Gaelic family. They lived in a very isolated district on the west coast of Scotland: the parents were not middle class, nor, so far as I can recall, was there only one child, but the father was considerably older than the norm for a parent in that community and was a formidable and demanding figure in the household life.) Reports of community-wide tip turn up with some frequency in the growing literature of language shift (see, for example, Gal 1979, Hinojosa 1980, Mertz 1980). Because parents in communities where transmission failure seems sudden often simply decline to raise their children as bilinguals, usually citing concern for the children’s success in school or ability to get ahead in the world as reasons (Denison 1971: 166–167; Dorian 1981: 104; Huffines 1980: 52; Pulte 1973: 426; Timm 1980: 30), some scholars have raised the question of whether the passing of such languages ought rather to be considered “language suicide” than language death (e.g. Denison 1977, Greene 1972). But this is to ignore the long history, usually stretching centuries into the past, of relentless pressure on the non-dominant language. The Zapotec case presented by Hinojosa (1980) is particularly interesting, because Zapotec in fact showed a relatively unusual degree of resistance to the spread of Spanish among Mexican Indian populations. The Zapotec had been doubly resistant to dominant language pressure, what is more, since in the pre-Conquest period they successfully fought off the Aztecs and retained their independence and identity to a unique degree (Hinojosa 1980: 28). The town of Juchitán, not served by the railroad and the loser in the rivalry for capital-city status in the Isthmus of Tehuantepec, remained strongly monolingual in Zapotec for a surprisingly long time. But in the early 1970s the discovery of oil in a nearby coastal area led to the creation of a new port, Salina Cruz, relatively near to Juchitán. For the first time there was strong economic incentive for Juchitán natives to acquire Spanish, since good jobs became available to those with control of Spanish (op. cit.: 28, 30). It would seem that the early

abrupt transmission failure

175

signs of language shift documented by Hinojosa on the basis of fieldwork done in Juchitán in 1979 might be taken as a rather unambiguous case of a sudden change in language behavior clearly motivated by dramatically new economic factors. The change was sudden, and the economic factors were new and dramatically different. Yet when one takes into account the long and complex history of language policy in Mexico meticulously traced by Heath in her volume Telling Tongues: Language Policy in Mexico, Colony to Nation (1972), it seems permissible and even necessary to doubt that the change from proud Zapotec language loyalty in Juchitán to sudden willingness to embrace Spanish as the language of economic opportunity could have taken place quite so rapidly without a long and sustained period in which first colonial and then state policy disvalued the Indian and his language. And if it is true, as Heath reports, that “the Indian had been locked in a caste-like system, which defined his position at the bottom of the nation’s socioeconomic hierarchy since Independence” (1972: 156), then how much stronger the attraction of a sudden and entirely unexpected opportunity to move out of the lock-in and upward within the socioeconomic hierarchy? Perhaps the abruptness of what looks like an impending tip, leading Hinojosa to speculate on the basis of her findings that “If this tendency continues, the whole community will soon be bilingual and the children will begin to be socialized in Spanish” (1980: 38), is abrupt in onset and potential outcome, but not in gestation. This was certainly what I found to be the case in Gaelic East Sutherland, where the tip clearly took place during the nineteenth century, but the negative attitudes which had prepared the way for that tip could be traced within Scotland for at least six centuries and readily documented for Sutherland itself for a period of over 300 years. Just as the discovery of oil “opened” Juchitán to outside influences and the attendant pressures in favor of Spanish, so the construction of railroads, bridges, and roads in the early nineteenth century and the institution of schools toward the end of the preceding century “opened” East Sutherland to outside influences and to massive pressures disfavoring Gaelic and favoring English. Remoteness had buffered East Sutherland, as it had Juchitán; but with the loss of that remoteness, the buffering rapidly proved inadequate and centuries of distaste for the indigenous language made themselves felt. In the East Sutherland case, I tried to express this by suggesting that “suddenly, around the beginning of the nineteenth century, Britain came to Sutherland” (Dorian 1981: 51). It is the existence of a long lead-in period which in the end effectively belies the apparent abruptness of transmission failure in communities where a language outside the national linguistic mainstream seemingly turns up its toes

176

chapter 8

so dramatically after persisting with anomalous strength for so long. The failure of linguistic will under these circumstances is a measure of the potency of long-brewing negative pressures and the fragility of isolation as a buffer, since a serious breaching of that isolative buffer can produce such rapid decline in a previously resistant population. It’s possible that less isolated communities, with longer experience of compromise (for example, such a compromise as the linguistic domain-separation discussed above) have an advantage in survival potential precisely because they have had a prolonged period in which to learn to cope with pressures for linguistic assimilation. Metaphorically speaking, the more isolated linguistic groups may resemble North American Indian tribes or South Pacific island populations exposed to measles for the first time and carried off in disastrous numbers by the unfamiliar contagion. The measles virus was long in existence, but slow to reach them: when it did, they succumbed with terrible swiftness. In something of the same fashion the “virus” of hostility to non-mainstream languages may gather strength for a very long time and when it finally breaks through to an isolated community, carry the minority language off in an equally swift and deadly wave of social contagion, producing the phenomenon of linguistic tip. References Denison, Norman. 1971. Some observations on language variety and plurilingualism. Social anthropology and language, ed. by Edwin Ardener, 157–83. London: Tavistock Publications. ———. 1977. Language death or language suicide? International Journal of the Sociology of Language 12:13–22. Dorian, Nancy C. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. ———. 1982. Linguistic models and language death evidence. Exceptional language and linguistics. ed. by Loraine K. Obler and Lise Menn, 31–48. New York: Academic Press. Gal, Susan. 1979. Language shift: Social determinants of linguistic change in bilingual Austria. New York: Academic Press. Greene, David. 1972. The founding of the Gaelic League. The Gaelic League idea, ed. by S. O’Tuama, 9–19. Cork: Mercier Press. Heath, Shirley Brice. 1972. Telling tongues: Language policy in Mexico, colony to nation. New York: Teachers College Press. HilI, Jane H. 1973. Subordinate clause density and language function. You take the high node and I’ll take the low node: Papers from the Comparative Syntax Festival, ed. by

abrupt transmission failure

177

C. Corum. T. C. Smith-Stark, and A. Weiser, 33–52. Chicago: Chicago Linguistic Society. Hinojosa, Maria de la Paz. 1980. The collapse of the Zapotec vowel system. Penn Review of Linguistics 4:28–39. Huffines, Marion Lois. 1980. Pennsylvania German: Maintenance and shift. Inter­ national Journal of the Sociology of Language 25:43–57. Kloss, Heinz. 1977. The American bilingual tradition. Rowley, Massachusetts: Newbury House Publishers. Mertz, Elizabeth. 1980. Economic ideas and the motivation of linguistic shift: A Cape Breton case. Paper presented at the 1980 American Anthropological Association meetings. Pride, J. B. 1971. The social meaning of language. London: Oxford University Press. Pulte, William. 1973. Cherokee: A flourishing or obsolescing language? Language in many ways, ed. by William B. McCormack and Sol Wurm, 423–32. The Hague: Mouton. Timm, Lenora A. 1980. Bilingualism, diglossia and language shift in Brittany. International Journal of the Sociology of Language 25:29–41. Tsitsipis, Lukas D. 1983. Narrative performance in a dying language: Evidence from Albanian in Greece. Word 34:25–36.

chapter 9

Age and Speaker Skills in Receding Language Communities: How Far do Community Evaluations and Linguists’ Evaluations Agree? 1

Age and Receding-Language Speaker Skills: The Assumption and The Exception

Linguists working with receding languages in a language-shift setting have generally begun their research with the reasonable working assumption that age will show a correlation with proficiency: that the oldest remaining speakers will represent the pinnacle of proficiency for the community they belong to, and that the youngest continuing speakers will demonstrate some degree of reduced proficiency as compared with the most senior individuals (Voegelin and Voegelin 1977). This expectation is often born out, in the sense that the oldest speakers show measurably higher expression of some traditional features of the language than their juniors (Bavin 1989: 281, Dorian 1981: 114–51, Jones 1998: 79–80, Schmidt 1985: 26). These are very general age-group patterns, however, and the life circumstances of individual speakers may result in deviations from the pattern, as in several cases discussed by Paul Kroskrity among Arizona Tewa speakers: he profiles age-atypical speakers whose childhood or young-adult experiences produced either greater proficiency or lesser proficiency than was typical of their age group (Kroskrity 1993: 113–41). Among younger speakers in particular, relatively large differences in speaker skills are a familiar phenomenon as language shift takes hold and community-wide use of the ancestral language can no longer be taken for granted. Some younger individuals engage more than others with traditional cultural practices (as in two of Kroskrity’s cases). Some parents continue to use the receding language with their children longer than others do. How completely the younger children acquire the receding language may also be affected by the number of older siblings who have already brought the expanding language into the home. In cash economies based on wage labor some young people have jobs that position them among linguistically conservative senior community members while others spend more of their working lives among speakers of the expanding language.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�11

age and speaker skills in receding language communities

2

179

Factors that Come into Play in Assessment of Speaker Skills

Several things need to be kept in mind in assessing the relationship between age and speaker skills in receding languages. One is that even in monolingual communities individuals differ in their language skills. Normal individuals may all be proficient in terms of a general ability to create intelligible sentences, but some speak with greater clarity, effectiveness, and expressiveness than others. With receding languages, differences in the ease and effectiveness of verbal self-expression may have an effect on evaluation of fluency and proficiency, whether by fellow community members or by outsiders and whether the speaker is at the high end or the low end of the proficiency spectrum. Another is that the criteria linguists typically use in assessing proficiency do not necessarily match the criteria used by community members when they evaluate one another’s ability to speak well. Linguists, for example, have preoccupations that the speaker community shares only partially or not at all, such as paradigmatic conservatism (an entry for every possible category in a maximally traditional paradigm). Native speakers, especially those whose language variety is unwritten, recognize some facets of their language more easily than others. Features that they have no ability to describe in their own language are difficult for them to discern because they exceed what Silverstein 2001 [1981] has called the “limits of awareness”; ordinary speakers are unlikely to note ongoing changes connected with such features, while linguists who identify such overlooked features pay them great attention. Additionally, the speaker community is inclined to value control of cultural content more highly than linguistic researchers do. Expert knowledge of indigenous place-names is often greatly prized by local people, for example, but isolated place-names outside of discourse context have less interest for most descriptive linguists. 3

The Role of Local Ideology in Community Assessment of Speaker Skills

Conceptions of what is entailed in a “good” use of the local language and in “speaking well” are specific to particular cultural settings and need to be considered before the position of individual speakers in the local proficiency spectrum can be properly assessed. An exemplary treatment of the connection between age and speaker skills in a receding language is provided by Alexandra Aikhenvald in her account of the two remaining Tariana-speaking villages of the Vaupés region of Northwest

180

chapter 9

Amazonia (Aikhenvald 2001). Before she offers characterizations of individual speakers’ language skills in one of the villages, currently a language-shift setting, she first discusses the community’s own notions of desirably “correct” or “good” Tariana speech. She examines the strong constraints against borrowing in that multilingual and obligatorily exogamous setting and then details certain morphological, phonological, and morphosyntactic features that fall within the awareness of Tariana speakers, causing them to reject certain words and structures as unacceptably “mixed” speech. She reports that borrowing from any other language spoken in the region is severely condemned, but most especially borrowing from Tucano (an unrelated language increasingly dominant in the area), and that speech considered imperfect or incorrect is openly ridiculed: those who introduce borrowed elements into their Tariana are subject to laughter and scorn (Aikhenvald 2001: 423; see also Aikhenvald 2003: 129). Strong as the constraints against borrowing and mixing are, a few elements that can be identified as borrowings have made their way into Tariana and become established as nativized bound verb roots or verbal enclitics; they escape condemnation because in addition to their use by the current oldest generation they are known to have been in use among that generation’s predecessors as far back as memory reaches (Aikhenvald 2001: 418). Forms taken to reflect older-generation speech serve as the community model: “The forms attributed to the older generation are considered correct, good Tariana” (Aikhenvald 2001: 417). At the same time, however, the community recognizes individual variation in lexicon, phonology, and morphosyntax, and the oldergeneration model is one of a “critical mass” of speakers who favor particular forms (2001: 419–20). Traditionally, in this exogamous region, children derived their linguistic identity by acquiring the language of their fathers. But with transmission of Tariana falling off and use of Tucano expanding, the children of Tarianaspeaking fathers were no longer routinely becoming proficient Tariana speakers by the time of Aikhenvald’s research in the community. Knowledge of a good deal of Tariana lexicon did not count for much; individuals with substantial lexical knowledge but no conversational ability were ridiculed for their inability to speak the language and given a label translated by Aikhenvald as “those who cannot speak and can only call names” (2001: 421). Aikhenvald sums up the Tariana position on speaking well as follows (2001: 422): What is appreciated is the ability to maintain a conversation in Tariana without language mixing and to tell a long coherent story, especially a culturally significant one such as a story of the

age and speaker skills in receding language communities

181

wanderings of Tariana ancestors, or any version of the origin myth. A competent speaker is also expected to have a clear pronunciation (e.g., no or little post-tonic vowel reduction and careful pronunciation of aspirated consonants), to speak “like our grandfathers did,” and to avoid innovative Tucano calques as much as possible. . . . 3.1 Tariana Speaker Profiles: Generational Seniority as a Central Value With Tariana ideology about good speech clearly set forth, Aikhenvald’s profiles of some of the speakers with whom she worked are readily interpretable. Candi, over 70 years of age, “is considered the model of a traditional speaker of Tariana”. He speaks only Tariana to the prescribed kinfolk (his children, his younger brother, his classificatory brothers) and insists on Tariana in reply (2001: 422). His Tariana usage is “archaic in all respects” (ibid.). He is gifted as a storyteller and has a rich store of traditional lore and culture, including knowledge of place-names. Aikhenvald’s use of the passive “is considered” indicates that evaluation of Candi as “the model of a traditional speaker” is the community’s rather than her own. There has been no previous mention of archaism as a valued speech trait in itself, apart from the implied archaism/traditionalism involved in avoidance of mixing (2001: 412–419) and in wariness about using innovative forms (2001: 420), but the term reappears, each time with approbation, in subsequent profiles of Leo, Candi’s younger brother, and of Ame, the oldest living speaker at the time. Leo, in his early fifties, is praised for his storytelling abilities, his use of some archaic forms, and his avoidance of mixing elements from other languages into his Tariana, but his intermittent use of Tucano as well as Tariana to his sons and his regular use of Tucano to his daughters are noted as failings (2001: 422). In Leo’s case the phraseology makes it less clear that the praise and dispraise are strictly the community’s own, and this is also true in the sketch of Ame, who is reported to deviate from traditional practice by speaking only Tucano with most of his children (ibid.). But one telling critique of Ame clearly comes from the community: on hearing the elderly Ame inserting Tucano kinship terms into his Tariana, community members “generally proclaimed” him to be “ ‘not all there any more’ ” (ibid.). Yuse Paiphe, in his seventies and thus considerably older than Leo, speaks Tariana to his sons and grandchildren. He has “some cultural knowledge and is considered reasonably competent”, and he also uses some archaic forms; but his pronunciation is faulty in that he reduces most unstressed vowels (2001: 423). He “is said to” have grown up among speakers of a related language (Baniwa) with only a single evidential and never to have learned the use of the

182

chapter 9

fuller set of Tariana evidentials, for which he is faulted, apparently by the community in this case, too (ibid.). Dika, also in his seventies, speaks both Tariana and Tucano to his children, but his Tariana is “known to be hard to understand and frequently ungrammatical” (2001: 423). He uses younger people’s forms “often” and inserts a Tucano morpheme “now and again”; as a result “he is constantly made fun of and called . . . ‘useless’ ” (ibid.). The Tariana of the next generation shows certain recognizable deviations from that of their fathers, with structural intrusions diffused from East Tucano languages (e.g. possession marked with juxtaposition of terms instead of possession marked with prefixes) and with certain phonological or structural changes (e.g. reduction of –wa to –a, use of –hipe instead of –hipita as the classifier for ‘land’). Aikhenvald tabulates the deviations in question (2001: 417– 418) and recognizes on the basis of their use a “younger speakers’ ” or “younger people’s” Tariana (2001: 423–424). Candi’s eldest son Maye is “the best Tariana language expert” and “the most respected representative” of this first-descending generation, with a very clear pronunciation and “morphologically complex” speech (2001: 423). He “only occasionally” uses younger speakers’ forms, and he has “good lexical and cultural knowledge” (ibid.). Maye is actually just one year younger than his uncle Leo, but Leo as a member of Candi’s own sibling set (and of Candi’s generation in that sense) “has more authority” than Maye (ibid.) Candi’s eldest daughter Oli is a fluent speaker “with a comprehensive knowledge of the traditional kinship system, for which she is respected”, but she speaks “a typical younger speaker’s Tariana” and the occasional Tucano morpheme creeps into her Tariana (ibid.). Maye, Leo, and Candi correct her speech and “behind her back she is called Yasenisado ‘Tucano woman’ ” (ibid.). Gara, another of Candi’s sons, is “a good and witty storyteller”, but his Tariana is “full of Tucano calques and ungrammaticalities” (ibid.). Aikhenvald reports that “his speech is condescendingly referred to as Gara yarupe ‘Gara’s thing’ ” (2001: 424). Candi’s youngest son Ze, a trained schoolteacher, is “very fluent in younger people’s Tariana” but aspires to more traditional speech: Aikhenvald describes him as “desperately trying to make his speech as archaic as possible” and as “the greatest partisan of the correct Tariana spoken ‘the way our fathers speak it’ ” (2001: 424). His effort to become a language authority “is appreciated by the community” even though his practice falls short (ibid.). Yuse Paiphe’s children are all fluent in younger speakers’ Tariana, but only his son Ñu “is considered ‘reasonable’ ” as a speaker, with a clearer pronunciation than the others (unreduced vowels and aspirated consonants). Ñu’s older brothers Saba and Kiri have reputations as “ ‘deficient speakers’ ”, their stories

age and speaker skills in receding language communities

183

full of Tucano calques, their speech lacking the traditional person marker and without the full set of Tariana evidentials: “their lack of language competence is constantly commented upon” (2001: 424). Yuse Paiphe’s daughters, like Candi’s daughter Oli, are “frowned upon as ‘Tucano women’ ” (ibid.), evidently for the same sort of Tucano influence in their Tariana. Dika’s sons Emi and Raimu both use younger speakers’ forms and insert Tucano morphemes “occasionally”. Emi, the elder, is “acclaimed as one of the best storytellers in the village”, but Raimu “does not feel confident enough to tell a full story in Tariana”. Emi and Raimu, like Saba and Kiri, are considered deficient speakers: Aikhenvald reports that “their inability to ‘speak correctly’ is constantly lamented by other members of the community” (2001: 424–25). These and other profiles offered by Aikhenvald are consistent with her depiction of the Tariana concept of “speaking well”, and overall it is clear that she intends to convey the community’s judgments of the speakers in question. Her portraits leave no doubt that age in the sense of generational membership has a bearing on speaker status, and her identification of particular linguistic features that distinguish a more traditional form of Tariana from a younger people’s Tariana demonstrate that age has a bearing on language structure as well, even though not all older-generation speakers represent traditional Tariana equally well (Yuse Paiphe and Dika falling particularly short of the ideal for their generation). Menomini Speaker Profiles: Age and Generational Seniority as Lesser Values Aikhenvald states that she is offering her sketches of proficiency and speaker status “in the spirit of Bloomfield’s (1927) linguistic profiles of his Menomini consultants” (2001: 421), but in actuality her renderings are considerably more interpretable than his. Bloomfield’s interest had been caught by the fact that members of an illiterate speech community proved to be just as capable of distinguishing good speech from bad as literate speakers of a standardized language like English. The Menomini Indians of Wisconsin formed a “compact tribe” of about 1,700 people at the time of Bloomfield’s writing; there were no dialectal differences and they did not write their language (Bloomfield 1964[1927]: 394). Bloomfield offers “sketches” of “the linguistic position” of some of the Menomini speakers whom he knew best, characterizing various aspects of their speech and noting languages other than Menomini that they were familiar with (ibid.). The sketches he offers were presumably intended to represent the community’s viewpoint, since the aim of his paper was to document the fact that Menomini community members “will say that one person speaks well and another badly, that such-and-such a form of speech is incorrect 3.2

184

chapter 9

and sounds bad, and another too much like a shaman’s preaching or archaic (‘the way the old, old people talked’)” (ibid.). But Bloomfield notes that he was surprised to find that despite his own relatively slight acquaintance with the language he was able to share in the judgments that native speakers made, just as a foreigner could come to share such judgments about English speakers’ usage: “[I]f he listened to us long enough and . . . fortune favored him”, he would pick up the social values (“normal good form”, “deliberate and elevated”, “unidiomatic, vulgar, pedantic”) that English speakers attach to competing but semantically equivalent structures (ibid.). The individual whose Menomini speech Bloomfield rates most highly is a woman in her sixties, Red-Cloud-Woman, who “speaks a beautiful and highly idiomatic Menomini”; fluent also in Ojibwa and Potawatomi, she may have spoken a little Winnebago as well but “knows only a few words of English” (1964: 395). Her husband, Storms-At-It, speaks Potawatomi as well as Menomini but knows no English at all. In Menomini he “often uses unapproved, – let us say, ungrammatical, – forms which are current among bad speakers”, yet he also on “slight provocation” shifts “into elevated speech, in which he uses what I shall describe as spelling-pronunciations, together with long ritualistic compound words and occasional archaisms” (ibid.). A man in his fifties, Stands-Close, son of a man known as “an oracle of old traditions”, speaks a “well up to standard” Menomini “though less supple and perfect than Red-Cloud-Woman’s”; but it is “interlarded with words and constructions that are felt to be archaic” (ibid.). Little-Doctor, in his sixties, had huge vocabularies both in English and in Menomini. In English he had “a passion for piling up synonyms” and his vast Menomini vocabulary made him an explicator of rare words to fellow-speakers. Bloomfield reports that “in both languages his love of words sometimes upset his syntax, and in both languages he was given to over-emphatic diction” of the spelling-pronunciation type (ibid.). At the very low end of the proficiency scale Bloomfield places WhiteThunder, a man of about 40, who speaks “less English than Menomini”, a serious condemnation in view of the fact that Bloomfield describes his Menomini as “atrocious”: “his vocabulary is small; his inflections are often barbarous; he constructs sentences of a few threadbare models” (ibid.). Because Bloomfield states that he is able to make the same judgments that Menomini community members make, it is not always clear whether the specific characterizations he offers are his own or the community’s unless, as in Stands-Close’s case, he uses a passive construction like “are felt to be archaic”. Unlike Aikhenvald he does not provide directly quoted renderings of

age and speaker skills in receding language communities

185

c­ ommunity judgments except in the one instance where he equates archaic with “the way the old, old people talked”, nor does he discuss the community’s linguistic ideology. The value of the terms “archaic” and “archaism” is particularly unclear in Bloomfield’s characterizations, suggesting overall more a negative than a positive value. (Describing an individual’s speech as “interlarded” with archaic words and constructions, for example, as Bloomfield describes Stands-Close’s speech, does not create a positive tone.) It does at any rate become clear by his characterizations of three particular speakers, Red-CloudWoman, Bird-Hawk, and Little-Jerome, that age does not correlate in any automatic way with superior speaker skills in Menomini. Red-Cloud-Woman, whose Menomini speech receives Bloomfield’s highest praise, is only in her sixties. Bird-Hawk, “a very old man” who had died by the time Bloomfield’s article was written, was either monolingual in Menomini or possibly spoke a little Ojibwa as well, but according to Bloomfield he spoke, “as soon as he departed from ordinary conversation”, “with bad syntax and meagre, often inept vocabulary, yet with occasional archaisms” (1964: 395). (Bloomfield does not specify the register Bird-Hawk was trying to use in such departures from ordinary conversation, but it seems likely that an elevated register was the target.) As a description of an apparently frequent type of speech performance from what was quite possibly a monolingual individual and certainly a very elderly man, this is a remarkably negative evaluation, and in this case especially it would be useful to know the extent to which this was a community-wide assessment. By contrast Little-Jerome, only in his fifties, is said to be “a true bilingual” who speaks both English and Menomini “with racy idiom, which he does not lose even when translating in either direction” (ibid.). Some of the terms that Bloomfield uses in specifying certain features of “bad” Menomini make it clear that not all such features can be correlated with age. He notes for example the phonological distortions of “older bad ­speakers”, and also the “over-elegant” speech of shamans, most of whom would presumably have been in the older and culturally more traditional half of the tribal population. Among the phonological faults that Bloomfield catalogs, he associates only anglicized pronunciations with younger people in particular. Difficulty using the obviative inflection and failure to use the quotative verb form in story telling are described as failings of the “bad” Menomini speaker but are not linked specifically with younger speakers in Bloomfield’s account. Overall, then, there does not seem to be the clear-cut connection between age and speaker skills, or between age and the reputation of being a “good” speaker, that Aikhenvald encountered in the Vaupés area of Northwest Amazonia.

186 4

chapter 9

Age-Based Proficiency Groupings and Individual Exceptions in East Sutherland

In ongoing work with the Gaelic dialect spoken by fisherfolk and their descendants in the village of Embo, on the east coast of the northern mainland county of Sutherland in Highland Scotland, I have had occasion to refine my original age- and proficiency-based speaker categories, developed on the basis of fieldwork in the 1960s and ’70s, in the light of resumed fieldwork in the 1990s and the first decade of the 21st century, and to attach less blanket importance to the role of age as a correlate of proficiency (Dorian 2010). This change was prompted above all by two conspicuous instances of discrepancy between my assessment of speaker skills, as a linguist focusing especially on grammatical conservatism, and the community’s own assessment of speaker skills, reflecting a different conception of “speaking well”. Material collected in Embo in the ’60s and ’70s showed a surprisingly large amount of change in progress in Embo speakers’ Gaelic, with older speakers by and large using more traditional grammatical forms (and to a lesser extent also phonology) than younger speakers, within the ranks of the fully fluent, and with a final group of imperfect speakers maintaining a limited use of Gaelic with noticeably deviant grammar and phonology. The community itself recognized these imperfect speakers, although they had no special term for them and even though the group members differed from one another in grammatical intactness and level of fluency. I termed these imperfect speakers semispeakers in order to distinguish them from the fully fluent (Dorian 1973: 417).1 The community did not recognize an older/more traditional vs. younger/less traditional division among fluent speakers, but my particular speaker sample happened to divide rather neatly between a group of speakers among whom a linguistically well respected same-age pair were the youngest, on one side of the divide, and a group of fluent speakers among whom the oldest were six years younger, on the other. In view of the age divide I termed the older part 1 The term semi-speaker does not necessarily translate well to other speech communities. As noted, the East Sutherland Gaelic semi-speaker group includes a variety of levels of fluency and proficiency, so that I have sometimes distinguished among high-proficiency, middleproficiency, and low-proficiency semi-speakers in characterizing degrees of grammatical intactness. The characteristics shared by all members of the group go beyond the presence of considerable grammatical deviation to the conspicuous use of certain strongly disfavored elements in their Gaelic and the absence of any conversation partners with whom Gaelic is their default language. Evaluation of speaker skills and the semi-speaker concept/category are discussed more fully in Dorian 2010.

age and speaker skills in receding language communities

187

of the sample older fluent speakers (OFSs) and the remainder younger fluent speakers (YFSs). In describing the differences between the Gaelic of these two fully fluent groups I then tabulated the data that reflected grammatical change in progress in terms of group-based characteristics. Taken as groups, the OFSs and YFSs were clearly distinguishable by their different degrees of grammatical conservatism (see Dorian 1981: 147, Table 7). As in the Tariana case, that is, it was possible to identify a more conservative older speakers’ fluent Embo Gaelic and a somewhat less conservative younger speakers’ fluent Embo Gaelic. But in one paper arising from that early fieldwork I noted the anomalous finding that in extensive translation-task testing the Gaelic of one young man considered fully fluent by the community resembled that of his OFS mother, 29 years older than he, mainly in one particular respect, namely the preservation of irregularities in four grammatical structures, while in a good many other respects his grammar more nearly resembled that of a semi-speaker sister only one year younger than he (Dorian 1977).2 Continuing fieldwork in the 1990s (conducted by tape recorder and telephone, since health difficulties precluded a return to the field site)3 confirmed my earlier view that “Seònaid”, the youngest individual among the OFS group, enjoyed a high reputation as an Embo Gaelic speaker and that the young man “Rory” was considered a fully fluent speaker, while his one-year-younger sister “Elsie” was not, in spite of the similarities in the siblings’ grammar apart from the different degree to which they preserved the four irregular structures. But ongoing fieldwork also confirmed that Seònaid was less conservative in regard to certain grammatical structures than most of her OFS peers, and that Rory was more deviant in certain grammatical structures than most other YFSs and even than some semi-speakers. That is, again as in the Tariana case, the

2 Translation-task testing was unusually easy and effective with fluent East Sutherland Gaelic speakers. In the fisherfolk communities the routine practice in producing reported speech was to use whichever language a conversation partner favored. With language shift well underway, all remaining Gaelic speakers had younger kinfolk, and sometimes neighbors, who were English monolinguals and for whose benefit remarks originally made in Gaelic were regularly “quoted” in English; the same was true in reverse of remarks originally made in English, which were “quoted” to fellow Gaelic speakers in Gaelic (Dorian 1997). In these small communities remarks of high interest were constantly passed along in the course of daily conversation, and the frequency of translation in both directions seemed to have made all speakers proficient translators. This was less true of semi-speakers, but the Embo semispeakers in particular (as compared with semi-speakers among the fisherfolk speakers of Brora village) proved surprisingly adept at translation and adventurous about undertaking it. 3 Telephone conversations were recorded with the permission of the other speakers.

188

chapter 9

usage of some individual fluent speakers embodied conservative linguistic norms more fully than others. The Community and the Linguist in the Assessment of Speaker Proficiency Level There were certainly some typically conservative OFS grammatical features in the Embo Gaelic spoken by Seònaid. She preserved the vocative case at a 100% level (28 of 28 instances), for example, and irregular adjective comparison also at 100% (10 of 10). Her negative imperative constructions were 94.7% traditional (18 of 19), and her feminine pronoun replacements were 93.8% traditional (15 of 16). Rory by contrast preserved vocative marking only 23.1% of the time (3 of 13, a lower retention rate than two of four semi-speakers), irregular adjective comparison 66.7% of the time (6 of 9), and like two of the semispeakers but none of the other YFSs he used no traditional negative imperatives at all (0 of 5).4 Also in sharp contrast to Seònaid, Rory retained no use at all of feminine pronoun replacement (0 of 8), but this particular departure from conservative grammar he shared both with nearly all of his fellow YFSs and with two of three semi-speakers. (His semi-speaker sister Elsie was exceptional among younger speakers, and especially among semi-speakers, in producing one feminine pronoun replacement in 13 opportunities, or 7.7%; this was one of four structures in the use of which her Gaelic was slightly more conservative than Rory’s.) In spite of her conservatism with regard to the grammatical structures just noted, Seònaid showed surprisingly low retention of other structures that most OFSs preserved well, or at least better than she. Nine other OFSs preserved the distinction between locational and directional adverbs at a level of 90% or better, but Seònaid only at a 47% level (31 of 66). Use of the traditional synthetic inflection for the 1st-person conditional was in the 90–100% range for three other OFSs, in the 80–89% range for two more, and at 66.7% for one, but Seònaid used only 32% (16 of 34). Seònaid also used the conservative structure for complex prepositions with pronouns at only a 50% level (22 of 44). In this last case her low level of use was actually exceeded by that of one of her fellow OFSs, but five others had higher levels ranging from 100% to 66.7%. Rory showed still lower use of the distinction between locational and directional adverbs (14.3%, 3 of 21); although his was not the lowest retention level among YFSs (two had lower levels still), it was nonetheless lower than that of 4.1

4 Rory died suddenly in his sixties early in 1993, whereas Seònaid died in her nineties early in 1999. The fact that she outlived him by six years during the latter phase of my fieldwork accounts for the fact that her token counts are considerably higher than his.

age and speaker skills in receding language communities

189

three semi-speakers, including Elsie. Like three of four semi-speakers he used no conservative structures at all for complex prepositions with pronouns (0%, 0 of 6), whereas no other YFSs who provided more than two instances of that structure showed less than a 15% level of traditional forms and four preserved them at 50% or better. The 1st-person-singular inflection (a lone synthetic inflection, whereas all other conditional forms are analytic in structure) was one of the four grammatical irregularities that Rory preserved unusually well: he actually exceeded OFS Seònaid (as well as every fellow member of the YFS group) in retention of the synthetic 1st-person-singular conditional inflection (90.9%, 10 of 11 instances). With these particulars noted it becomes clear that while age-group profiles corresponding to relative degrees of grammatical conservatism have a certain overall validity, individual speakers do not reliably conform to the group profile. It is also clear that community evaluation of speaker capabilities is not based on consistent grammatical conservatism. Seònaid demonstrated atypically low levels of vocative retention, locational vs. directional adverb distinction, and conservative complex prepositional constructions with pronouns for her age-group, but her reputation was that of a superior high-proficiency speaker of Embo Gaelic. Rory demonstrated lower retention of locational vs. directional adverb forms than several semi-speakers, less retention of vocative marking than two of four semi-speakers, and the same complete absence of conservative complex prepositional constructions with pronouns and of traditional negative imperatives as members of the semi-speaker age-andproficiency group, yet his reputation in Embo was that of a fully fluent speaker. Seònaid’s strong reputation is understandable in terms of some of the same criteria Aikhenvald found to be operative among the Tariana. First of all she was in fact a little older than all of the Embo YFSs in my sample (six years older than the eldest of them). More importantly still, she had a superior knowledge of some kinds of cultural material (proverbs especially, but also rhymes and ditties as well as by-names and the kin lines they represented).5 Finally, she 5 The small number of surnames in Embo village, together with traditional naming practices that resulted in repetition of certain given names across adjacent generations of any family, gave rise to multiple instances of identical official names within the village. The official names were of little practical use, consequently, and individuals were identified instead by by-names. Superficially these resemble nicknames, but because a great many of them had humorously mocking elements they were used only in reference and not in address. By-name elements were frequently shared by some or all of an individual’s children and often carried over into a third generation as well. They therefore served to identify lines of descent and relationship, just as family names would in a community with a wider selection of surnames.

190

chapter 9

was a gifted anecdotalist, a fount of comic stories based on incidents in her own and others’ lives, and her ability to keep the company entertained by her verbal skill was much appreciated in the village. Rory’s reputation as a fully fluent speaker certainly did not rest on the conservatism of his Gaelic, and he was a relatively shy man, not the conversationalist or entertaining storyteller that Seònaid was by any means. But he had one attribute very highly valued in Embo, namely language loyalty. More than any other individual close to him in age, Rory showed a preference for Gaelic. He was fluent in English as well as Gaelic, but his self-presentation was easier and more confident in Gaelic. With any conversation partner at all who offered the possibility of either Gaelic or English, Rory opted for Gaelic. The fact that he was heard to speak Gaelic at every opportunity, and that he spoke it comfortably, without disfluencies of any kind, established him as a fully fluent speaker in Embo terms regardless of the grammatical forms he used. Elsie, though only a year younger than Rory, had no conversation partners with whom she routinely spoke Gaelic (as was equally true of all the other semi-speakers). She was primarily a short-burst speaker of Gaelic, and her overall preference for using English was apparent: she did not always reply in Gaelic when spoken to in Gaelic, a breach of local courtesy rules. Limited use of Gaelic was not by itself a determinant of imperfect speaker status, but when as in Elsie’s case it coincided with an evident preference for English and with use of a pair of structural deviations, one phonological and one grammatical, that puristically inclined fluent speakers considered stereotypical of “bad” Gaelic, it prompted those more puristic fellow villagers to consign her to a less-than-fully-fluent speaker category. 5 Conclusion Direction of structural change – movement from synthetic to analytic structure, merger or loss of grammatical categories, altering or fixing of word order, and the like – can often be detected on the basis of age-related speaker categories such as the OFS, YFS, and semi-speaker categories I originally applied in formulating a speaker typology for receding East Sutherland fisherfolk Gaelic. But the utility of age-and-proficiency groupings in identifying directionality in change processes should not obscure the fact that individual speakers deviate, sometimes to surprising degrees, from the general profile for their agegroup, and such groupings will not in any case encompass all the criteria that the community itself applies in assessing speaker skills. The brief individual sketches offered by Aikhenvald and Bloomfield, like the fuller profiles provided

age and speaker skills in receding language communities

191

by Kroskrity, are useful correctives in this respect. The value of the briefer sketches is greatest, however, if, as in Aikhenvald’s Tariana case, community language ideology is fully considered and local views of speaking well are clearly identified. This sort of consideration has proved essential in the Embo case to explaining the sharp difference in the way age-adjacent brother and sister Rory and Elsie are evaluated as speakers by their own speech community, and it may prove to be important more generally in revealing differences between the criteria linguists apply and the criteria community members apply in appraising individuals’ ability to speak well. Linguists may therefore need to be more cautious about attaching importance to the role of age as a correlate of proficiency. References Aikhenvald, Alexandra Y. (2001). Language awareness and correct speech among the Tariana of Northwest Amazonia. Anthropological Linguistics 43:411–430. ——— (2003). Teaching Tariana, an endangered language from Northwest Amazonia. International Journal of the Sociology of Language 42:125–39. Bavin, Edith L. (1989). Some lexical and morphological changes in Warlpiri. In Investigating Obsolescence, Nancy C. Dorian (ed.), 267–286. Cambridge: Cambridge University Press. Bloomfield, Leonard ([1927] 1964). Literate and illiterate speech. In Language in culture and society: A reader in linguistics and anthropology, Dell Hymes (ed.), 391–396. New York: Harper and Row. Dorian, Nancy C. (1973). Grammatical change in a dying dialect. Language 49:413–438. ——— (1977). The problem of the semi-speaker in language death. International Journal of the Sociology of Language 12:23–32. ——— (1981). Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. ——— (1997). Telling the monolinguals from the bilinguals: Unrealistic code choices in direct quotation within Scottish Gaelic narratives. International Journal of Bilingualism 1:41–54. ——— (2010). Investigating variation: The effects of social organization and social setting. New York: Oxford University Press. Jones, Mari C. (1998). Language obsolescence and revitalization: Linguistic change in two sociolinguistically contrasting Welsh communities. Oxford: Oxford University Press. Kroskrity, Paul V. (1993). Language, history, and identity: Ethnolinguistic studies of the Arizona Tewa. Tucson and London: The University of Arizona Press.

192

chapter 9

Schmidt, Annette (1985). Young people’s Dyirbal: An example of language death from Australia. Cambridge: Cambridge University Press. Silverstein, Michael. ([1981] 2001). The limits of awareness. In Linguistic anthropology: A reader, Alessandro Duranti, ed., 382–401. Blackwell: Oxford and Malden, MA. Voegelin, C. F., and Voegelin, F. M. (1977). Is Tübatulabal de-acquisition relevant to theories of language acquisition? International Journal of American Linguistics 43:333–338.

chapter 10

Linguistic Lag as an Ethnic Marker1 It is commonplace in the body of literature on ethnicity to find language identified as one of the chief markers of ethnic identity (e.g., Peterson 1975: 178; De Vos 1975: 15; Giles, Bourhis & Taylor 1977: 325; Chapman, Smith & Foot 1977: 141). There are even cases reported in which language seems to be almost the sole marker of ethnic identity. Jackson (1974: 57) describes something of the sort for a complex multilingual setting in the southeastern Colombia rainforest: Another point is that in the Vaupés, language is by far the most important marker distinguishing language-aggregates and their members. It is primarily the Bará language which all Bará Indians share and which separates them as a category from Indians affiliated to other languages. In most other multilingual situations which have been reported on, language is but one of several such markers, others such as physical characteristics, dress, differences in technology, eating patterns, etc., being of equal or greater importance, at least in the eyes of the natives. As stated above, formal language affiliation in the Vaupés is determined by membership in a named patrilineal descent group, which also confers the right to manufacture certain ceremonial artifacts (that is limited to adult men) and to use various chants and names associated with the language in its role of father-language. No other differences exist which coincide with language-aggregate membership, regardless of whether one is looking for markers used by the Indians themselves to classify one another or looking for more subtle differences the Indians may not be aware of or choose not to acknowledge. So profound is the connection between language and ethnicity that it is possible to find a people using a language which few of them actually speak as a symbol of their separate ethnic identity, as in Ireland (De Vos 1975: 15). There are, to be sure, counter-cases. Hymes (1966: 126), citing Hohenthal and McCorkle 1 The research reported in this paper was supported by grants from Bryn Mawr College (1974), the American Philosophical Society (1976), and the National Science Foundation (1978, grant BNS 77-26295). I am grateful to Dell Hymes and Michael Silverstein for comments on an earlier draft of this paper.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�12

194

chapter 10

(1955), contrasts two South American Indian groups in terms of retention of language and retention of identity: The Fulnio of Brazil have given up their lands several times during the last three centuries, moving in order to preserve their language and annual religious ceremony (to which proper use of the language is essential) as basis of their identity. In contrast, the Guayqueries of islands off Venezuela seem to have given up their aboriginal language and native religion so early that no trace remains. They have preserved their identity on the basis of a special socio-economic structure, as demonstrated by the fact that a small group who abandoned the special structure are today indistinguishable from other Venezuelans. Language, together with religion, has served as separatist and unifying function in the one case, but not in the other. Despite such counter-cases, the expectation continues to be that language and ethnicity will show a strong correlation, as in the following quotation from Fishman (1977: 25): Anything can become symbolic of ethnicity (whether food, dress, shelter, land tenure, artifacts, work[,] patterns of worship), but since language is the prime symbol system to begin with and since it is commonly relied upon so heavily (even if not exclusively) to enact, celebrate and “call forth” all ethnic activity, the likelihood that it will be recognized and singled out as symbolic of ethnicity is great indeed. It would seem that the fisherfolk of coastal East Sutherland, in the far north of the Highland Scottish mainland, are simply one more such case of a people preserving their ethnicity largely through the use of a distinctive language. Up to at least World War I, the fisherfolk were distinctive in a rich variety of ways. Their occupation was itself distinctive, and they (especially the men) had a distinctive style of dress which accompanied it. Because of the way they earned a living, their diet was different from that of the rest of the population; they not only ate fish with greater frequency, but they also ate more varieties of fish than other people, and they had a number of specially-prepared fish dishes peculiar to themselves. Residentially the fisherfolk were almost wholly segregated. Two of the villages concerned had a separate district where the fishers lived (“Lower Brora” as opposed to “the upper village” in Brora; “the West End” as opposed to “the East End” in Golspie); the third, Embo, was a satellite settlement, composed wholly of fisherfolk, to the nonfishing village of Dornoch. All

linguistic lag as an ethnic marker

195

three of these fisherfolk residential areas were known, unofficially and impolitely, as “Fishertown”. The fisherfolk were also entirely endogamous. Before World War I, marriages between fisherfolk and other groups in the population were almost unheard-of. Today nothing remains of these patterns but some residential clustering (very weak now in Brora and Golspie, quite strong still in Embo). But the former fisherfolk – who no longer fish at all – are still highly distinctive in terms of their language: a local variety of Scottish Gaelic is their mother tongue, though all are also fluent in English. Even younger speakers whose control of Gaelic is imperfect (“semi-speakers”: Dorian 1973, 1977) explicitly claim Gaelic as their mother tongue despite the fact that they are more proficient in English. They define “mother tongue” temporally, as the language used in their homes in infancy and childhood, and express no doubt as to the claim of Gaelic to mother-tongue status. The ethnic identity “fisher” was a severely stigmatized one in East Sutherland at the time when the current bilinguals were growing up. It is clear that they understood the connection between speaking Gaelic and being identified as a “fisher”, since the automaticity of that connection was a factor in the abandonment of Gaelic by some individuals: I think myself, as the children from Lower Brora got older, they . . . were ashamed to speak the Gaelic, in case they would be classed as a – a fisher. (Brora bilingual, 1974) At the present time the connection is wholly warranted: any fluent speaker of any variety of East Sutherland Gaelic is “of the fisherfolk”. There are no exceptions. Language and a residual ethnicity are perfectly correlated. And apart from the remaining residential patterns, which are very weak in two of the villages, there are no other markers of fisherfolk ethnicity.2 The perfect correlation between language and ethnicity works only synchronically, however. As soon as one goes back a bit in time, the picture becomes complicated by the fact that other segments of the population also spoke Gaelic in the past. The crofters, sub-subsistence agriculturalists who lived in the country districts around the villages we are concerned with, were still fairly strongly Gaelic-speaking into the period between the two World Wars; the last Gaelic-speaking crofters died only in the 196os. Earlier still, 2 In a few homes, some old-style fisherfolk dishes are prepared, but even this small difference in cuisine is not general, and nowadays people of fisherfolk descent eat no more fish than does the rest of the population.

196

chapter 10

before World War I, some Gaelic speakers remained even in the villages themselves. Yet despite the presence of other Gaelic-speaking groups, informants continue to talk as though Gaelic were important to fisherfolk ethnicity in the early years of this century. The information one gleans can be confusing and contradictory. In one interview about the roots of the prejudice against “fishers”, a septuagenarian Golspie native, monolingual in English, identified the fisherfolk’s Gaelic as a possible source of hostility, and then in the next breath denied it, when I pressed the issue, on the very reasonable grounds that other people, towards whom there was no prejudice, spoke Gaelic, too: English monolingual: I don’t know just what it was. Was it people despised them for their Gaelic, and then some of them weren’t very clean, y’ know. Some of them. I don’t know just what it was. . . . Investigator [a few seconds later]: You say they might have been despised for their Gaelic. English monolingual: Oh, no, I don’t think for the Gaelic. I don’t think so, because a lot of people that I knew in the East End had Gaelic, y’ see? My own mother had it, and a lot more but her had it. No, it wasn’t that. (Golspie monolingual, 1976) After bringing up fisherfolk Gaelic on her own, without any prompting from me, she then categorically denied her own suggestion when confronted with it. The confusion experienced by this informant clears away, however, if one accepts that she was completely correct in identifying a linguistic component to fisherfolk ethnicity, and also accepts that that linguistic component did not consist of the Gaelic language as such. That is, language can be an important part of ethnicity without taking the form of the presence or absence of some particular language or dialect. In East Sutherland this becomes clear in a diachronic view of the situation. As far back as 1897, the fisherfolk of the area were being singled out for a distinctive linguistic behavior. The writer quoted below was a Gaelic scholar who was comparing the relative purity, in the sense of lack of adulteration by English loanwords, of the eastern, northern (Reay Country) and western (Assynt) dialects of Sutherland: Comparing these three sub-dialects then, it will be granted that the language of the people of Machair-Chat, as the low-lying east coast of Sutherland is called, is less pure than that of the Reay Country; and the dialect of the latter is less pure than that of Assynt. We may except one

linguistic lag as an ethnic marker

197

or two fishing communities from this comparison; for example, Embo, and Brora, whose natives can express themselves in fairly good idiomatic Gaelic (Gunn & McKay 1897: 177) The fishing communities were out of line linguistically, not in the sense that they spoke a language which the rest of the population did not speak, but in the sense that they spoke it differently (in this case, better). Here we have what can be recognized as a first form of linguistic lag. Correlated with the relative excellence of the fisherfolk in Gaelic was a relative weakness in English. The expansion of the English language into East Sutherland was evident by 1841, the date of the second Statistical account of Scotland. The Statistical accounts were sketches of each parish in the country, in which the parish minister dealt with a variety of subjects such as topography, climate, population, economics, and education. The first Statistical account of Scotland appeared in 1793, and there was no mention whatever of language by the ministers of the three parishes where Brora, Golspie, and Embo are located. The second Statistical account, published 48 years later, contained comment by all three of the ministers from these same parishes on the rise of English; two of the three even predicted the death of Gaelic in the area, so vigorous was the expansion of English. The reasons for this need not concern us here (see Dorian 1978: 6–10 for the historical and social causes). We can take the comment of the Golspie parish minister as typical: Forty years ago, the Gaelic was the language generally spoken in the parish. But . . . that language is now fast on the decline; and among the young there is now hardly an individual who does not understand and speak English (Statistical account of Sutherlandshire 1841: 35). The Golspie minister might, like Gunn, have excepted the fishing communities, since Gaelic monolingualism was still to be found among the fisherfolk into the twentieth century. An Embo septuagenarian said in 1976: “My mother couldn’t speak English. Couldn’t speak any English. Very few words.” Pre-school monolingualism also lingered late, most especially in Embo. Here are two Embo bilinguals in their mid-sixties discussing the subject in 1978: 1: We didn’t have a word of English [before going to school]. 2: (her husband): None of us had any English. We didn’t know – 1: We – if anybody spoke English to us, we’d run away a mile. We wouldn’t know what it was. 2: Just like a foreign –

198

chapter 10

1: Oh, it was like a foreign language to us. 2: Well, it was a foreign language. 1: It wasn’t a word of English being spoken in any of the houses in Embo. The last instances of pre-school monolingualism in Embo date from the 1940s, a full century after the Golspie minister asserted that bilingualism was the norm. Here again the fisherfolk were out of line linguistically, continuing to speak only Gaelic when all other segments of the population knew English, or both English and Gaelic. This represents a second form of linguistic lag. It is undoubtedly as early as the “Gaelic purity” lag; both may be assumed to have begun in the second half of the nineteenth century. Naturally enough, the second form of lag led to deficiencies in English relative to the rest of the population when the fisherfolk did become bilingual: And, their English was very much poorer than ours. . . . [By-name of a fisher woman], she used to come into the shop, and she used – ‘How much are your tuppenny sponges?’3 Y’ see? . . . They would hear these [English] words, y’ see, and – and they didn’t really know. (Golspie monolingual, 1976) Many present-day bilinguals tell amusing stories about the difficulties they had with English as youngsters and laugh at the “howlers” that they came out with through insufficient experience with English. One could even acquire a temporary by-name thanks to an English blunder. One Golspie bilingual, now a septuagenarian whose English is fluent, was for years called “That’s-a-dog-of-me” by a local shopkeeper because of the Gaelic-patterned English she mustered as a young girl to claim the dog that the shopkeeper was trying to chase out of his shop. Again we see that the fisherfolk were linguistically distinct, this time for their flawed English when all the rest of the population was thoroughly proficient in English. This represents a third form of linguistic lag. Finally, at the present day, a fourth form of linguistic lag appears in an anachronistic GaelicEnglish bilingualism which characterizes the fisherfolk when all other segments of the population are monolingual in English. The pattern seems clear. For a hundred years or more, the East Sutherland fisherfolk have had a linguistic component to their ethnic identity, but it has not consisted of the use of any particular language or dialect as such until very recently. Rather, their 3 Taking “tuppeny” as a kind of sponge, but not knowing that it was itself an answer to the question, “How much?”

linguistic lag as an ethnic marker

199

linguistic distinctiveness has consisted in a marked linguistic lag, such that they were always out of step linguistically with the surrounding population. The lag has taken a variety of forms over time: a Gaelic that was unusually pure; a monolingualism that was peculiar to them; an imperfect English at a time when mastery of English was the local norm; and finally, persistent bilingualism in an otherwise monolingual population. Whatever the linguistic norm at a given point in time, the fisherfolk have been deviant in terms of that norm. The social separateness of the fisherfolk, marked though it was, was not of sufficiently long standing to produce a notably different form of either Gaelic or English; the ancestors of the East Sutherland fisherfolk were settled in their “fishertowns”, from nearby areas, only from about 1810. Prior to that time they did not fish and were not a socially separate group. Though there were, by the mid-twentieth century, some minor differences between “crofter Gaelic” and “fisher Gaelic” (Dorian 1978: 145–47), all forms of Gaelic in East Sutherland were clearly one as opposed to, say, the northern and western dialects of Sutherland Gaelic. The English of the fisherfolk, though accented somewhat by Gaelic, is also of a generalized East Sutherland type. What the social separateness of the fisherfolk produced was not a separate dialect, but the linguistic lag we have been following through its various forms and stages. The “ideal” correlation between language and ethnicity which marks fisherfolk identity at the present time made its appearance as recently as the beginning of the 1970s (though it will last out the present century). In the view both of bilinguals within the fishing communities and of monolinguals outside them, the fisherfolk have always shown some difference in linguistic behavior as compared with the rest of the population. This difference is usually expressed in terms of “poor English” and “speaking Gaelic”, and correlates with a lesser social standing. With the help of the documentation provided by the mid-nineteenth-century second Statistical account and by Gunn’s latenineteenth-century appraisal of eastern Sutherlandshire Gaelic, and then by the first-hand testimony of twentieth-century monolinguals and bilinguals, the difference in linguistic behavior can be seen to have been a time-differential, or lag, in which the fisherfolk showed at a markedly later date linguistic behavior which had been characteristic of most of the rest of the population at an earlier time. The case of the East Sutherland fisherfolk is noteworthy as an extension of the concept of the language component of ethnic identity. A marked linguistic lag can operate as a perfectly adequate marker of in-group membership even when the language spoken is not peculiar to the group in question. The lag can be either positively valued or negatively valued and still be effective. Any Gaelic speaker would have agreed with Gunn that the unadulterated

200

chapter 10

Gaelic of the late nineteenth century East Sutherland fisherfolk was superior to the adulterated Gaelic of the surrounding population. Judging by the invidious comparisons I have heard among Gaelic speakers in Sutherland and in other parts of Highland Scotland, it is safe to say that the non-fishers of East Sutherland would themselves have agreed with Gunn; the number of English loanwords in one’s Gaelic has long been a sensitive matter among Gaels. On the other hand, the “broken” English of the imperfectly bilingual fisherfolk during the early twentieth century was certainly negatively valued and subject to much criticism. The positive and the negative lag were equally distinctive for fisherfolk identity. Giles, Bourhis, and Taylor (1977: 325) make what can be taken to be a “­ classic” statement about language and ethnicity: Ethnic groups are an example par excellence of linguistic categorization since they are often found to manifest their distinctiveness from each other by means of separate languages or dialects. To this formulation I would add, in the light of the East Sutherland fisherfolk case, that an ethnic group may also manifest its distinctiveness from others by means of a linguistic behavior which may be popularly perceived in terms of a language, but can be seen in diachronic perspective to consist of a time differential: a persistent lag in linguistic habits as compared with the linguistic habits of neighboring groups. The precise nature of the out-of-step behavior may vary over time, as it has done in Sutherland, but the constant is the presence of some difference in linguistic behavior, and not a particular different language or dialect as such. The ethnic marker is rather the lag than the language. The traditional notion of the form the “language” component takes, in the relationship between language and ethnicity, needs to be broadened to allow for such cases. References Chapman, A. J., J. R. Smith, & H. C. Foot. 1977. Language, humor, and intergroup relations. In H. Giles (ed.), Language, ethnicity and intergroup relations. N.Y.: Academic Press. 137–69. De Vos, George. 1975. Ethnic pluralism: Conflict and accommodation. In G. De Vos & L. Romanucci-Ross (eds.), Ethnic identity: Cultural continuities and change. Palo Alto, California: Mayfield Publishing Co. Dorian, Nancy C. 1973. Grammatical change in a dying dialect. Language 49, 413–38.

linguistic lag as an ethnic marker

201

———. 1977. The problem of the semi-speaker in language death. International Journal of the Sociology of Language 12, 23–32. ———. 1978. East Sutherland Gaelic. Dublin: Dublin Institute for Advanced Studies. Fishman, Joshua A. 1977. Language and ethnicity. In H. Giles (ed.), Language, ethnicity and intergroup relations. N.Y.: Academic Press. 15–57. Giles, H., R. Y. Bourhis, & D. M. Taylor. 1977. Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, ethnicity and intergroup relations. N.Y.: Academic Press. 307–48. Gunn, A. & J. MacKay, (eds.). 1897. Sutherland and the Reay Country. Glasgow: John MacKay. Hohenthal, W. D., & T. McCorkle. 1955. The problem of aboriginal persistence. Southwestern Journal of Anthropology 11, 288–300. Hymes, D. 1966. Two types of linguistic relativity (with examples from Amerindian ethnography). In W. Bright (ed.), Sociolinguistics. The Hague: Mouton. 114–67. Jackson, J. 1974. Language identity of the Colombian Vaupes Indians. In R. Bauman & J. Sherzer (eds.), Explorations in the ethnography of speaking. Cambridge: Cambridge University Press. 50–64. Petersen, William. 1975. On the subnations of western Europe. In N. Glazer and D. P. Moynihan (eds.), Ethnicity: Theory and experience. Cambridge, Mass.: Harvard University Press. 177–208. The statistical account of Sutherlandshire, by the ministers of the respective parishes. (1841). n.p.

part three Language Shift and Language Maintenance

∵

CHAPTER 11

Language Loss and Maintenance in Language Contact Situations There are a good many possible outcomes of extensive language contact, and by no means all of them include language loss, whether on the community level or on the individual level. By definition, though, we are interested on this occasion in the language contact situations which do result in same form of loss, and my remarks will be confined to such cases. But because there is some tendency in North American cultural and intellectual life to assume that extensive language contact produces one form or another of loss, I feel obliged to stress that this needn’t be so, even though I will be dwelling on the cases where it is so. One common result of extensive language contact on the community-wide level is language shift, the gradual displacement of one language by another in the lives of the community members. This occurs most typically where there is a sharp difference in prestige and in levels of official support for the two (or more) languages concerned. Where such differences in prestige and official support exist, there are usually also marked differences in the utility of the two (or more) languages for the speakers, (To simplify terminological matters, I will speak henceforth of two languages and of bilingualism, but my remarks should be understood to apply also to situations in which more than two languages are spoken and to multilingualism.) The two “classic” settings in which this phenomenon has been relatively well studied are the indigenous minority language and the transplanted immigrant language. On the level of the individual, the result of extensive language contact may also be a complete shift, over a lifetime or over a briefer transition period; but at least as common, and perhaps more so, is a partial shift or even a partial merger, so that at least one of the two languages does not retain its full complement of functions or perhaps even of forms. Sometimes neither language retains its full complement of functions or of forms. To take up first the case of shift on a community-wide level, there are several questions in connection with that phenomenon which are of potential interest to us here. It seems usual for members of any given speech community to have very little difficulty ranking two coexisting languages in terms of prestige. There are societal attitudes toward specific languages and to their native speakers. How are these attitudes formed? Are they usually local in origin and

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�13

206

chapter 11

focus, or are they affected by conditions of a much larger scope? Students at schools and colleges can often rank foreign languages by differential prestige, too. Are such student rankings a totally different matter (based, say, on how easy or hard a foreign language is said to be to learn), or do they share certain factors with “real-world” rankings? More importantly, are the factors manipulable at all, in either case? That is, can one make a language seem more or less attractive to its potential population by changing its “reputation”? Further, there is the matter of language loyalty, or to put it in more general terms, positive language orientation in an essentially negative situation.1 Some individuals are in the vanguard of any language shift (relative to their generation as a whole), whereas some lag behind (again, relative to their generation as a whole). Are the “laggards”, the language loyalists, in any way comparable to the likewise relatively unusual North American student who enjoys foreign languages and elects to learn them in an educational milieu where the study of foreign languages is not highly regarded or rewarded, on the whole? Are there attitudes characteristic of whole populations which favor foreign language acquisition or, alternatively, disfavor it? I propose to review some of the literature of language shift for tentative answers to a few of these questions. Certainly the available treatments of language shift were in no case intended to provide parallels to classroom foreign language acquisition or loss processes, but I think it will prove possible to suggest some parallels, nonetheless, although scarcely to “answer” many of the questions I’ve just raised. One of the most obvious facts about the rise and fall of linguistic fortunes is that they are linked to rises and falls in political fortunes. Where an empire appears, it is almost certain that the official language of that empire will spread at the expense of the languages of lesser powers which are absorbed by, or even just administered by, the imperial power. Examples are plentiful from both the Old World and the New: Persian and Greek are relatively minor languages in the world’s linguistic arsenal today, but were at one time the aggressively expanding languages of great political powers. The same was true of Ottoman Turkish. Aztec (Nahuatl) and Quechua experienced similar expansion and subsequent contraction in the western hemisphere. English and Russian, at one time rather minor languages of the European periphery, are currently in a phase of expansion commensurate with the political fortunes 1 One can also have a positive language orientation in a positive situation, but the concept of language loyalty is rarely introduced in such cases since the assumption is that loyalty to the language in question will be high. In negative situations there can be no such assumption, and where loyalty is nonetheless encountered, it is consequently worthy of comment.

language loss and maintenance in language contact situations

207

of the United States (and, of course, Great Britain before it) and of the Soviet Union. Occasionally, a politically dominated area possesses a cultural tradition of such strength that its language survives conquest and perhaps even prevails over the language of the conquering power; Greek maintained a strong position under the domination of the Romans, and the Norsemen adopted a Romance tongue in France. Various compromises are also possible, such as the Arabicand Persian-encrusted “Turkish” of the Ottoman Empire, or the Normanized lexicon of resurgent English in the thirteenth and fourteenth centuries. Two factors which have undoubtedly played a role in the shift characteristic of historic conquest situations have been identified in contemporary cases of impending language extinction as well: pragmatism, and a cultural stance which either does or does not favor language maintenance. The two may or may not be related. That is, pragmatism is sometimes a part of a cultural stance which favors language shift, but in other cases it may be a separate issue and may favor bilingualism rather than language shift. Examples abound of communities where the local language was deliberately not transmitted to the children because it had no practical value in the national or even local setting. That is, no jobs and no social advantages accrued through mastery of the language in question. A native speaker of East Sutherland Gaelic, a moribund variety of Scottish Gaelic, told me that she hadn’t taught the language to her children because “Gaelic’s no use to you through the world.” Denison reports for Sauris, an originally German-speaking area in northeastern Alpine Italy, a pattern of parental use of German among themselves but Italian to their children: The reason given by informants for this use of Italian is in almost all cases the desire to ease the path of their children at school: a few have mentioned the general usefulness of Italian as compared with the other languages (the local German dialect and Friulian) . . . (Denison 1971: 166–67). A change in patterns of language maintenance can often be linked directly to changing economic and social conditions, and can occur with rather dramatic rapidity in a given setting. Gaelic East Sutherland “went English”, in large part, in the 50-year period after the infamous Highland Clearances introduced large-scale sheep farming and a small but highly visible English-speaking middle class. The change-over was so swift and drastic that only 33 years after the Clearances began in East Sutherland, a well-informed bilingual insider was predicting the extinction of Gaelic in the area (Statistical Account, 1841, vol. 15: 156). An even more rapid and sudden change is now occurring in Juchitan (Oaxaca state), Mexico, where the discovery of offshore oil has

208

chapter 11

brought about radically different economic and social conditions. Before 1974, “socio-economic conditions were such that there was no advantage identified with the learning of Spanish for those who did not plan to migrate”; but the greater utility of Spanish for jobs within Oaxaca, since the development of a port near Juchitan, has already produced a weakening of language loyalty to Isthmus Zapotec in Juchitan in the few years since 1974 (Hinojosa 1980). Susan Gal’s masterful study of the language-shift processes in Oberwart, Austria, shows that the current shift to German on the part of the originally Hungarianspeaking segment of the population had its counterpart earlier, when the district was part of Hungary, in a tendency among some German speakers to shift toward Hungarian (Gal 1979: 42). In short, people know very well on which side their linguistic bread is buttered. For the transplanted immigrant, there is often a major reversal in the utility of the mother tongue (unless, of course, he emigrates to a colony or former colony where his mother tongue is still spoken). In the new environment, his mother tongue will be useful only insofar as he interacts with other immigrants from the same home country. Otherwise, a quite different language will become much more useful. Karttunen considers the homogenous settlements peopled by Scandinavian immigrant groups in the northern forest and plains areas of the United States and Canada to be particularly favorable for language maintenance: rural immigrants there continued to speak the languages of their home countries longer than did immigrants in urban centers. But even in these ethnically homogeneous settlements, language shift was soon underway: “For a while, such settlements had rather limited contact with English, and it was possible to remain monolingual in North America. Once contact began, however, English rather quickly took over.” (1977: 174). In such cases as these, it’s clear that pragmatism is a major factor. Language loyalty persists as long as the economic and social circumstanees are conducive to it, but if some other language proves to have greater value, a shift to that other language begins. Pragmatism is not the only value operating, however. There seem also to be cultural dispositions which do or do not favor language maintenance, or cultural-historical dispositions. Hymes drew attention more than a decade ago to a study by Hohenthal and McCorkle (1955) which contrasted two South American Indian groups in terms of retention of language (and identity). One group, the Fulnio of Brazil, have given up their lands several times during the last three centuries, moving in order to preserve their language and annual religious ceremony (to which proper use of the language is essential) as the basis of their identity. (Hymes 1966: 126)

language loss and maintenance in language contact situations

209

The other group, the Guayqueries, who live on islands off the coast of Venezuela, apparently abandoned their language and their pre-Christian religion very early, since no trace of the original language or religion survives. The Guayqueries maintain their identity by way of a particular socio-economic structure; in the case of one small group of the Guayqueries who gave up that social structure, their distinctiveness was lost, and they became indistinguishable from other Venezuelans. Hymes comments: “Language, together with religion, has served a separatist and unifying function in the one case, but not in the other” (ibid.). McLendon, in studying a Californian Indian language, found that linguistically exogamous marriages coincided with a thoroughly pragmatic approach to language learning and unlearning. Eastern Pomo Indians acquired whatever languages might be spoken in their homes very readily but forgot them just as readily if the use for them passed (via the death of a grandparent, for example). Characteristic of the Eastern Porno, also, is “a matter-offact valuation of language and lack of romanticism about or idealization of only one language” (McLendon 1978: 143). Clearly this is not a cultural stance which fosters a language-centered ethnic symbolism, and McLendon suggests that English entered the Eastern Pomo speech community on the same pragmatic basis as other useful languages in the past, but outweighed all the rest, ultimately, in its long-run utility – including Eastern Pomo itself, which is now approaching extinction (op. cit.: 146). The strongest statement on the role of cultural dispositions is probably to be found in a recent paper of Eric Hamp’s. He points out that “cultures . . . have as a part of their complex of traits a specific view or policy concerning integrity of the culture and the role of diversity” with “important consequences for language contact, and the more so to a dramatic degree in the case of small communities” (1978: 160). His most striking case in point is the very dissimilar fates of Albanian· language enclaves in Greece and in Italy. In Italy, where a tradition of cultivating and valuing localism (including local dialects) prevails, the Albanian enclaves are preserving their language; while in Greece, with an ethnocentric reverence for all things Greek as the dominant feature, the Albanian areas – once quite extensive – are losing their distinctive language. Hamp maintains that it is entirely in keeping with, even predictable from, these different cultural traditions, that Italian immigrants in the United States usually give up Italian after the first generation in favor of the language of the new locality, while “immigrant Greeks in the United States will cling to their language to the third generation” (op. cit.: 161–62). As a veteran of two and a half years of “Greek school”, I can testify to the determination of Greek Americans to maintain their language; the very numerous Italian Americans in my home town, on the other hand, provided no language-school classes for their children at all.

210

chapter 11

So very many languages are dying, during our time, in so very many localities, especially in the face of a vast expansion of a relatively few languages of enormous political and cultural potency like English, Russian, Spanish, and Mandarin Chinese, that one might well wonder whether even the most resistant linguacentric cultural stance would be enough to preserve a language with limited geographical, cultural, or demographic scope today. The dying languages are typically within the boundaries of a country in which a “major” language is spoken. That is, English is not displacing indigenous languages in India or Africa (see Fishman 1977: 114), but it is doing so, rapidly, in Anglophone Canada, the United States, and Australia; Russian is completely displacing a good many languages within the Soviet Union (Lewis 1972), but not in the rest of Eastern Europe. This phenomenon highlights again the political aspect of language maintenance. And, indeed, where threatened languages have made a comeback, it has usually been in connection with a sharp rise in nationalistic sentiment. Thus, there were some notable success stories in Europe as nationalism gained strength: Finnish successfully replaced the prestige language which had threatened to displace it in its own homeland (that is, Swedish), and Czech ceased to lose ground to German with the dissolution of the Austro-Hungarian Empire and the establishment of an independent Czechoslovakia (Ellis and mac a’ Ghobhainn 1971). German itself was once actually endangered by a fashion for French which spread beyond the aristocracy and beyond even the middle class, threatening to become truly “popular”. Historians of the German language often quote Voltaire, who claimed that during his rather extensive experience of eighteenth-century Germany (he was a guest at the court of Frederick the Great), he needed German only for travelling: “Je me trouve içi en France. On ne parle que notre langue. L’allemand est pour les soldats et pour les chevaux: il n’est nécessaire que pour la route” (quoted in Waterman 1966: 138). But the German language survived to emerge later in the eighteenth and nineteenth centuries as a stable standard (and eventually national) language. Hebrew, too, is the great exception to the requirement that a reviving language must have a certain numerical strength (of as yet undetermined size) in the population in order to succeed: for, although it seems clear that large numbers of speakers will not save a language (cf. Irish, which gave way to English during the eighteenth and nineteenth centuries, despite the fact that it boasted great numerical superiority at the start of that period over English, the language of a relatively small elite; see Macnamara 1971), it also seems that a comeback on the part of a threatened language rarely occurs without a substantial demographic base, even if that base is largely among the peasantry. If we now look at all these factors – the utility and prestige of a language, cultural values which do or do not support identification and maintenance

language loss and maintenance in language contact situations

211

of the group through its language (including ethnocentric or nationalistic feeling), and a demographic base – we can and should ask what clues may be lurking in “national” language settings for foreign language learning and retention. I think, speaking now as a classroom foreign-language teacher, as well as as a specialist in language extinction, that there are some real messages in cases like the sampling I’ve provided. Utility is an obvious candidate: every foreign-language teacher is familiar with the transformation which takes place in the lackluster language student who learns that he is to have a chance to visit the country in which that language is spoken, the following summer or year. Suddenly he is not only doing fine classwork, but turning up at the weekly foreign-language table in the dining center, attending all the functions of the language club, and assiduously cultivating every student of the foreign nationality in question who happens to be on the campus. This kind of transformation is so commonplace, and so obvious, that it seems equally obvious that our universities, our businesses, and our government could change the climate of foreign-language classrooms, and make a start on a change in student attitudes in general (thereby producing a significantly higher achievement level in foreign-language mastery), by providing more scholarships abroad, more internships abroad, and more jobs abroad for candidates with the requisite language skills. Whether a higher level of foreign-language mastery correlates directly with a better level of retention is, of course, one of the things we need to find out much more about. It certainly is true that classroom students who do very well at foreign languages are more likely to seek out, or fall into, situations in which they can make use of their language skills, thereby reinforcing them; whether it is also true that a foreign-language student who earns top classroom grades but makes no subsequent use of the language in question retains more than his fellow student who earns mediocre grades and also makes no subsequent use of the language, is a much broader question, and one about which less is known. In the sense that job and career opportunities, or pay scales which reward foreign-language skills, create a more favorable climate for foreign-language learning in the first place, one can perhaps assume the payoff that usually accompanies stronger motivation and a recognized reward system. Noting the weakness of Scottish Gaelic in the competition with English for the loyalty of its speakers, Derick Thomson writes, “Neither in education nor in government service is there a secure career structure linked to Gaelic . . .” (Thomson 1979: 17). It is unfortunately possible for the United States (as it would not so easily be for, say, Denmark or the Netherlands) to send officials abroad without special language skills, and both diplomats who are political appointees and experts who represent the United States in technical areas like agriculture or military

212

chapter 11

training are all too often dispatched to non-English-speaking countries with no language preparation at all. The U.S. government can be said to remove incentive for foreign-language mastery (and to contribute to the ill-will that citizens of other countries often feel toward a people who make no linguistic gestures of respect where other nationalities and cultures are concerned) with each such appointment. United States businesses, as far as I can tell from anecdotal evidence, are more likely to settle for a bilingual secretary than to train a bilingual overseas representative. Ethnocentric attitudes are surely at least as much responsible for the Anglophone North American failure to master and retain foreign languages as is the absence of economic motivation, however. In particular the assimilative tradition and its pressures, in nations populated primarily by immigrants, are major forces to be reckoned with. The question of attitudinal and social factors in second language acquisition and retention is discussed in its own right in this volume however (see the paper by Gardner), and I will not deal with it further here, except to say that if popular journalism is anything to go by, the assimilative climate may be weakening a little in the U.S.2 and even the deeply entrenched traditional Anglo-Saxon complacency about ignorance of foreign languages may be changing.3 Much of what I’ve just said seems to deal primarily with foreign-language acquisition rather than with foreign-language retention, but obviously one must persuade the individual to learn the language before one worries about his retaining it; and there is the likelihood, further, that attitude affects retention just as it does acquisition (again, see Gardner, this volume). Where retention is concerned, we are no doubt dealing with factors as mysterious as those which make up language aptitude in the first place, and I should think one of the most basic things we need to discover (in terms of research goals arising from this conference) might be whether the tests which are predictive of language-learning aptitude are equally predictive of language retention in the wake of equal periods of training and equal subsequent periods without training. If so, we may suppose that the abilities and attitudes which foster learning also foster retention; if not, we need to find out which other abilities (such as long-term-memory capacity) may be operative.

2 See the spate of ethnic-awareness articles. 3 In recent weeks, I’ve read an article pointing out the weakness of an American business community in which perhaps six or seven members speak Japanese, confronting a Japanese business community in which 10,000 members speak English, and heard on the radio an interview warning of the helplessness of diplomatic personnel who can’t understand what is being said by natives of the possibly hostile community in which they have to function.

language loss and maintenance in language contact situations

213

Moving now from the very general to the very particular, I would like to draw on the in-depth study which I’ve been making of dying East Sutherland Gaelic (ESG), on the extreme northeast coast of mainland Scotland, for information about attrition in the language extinction process. Among the 13 imperfect speakers4 of ESG whom I’ve tested since 1974, three different acquisitional histories are identifiable. Most of the imperfect speakers have always been exactly that: they lacked sufficient exposure or motivation to become fully proficient speakers of ESG and have never at any time spoken a grammatically normal form of the language. One imperfect speaker has a very different history, however: she was fully fluent (and almost certainly Gaelic-dominant, to judge by other members of her family whom I know) up to the age of 18, when she left her home village for the Lowlands. There she ultimately married a monolingual English speaker and raised her family. At the time that I tested her (1974), she had been away from the home village for 40 years with very little opportunity to use Gaelic, apart from vacation trips home roughly once a year for about two weeks at a time. Her passive knowledge of Gaelic was virtually perfect (as in fact is that of all of the imperfect speakers, including those who have never been fully profieient),5 and her phonology was essentially normal for her age group, but her spoken Gaelic was halting and her grammar fairly deviant. Two other imperfect speakers, a brother and sister pair, have yet another acquisitional history, in that they were fully f1uent child speakers of ESG up to five or six years of age, when they entered school. It is abundantly clear from family anecdotes that they were Gaelic-dominant at school entrance, and the most convincing evidence of that fact today is that their phonology is unusually good compared to that of other imperfect speakers. In this last respect, they resemble the exiled speaker mentioned above. For comparison with the imperfect speakers, eight fully proficient speakers were also tested: four “older fluent speakers” relatively close to the ­conservative norm for the dialect[,] . . . established on the basis of the Gaelic used by the oldest and best f1uent speakers available to me in the early 1960s when my 4 I use the term “imperfect speaker” throughout this paper rather than the term “semi-speaker”, which I have frequently used in other papers (e.g., Dorian 1977, 1978b) because the sample of speakers reported on here includes one individual with a different acquisitional history and a different level of proficiency from the semi-speakers on whom I have previously reported. . . .  5 The perfect passive bilingualism of the imperfect speakers is apparent from: their flawless interactions with fluent speakers; their ability to provide instant English translations of anything said in ESG; their very evident comprehension of jokes, teasing, and rapid-fire banter; and so forth. They also state that they understand everything said in their presence in the local Gaelic.

214

chapter 11

study of ESG began (see Dorian 1978a); and four “younger f1uent speakers” whose Gaelic departed from the conservative norm in quite a few respects (see Dorian 1973, for example), but in ways and degrees not generally apparent to the community. The tests consisted of sentences in English presented for translation into Gaelic, a task made less artificial and less difficult by the following factors: translation into and out of Gaelic is a frequent activity in the community due to mixed networks of monolingual and bilingual kin and friends; the tests were conducted in the speakers’ own homes, in a friendly atmosphere and with much encouragement; I had worked in the community for eleven years before these tests were undertaken and was well known to almost all of the speakers, and very well known to a good many. Very much by herself, in terms of test results, was the exiled speaker. She was neither fish nor fowl, linguistically speaking. She was markedly deficient in morphophonology, morphology, and syntax by comparison with the two groups of fully proficient speakers, yet she was notably better than the other imperfect speakers in some (though not all) respects. Some examples: (1) she chose the conservative synthetic form of the first-person singular conditional verb 100% of the time, as did seven out of the eight fully proficient speakers, but none of the other imperfect speakers did (only one of whom ever used it); (2) she and two other imperfect speakers were the only speakers (out of 16 people tested on this point) to use fewer conservative placements of a pronoun object than nonconservative placements; (3) like all other imperfect speakers, but only one fully proficient speaker, she lacked the imperative plural morpheme entirely; (4) in the negative imperative, she used none of the most conservative possible forms, but also none of the least conservative forms favored by most of the imperfect speakers, producing only intermediate forms (except for one ambiguous form which may either have been an out-and-out mistake of a kind made otherwise only by imperfect speakers or have been a slightly deviant rendering of a conservative form); (5) all of her tenses were intact, a record matched by only three other imperfect speakers; (6) she controlled all of the subordinating conjunctions except the most infrequent ones (e.g.; ‘unless’, ‘although . . . not’), which was true of no other imperfect speaker: (7) in a test for retention of irregular noun plurals, she gave exactly as many analogical plural formations as conservatively irregular ones, a much weaker performance than any fully proficient speaker, and also weaker than four out of eight other imperfect speakers, but stronger than the remaining four tested on this point. Since this speaker’s acquisitional history is unique in my sample of subjects, no firm conclusions can be drawn about what remains and what disappears in the speech of the ESG-speaking individual who is totally fluent in late ado-

language loss and maintenance in language contact situations

215

lescence but has very little subsequent use for the language. All the same, it is clear from her test results that she has suffered marked attrition compared to her peer group who remained in the home community. Her younger sister and her same-age brother-in-law were among the younger fluent speakers tested. They outperformed her on every measure, since even where her morphology and morphophonology were intact, as in tense formation, she made mistakes which they did not make (such as the choice of the wrong root-form on which to build the tense-form itself). Although I have not had the opportunity to broaden my sample of exiles who have suffered attrition,6 there is a wealth of anecdotal material available in East Sutherland on the subject. The reason for this abundance of anecdotal material is itself of interest: it seems that there are enormous individual differences in the retention of the language among exiles, both in Britain and overseas, and the differences are so great that the community is at a loss to account for them except by attributing them to differences in loyalty to the home community and to its language. Since the sole remaining group which speaks Gaelic in East Sutherland today is a stigmatized ethnic group, willingness to use Gaelic is valued as a sign of group solidarity, and unwillingness to use Gaelic is resented as a sign of rejection of one’s origins and, hence, of social climbing. For this reason, a great deal of attention is paid to the state of any formerly fluent exile’s Gaelic if he or she comes home for a visit. There are, as I’ve noted, vast individual differences in retention. Families which have siblings who have came home after thirty or forty years in Australia with their Gaelic essentially intact boast of this for years afterwards, and families whose siblings are unable to converse in Gaelic after an equal length of time abroad make excuses miserably (“There was no one else from our part of the Highlands in his area”, and “He was married to an English speaker, you know.”). Extremely harsh condemnation is the lot of individuals who claim no longer to be able to speak Gaelic, or whose Gaelic in fact measures up poorly, after only 6 In actual fact there were two other exiles in my test sample, ESG speakers who had been living in London for a great many years. They were married to each other, however, and used Gaelic habitually in the home. I have often stayed with them in London and have had ample opportunity to observe that not only they but almost all adult visitors to their home are habitual Gaelic speakers. If one avoids looking out the window, there is nothing to remind one that one is not in East Sutherland. These two London residents have some relatively non-conservative usages, and some skewing of lexicon and idiom, but they are really fully proficient speakers of ESG and hence not useful subjects for the study of attrition despite their removal over a long period of time from the home community.

216

chapter 11

a ­relatively short time away – say, two to five years in an English-speaking area. The ESG speech community members emphatically do not believe that the performance differences they have observed over the years are due to actual differences in retention. They attribute all of the performance differences to attitude differences, that is, to differential language loyalty. As an outsider, I disagree. It seems evident to me, after witnessing from the inside of several different households the return of an exile-sibling, that there are real individual differences in the capacity to retain a once-dominant language in conditions in which there is no opportunity to speak it. I don’t doubt that same individuals work at it in some fashion; as with child learners, there are apparently some exiles who literally “practice”, whether by carrying on internal monologues, by talking to themselves when working alone, by praying regularly in the original home language, or by cultivating a repertoire of songs in it. (I heard tell of one man in Australia who had kept his Gaelic alive at least partly through songs.) But some people apparently don’t have to work at it very much. The language just stays with them, through no special effort on their part. Other people, like the exile whom I tested, retain full receptive knowledge of the language, but lose a good deal of their active control. Whether they could prevent this by special effort is entirely unknown. It may be as it is among adults with adopting the accent of the place you’ve gone to live: some people can’t help it, and others couldn’t do it if they tried. The issue of whether retention can be achieved by any individual through the exercise of will, or with the spur of highly positive attitudes, is one which interests me intensely because of the vehement feelings which attach to the Gaelic performance of returned exiles in East Sutherland. It should not be difficult to approach the matter objectively. Many American towns have some exiles from the smaller language groups of Europe – for example, Gaels, or Frisians, or Friulians – who have little opportunity to speak their original home language because they are too few and too scattered. Questionnaires could be devised to evaluate the strength of their attachment to the mother tongue and the degree of integration into any accessible exile community, as well as selfevaluation of the use they make of the mother tongue and their proficiency in it (both for ordinary purposes and for inner speech). A single investigator could administer such questionnaires and also some measures of vocabulary retention and productive abilities to a sample of just 50 to 100 exiles of identical mother tongue in various localities, and we would very quickly have some idea of the range of individual variation in these matters. Ideally, one would like also to administer a foreign language aptitude test, in order to determine whether language retention capacity is linked to language learning abilities.

language loss and maintenance in language contact situations

217

To return now to the other two groups within my imperfect-speaker sample, the most interesting finding for our purposes is that the speakers with a history of childhood fluency do not necessarily show an advantage in language production as older people. I say “not necessarily” because I have full test results only for the sister in my brother-and-sister pair; the brother’s very partial results do suggest a better than average control of ESG. The sister, however, is one of the very weakest imperfect speakers in my sample, with test results well below those of a number of speakers who have never spoken Gaelic fluently. In tests for the three tenses of Gaelic, for example, she was missing one entirely and produced only two out of sixteen instances of a second (i.e., a 12.5% success rate). In tests for three subordination structures (relativization of subjectand object-pronouns, ‘that’-subordination, and ‘if’-subordination), she had a flat zero success rate: she was unable to produce a single correct structure. She has her areas of somewhat greater control: for example, she used conservative placement of the pronoun object three-quarters of the time, her negative imperatives were about 50% correct (and were almost all of the intermediate level of conservatism in construction), and she showed a greater than average use of the conservative possessive adjective with inalienables (as opposed to an alternative construction with a preposition, originally used mainly for alienables in the dialect, but increasingly used to express all possession by younger fluent speakers and especially by imperfect speakers). Nonetheless, she is an extremely weak speaker overall and had more trouble producing responses to my tests than any other individual tested (possibly for reasons ·of personality structure as well as low proficiency). Outside the testing situation, she remains a halting speaker of very limited active skills. This woman’s case is interesting, not only because it is certain that she was once both fluent and Gaelic-dominant, but also because she is well integrated into the Gaelic-speaking minority group in her village. That is, it is not the case that her poor control of ESG results from the sort of emotional self-distancing that some of the exiles are suspected of. This woman did leave the home village when she was young and worked for some years in large Lowland cities. But when asked how she has managed to keep her Gaelic, despite being married to an English monolingual and despite the fact that many of her age group turned their backs on Gaelic and do not now speak it, she attributes her retention of Gaelic precisely to her period of exile and the fact that two other women from her own village were working in the same city: . . . I remember when we were working away, when I was in Edinburgh, and there were girls there from Brora, and we always went out and we

218

chapter 11

spoke together [in Gaelic]. You know, the three of us. Because – y’know, we just liked speaking. I’ve asked her the same question more than once, several years apart, and her answer was the same each time: that spending time away from the community, in company with other exiles, led her to value her mother tongue more and to try to hold on to it. Furthermore, since her return home to marry, she has lived for many years immediately next door to an older Gaelic-speaking kinswoman and her husband, with other Gaelic-speaking relatives a little farther down the street. She interacts regularly with such of these relatives as are still alive and with still others a little farther away in the village, and fully acknowledges, both publicly and privately, her membership in the stigmatized Gaelic-speaking group. That is, she does not make the attempt to “pass”, socially, by suppressing aspects of her identity (such as knowledge of Gaelic) which identify her as a member of the group in question. The issue of emotional distance seems worthy of special attention because it has been found to operate in second language learning (cf. Schumann 1978; Fillmore 1979). But where retention of the mother tongue is concerned (and almost all of the imperfect speakers consider Gaelic their mother tongue, interestingly enough, regardless of the level of proficiency they display), I find that not only the stronger imperfect speakers, but also the weaker imperfect speakers, are relatively loyal to Gaelic: favorable to the language in their attitudes, positive in their assessment of its beauty and richness, desirous of its being taught in the schools, and so forth. That is, a positive attitude toward Gaelic does not characterize only people who have succeeded in retaining it quite well, but rather it characterizes all imperfect speakers, even the very weak ones. It is probably a minimum condition for some retention of a language in what are clearly the final decades of that language’s existence; but it is not in any way predictive of the degree of success in maintaining control of the language. This contrasts sharply with studies of second language acquisition, where attitude often is predictive of success, both inside and outside the classroom. (In addition to Schumann and Fillmore, cited above, see Gardner and Lambert 1972, and the extensive bibliography which it contains, and Schumann 1975). One possible explanation for the difference between second language learning and mother tongue retention in this respect is the total receptive bilingualism of the imperfect mother tongue speakers, to which I’ve already referred above. They interact with fully fluent speakers superbly by drawing on their perfect comprehension and their knowledge of the sociolinguistic norms of the community (i.e., when to speak, when to be silent, whether to phrase something as a statement or a quotation, etc.), and by using their sometimes very limited

language loss and maintenance in language contact situations

219

productive skills judiciously. Since the second language learner normally has everything to learn, receptive as well as productive skills and sociolinguistic norms, the forward thrust provided by emotional involvement (integrative motivation; see Lambert 1967: 102) may be crucial to successful mastery. And, of course, the second language learner does not in any case have the option of leaning on passive and sociolinguistic skills to lessen the need for active control. One of the more suggestive findings of my Gaelic language-death studies for linguistic attrition is that the imperfect speakers in my sample show certain kinds of reductive phenomena in common in their Gaelic, regardless of acquisitional history. For example: a greater or lesser use of analogically regularized allomorphs in place of irregular allomorphs; complete loss of morphemes that are already showing weakness in the fully f1uent population’s Gaelic; loss of “inventory”, as in both the examples already mentioned and also in loss of vocabulary from both open and closed classes. On the other hand, my oneperson sample of the “formerly f1uent adult” category showed the following two differences from all of the other imperfect speakers: she did not show any certain evidence of loss of syntactic options except by loss of inventory (she retained forms appropriate to her age group for the two ESG passive structures, for example); and she did not show a tendency to make synthetic structures analytic, as all the other imperfect speakers quite strongly did. Thus, there seem to be both some important similarities and some important differences between the once-fluent adult who ceases to use a language and the speaker who has never actually achieved full proficiency. Analogical regularization is one of the former; resistance to analysis of synthetic forms is one of the latter. Obviously we need much richer evidence before we speculate much on why this should be so – most of all, we need to know whether these tendencies hold up over additional test cases. One other phenomenon which deserves mention is the amount of admixture or interference that shows up in the imperfect speaker’s weaker language. Like the second language learner, the imperfect speaker of a dying language controls some other language much better than the home language he has either partially forgotten or never fully mastered (or both). It has surprised me, in working with ESG, how little structural interference there is from English in the speech of imperfect speakers, all of whom are, of course, fully f1uent in English. There is a great deal of lexical interference, and some of the weakest imperfect speakers introduce unnatural syntactic patterns from English (perhaps a good deal of this promoted by the artificiality of the translation tests). But I’m struck by the fact that, although both ESG and English form most plurals by suffixation, and imperfect speakers greatly extend, through ­analogy,

220

chapter 11

the number of plurals formed by suffixation, the English sibilant plural is rarely borrowed into Gaelic by imperfect speakers (though all speakers often use it on English nominal loanwords embedded in a Gaelic sentence). Many accounts of threatened languages do stress the influence of the national language on the minority language (e.g., Costello 1978; Hinojosa 1980). But there are other reports of contrary findings; for example, Rankin’s discovery that the complex inventory of consonant phonemes in Quapaw, a Siouan language, was simplifying as the language moved toward extinction, but not in the direction of English (1978: 51). This suggests that we cannot simply assume that the person who is forgetting a language will substitute his dominant language’s structures for whatever he has forgotten in the second language. It is well established now that by no means all errors in the classroom learning of a second language are the result of interference from the native language (e.g., Richards 1973; Stenson 1974). Perhaps the errors in a half-forgotten language have a logic of their own too (that is, arise from properties of the language being forgotten or from the structure and order of the forgetting process itself), and are not simple interference phenomena . . . Where studies of attrition in the language-death setting are concerned, the first priority must be the accumulation of more data: we do not have enough evidence in hand to permit firm conclusions. Many more studies, especially more richly detailed studies, are needed from many more types of languages and language communities. Perhaps interest in language loss from both the theoretical and the practical points of view will result soon in an advance in the amount of rich and reliable data available. References Costello, John. 1978. “Syntactic change and second language acquisition: The case for Pennsylvania German.” Linguistics, 213:29–50. Denison, N. 1971. “Some observations on language variety and plurilingualism.” In: Edwin Ardener, ed., Social anthropology and language. London: Tavistock Publications. Dorian, Nancy C. 1973. “Grammatical change in a dying dialect.” Language, 49:413–38. ———. 1977. “The problem of the semi-speaker in language death.” International Journal of the Sociology of Language, 12:23–32. ———. 1978a. East Sutherland Gaelic. Dublin: Dublin Institute for Advanced Studies. ———. 1978b. “The fate of morphological complexity in language death: Evidence from East Sutherland Gaelic”. Language, 54:590–609.

language loss and maintenance in language contact situations

221

Ellis, Peter Berresford, and Seumas mac a’ Ghobhainn. 1971. The problem of language revival. Inverness: Club Leabhar Limited. Fillmore, Lily Wong. 1979. “Individual differences in second language acquisition.” In: Charles L. Fillmore, Daniel Kempler, and William S.-Y. Wang, eds. Individual differences in language ability and language behavior. New York: Academic Press. Fishman, Joshua A. 1977. “The spread of English as a new perspective for the study of ‘Language maintenance and language shift’.” In: Joshua A. Fishman, Robert L. Cooper, and Andrew W. Conrad, [eds]. The spread of English. Rowley, Mass.: Newbury House. pp. 108–33. Gal, Susan. 1979. Language shift. New York: Academic Press. Gardner, R. C., and W. E. Lambert. 1972. Attitudes and motivation in second-language learning. Rowley, Mass.: Newbury House. Hamp, Eric. 1978. “Problems of multilingualism in small linguistic communities.” In: James E. Alatis, ed. International dimensions of bilingual education. Georgetown: Georgetown University Press. pp. 155–64. Hinojosa, Maria de la Paz. 1980. “The collapse of the Zapotec vowel system.” Penn Review of Linguistics, 4:28–39. Hohenthal, W. D., and T. McCorkle. 1955. “The problem of aboriginal persistence.” Southwestern Journal of Anthropology, 11:288–300. Hymes, Dell. 1966. “Two types of linguistic relativity (with examples from Amerindian ethnography).” In: W. Bright, ed. Sociolinguistics. The Hague: Mouton. Karttunen, Frances. 1977. “Finnish in America: A case study in monogenerational language change.” In: Ben G. Blount and Mary Sanches, eds. Sociocultural dimensions of language change. New York: Academic Press. Lambert, Wallace E. 1967. “Social psychology of bilingualism.” Journal of Social Issues, 23:91–109. Lewis, E. Glyn. 1972. Multilingualism in the Soviet Union. The Hague: Mouton. Macnamara, John. 1971. “Successes and failures in the movement for the restoration of Irish.” In: Joan Rubin and Bjorn H. Jernudd, eds. Can language be planned? Hawaii: University of Hawaii Press. McLendon, Sally. 1978. “How languages die: A social history of unstable bilingualism among the Eastern Pomo.” In: Margaret Langdon, Shirley Silver, and Kathryn Klar, eds. American Indian and Indo-European studies. The Hague: Mouton. pp. 137–50. Rankin, Robert L. 1978. “The unmarking of Quapaw phonology: A study of language death.” Kansas Working Papers in Linguistics, 3:45–52. Richards, Jack C. 1973. “Error analysis and second language strategies.” In: John W. Oller, Jr., and Jack C. Richards, eds. Focus on the learner. Rowley, Mass.: Newbury House. pp. 114–35. Schumann, John. 1975. “Affective factors and the problem of age in second language acquisition.” Language Learning, 25:209–35.

222

chapter 11

———. 1978. The pidginization process. Rowley, Mass.: Newbury House. The statistical account of Sutherlandshire, by the ministers of the respective parishes. 1841. (Vol. 15 of the Statistical Account of Scotland.) n.p. Stenson, Nancy. 1974. “Induced errors.” In: John H. Schumann and Nancy Stenson, eds. New frontiers in second language learning. Rowley, Mass.: Newbury House. pp. 54–70. Thomson, Derick S. 1979. “Gaelic: Its range of uses.” In: A. J. Aitken and Tom McArthur, eds. Languages of Scotland. Edinburgh: Chambers. pp. 14–25. Waterman, John T. 1966. A history of the German language. Seattle: University of Washington Press.

chapter 12

The Value of Language-Maintenance Efforts which are Unlikely to Succeed1 In the fifteen years during which I was constantly visiting the East Sutherland district of the northeast Scottish mainland or living there, actively studying the dying Scottish Gaelic dialect of the region and clearly deeply engaged by its distinctive history and unique flavor, I was asked many times whether my activities or indeed anyone else’s could make any difference to the ultimate fate of the dialect. But I was never asked that question by a native of the region. Both academics and laypeople in other parts of the world thought it theoretically possible that the ebb tide of linguistic retreat might turn or be turned. And theoretically of course such a thing is perfectly possible and has in fact happened in the case of a good many languages. The heroic, near miraculous example of Hebrew, regaining vernacular status after centuries of more or less fossilized existence, is always to the fore, and the number of other languages once threatened but now perfectly secure is sufficient to make the question reasonable enough. For a compilation of success stories, see EIlis and mac a’ Ghobhainn (1971); for examples of enclaved peoples who have survived against all odds, sometimes with their languages as part of the persisting identity, see Castile and Kushner (1981). Nonetheless, at the risk of casting a dark shadow over a subject already rendered gloomy by the sheer number of languages known to have died or acknowledged to be in the gravest danger of doing so very shortly, I wish to deal here with some of the more knotty problems which pose severe obstacles to well-intended and even well-funded efforts to promote the survival (whether in terms of true maintenance or of revival from a barely existing population base) of threatened languages. I will deal first with the East Sutherland Gaelic case and then add examples from the Irish experience. My reason for emphasizing the negative rather than the positive in this paper is not a wish to deny the value of language-maintenance programs in general (see the closing sections of this paper in evidence), but rather an uneasy sense that maintenance programs are too easily and comfortably invoked as a solution to the decline of any speech form. The reality, as usual, is more complex and difficult.

1 I am indebted to Mr. Séamus Ó Ciosáin of the Roinn na Seirbhíse Poiblí, Dublin, for his reading of the prepublication form of this paper and for several corrections and suggestions.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�14

224

chapter 12

Marked political and/or economic change is the scenario most often suggested as favorable for a corresponding change in linguistic fortunes. Examples are not lacking. A case not only in point but also in progress is Faroese, where a new degree of relative freedom from political and economic subservience to Denmark has made possible a resurgence of Faroese linguistic activity, in schooling, journalistic and creative writing, scholarship, and popular speech (Wylie and Margolin 1981). The outlook for Faroese is vastly more positive than it was only fifty years ago, and Faroese ethnic identity has gained notably in conjunction with the progress made in establishing two forms of the language as valid “national” written and spoken norms (ibid.). The question then arises: could not “devolution”, partial or total disengagement from a general British polity in which Scottish and particularly Highland interests are given disgracefully little attention, create a similarly positive climate for Gaelic? If the revenues from North Sea oil and whiskey export were to be kept “at home”, as Scottish Nationalists have often enough urged, could huge infusions of financial support for Gaelic make a significant difference? Where Scottish Gaelic generally is concerned, such an outcome is not impossible, however unlikely. For East Sutherland Gaelic, the answer must be negative. The reasons are all too numerous. First, East Sutherland Gaelic is an isolated dialect, a speech island cut off from other dialects of the same language. Second, as one might expect in such circumstances, it is an unusual dialect, quite unlike demographically better represented, less widely separated dialects in the western part of the country. Third, it is an unwritten dialect, lacking any tradition for rendering it visually. Fourth, such fluent speakers as survived even in the 1960s, when I began my work there, were on the whole poorly educated and elderly, ill suited to take on leadership roles in promoting or codifying the dialect. Fifth, by reason of the deviance of the dialect and the lack of a written tradition, well-meaning outsiders would find it difficult to learn or (if speakers of other dialects) adapt to the local norms and uncongenial to promote a dialect so far from the more usual or better known varieties. Any promotion of Gaelic in East Sutherland would have to mean, and would have had to mean even in the early 1960s, when the speaker population was still 200 or more, promotion of a more nearly standard form of Gaelic, somewhat in keeping with dominant western dialects such as those of Skye or Lewis or the Outer Hebridean islands south of Lewis. In sophisticated populations with a tradition of literacy and therefore usually some experience of standardization or of compromise in the general direction of standardization, such an introduction of outside norms might be welcomed, or at least tolerated, as preferable to loss of the language altogether in the local area. In unsophisticated populations lacking such tradition or experiences, a response of that kind is unlikely. This is the more true because unwritten, nonstandard minority

value of language-maintenance efforts

225

dialects in isolated pockets are characteristically poorly valued even by their own speakers. Attitudes tend to be apologetic and negative (Dorian 1981: 28–29, 86–87), and the introduction of nonlocal norms only has the effect of reminding local speakers of just how deviant their own everyday speech is. Far from contributing to pride in the native ethnicity and traditions, linguistic or otherwise, the effect is typically to undermine further any native self-confidence and to erode what little self-esteem may have survived. Characteristic responses of East Sutherland Gaelic speakers confronted by teachers obliged, either by their own notions of ‘correct’ Gaelic or by their unfamiliarity with local norms and the lack of a writing system which would have accommodated those norms, to promote more nearly standard Gaelic among speakers of East Sutherland Gaelic, included anger, frustration, and even disbelief: . . . (I)f you’re not satisfied with that [the local version of a sentence given out for translation], go and take your own Gaelic. [From an adult woman who abandoned an adult Gaelic class because of dialect conflict.] Everything has to be grammatically correct. . . . If I – I read over a sentence to her, and my – my own Gaelic is bound to crop up, y’ know? And she just stops me at once. And tells me the proper way to say it. And it’s – it – to me, it sounds foolish, y’ know? You feel – y’ know? – that you’re – you’re – it’s . . . not right. Well, not in your estimation, y’ know? [From an adult woman who continues to take evening Gaelic classes because of her devotion to competitive Gaelic singing.] /ma’khwi:/ [‘Mackay’] is my name and it couldn’t be anything else. [From a woman told as a schoolchild that her family name in Gaelic ought to begin with nic ‘daughter’ instead of mac ‘son’, because she was female.] (All quoted in Dorian 1981: 88–89.) The effect of these efforts to promote Gaelic by teaching it can be seen overall to have been alienating rather than reinforcing, and this is a common enough result where a threatened language displays pronounced dialect differences and no established standardization tradition exists to temper the alienating effect of encountering authority figures who attempt to inculcate nonlocal norms. Languages struggling to survive have often had to make heroic efforts to bridge dialect differences in creating written forms which could be promoted without prejudice to one or another region and could therefore enlist the sympathy of the speaker population as a whole. For the ingenious and unusual solution devised for Faroese, see Wylie and Margolin (1981: 82–94). See Timm

226

chapter 12

and Kuter (1982: 12–14) for the ongoing difficulties in devising an acceptable pan-Breton orthography. For members of societies with relatively long histories of literacy, even though dialect variation may be pronounced (as in Germany or Italy, for example), it can be hard to appreciate the aversion among self-conscious and underconfident dialect speakers to “strange” forms of “the same” language. Because of inadequate or absent literacy, experience in relating local forms to other forms via an intervening standard form to which each can be referred is also lacking, and the effort involved in working out the equivalences necessary to easy understanding is correspondingly greater. Thus even a west-coast speaker of Scottish Gaelic complained to me once that she wished a certain woman in her neighborhood would just speak English to her instead of insisting on using a nonlocal (but also western) Gaelic dialect, as it was such a bother trying to “translate” the other dialect forms into local equivalents at high-enough speed for easy communication. In beleaguered speech communities where there is competition from and heavy pressure in favor of some language of wider currency, it often seems also to be the case that tolerance of dialect differentiation in the threatened language is low. Perhaps the awareness that dominant-language speakers typically already have a negative attitude to the minority language makes the minoritylanguage speakers hypersensitive to what seem to be aberrations within the minority-language varieties. Scottish Gaelic speakers generally are severely critical of their fellows, in my experience, for such failings as heavy use of English loanwords or calques of English constructions; they also react rather vehemently, again in my experience, to dialects of Gaelic which show collapse of grammatical categories maintained in the most conservative dialects, likewise to loss of lexical richness and to loss of phonemic contrasts – this without regard to (or interest in) the possible long-standing history of such changes for the dialects in question. A classic example of interdialect intolerance, comical really because of the historical absurdity of the attitudes when one is aware of the actual etymologies involved, is the contempt sometimes expressed by Scottish Gaelic speakers whose dialects use the word coinean to mean ‘rabbit’ when they encounter speakers whose dialects use the word rabaid instead for the same meaning. Coinean users heap scorn on the rabaid users, taunting them for merely putting a Gaelic pronunciation on so obviously English a word. The irony, and the absurdity, in this case, is that coinean is equally an English loanword, but borrowed at a somewhat earlier period when the English dialect form coney was in common use in Scotland with the meaning ‘rabbit’. Sic transit historia verbi.

value of language-maintenance efforts

227

Still the question remains whether large-scale changes in the political climate (with attendant changes in attitude toward indigenous minority languages), especially if accompanied by favorable economic developments, could work enough magic to bring a struggling minority language back from the brink. For a language as a whole, the answer is surely yes, since there are well-documented cases to prove the point. For local forms of a given language, the answer may still be no, if decline has progressed beyond a certain point. One might imagine a modern-day Carnegie emerging among the many hundreds of Gaelic-speaking East Sutherlanders who went abroad in search of opportunity. No matter how large the infusion of money and goodwill provided by such a benefactor, the wherewithal for revival would not be available. Not only would personnel who could speak, or learn to speak, the local dialect fluently be virtually absent at the outset, but the time period required to develop an orthography, produce texts, create curricula, and train teachers would almost certainly coincide with the time period during which the last native speakers fluent enough to act as adequate source people for these undertakings died out entirely. Whatever might emerge from such an effort, it would not be East Sutherland Gaelic as we now know it. One could just possibly reinstate Gaelic in East Sutherland, but the outcome would resemble the introduction of Oxford English into an Alabama or a Vermont where the local American English had disappeared. It is not of course by any means certain that any sort of Gaelic could be reinstated in East Sutherland, even given a riotously wealthy benefactor dedicated to the revival of his ancestral tongue. It is here that the experiences of the Irish revival effort are relevant, and they certainly do not provide grounds for unbounded optimism. Four accounts bearing on the Irish experience, dating from 1971 to 1985 and offering therefore the perspective of fifty to sixty years since the establishment of an Irish state with the avowed objective of reviving Irish as a national language, may be cited in evidence. ‘The Irish language as the national language is the first official language. The English language is recognized as a second official language’, according to the 1937 constitution, still in effect (quoted in Ó Ciosáin 1983: 12). Most poignant of these accounts is John Macnamara’s little anecdote of being chidden as a boy of about eight for not speaking Irish with his older sister when they went into a shop to buy sweets. Though speechless before the shopkeeper’s reproach, the young boy asked his sister afterward why they had been reproved, and on learning that the sweetshop lady expected them to speak Irish because they were being taught it at school, he could only ask in surprise, “Is Irish for talking?” His judgment on the episode, looking back on it in adulthood, is that this was “the inevitable effect when society at large

228

chapter 12

disclaims responsibility for a social enterprise and leaves it to the schools” (Macnamara 1971: 73). In a somewhat similar vein Desmond Fennell considers that the costly and relatively large-scale efforts to preserve the few remaining areas of Ireland where Irish was genuinely the language of people’s daily lives as of the 1920s (the so-called Gaeltacht), and to maintain both the population size and the regular use of Irish in such districts over the past sixty years, failed because the government neglected to recruit the people of the Gaeltacht themselves to the effort – to mobilize their enthusiasm, energy, and will: The attempt [to save the Gaeltacht] was based on a false assumption which was made not merely by the Irish government, but also by the language movement and the Irish people generally. This assumption was to the effect that the state bureaucracy, the semi-state companies, and particularly Gaeltarra Eireann [a state company for the economic development of the Gaeltacht, set up in 1958], could stop the Gaeltacht shrinking. Acting on this assumption, the government gave that task to these agencies. But the assumption that these agencies could perform that task was quite mistaken. . . . If there is a territory in which a particular language is usually spoken, and it is contracting continually through language change on the fringes, who can stop this contraction? Clearly, only the people of that territory – by deciding to do so and by taking appropriate measures. So another way of explaining why the state failed to save the Gaeltacht is by saying that the government failed to perceive this fact, and failed therefore to take action accordingly. It made no serious attempt to persuade the people of the Gaeltacht to decide to end the erosion – it never even asked a representative assembly of them whether they would try to end it – nor did it establish a representative regional institution which would have enabled them to “take appropriate measures” (Fennell 1981: 36–37). In a very specific study of three industrialization projects created in order to help halt the “migration hemorrhage” in the Gaeltacht by Udaras na Gaeltachta, a government agency which succeeded Gaeltarra Eireann in 1980, three authors recently found that commercial ventures intended to stabilize population and promote economic growth in the Gaeltacht needed to be very carefully controlled if they were to have the desired effect. Because special skills were inevitably required, outsiders had to be brought in at some of the upper levels. Under these circumstances, and depending on the size of the work force, the vigor of the local tradition of Irish use (including the favorability of attitudes

value of language-maintenance efforts

229

toward the lrish language generally), and especially the ratio of native Irish speakers to native English speakers both within the work force and within the district (where the spouses and children of outsiders were also a factor), industrialization projects might or might not work to favor retention of Irish in a given district. Two of the projects investigated had been successful in terms of the intended goal; the third had actually had an adverse effect on use of Irish. The third project had been particularly likely to fail in that respect, in the judgment of the authors, because it was too highly technological in nature and therefore required not only that more than the average number of outsiders be brought in to run it, but also that too many English technical terms be pressed into service in order to discuss plant operations (O’ Cinneide et al. 1985). Looking at the results of Irish educational efforts overall as of 1975, David Greene, basing his figures on a large-scale research survey (Committee on Irish Language Attitudes Research 1975), notes that an outcome of roughly 220,000 at least fairly fluent non-native speakers is a moderately respectable result for a government-sponsored language-promotion effort (and he might have added that this is especially so when the dominant rival language is one of worldlanguage status, namely English, and the only near neighbour states are officially English-speaking). But he points to two serious weaknesses which almost nullify the seeming numerical success: the new speakers are largely from the well-educated middle c1ass, leaving the working c1ass unaffected, and they are not residentially clustered in any way, so that no actual neighborhoods have been created where transmission of Irish to a new generation on more than an individual family basis is likely to take place (Greene 1981: 6). The Action Plan for Irish 1983–1986 clearly recognizes this latter problem and proposes – with what success remains to be seen – that “community schemes” be put into operation throughout Eire, plus two new city projects in north and south Dublin, in order to bring Irish speakers into regular contact with each other and to promote larger-scale, more systematic, and more group-oriented use of the language (Bord na Gaeilge n.d.: 15). The emphasis in this chapter thus far on the difficulty and even in some circumstances the impossibility of promoting language maintenance on an official or individual-benevolent basis is not, despite appearances, intended to suggest that such efforts are altogether without value. There are a number of possible reasons for undertaking efforts of this kind, sometimes in the face of almost certain failure, or in the certainty that one will be in effect introducing Oxford English into a region which once spoke Alabaman American English. The first is that one of the commonest reasons for failure – negative attitudes internalized by the speakers or potential speakers themselves – is in itself a serious reason for attempting to promote the language. A middle-aged

230

chapter 12

native speaker of Berks County Pennsylvania Dutch (a German dialect with extremely negative loading in that part of Pennsylvania, where dialect speakers were commonly stigmatized as “dumb Dutch”) told me that he thought his age-mates would not have chosen to take Pennsylvania Dutch in school when they had reached 7th or 8th grade, even if it had been offered then as a school subject instead of being (as it was) strictly excluded from either use or study in the school setting: “I . . . feel it was something most of them were kind of ashamed of because they were ridiculed with it. I think most of them wanted to get away from it” (Dorian 1978: 651–652). In such a climate the gesture of school and community support can act as a corrective in a psychological sense, even when the practical consequences of promotion are unlikely to be significant. Speakers, or at any event their children and their children’s children, might possibly derive some compensation for the pain of stigma and ridicule, or at the least some basis for mitigating negative family attitudes, by witnessing a reversal of official attitude and a possible concomitant lessening of general hostility to the minority culture (even if the language were lost) in the community at large. A second and related benefit from promotion efforts is the fact that they nearly always carry with them, if only because of the need for appropriate instructional materials, some emphasis on traditional lifeways and some transmission of ethnic history. Quite typically the threatened language community is also dispossessed of its heritage, often astonishingly ignorant of such basic information as where the ancestral population came from, what the original nature of their means of livelihood was, how their cultural institutions functioned, and what their traditional lore (songs, stories, proverbs, humor, satire or invective, artwork, crafts) was like. The self-awareness and self-confidence which can be regained through the recovery of such information have value in themselves, as the Black community in the United States has found and has forcefully proclaimed. A third potential benefit to be derived from promotion efforts, whether they qualify literally as ‘maintenance’ efforts or (if the language is as good as dead) ought to be considered “revival” efforts, is economic. Spolsky discusses this issue with refreshing candor in reviewing a variety of Native American bilingual education undertakings. One of the most important economic effects of a bilingual education program is in its potential for immediate benefit to the local community. The size of this benefit varies from the possible thousand well-paying teaching jobs on the Navajo reservation to the part-time job for an older

value of language-maintenance efforts

231

speaker of a dying language, but its impact on a local poor community cannot be underestimated (Spolsky 1978: 357). He points out that the establishment of “even a minimal transitional program for the first three grades” on the Navajo reservation would require a thousand Navajo-speaking teachers, whereas the long-standing Bureau of Indian Affairs policy of hiring only “fully qualified, college-trained teachers for its schools” had resulted in only 200 Indian teachers among the 3000 teaching in various types of schools on the Navajo reservation as of 1974 (Spolsky 1978: 355). It has been said of the Irish that they “use Gaelic as a symbol of their Celtic identity . . . but speaking Gaelic is not essential to group membership” despite that fact (De Vos 1975: 15). Indeed, one can look at the history of the Irish experiment in language maintenance and revival as an appalling waste of money and a colossal failure, drawing the cynical conclusion that for sixty years the Irish have chosen for purely symbolic reasons to throw good money after bad in a classically impractical and romantic way. It has been an expensive exercise, and it has not been a notable success in terms of securing a future for Irish as a national vernacular side by side with English (which – at least in recent decades – it was not intended to replace but only to join as a coequal language). But when one considers the cultural climate overall, there have surely been substantial gains. The contrast with Scotland is instructive. Highlanders usually know next to nothing about their ethnic heritage. It is the exceptional intellectual, not the ordinary intelligent layman, who can speak at all of the bardic tradition and the extraordinary role of satire and eulogy within it, of the institutionalized fostering of children and the societal bonds it created, of the role of the king as the “spouse” of his territory responsible in large measure for its prosperity or decline, or of the legal system so unlike the Anglo-Saxon common law. The Scottish Celt scarcely knows that his people had in ancient times a tradition radically different from the Anglo-Saxon one. Highland schoolchildren are fortunate if they are even given history texts which represent so much as the Scottish viewpoint, never mind the Celtic viewpoint, on British history from time to time; texts which present Celtic cultural tradition are only now being created, and then only in the few locations where bilingual education has been introduced since the 1970s. Ireland, by contrast, produces mountains of scholarship on its Celtic past at every possible level of accessibility. The esoteric researches of its scholars in language, law, literature, pre-Christian religion, and so forth are published in quantity, often with government support. Many of the same scholars have

232

chapter 12

also become adept, in the favorable climate which prevails, at reworking their findings for popular radio and TV broadcasts – and these more widely “consumable” versions are often also published in inexpensive, readily available editions. Schoolchildren have copious materials to acquaint them with their ancestral culture; an official Curriculum Development Unit puts out richly illustrated readable books for that purpose. Not all Irish, adults or children, are especially interested in their Celtic heritage. The point is, however, that if they should be, there are no obstac1es whatever to learning about it. Irish schoolchildren, however distant from the Gaeltacht their homes, are most unlikely to be denied the opportunity to study Irish if they wish it; over most of Highland Scotland, schoolchildren and their parents are still told either that there are no teachers available to teach Gaelic or that there is no room in the curriculum for the subject. It is not that members of the Irish public must engage with their ethnic past, but that there is ample opportunity to do that if they are so inclined. In Scotland, except of late in the areas where bilingual education projects have begun (the Outer Hebrides, the Isle of Skye), the dedicated intellectual or the tenacious amateur historian can contrive to acquaint himself with his ethnic past, but only by dint of major personal effort. The legacy of language-maintenance efforts in Ireland is far from uniformly positive. Active aversion to Irish, nurtured mostly in the years of premature maintenance efforts when the policy of “compulsory Irish” was in force despite woeful lack of personnel with the requisite language skills or the requisite methodological training, is still quite widely met with. On the other hand, cultural disinheritance is at least potentially a thing of the past, writers in Irish have a modest reading public, and fluent speakers have job preference in some branches of governmental service where a knowledge of Irish is highly useful. It is by comparison with Highland Scotland, where none of these conditions prevail (except active aversion to Gaelic, in this instance because of generations of suppression and negative stereotyping), that the Irish situation gains a certain luster. Maintenance efforts on behalf of East Sutherland Gaelic cannot work. Maintenance efforts on behalf of Irish have barely worked, speaking strict1y in terms of the ongoing transmission of the language. Yet if asked whether maintenance efforts nonetheless serve some useful purpose, have some value, one can still find reasons for answering in the affirmative.

value of language-maintenance efforts

233

References Bord na Gaeilge (n.d.) Action plan for Irish 1983–1986. Republic of Ireland: Bord na Gaeilge. Castile, G. P., and Kushner, G. (eds.). 1981. Persistent peoples: cultural enclaves in perspective. Tucson: University of Arizona Press. Committee on Irish Language Attitude Research. 1975. Report. Dublin: Oifig Dhíolta Foilseachán Rialtais. De Vos, G. 1975. Ethnic pluralism: conflict and accommodation. In Ethnic identity: Cultural continuities and change, G. De Vos and L. Romanucci-Ross (eds.), 5–41. Palo Alto, Calif.: Mayfield. Dorian, N.C. 1978. The dying dialect and the role of the schools: East Sutherland Gaelic and Pennsylvania Dutch. In International dimensions of bilingual education, J. E. Alatis (ed.), 646–656. Georgetown University Round Table on Languages and Linguistics 1978. Washington, DC: Georgetown University Press. ———. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. Ellis, P.B., and mac a’ Ghobhainn, D. 1971. The problem of language revival. Inverness: Club Leabhar. Fennell, D. 1981. Can a shrinking linguistic minority be saved? In Minority languages today, E. Haugen, J. D. McClure, and D. Thompson (eds.), 32–39. Edinburgh: Edinburgh University Press. Greene, D. 1981. The Atlantic group: Neo-Celtic and Faroese. In Minority languages today, E. Haugen, J. D. McClure, and D. Thompson (eds.), 1–9. Edinburgh: Edinburgh University Press. Macnamara, J. 1971. Successes and failures in the movement for the restoration of Irish. In Can language be planned? J. Rubin and B. H. Jernudd (eds.), 65–94. Honolulu: East-West Center, University of Hawaii. Ó Cinneide, M. S., Keane, M., and Cawley, M. 1985. Industrialisation and linguistic change amongst Gaelic-speaking communities in the west of Ireland. In Language Planning and Language Problems 9 (1), 3–16. Ó Ciosáin, S. 1983. Bilingualism in public administration: the case of Ireland. Revista de Llengua i Dret 1, 11–19. Spolsky, B. 1978. American Indian bilingual education. In Case studies in bilingual education, B. Spolsky and R. L. Cooper (eds), 332–361. Rowley, Mass.: Newbury House. Timm, L. A., and Kuter, L. 1982. Language problems and language promotion in Britanny. Unpublished manuscript. Wylie, J., and Margolin, D. 1981. The ring of dancers: Images of Faroese culture. Philadelphia: University of Pennsylvania Press.

chapter 13

The Ambiguous Arithmetic of Language Maintenance and Revitalization 1 Introduction If you were to ask a random group of people what language is for, most would probably reply that language is for communication: it’s what we use to express our thoughts and feelings. This is a good and sensible reply, but not a complete one.1 If we used language only for communication, we would not maintain more than one language, or try to acquire more than one, unless we had (or foresaw) some practical need to communicate across a language barrier. But many people who already speak a language of wider communication (LWC) are willing to maintain or acquire some other less widely spoken language for reasons that are not related to practical need. Typically those other reasons are related to identity. Not that language is the only available signal of identity, of course: any number of other behaviors or traits can serve the same purpose, and in any case not all groups that are considered ethnically distinctive have a language of their own. Still, language is well recognized for its special culturecarrying capacity, so that learning another language, or keeping up one from your home, gives you a gateway to the culture embodied in that language. 2

Utility-Based Language Shift

It’s a general truth that when a language gains in practical value, it also gains in speakers: people see a need for it and are willing to go to the trouble of acquiring it. In our time a relative handful of the six or seven thousand languages spoken around the world have gained exceptionally wide distribution, making them exceptionally useful. These few very useful languages have been adding more and more new speakers whose parents and grandparents were mothertongue speakers of some other language. Where smaller languages have been swept into the orbits of expanding languages, language shift has followed in 1 I am grateful to Alasdair MacMhaoirn for reviewing and commenting on my remarks about contemporary Scottish Gaelic instruction in East Sutherland and to Magnus Pharao Hansen for constructive suggestions for shortening and sharpening an earlier draft of this chapter.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�15

the ambiguous arithmetic of language maintenance

235

what seems like an irresistible process, creating what many consider a crisis of language endangerment. But in actuality this process is interruptible and even potentially reversible. Political, demographic, and cultural conditions can change, and when they do, changes in relations between ethnic populations and their languages may follow. The breakup of the former Soviet Union created new conditions for languages such as Kazakh and Uzbek in Central Asia, for example, and in Solomon Islands, in the South Pacific, the advantage enjoyed by Bugotu, favored by missionaries, was lost when the influence of the missionaries was replaced by other cultural forces (Terrill 2002, 207–208). 3

Resistance to Utility-Based Shift

Practical usefulness confers a clear advantage, yet in a wide variety of locations around the world people have also demonstrated a surprising willingness to expend time, effort, and financial resources on maintaining or acquiring a low-utility, often purely local, language, in spite of the fact that they are already speakers of an LWC. Their efforts are a testimony to the strength of non-economic factors in language acquisition and maintenance, since their LWC already provides access to economic advantages. Strikingly, people can be found making this kind of effort all across the economic-development spectrum: in first-world France, for example, where members of a minority population are struggling to revitalize Breton (Kuter 1989) and also in Amazonian Peru, where villagers of Indian descent are working to encourage the use of their nearly vanished language Shiwilu (Valenzuela 2010). Almost two decades ago Joshua Fishman pointed out that the number of people engaged in RLS (Reversing Language Shift) efforts ran into the millions (Fishman 1991, 381), and the numbers have only grown since then. It seems that there must be great psychosocial rewards from such undertakings, since there is seldom much in the way of economic reward. The sheer number and variety of ethnolinguistic groups currently engaged in a support effort for a local heritage language is remarkable, but there are great differences in scale among these undertakings. The differences reflect such things as the history and size of the ethnic group, the proportion of the ethnic group taking an active part in the work, the resources available to the activists, the degree of cooperation or resistance from national or regional governments, and the friendliness or hostility of neighboring ethnic groups. At the very high end of the spectrum are movements that represent a numerically strong people, concentrated in a geographical area where mother-

236

chapter 13

tongue fluency is widespread and political mobilization is already strong. These favorable conditions prevailed, for example, when the position of the Slovenian language underwent a dramatic transformation as recently as 1991: following a brief armed struggle, a minority language of the former Yugoslavia became the national language of an independent Slovenian nation (Tollefson 1997). A linguistic transformation of this sort is so large-scale that we scarcely think of it in terms of a support effort for a minority language, yet that is one of its aspects. At the other extreme are languages without any surviving fluent native speakers. A surprising number of language-support movements have come into being long after the language in question has passed out of use as an everyday spoken language. The Celtic language Cornish in Britain, the Aboriginal language Kaurna in Australia, and the Native American languages Wampanoag in Massachusetts and Mutsun in California – all have reclamation movements that began without the benefit of a contemporary speaker population. In committing their time and effort, local-language supporters are testifying to the value that an ancestral, non-mainstream identity has for them. In many such cases the strength of the commitment reflects a deep sense that the course of history has worked harshly against their group, targeting their ancestral identity for suppression or even obliteration. Language restoration and revitalization efforts have intense personal value for some of those involved, and testimony to that special personal value has made its way onto the pages of academic books and journals. Linda Yamane, a Californian who is by ethnicity a Rumsien Ohlone Indian, writes of her slow, patient struggle to put together from scattered linguistic and anthropological records the grammar and phonology of her ancestral language and some of its traditional lore: “I thought I would never know the sound of our language or songs, . . . or the richness of our stories. But now, incredibly, . . . I can speak it. . . . I did not make an academic decision to learn our language. It evolved into something I could not ignore or stop” (Yamane 2001, 432). Yamane’s account of the painstaking work needed to recover the structure and some of the cultural content of her ancestral language demonstrates the uphill struggle involved in reclaiming a language that is no longer spoken. But even if the language in question is well studied, with at least scholarly documentation readily available, and even if the language still has native speakers, the potential obstacles to successful language-support efforts can be formidable. They may include, for example, too small a core of active supporters in the overall ethnic population, a deep-seated shame about an ancestral language that once carried a social stigma, and disagreement among leaders about goals or about ways of achieving their goals.

the ambiguous arithmetic of language maintenance

237

Practical difficulties abound. Restoring a missing transmission bridge between a few remaining high-proficiency elders and young people who have no knowledge of the heritage language can be a major challenge because adult second-language learning to the level of fluency is so difficult to achieve. Devising a writing system where none exists poses a challenge, as does replacing clumsy or competing writing systems, and putting forward one particular speech form (or an engineered compromise form) for adoption without creating fatal dissension or perceptions of inauthenticity is an equally daunting challenge. Political forces often work against support efforts: regional or national governments may not wish to see an ethnic population assert itself by promoting its own language. Resistance from the top is also likely if what was previously considered a regional dialect asserts a claim to independent and equal status within a country that already has an established standard form. Other chapters in this volume offer ample testimony to the many impediments that can prevent language-support movements from reaching their goals. 4

The Results of Language-Support Efforts: Success of Failure?

When it comes to evaluating the results of efforts on behalf of receding languages, the difference between a positive or a negative evaluation sometimes resides in the famous eye of the beholder. Some observers survey the scene and note, accurately, that only two or three language-support movements have succeeded in moving a previously sidelined, undervalued, or suppressed language into a secure position in recent times: Hebrew in Israel, French in Quebec, and perhaps Catalan in Spain. A few other languages are usually acknowledged to have made notable gains: Welsh in Wales, Ladin in northern Italy, Greenlandic in Greenland, Maori in New Zealand, Hawaiian in the United States. (Partisans of various language groups would certainly make additions here.) A good many support efforts are considered to have improved the standing of the languages in question in a legal sense; the minority languages of Scandinavia, Italy, and Canada, for example, enjoy enhanced legal standing today. In many cases the status of the ethnic language is improved in the eyes of its own speakers, who had long been made to feel that their speech forms were of little or no value; this kind of revalorization is likely to take place at least among the activist core, if not among all ethnic-group members. Doubts about the effectiveness of language-support movements or the value of RLS efforts are most commonly expressed when “success” is taken to mean that a receding language has been restored to full daily use and is now transmitted to ethnic-group children in the home. In spite of the proliferation

238

chapter 13

of language-support movements, and in spite of improved language-rights legislation in a good many countries, this remains a rare outcome. Historically it is certainly not unheard of: it is known to have happened in cases where a local language kept a strong speaker base in the general population even though a governing elite spoke some other expanding language (as with English during the Norman period and Finnish into the late nineteenth century). Although relatively few small, indigenous languages enjoy such favorable prospects today, Greenlandic represents one likely case. Speaker numbers have risen and the functions for which the language is used have expanded greatly as the island moves toward greater political autonomy in its relationship with Denmark, the previous colonial ruler. 5

Irish: The Half-Full, Half-Empty Glass

The very embodiment of the ambiguities involved in assessing language revitalization and restoration is surely Irish, a case which some see as evidence that language-support movements can achieve little and others see as evidence that they can achieve much. Irish speakers and their language suffered severe suppression under British rule during the seventeenth and eighteenth centuries, and the language then also suffered a drastic reduction in speaker numbers during the nineteenth century; deaths during the terrible potato famine of the nineteenth century, together with the emigration of many more who were fleeing desperate conditions, reduced the percentage of Irish speakers from about 45 percent at mid-century to a little more than 19 percent in 1891 (Ó Riagáin 1997, 4–5). The spatial and generational distribution of the remaining speakers suggested serious problems for the future of Irish at that point, with most speakers, and especially most young speakers, located in the rural western areas where poverty-driven emigration remained very high. A strong language-support movement came into being with the founding of the Gaelic League at the end of the nineteenth century, however, and with Irish independence in 1922 the Irish language gained official-language standing and received strong governmental support, especially in education. Very few longsuppressed languages have come to enjoy decades of political and financial support from a national government, as Irish did after 1922, and because its post-independence advantages have seemed so great, the relatively modest gains made by Irish are sometimes seen more nearly as an indication of failure than as measures of success. On the one hand, it is certainly true that less has been achieved for Irish than turn of the twentieth century activists had in mind when they promoted

the ambiguous arithmetic of language maintenance

239

a nationwide restoration of Irish. But one scholar has pointed out that when the Gaelic League was founded in 1893, 99 percent of the population of Ireland spoke English and 85 percent of those did not speak Irish; he suggests that successfully reawakening a national consciousness after centuries of relentlessly hostile colonial rule was more surprising than failing to persuade 85 percent of the population to change their language (Ó Cuív 1969, 128–129). Another scholar describes Irish-language policy since 1922 as a struggle to find a fair and suitable balance between two objectives: maintaining Irish in the discontinuous western districts where it was still a first language in 1922 (the Gaeltachtaí), and reviving Irish elsewhere in the country by increasing the number of speakers through education (Ó Riagáin 2008, 56). No dispassionate assessment can assert that these policy aims have since been fulfilled. Changing employment and social network patterns in recent decades have weakened Irish as a community language in the Gaeltachtaí (Ó Riagáin 2008, 57); at the same time, national economic development has created pressures leading to a relaxation of language requirements that promote Irish by linking job opportunities to proficiency in Irish (Ó Riagáin 1988, 45–47). Worrisome to such knowledgeable Irish scholars as Ó Riagáin and Gearóid Ó Tuathaigh is the fact that the largely middle-class networks of active Irish users that have emerged outside the Gaeltachtaí via the education system have not produced any self-sustaining urban communities, nor has the Irish state promoted the development of such communities (Ó Riagáin 2008, 59–60; Ó Tuathaigh 2008, 41). Ó Tuathaigh notes the connection many urban Irish-speaking networks have with one or another of the Gaeltachtaí, whether as place of origin or place of fluency acquisition, and wonders whether those networks could survive if the shift to English evident among Gaeltacht schoolchildren were to result in the disappearance of Gaeltacht base communities for Irish (Ó Tuathaigh 2008, 40–41). On the other hand, the glass can appear half-full when the Irish situation is compared with many others. Irish has weakened, especially as a first language, but it has not disappeared, even as a first language, despite the fragile demographic position it was already in well over a hundred years ago. If state support for Irish has not produced a nation of fully fluent bilinguals (and it has not), it has produced an education system with options by way of which new high-fluency Irish speakers can be generated (and continue to be generated, albeit not routinely). Certain desirable occupational avenues are still reserved for those with an advanced knowledge of Irish. Irish-language broadcast media, now especially television, continue to develop, providing jobs that are particularly attractive to young people. All-Irish schools, in which the medium of instruction is Irish throughout the curriculum, continue to increase in

240

chapter 13

number and popularity as an educational option. In addition, the international status of Irish has been enhanced by its recent recognition as an official working language of the European Union. In the light of all this, Suzanne Romaine’s moderately positive appraisal, in a review of Irish in the global context, seems justified. She notes that while active users of Irish remain a very modest minority, perhaps as many as one in three people in Ireland have some understanding of Irish (Romaine 2008, 24): “This means that the world in Irish will not be lost and the world can indeed still be lived in Irish by those who choose to learn and use it. That is hardly failure.” 6

Language-Support Efforts: Who Benefits?

Efforts in support of languages perceived as threatened by their speakers are by now so widespread that the absence of native-speaker support for a sharply receding language has become a reportable matter. Researcher Yaron Matras, objecting to the notion of linguists as quasi-heroic figures who want to “save” languages, goes out of his way to report on two speech communities in which little or no regret over the impending loss of the heritage language was expressed (Matras 2005). In my own forty-five years of fieldwork with the Sutherland-shire Gaelic of fisherfolk sub-communities in the northeastern Scottish Highlands, I have found expressions of regret very common in connection with the foreseeable end of this local language, but I have also found acquiescence in its demise very nearly universal, just as in the cases Matras discusses (Dorian 1981, 1987). It is not purely coincidental, however, that Matras’s chief example, the Dom (“gypsy”) community of Jerusalem, and the fisherfolk Gaelic speakers of East Sutherland both represent strongly stigmatized identities. Holding to your ethnic language in the face of powerful, long-standing denigration of that language and the identity it represents is no small matter. The language of linguistic intolerance can be just as vitriolic, and just as damaging, as the language of racial intolerance (with which, furthermore, it is often linked, either explicitly or inferentially). But even while flight from a stigmatizing ethnic identity is often the cause of a language shift, the flight itself sometimes plants the seeds of later language revitalization or reclamation efforts. When the linguistically assimilated children of parents who protectively withheld a stigmatizing language later come to recognize a lost linguistic heritage, some of them emerge as important figures in the effort to reclaim it. In the justly celebrated Master-Apprentice Program pioneered among Native Californian groups, for example, the younger people who are “apprenticed” to fluent-speaker elders in order to learn their

the ambiguous arithmetic of language maintenance

241

ancestral language are often relatives of those same elders, young people who lost their chance to acquire the language because of a break in the transmission process; aware of how much escaped them, they are now actively trying to overcome the loss (Hinton 2001). In East Sutherland, apart from the disappearance of the livelihood that had kept fisherfolk separate and the powerful social stigma that that low-income occupation had produced, there was yet another reason to foresee that the local form of Gaelic had a limited life expectancy. It differed a good deal from most other forms of Scottish Gaelic and was somewhat reduced by comparison with more mainstream dialects in terms of its sound system and its grammar. The Gaelic language has a codified written form that has long been used for writing more mainstream dialects, but local East Sutherland pronunciations and grammatical features did not correspond well to those of the written language. Because English had been introduced earlier there than in most other parts of the Highlands, the local Gaelic also showed certain unusual features reflecting long contact with English. It was easy to predict that if Gaelic were to enjoy an unforeseen rise in support, locally or nationally, it would not be the local Gaelic that was promoted but a form closer to the written language. And this is in fact exactly what has happened. Adult Gaelic classes are now fairly regularly on offer in eastern Sutherland, and Gaelic-medium primary education exists both in an East Sutherland village a little to the south of the former fishing villages and in a neighboring county a relatively short drive away. But the instructors all speak, and teach, a non-local Gaelic that conforms better to the written language; only once has an instructor made an effort to include local material, on an occasion when he had several adult students of fisherfolk descent. Even so, it would not be true to say that local East Sutherland speakers have derived no benefit from the heightened visibility of Gaelic in Scotland and its new instructional availability in Sutherland and other parts of the country. During the greater part of their lifetimes Gaelic – especially the local Gaelic, but to some extent Gaelic in any form – was regarded as a sign of cultural backwardness, a survival of an outdated way of life that had little or no value for the future. It was rare for local speakers to encounter people who were serious about learning Gaelic or showed a strong interest in its history and tradition; the occasional dialectologist, linguist, or folklorist might visit the region, some of them taking an interest in the local form of the language, but normally there was no prestige to be gained through a native knowledge of Gaelic and frequently there were distinctly negative overtones attached to that status. As activism on behalf of Gaelic took hold in the last two decades of the twentieth century and the first decade of the twenty-first, this began to change. There were more Gaelic learners on the scene as Gaelic classes multiplied, and a

242

chapter 13

few local Gaelic speakers were able to enjoy, for the first time in their lives, a certain special standing as native speakers of a suddenly more desirable language; they found themselves consulted by learners about pronunciations or constructions and occasionally approached by hopeful learners for practice in conversation. This revalorization of a language that was once disdained can be one of the worthwhile by-products of language revitalization and restoration efforts, finding expression not just in an improved self-regard among speakers and the ethnic population but in a general upsurge of interest in the language, its history, and its culture. Despite its psychological benefits, however, Gaelic-language activism has not “saved” East Sutherland Gaelic, which no longer has any fully fluent speakers. Both for the skeptical and for the hopeful, the long-term effectiveness of support efforts for Gaelic remains an open question. On the one hand, recent developments favorable to Gaelic have been notable, including the opening of Gaelic-medium units in a rising number of schools and even a few free-standing all-Gaelic schools, with 2,766 pupils receiving some or all of their education through Gaelic in 2008 (Gaeliconline 2008). A Gaelic Language Act passed by the Scottish Parliament in 2005 created a policy body, Bòrd na Gàidhlig (the Gaelic Board), committed to the protection and promotion of the Gaelic language. A digital television channel dedicated to Gaelic came into being late in 2008. Government funding for these and other Gaelic initiatives is currently estimated at many millions of pounds. On the other hand, Gaelic is spoken by a tiny minority of the country’s population (1.2 percent as of the 2001 census), many of whom live in peripheral rural communities at a great distance from the political and economic centers of Scottish life. Speaker numbers are expected to drop again at the next census, since the deaths of elderly speakers still far outstrip gains in new speakers via home transmission and Gaelic-medium schools. Gaelic, in competition with other languages (originally Pictish and British, subsequently Scots and English) throughout its entire existence, does not serve as a symbol of Scottish identity in the way that Irish serves as a symbol of Irish national identity. In a time when cost-effectiveness and the bottom line are always a factor in official policies and funding decisions, the relatively favorable current position of Gaelic is very precarious. 7

Conclusion: Prospects for Language-Support Efforts

There is no likelihood that movements to reclaim, revitalize, or otherwise promote ancestral languages will disappear. If anything, they are likely to increase,

the ambiguous arithmetic of language maintenance

243

and it may be that those who will benefit most from these efforts are those who have lost the most, members of ethnic groups whose languages are sometimes spoken of as “sleeping,” that is, viewed as a central part of a cultural legacy but unrepresented by active fluent speakers at the time when restoration movements begin. In spite of the colossal effort that goes into recovering the structure, the lexicon, and the sounds of such languages, every word or sentence that is discovered or reconstructed and made available in some form to the ethnic community is a precious heritage to those who wish to reclaim their languages. Improving the position of still spoken but sharply receding languages is not a simple matter, and in particular it is not a matter of language policy alone. Even the most supportive-seeming language policies can be ineffective if they are formulated in isolation from other far-reaching social and economic policies that might have contrary effects. In the Irish case, where this has become particularly apparent, Ó Riagáin points to the counterproductive effect on language revitalization of certain policies undertaken in the economic sphere, in regional planning, and in education. These other policies were rarely considered in terms of their impact on language patterns, he notes, yet they may in the end have more effect on those patterns than explicit language policies themselves (1997, 170–171). For example, if school districts are created that send Irish-speaking young people from adjacent areas to separate secondary schools in each of which English is the dominant language, this will have a negative effect on maintenance of Irish when compared with a policy that would have funneled those Irish-speaking young people into a single secondary school where they constituted a majority (Ó hIfearnáin 2007, 514–515). A more narrowly language-related policy – awarding prizes for school achievement in Irish, say – is of limited impact by comparison. What is more, some students of the Irish revitalization movement argue that elements of the official languagesupport policy itself are counterproductive: in adopting an “official” Irish that is no one’s actual spoken language for the purposes of country-wide education in Irish, schooling in Irish actually weakens the use of authentic local dialects without in return winning the allegiance of young Gaeltacht residents (Ó hIfearnáin 2008). There is a startling lack of consensus about even some of the most basic questions in connection with language endangerment and language-support efforts, as seems clear from one recent review of the prospects for indigenous language survival: the author surveys the field and concludes that opinions differ on the threshold of endangerment (in terms of some critical mass of speakers necessary for language survival, for example), on just what level of linguistic and cultural proficiency is involved in being a speaker (and so in determining

244

chapter 13

how many speakers a language does or does not have), on the methods and prospects for stemming language loss, on the reasons that some groups succeed better than others, on whether outside expertise can or should contribute significantly to indigenous revitalization efforts, and so forth (Walsh 2005). There is by now some very solid, nuts-and-bolts advice available to would-be revitalizers (for example, Brandt and Ayoungman 1989 and Hinton et al. 2001), but among indigenous, small-language populations there has also long been a disturbing tendency to concentrate on schooling and to neglect parent-tochild transmission in the home, even among the ranks of indigenous-language teachers themselves (Hinton 2009). Fishman pointed long ago to the limitations of schooling where transmission and language continuity are concerned (1991, 368–373) and to the need for intimate, small-scale network processes “too gratifying and rewarding to surrender” if a small language is to be maintained and transmitted (1989, 399). Such network processes are difficult to bring into existence. There is some evidence that coherent residential or religious spheres of linguistic influence (residential neighborhoods or religious congregations, the latter sometimes also residentially grouped), where use of a particular, non-mainstream language can be cultivated, are particularly effective approaches (Maguire 1991, Burridge 2002, Al-Khatib and Al-Ali 2005, Hansen 2010), but even the Amish and Mennonites, whose most conservative communities are famous for language retention, have difficulty securing ownership of enough land to support residentially coherent communities. There is also some evidence that situating language revitalization efforts within an overall context of cultural revitalization is a relatively effective approach: this strategy is part of the Master-Apprentice Program, for example. And always, political and social contexts can change, sometimes very quickly, making strategies and technologies obsolete almost as soon as they are devised. But if it would be unwise to declare many particular language movements successes, it would also be unwise to declare them all unequivocal failures. Even when it appears to be stark, the arithmetic of language maintenance and revitalization is always to some degree ambiguous. References Al-Khatib, Mahmoud, and Mohammed N. Al-Ali. 2005. Language and cultural maintenance among the Gypsies of Jordan. Journal of Multilingual and Multicultural Development 26: 187–215. Brandt, Elizabeth A., and Vivian Ayoungman. 1989. Language renewal and language maintenance: A practical guide. Canadian Journal of Native Education 16: 42–77.

the ambiguous arithmetic of language maintenance

245

Burridge, Kate. 2002. Steel tyres or rubber tyres – maintenance or loss: Pennsylvania German in the “horse and buggy” communities of Ontario. In Language endanger­ ment and language maintenance, David Bradley and Maya Bradley (eds.), 203–229. London: Routledge Curzon. Dorian, Nancy C. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. ——— 1987. The value of language-maintenance efforts which are unlikely to succeed. International Journal of the Sociology of Language 68: 57–67. Fishman, Joshua A. 1989. Language and ethnicity in minority sociolinguistic perspective. Clevedon: Multilingual Matters. ———. 1991. Reversing language shift. Clevedon: Multilingual Matters. Gaeliconline. Feb. 26, 2009. Retrieved from http://www.gaeliconline.co.uk. Hansen, Magnus Pharao. 2010. Nahuatl among Jehovah’s Witnesses of Hueyapan, Morelos: A case of spontaneous language revitalization. International Journal of The Sociology of Language 203: 125–37. Hinton, Leanne. 2001. The master-apprentice language learning program. In The green book of language revitalization in practice, Leanne Hinton and Ken Hale (eds.), 217– 226. San Diego: Academic Press. ———. 2009. Language revitalization in the home. Paper presented at the First Inter­ national Conference on Language Documentation and Conservation, University of Hawaii, Manoa, March 2009. Retrieved from http://scholarspace.manoa.hawaii .edu/bitstReam/10125/5165/9/5165script.pdf. ———, Matt Vera, and Nancy Steele. 2001. How to keep your language alive. Berkeley: Heyday Books. Kuter, Lois. 1989. Breton vs. French: Language and the opposition of political, economic, social, and cultural values. In Investigating obsolescence: Studies in lan­ guage contraction and death, Nancy C. Dorian (ed.), 75–89. Cambridge: Cambridge University Press. Maguire, Gabrielle. 1991. Our own language: An Irish initiative. Clevedon: Multilingual Matters. Matras, Yaron. 2005. Language contact, language endangerment, and the role of the “salvation linguist.” In Language Documentation and Description, Vol. 5, Peter K. Austin (ed.), 225–251. London: Hans Rausing Endangered Languages Project. Ó Cuív, Brian. 1969. The changing form of the Irish language. In A view of the Irish lan­ guage, Brian Ó Cuív (ed.), 22–34. Dublin: Stationery Office. Ó hIfearnáin, Tadhg. 2007. Raising children to be bilingual in the Gaeltacht: Language preference and practice. International Journal of Bilingual Education and Bilingualism 10(4): 510–519. ———. 2008. Endangering language vitality through institutional development. In Sustaining linguistic diversity, Kendall A. King, Natalie Schilling-Estes, Lyn Fogle,

246

chapter 13

Jia Jackie Lou, and Barbara Soukup (eds.), 113–128. Washington DC: Georgetown University Press. Ó Riagáin, Pádraig. 1988. Bilingualism in Ireland 1973–1983: An overview of national sociolinguistic surveys. International Journal of the Sociology of Language 70: 29–51. ———. 1997. Language policy and social reproduction: Ireland 1893–1993. Oxford: Oxford University Press. ———. 2008. Irish language policy 1922–2007: Balancing maintenance and revival. In A new view of the Irish language, Caoilfhionn Nic Pháidín and Seán Ó Cearnaigh (eds.), 55–65. Dublin: Cois Life. Ó Tuathaigh, Gearóid. 2008. The state and the Irish language: An historical perspective. In A new view of the Irish language, Caoilfhionn Nic Pháidín and Seán Ó Cearnaigh (eds.), 26–41. Dublin: Cois Life. Romaine, Suzanne. 2008. Irish in the global context. In A new view of the Irish language, Caoilfhionn Nic Pháidín and Seán Ó Cearnaigh (eds.), 11–25. Dublin: Cois Life. Terrill, Angela. 2002. Why make books for people who don’t read? A perspective on documentation of an endangered language from Solomon Islands. International Journal of the Sociology of Language 155/56: 205–219. Tollefson, James W. 1997. Language policy in independent Slovenia. International Journal of the Sociology of Language 124: 29–49. Valenzuela, Pilar. 2010. Ethnic-racial reclassification and language revitalization among the Shiwilu from Peruvian Amazonia. International Journal of the Sociology of Language 202: 117–30. Walsh, Michael. 2005. Will indigenous languages survive? Annual Review of Anthro­ pology 34: 293–315. Yamane, Linda. 2001. New life for a lost language. In The green book of language revital­ ization, Leanne Hinton and Ken Hale (eds.), 429–432. San Diego: Academic Press.

chapter 14

Purism vs. Compromise in Language Revitalization and Language Revival Speakers of well-normed wide-currency languages commonly meet strong puristic attitudes for the first time at school, where they are likely to be exhorted to mind their grammar, avoid slang and excessive colloquialisms (at least in settings of any degree of formality), and in some cases to foreswear the overuse of foreign loanwords. When the norms for such languages are of long standing, as is the case with English, many speakers view the puristic attitudes of their schoolteachers as a little laughable – a variety of fuddy-duddy conservatism about language which is consistent with a similar schoolteacher conservatism where clothing and hairstyles are concerned. In more recently normed languages, the situation can be very different; puristic attitudes may threaten the very success of the effort to promote a standard language. In India, where a popular form of “Hindustani” had come into being before independence through natural interactions among people of various backgrounds, standardizers promoted a policy of Sanskritization and deplored the many loanwords of Persian, English, and other origins which characterized Hindustani. They had enough influence to put in place a conservative and puristic policy which expanded the vocabulary of Hindi, as a new national language, by drawing on Sanskrit. This policy has suited the educated urban elite well enough, but it risks “alienation of the language from the masses” (Coulmas 1989: 11). Purism has also been a problem in the Arabic-speaking world; here, however, the problem is not recent norms, but ancient ones. Classical Arabic, codified in the eighth century CE, has led a rarified life as a fossilized language form, growing increasingly distant from all spoken forms of Arabic. Although a Modern Standard Arabic exists, it is not the variety of Arabic represented in the grammar books, nor are the modern lexical items of Modern Standard Arabic entered in dictionaries. Grammar constitutes a particular problem: The grammar books teach a lot which long ago ceased to be of any relevance to standard Arabic as it is practiced today . . . Since the rules of Arabic grammar are based on prescriptive rules instead of actual usage, they will remain hopeless and unattainable goals for the vast majority of Arab learners. (Ibrahim 1989: 42)

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�16

248

chapter 14

In all of these contexts, purism can be seen to represent a form of conservatism, a harking back to the favored forms or styles of earlier times: those of relatively recent generations in the case of English, but those of a truly ancient ancestral language in the case of Hindi, and those of a less ancient but still very remote period in the case of Arabic. The norms invoked in these cases are not truly those of the community at large, but rather those of a small segment of it: an educated elite of teachers, writers, broadcast journalists, intellectuals, and the like. It would be possible to imagine, on the basis of such cases, that puristic attitudes are typically imposed on the general community of speakers; and that those speakers, if not reined in by the exhortations and warnings of conservative language monitors, would soon kick over the linguistic traces and abandon traditional norms. But the regularity with which puristic attitudes appear in small language communities – even those with low literacy levels, or with only very recent experience of literacy – speaks against this notion. Everyday speakers of languages large and small often subscribe to puristic notions. Whether or not puristic attitudes are universal, they are widespread enough to create problems for efforts to support minority languages with a small native-speaker base, when these come under heavy pressure from neighboring languages of wider currency with larger speaker populations. I suggest that a common challenge for language revitalization and language revival is to limit the restrictive role which puristic attitudes are likely to play in the communities in question, or to channel such attitudes into forms which are useful rather than harmful. In what follows, I distinguish revitalization from revival. In contexts of revitalization, the language survives, but precariously. Efforts on its behalf require the mobilization of remaining speakers, as well as the recruitment of new speakers; in fact, the mobilization of at least some of the remaining speakers is typically crucial to the recruitment of new ones. In contexts of revival, the language is no longer spoken as a vernacular; it may have ceased to be spoken rather recently, or it may have been out of use as a vernacular for a long time. In either case, there may still be some fossilized use of the language, with the users either aware of the precise meaning of the fossil forms or unaware of it. Recruitment in this sort of context can perfectly well be undertaken by individuals who have not originally been among those most involved in traditional cultural life and have not been among the leading users of whatever fossilized language forms remain. Language revitalization efforts are much more common than language revival efforts. For one thing, there is a large – distressingly large – number of languages which still have a modest number of proficient elderly speakers, but

purism vs. compromise in language revitalization

249

far fewer middle-aged speakers, and perhaps none at all among young people.1 For another thing, introducing a language which can still be modeled for potential new speakers by remaining fluent speakers is considerably easier than introducing to them, in any convincing way, a language which exists in recorded texts or in books, but is not in ordinary use by any living person. Puristic attitudes should, in theory, be more of a potential problem in revitalization than in revival, since bringing about alterations in what people are already saying could be expected to produce more resistance than prescribing certain ways of speaking a language they have yet to learn. Yet in actuality puristic attitudes are likely to cause problems in both sets of circumstances, as a few illustrations will indicate. The Tiwi language, spoken on Melville and Bathurst Islands off the north coast of Australia, shows an all too typical profile for an indigenous language overtaken by the rapid expansion of a wide-currency language, in this case English (Lee 1987, 1988). With exposure to intense pressure from English, quite radical changes have taken place in the structure of the language over a short time period; thus an older, already largely bilingual generation – which knows (and among its own members still uses) a conservative traditional form of the language – co-exists with a younger and wholly bilingual generation, which uses a much modified form of the same language. The traditional language is polysynthetic, with a particularly complex verb structure. The elderly still control this form of the language; but younger speakers’ Tiwi shows changes in phonology, lexicon, noun classification, syntax, and, above all, in verbal constructions. The verbal construction in N[ew] T[iwi] comes from the traditional T[raditional T[iwi] verbal construction . . . but there are fewer inflections on the auxiliary. Also the small class of free form verbs has been expanded by a greater use of loan verbs from English and also some simple imperative forms from TT. In NT basically the only inflection on the auxiliary is the prefix(es) though these are normally changed phonologically. (Lee 1988: 82)

1 Languages of this sort correspond to Stage 7 in Fishman’s typology of threatened statuses (1991: 86–7).

250

chapter 14

Examples from Lee include the following:2 (1) TT: yi-p-angurlimay he.P-con-walk NT: wokapat yi-mi walk he.P-do ‘he walked’ (2) TT: ngi-rri-min-j-akurluwunyi (nginja) I-P-you(sg.)-con-see you(sg.) NT: lukim ngi-ri-mi nginja see I-con-do you(sg.) ‘I saw you’ In theory, the Tiwi language is in a relatively favorable position for a small indigenous language in a region colonized by Europeans: rather than being wholly abandoned by all but the very elderly, it has continued to be spoken by younger people, even though it has undergone drastic changes in the usage of those younger people. In terms of revitalization efforts, however, the situation is actually a very difficult one: with a steep continuum of varieties of Tiwi stretching from the fully traditional (and agglutinatively complex) language to a much simplified language, with many free forms introduced from (Pidgin) English, what form of Tiwi can or should realistically be supported? When a bilingual program was begun, with the approval of the Tiwi, at the Roman Catholic school on Bathurst Island in 1975, the intention was to use the traditional language as the medium of instruction in the early grades, with a gradual transition to English to follow. Lee reports that “this is what was desired by the community” (1987: 7). The primers and readers designed for school use are produced by what has grown into the Nguiu Nginingawila Literature Production Centre, associated with St. Therese’s School. They are beautifully illustrated and are geared very much to the children’s own culture: the human figures are those of Aboriginals, the flora and fauna are local, and the content deals with Tiwi legends, history, and ways of doing things. These texts are not in any way translation equivalents of typical English-language texts, but fully Tiwi-oriented originals. But while subject matter and illustrations seem clearly appropriate to the Tiwi children for whom they are intended, the language of the texts is unavoidably problematic. 2 The abbreviations are: P ‘past’, CON ‘concomitative’; the parenthesized element in the second TT line is optional. Note that in each case the NT verb is a borrowing from (Pidgin) English: wokapat derived from walkabout and lukim from look.

purism vs. compromise in language revitalization

251

The older people within the Tiwi community of Bathurst wanted the traditional language used, passing to the school the job of teaching the children this conservative form of the language which was no longer being transmitted naturally via the family. This is in itself dubious policy. Fishman 1991 devotes a full chapter to “limitations on school effectiveness in connection with mother tongue transmission,” pointing out that “without considerable and repeated societal reinforcement schools cannot successfully teach either first or second languages” (371). The Tiwi situation was more difficult than average, since attempts to use Traditional Tiwi for school purposes meant not just teaching literacy, but also providing oral Tiwi instruction: the children were being introduced to a form of Tiwi quite different from the range of Tiwi styles they were most familiar with, and sharply different from the kind of Tiwi they actually spoke. The difficulties were such that the bilingual program shifted over time away from purely Traditional Tiwi toward the various styles of what Lee (1987: 80) calls Modern Tiwi: “a modified/simplified traditional Tiwi.” Lee herself, in her teaching efforts on Melville Island for the Summer Institute of Linguistics, tried to lessen the gulf between what young people actually spoke and what they would encounter in their first efforts at reading Tiwi by preparing comics (a genre she thought might seem acceptable for use of the modified Tiwi of the young) with the text in “a formal style of N[ew] T[iwi]” (91); she also encouraged teacher trainees to write stories initially in language they would ordinarily speak, and subsequently to put that material into a more “proper” style if they wished, as most did (1988: 92). She encountered difficulty, however, when a wave of puristic conservatism greeted the texts she had prepared. By writing down NT the author [i.e. Lee] seems to have inadvertently “stirred up a hornet’s nest.” Although the materials have been apparently accepted and enjoyed by some people, a number have objected to seeing NT in writing . . .  Because of the strong reaction from influential members of the community, the author is drawing back from producing materials in NT. She and her SIL colleague hope to act as catalysts in helping the school, church and community to work out suitable forms of expression allowing individual expression. (1988: 92) Lee worries, not without grounds, about the utility of all the support work for Tiwi. She fears that the language will not derive realistic benefit from text preparation, school programs, the compiling of a dictionary, teacher-training programs, or anything else that might be undertaken, if the children who are the targets of all these efforts do not find the Tiwi they are exposed to enough like the Tiwi they speak to interest and encourage them to use the language:

252

chapter 14

. . . the Tiwi language situation is an extremely complex one. It is a very sensitive issue for many of the Tiwi people. The traditional language can only be acquired in all its intricacies through the regular and consistent use of it in the home and camp environments. However, this is an impossible situation as many of the parents of the children, young adults themselves, do not speak the traditional language as their first language. The situation may be saved if older people are willing to concede to a simplified form of T[raditional T[iwi] as being acceptable. Even so, a concerted effort with the support of the community as a whole would be needed for such a style of Tiwi to be accepted as the norm. (1988: 93) In the Tiwi situation, there is an echo of the problem that plagues Arabic: puristic norms militate against the teaching of the students’ actual language forms, and instead promote unrealistic norms from an earlier version of the language. Tiwi, like other hard-pressed minority languages in intense contact situations (Dorian 1981: 151–53, Schmidt 1985: 213–14), has undergone major change in a sharply foreshortened time frame. In the span of three or four co-existing generations, it has reached something resembling the distance from a traditional model which Arabic has reached over many centuries. But in the case of Tiwi, it is neither the teachers at the Roman Catholic school nor the linguists of the Summer Institute of Linguistics who chose to promote the most conservative form of the language, but “influential members” of the community itself. The problems related to purism which arise in the case of Irish are different, but no less difficult. Puristic attitudes operate in two different directions in the Irish case. One set of conservative attitudes exists in the tiny remaining heartlands of the Irish language, consisting mainly of several non-contiguous extreme-western pockets of Donegal, Mayo, and Kerry; the other exists in the positions taken by the standardizers who were responsible for arriving at a normalized Irish suitable for country-wide use in textbooks, official documents etc. The two seem irreconcilable. Irish, in contrast to Tiwi, has an exceptionally long literary history, with a written tradition dating to the seventh century. All the same, when the revitalization movement took hold in the late 19th century, the last great period of the written language (Early Modern Irish, also known as Classical Irish) lay more than 200 years in the past. In the interim stretched a period of drastic decline in the number of speakers, of restriction in geographical distribution, of contraction of spheres of use, and of repression or neglect of the language by the governing powers. What had been a brilliant literary language survived in the monuments that had been produced by its practitioners; but so far as the spoken language was concerned, what remained was rustic in character, surviving

purism vs. compromise in language revitalization

253

in daily use almost exclusively among a peasantry. There were marked dialect distinctions from region to region. When political independence lent practical means to the hopes of those who wished Irish Gaelic to become once again the distinctive language of the Irish people, and to come into regular use among them, modernizing and standardizing tasks loomed large. It was not feasible to take one of the living Irish dialects as the clear-cut basis for a modern standard language, since none of the three main dialects had any obvious superiority in prestige or numbers (Ó Baoill 1988: 111); without such a realistic basis for selection, the speakers of each dialect were certain to object to any one dialect among the three being singled out for official favor. Compromise was necessary, but the result was inevitably artificiality. Revitalization required a single, normalized form of Irish which could transcend dialect differences. To speakers of living Irish dialects, however, the result is Gaeilge B’l’ Ath’ ‘Dublin Irish’, a stilted, unnatural form of Irish (Hindley 1990: 60). The puristic conservatism of native Irish dialect speakers takes authenticity as its chief virtue, and “Dublin Irish” fails the test. If authenticity is the form which conservative attitudes take in the rural Gaeltachtaí – the scattered residual areas where one of the living dialects is still spoken natively – historicity might be said to be the form which conservative attitudes take in the “official” Irish ultimately produced by the long labors of the standardizers. The standardized form of Irish steers clear of extreme regionalism, but makes less effort to steer clear of the grammatical complexities of conservative forms of the language. Literature written in the standard form creates very few problems for the average reader. In trying to use the system, however, even the most competent users have to often consult the dictionaries or handbooks. This is due mainly to the complicated morphological system of Irish, and the standard now evolving has not succeeded to any great degree in reducing the complex system of grammatical rules involving the use of inflections and the mutations of initial consonants of Irish to express different shades of meaning. Many of the rules and forms advocated by the standard have been simplified in the speech of native speakers . . . (Ó Baoill 1988: 117). To each his own form of conservatism, it seems. The standardizers, who were by necessity men of some erudition, found it possible to dispense with regionalism and idiomaticity, but not with traditional grammar. Native speakers, for their part, had found it possible over the centuries to dispense with

254

chapter 14

some of the more complex features of the traditional grammar; but in each locality they preserved the distinctive speech of their own region with its own forms, phraseology, and idioms. While each form of Irish has certain clear advantages – supra-regionalism and uniformity in the case of the standardized Irish, realistic local vividness in the case of the regional dialects – each has faults that limit its overall usefulness. By far the fullest and most penetrating account of the threat which purism can pose to a small language community is Jane Hill and Kenneth Hill’s study (1986) of Mexicano (Nahuatl, i.e. modern Aztec) in the Malinche region of Mexico. Mexicano and Spanish have co-existed for centuries, with contact phenomena appearing in both languages as spoken regionally; but Spanish influences in Mexicano are the more evident. Although the Hills take the view that “an ecological perspective can see linguistic syncretism as having a positive, preservationist effect on a language when its speakers must adаpt rapidly to changing circumstances” (59), some local attitudes are less tolerant. As part of a heightened attention paid to the ethnic boundaries of the local population and their towns, native purists reject the syncretic form of Mexicano, with its high Hispanic content, and insist on an artificial variety which is not only unrealistic, but in some instances even inauthentic (140). The result is a selfconsciousness and exclusivity with potentially harmful consequences. Purism in the Malinche towns may work against the survival of the Mexicano language. Since Mexicano is considered to be of very little economic utility, many people question the instrumental value of the language. Purism, which deprecates all modern usage, inspires speakers to question the moral and aesthetic value of Mexicano as well . . . Since no formal education about Mexicano is available in the Malinche towns, it is unlikely that young speakers can be educated to a purist standard, and when young speakers feel that their Mexicano is inadequate, they may choose to use only Spanish. (Hill & Hill, 140–41). The Hills note (122–23) that Mexicano purism focuses most zealously on lexicon as a particularly salient locus for contamination. External lexical influence is usually conspicuous in contact situations; in most revitalization efforts, the problematics of conservatism is very clearly to be seen in connection with attempts to nativize the lexicon while also updating it. There are normally two obvious options: to borrow or to coin. The difficulty is that the remaining native speakers often reject both. By way of illustration I offer anecdotal evidence from my personal experience with the reception of a coinage, on the one hand, and a pair of borrowings, on the other, by Scottish Gaelic speakers (cf. Dorian 1978).

purism vs. compromise in language revitalization

255

One friend, a telephonist in a family of telephonists and a native speaker of a variety of Wester Ross Gaelic, came across the coinage being promoted in one or another Scottish Gaelic publication for ‘telephone’. This took the rather straightforward form guthan, built on the noun guth ‘voice’. She delighted for a time in quizzing every native speaker who came along as to the Gaelic word for ‘telephone’ – stumping them all, of course, since the word was completely alien to ordinary speakers. Most tellingly, she did not actually adopt the word for regular use herself. Another friend, a native speaker of an Inner Hebridean Argyllshire Gaelic dialect, a number of times told scornfully of hearing a broadcaster ask the fellow whom he was interviewing De an team a tha thu a’ support-achadh? ‘What team are you supporting?’, with the English words team and support painfully obvious in the otherwise Gaelic sentence. It occurs to me in retrospect, however, that my friend never, in all the times he brought that interview up, suggested anything the broadcaster might have said in Gaelic instead in order to avoid the borrowings. Fishman notes (1991: 347–48) that pilot testing of proposed neologisms has been carried out successfully in Sweden – a tactic which can prevent coinage from being too exclusively the creature of language planners with esoteric knowledge and philological biases. However, this approach may require a level of finance and technology not readily available to many small languages. In some language settings, especially those with a certain residual vigor, semantic extension may be partial solution to the problem of updating the lexicon. Words meaning ‘a band (of men)’ and ‘prop, hold up’ might be pressed into service for team and support, or modified slightly to take on those meanings. Slightly archaic words could be reintroduced in the new meanings; or dialect forms might be given the new meanings, and introduced into general use. If another Celtic language already had words for these concepts, analogs could be created in Scottish Gaelic. All of these methods were used in the creation of Modern Turkish lexicon during the language reforms of Atatürk’s time, with considerable success (Heyd 1954), and they are part of the arsenal of language planning generally. But in the case of Scottish Gaelic – with few country-wide communications links among speakers, with no generally accepted spoken norm, and with full literacy not yet wide-spread – language planning efforts have been limited, and have had correspondingly limited success. To be effective, coinage and semantic extensions both require the support of a lively broadcasting industry, educational system, and publishing industry; these phenomena are only just appearing in Scottish Gaeldom. Speaker conservatism has usually been profound, and novel usages have been the object of derision and rejection. These reactions are hardly limited to Gaelic-speaking Scots, of course; they are responses already familiar from other language-planning

256

chapter 14

efforts. Yet today’s widely used coinage or extension was often yesterday’s laughingstock; transition from the latter to the former can seem random and mysterious, and the observer may see no reason to account for the successes, as opposed to the failures.3 It’s reasonably easy to appreciate the difficulties posed by conservative attitudes in instances of revitalization. By comparison, revival settings look invitingly free of potential resistances, since speakers have no entrenched habits to overcome. In revival settings, however, the hazards of rival proposals, giving rise to rival factions, pose just as great a threat; purism of one sort or another is quite likely to be at the heart of the rivalry. If the language to be revived is well preserved, even though not conversationally spoken, there may be disparate traditions for rendering it phonologically, as was the case with Ashkenazic and Sephardic pronunciations of Hebrew. If the language to be revived is not well attested, or is attested in more than one earlier form, there may be disparate reconstructions of the language itself, forming the basis of rival teaching materials for modern-day learners. This is the unhappy situation of Cornish, the Celtic language of Cornwall, which survived as a spoken language up to the late eighteenth century but is only moderately well attested as a written language. Efforts to recover Cornish, and to create the texts and reference works which would make it accessible to those who wanted to become acquainted with it, began in the very early twentieth century. Antiquarian interest in the language gradually gave way to a more active involvement, with some individuals and groups espousing revivalist sentiments. Two different approaches to locating “true” Cornish have led in very recent years to rival revivalist factions, each promoting its own version of the language. One faction favors Dr. Kenneth George’s version of Cornish, based on meticulous computer analysis of the grammar, lexicon, and orthography of the available Cornish texts, which span the period roughly from the twelfth to the seventeenth centuries (Ellis 1974, O’Callahan 1989: 27). The other favors Richard Gendall’s version of Cornish, 3 The struggle to eliminate foreign loanwords from the German language produced vast numbers of coinages and extensions from the 17th century on. Browsing through a few Verdeutschungswörterbücher produces some smiles at the apparent absurdities that zealous Germanizers urged on their compatriots; but these are quickly balanced by surprise at the realization that many words, utterly respectable and ordinary now, were coinages no less novel and curious in their time than the coinages which failed to gain acceptance and so produce smiles. In the 19th century, Arthur Schopenhauer found the coinage Stickstoff, recommended by the writer Campe as a replacement for Nitrogen, so ugly that he suggested Azot instead (Tschirch 1969: 260). But Campe prevailed; modern German dictionaries offer Stickstoff for ‘nitrogen’, but no Azot.

purism vs. compromise in language revitalization

257

based on Edward Lhuyd’s Cornish grammar of 1707 and on a modest amount of fossilized Cornish which survived in one or another enclave (e.g. among fishermen in West Cornwall) into the nineteenth and even the twentieth centuries. Both factions claim authenticity; the George faction claims the authenticity of grounding in scrupulous analysis of all established written texts, and the Gendall faction the authenticity of oral tradition (Lhuyd having based his grammar, according to them, on direct work with surviving Cornish speakers). The rivalry is acute and unfriendly, with each faction competing for the loyalties of prospective learners. There seem to be a fair number of learners and would-be learners, and there are perhaps 50 to 100 fluent speakers (O’Callahan 1989: 30, Anonymous 1990: 19). Some of the fluent speakers are now raising Cornish-speaking children, so that Cornish can once again claim to be a living mother tongue. On the one hand, the dedication of Cornish enthusiasts in bringing the language to a genuine “life,” if only in the mouths of a few child speakers, can be admired. On the other, the respective purisms, one textual and the other folkloristic, must be regretted insofar as they siphon off the energies of revival workers and alienate the sympathies of potential supporters. Bro Nevez, the newsletter of the US branch of the International Committee for the Defense of the Breton Language, published a rancorous letter by Richard Gendall, in response to an article in an earlier issue which appeared to favor the George version of Cornish; and the editor commented, in an appended note: “If some of the tremendous energy Celts have used to belittle each other’s ideas of ‘the truth’ was directed towards working for more resources to support research, teaching, and media use of Celtic languages and arts, people would not need to talk so much about survival” (Kuter 1989: 40). Revival leaders might do well, in the spirit of Kuter’s suggestion, to concede that more than one kind of authenticity exists, and to begin the more productive work of establishing a compromise version of Cornish which sacrifices a modicum of each form of authenticity in favor of learnability. If declensional patterns should be more regular in one of the versions of Cornish, but lexical coherence best reflected in the derivational patterns of the other, then it would serve potential learners well to promote a single form of Cornish which incorporated both of these features, even though they might derive from different approaches to reconstructing the language. This sort of compromise, if feasible, might achieve a channeling of the energies of linguistic conservatism for useful purposes, as Kuter urged. There has recently been some actual evidence in the literature of language obsolescence to suggest that, in cases where a small or otherwise precariously placed language has survived longer than might have been expected, an

258

chapter 14

absence of puristic attitudes may have characterized some speakers. Hamp (1989: 198–99) found, for instance, that phonological intactness was no measure of survival potential among the pockets of Albanian, known as Arvanitika, scattered through Greece. On the contrary, the youngest speakers with serviceable (if very incomplete) Arvanitika came from Attica and spoke with a substantially hellenized phonology. In Eleia, at the same period, he found by contrast only one old man who could attempt some minimal bits of Arvanitika – to be sure, with a phonology which preserved more of the original phonological characteristics of an Albanian speech form than the relatively serviceable Arvanitika of the university students from Attica.4 The suggestion in the Arvanitika case that structural compromise is not necessarily deleterious to the continued use of a small language is supported and strengthened by the work of Huffines in her research with two different groups of Pennsylvania German (PG) speakers. Among what she calls the nonsectarians (i.e. the non-Anabaptists), she found that the older generation for whom PG was the native language spoke a relatively conservative PG, showing little convergence with English; yet within this linguistically conservative community, “the death of PG . . . is rapid once it begins and is complete across three generations, often across two” (Huffines 1989: 225). Among the Mennonite and Amish sectarians in her study, shifting into English is impermissible within the community itself (though not in dealing with outsiders), but convergence with English and incorporation of English loanwords is commonplace. The sectarians’ German speech is not in immediate danger of disappearing, but it is noticeably less conservative than that of the non-sectarians. Huffines concludes, regarding the sectarians’ flourishing PG, that “sociolinguistic norms prescribe its use but not its form” (225). The greater conservatism of nonsectarian PG is not necessarily causal in its decline, but it is at any rate clear that it has not operated to preserve the language. That is, structural or lexical purity is not in itself a key to survival, nor does “impurity” necessarily represent an opening of the floodgates to external influences which must inevitably swamp a small language. A perceived need for linguistic integrity may in some cases offer a rallying point for revitalization or revival, especially among the intellectuals who are often the spearhead of such a movement (cf. Fishman 4 In the village of Embo, East Sutherland (Scotland), where Gaelic is dying out, it surprised me originally to find that the two mothers whose children are today the youngest Gaelic speakers were less conservative in their Gaelic than many of their older-fluent-speaker peers. It’s possible, however, that the willingness of these two women to adapt their Gaelic somewhat toward young-fluent-speaker norms may have been a factor in producing a home environment which encouraged their children to speak Gaelic, when most young people of comparable ages were not actively acquiring the language.

purism vs. compromise in language revitalization

259

1989: 229–301); but a sense of differentness, sufficient to sustain a separate identity, need not rest on a native linguistic purity that will stand up to strict etymological and grammatical scrutiny. In an instructive case of conscious language revival from Belfast, North­ ern Ireland, it appears that movement away from conservative norms may represent a price to be exacted in return for the emergence of young native speakers. Late in the 1960s a group of couples, learners of Irish, deliberately formed a community in Belfast where they could raise their children as Irish speakers. Members built their own homes on a site secured by a company which they established, and in 1971 a primary school using Irish as the medium of instruction was opened. The children of the 11 Shaw’s Road families formed the original nucleus of the school population; but in 1985–86 there were 194 pupils enrolled, including many from neighboring areas (Maguire 1987: 74). For the Shaw’s Road children, Irish was the first language of their homes, their neighborhood, their primary school, and their church. They are effectively Irish speakers, as their parents had hoped; but their Irish turns out to deviate markedly from the norms of the grammar books to which they were exposed in school. In particular, quite a number of grammatical features which are conveyed wholly or in part by changes in the initial consonants of nouns, verbs, and adjectives are often compromised – e.g. gender, certain tense forms, and genitive possessive constructions. Analogical verb forms appear in place of traditional irregular ones, and analytic ones in place of synthetic ones. In addition, English syntactic influence is evident, and English lexical items are plentifully borrowed. Gabrielle Maguire, who reports on the Shaw’s Road community, shows a keen awareness of the difficult balancing act called for in revival settings when she presents her findings on linguistic developments in [the] children’s Irish in her book-length study of that community (1991: 186–228). She quotes Haugen’s warning (1977: 101) about the risks arising from “linguistic straitjacketing”: “It may be better to bend than to break. Acceptance of useful convergence between codes is better than a total rejection of the mother tongue, which is likely to result if one always and everywhere insists on [the] rigid rhetorical norms of the academicians.” But at the same time (191) she contends that Haugen was not endorsing “allowing the language to landslide into the system of the D[ominant] L[anguage],” and she urges “a firm grasp on the reality of what constitutes healthy, inevitable change within a particular set of circumstances.” The problem, of course, is to identify what constitutes “healthy” change. Some adopt the view that the end-product is healthy by definition if it survives, as opposed to disappearing, regardless of its form. Others consider a highly convergent outcome too poor a representative of the original language

260

chapter 14

to count, and so disdain it. This is a value judgment and should be recognized as such. When the convergence in question lies safely in the past, the disagreement is innocent enough. English emerged from its period of social subordination to French, in the wake of the Norman Conquest, quite different in form from the English which had existed before the Conquest. Arguments can be made (and have been) for its continuation as a distinctly Germanic language – or for its latterday emergence as a mixed language, or even a creole. When convergence features are evident in a present-day speech form, the debate can take on more than academic interest. Maguire notes that the Shaw’s Road children are capable of some degree of grammatical monitoring and avoidance of English loanwords in more formal situations (1987: 87, 1991: 228). But overall she finds that they “adapt their [Irish language] system to suit their own needs,” and her summarizing comment (1987: 88) leans to the conservative side: “Although communicative competence and functional adequacy are mastered, a language which is very much on the defensive must aim higher in order to ensure its own separateness from the dominant language.” This is at heart a puristically inclined evaluation. In its absolute form, it is belied by the evidence of the sectarian speakers of Pennsylvania German; but it is certainly true that a sense of separate identity is a valuable sustaining feature in ethnic language revival and revitalization efforts. Ó Baoill, considering the outlook for preservation of traditional Irish phonological contrasts among speakers of whom many or most will be learners, in the context of the Irish Republic, considers compromise a likely necessity: “If Irish is to become a viable means of communication among the general population, I fear that much leveling will take place, and it is certain that many of the contrasts now existing in Irish will be lost. If the revival of Irish were to succeed, then it might all be worthwhile” (1988: 125). Ó Baoill’s is a slightly toothgritting embrace of revival, since he suspects that it can only come at a cost of phonological leveling in the original language. His predictions for Irish may have been embodied for Greek Albanian in the young Attican semi-speakers of Arvanitika whom Hamp encountered. Maguire and Ó Baoill both hope for the preservation of Irish and the proliferation of speakers of Irish. Maguire, like many language loyalists before her, is asking how dilute a language can become while still remaining the linguistic entity it was – distinct from all others, including (and especially) the neighboring language of wider currency. Ó Baoill, in a more pragmatic tradition, is asking how traditional a threatened language can afford to remain if its traditional forms pose obstacles to learnability and hence dissemination. In the best of all

purism vs. compromise in language revitalization

261

possible worlds, one would not need to choose, of course, and Fishman’s characterization of enlightened planning in a nationalist framework would prevail: “The enlightenment of nationalist purism in language planning . . . proceeds along many well-trodden paths: the differentiation between ethnic core and nonethnic periphery, between technical and nontechnical, the differentiation between preferred and nonpreferred sources of borrowing, and, finally, the appeal to common usage among the masses” (1989: 309). But in very small language communities which have no nation to their name, and little immediate prospect of acquiring anything deserving the term “masses,” a choice may be unavoidable. The rapidity of change and the expansion of contacts with other peoples add to the pressures. Drapeau (1992: 3) points this out in connection with Betsiamites Montagnais, an Algonquian language with a moderately solid speaker base but extensive exposure to French, despite its geographical isolation in Northern Quebec. The need for lexical elaboration is so high in persistent linguistic enclaves . . . confronted with the communicative demands of modern life, that there is no way for these communities to cope with this problem without importing massively, overburdened as they are by the sheer number of items to create [by coinage]. On the evidence of the difficulties posed by puristic stances for even very large modernizing languages, like Hindi and Arabic, and with the suggestive findings of Hamp and Huffines in cases at the other end of the spectrum as an encouragement, it may prove the wiser course to accept considerable compromise rather than make a determined stand for intactness, where threatened languages are at issue. If a language survives, after all, it has a future. If it can never again be exactly what it once was, it may yet be something more than it now is. Gifted speakers and writers may eventually appear who will coax new richness of expression from it, and tease it into forms that will be uniquely its own, even if not those of its past. Ælfric might well have been horrified at what Chaucer called English, had he lived to see it, since English emerged in a markedly altered state, both lexically and grammatically, from two centuries of domination by the Norman French and their language. But if Chaucer wrote in a sharply modified and even gallicized form of English, by comparison with that of Ælfric, that did not prevent Chaucer from writing masterful and enduring literary works. Purity need not be a requirement for persistence, and compromise need not be the death knell, for small languages any more than for larger ones.

262

chapter 14

References Anonymous. 1990. Le cornique: Renaissance d’une langue celtique. Ar Men 29: 3–19. Coulmas, Florian. 1989. Language adaptation. In Florian Coulmas (ed.), Language adaptation, 1–25. Cambridge & New York: Cambridge University Press. Dorian, Nancy C. 1978. East Sutherland Gaelic. Dublin: Institute for Advanced Studies. ———. 1981. Language death. Philadelphia: University of Pennsylvania Press. Drapeau, Lynn. 1992. Language birth: An alternative to language death. Paper presented at the XVth International Congress of Linguists, Québec. Ellis, P. Berresford. 1974. The Cornish language and its literature. London: Routledge & Kegan Paul. Fishman, Joshua A. 1989. Language and ethnicity in minority sociolinguistic perspective. Clevedon, England: Multilingual Matters. ———. 1991. Reversing language shift. Clevedon, England: Miltilingual Matters. Hamp, Eric P. 1989. On signs of health and death. In Nancy C. Dorian (ed.), Investigating obsolescence, 197–210. Cambridge & New York: Cambridge University Press. Haugen, Einar. 1977. Norm and deviation in bilingual communities. In Peter A. Hornby (ed.), Bilingualism: Psychological, social and educational implications, 91–102. New York: Academic. Heyd, Uriel. 1954. Language reform in modern Turkey. Jerusalem: Israel Oriental Society. Hill, Jane H., & Hill, Kenneth C. 1986. Speaking Mexicano: Dynamics of syncretic language in Central Mexico. Tucson: University of Arizona Press. Hindley, Reg. 1990. The death of the Irish language. London: Routledge. Huffines, Marion Lois. 1989. Case usage among the Pennsylvania German sectarians and non-sectarians. In Nancy C. Dorian (ed.), Investigating obsolescence, 211–26. Cambridge & New York: Cambridge University Press. Ibrahim, Muhammad H. 1989. Communicating in Arabic: Problems and prospects. In Florian Coulmas (ed.), Language adaptation, 39–59. Cambridge & New York: Cambridge University Press. Kuter, Lois. 1989. A note from the editor. Bro Nevez 32: 40. Lee, Jennifer. 1987. Tiwi today: A study of language change in a contact situation. Canberra: Australian National University. ———. 1988. Tiwi: A language struggling to survive. Australian Aborigines and Islanders Branch, Summer Institute of Linguistics, Work Papers 13 (series B), 75–96. Maguire, Gabrielle. 1987. Language revival in an urban neo-Gaeltacht. In Gearóid MacEoin, Anders Ahlqvist, and Donncha Ó hAodha (eds.), Third International Conference on Minority Languages: Celtic papers, 72–88. Clevedon, England: Multilingual Matters. ———. 1991. Our own language: An Irish initiative. Clevedon, England: Multilingual Matters.

purism vs. compromise in language revitalization

263

Ó Baoill, Donall P. 1988. Language planning in Ireland: The standardization of Irish. In Padraig Ó Riagáin (ed.), Language planning in Ireland (International Journal of the Sociology of Language, 70), 109–126. Berlin: Mouton de Gruyter. O’Callahan, Joseph. 1989. Cornish. Bro Nevez 31: 27–31. Schmidt, Annette. 1985. Young people’s Dyirbal. Cambridge & New York: Cambridge University Press. Tschirch, Fritz. 1969. Geschichte der deutschen Sprache, 2: Entwicklung und Wandlungen der deutschen Sprachgestalt vom Hochmittelalter bis zur Gegenwart. Berlin: Schmidt.

chapter 15

Western Language Ideologies and Small-Language Prospects* It might be said with a certain metaphoric license that languages are seldom admired to death but are frequently despised to death. That is, it’s relatively rare for a language to become so exclusively tied to prestigious persons and high-prestige behaviors that ordinary people become too much in awe of it to use it or are prevented by language custodians from doing so. By contrast, it’s fairly common for a language to become so exclusively associated with lowprestige people and their socially disfavored identities that its own potential speakers prefer to distance themselves from it and adopt some other language. Parents in these circumstances will make a conscious or unconscious decision not to transmit the ancestral language to their children, and yet another language will be lost. The power of the social forces involved is evidently considerable, since under better circumstances attachment to an ancestral mother tongue is usually strong. The phenomenon of ancestral-language abandonment is worth looking at, then, precisely because a good many people, especially those who speak unthreatened languages, are likely to have trouble imagining that they themselves could ever be brought to the point of giving up on their own an ancestral language and encouraging their children to use some other language instead.  Unless they become fossilized so that they persist in specialized uses without ordinary speakers, as sometimes happens in connection with religious practices (Latin, Sanskrit, Coptic Egyptian, Ge’ez, etc.), languages have the standing that their speakers have. If the people who speak a language have power and prestige, the language they speak will enjoy high prestige as well. If the people who speak a language have little power and low prestige, their language is unlikely to be well thought of. Because the standing of a language is so intimately tied to that of its speakers, enormous reversals in the prestige of a language can take place within a very short time span. The arrival of the Spaniards brought about precipitous changes of this kind in the fortunes of two major New World languages, that of the Aztec empire in North America and that of the Inka empire in South America. Both * I’m indebted to Christina Bratt Paulston for helpful criticisms of the first draft of this chapter and for suggestions for its improvement.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�17

western language ideologies and small-language prospects

265

had achieved great dominance, expanding at the expense of neighboring languages for some centuries as the Aztecs and Inkas conquered new territories and made ever more peoples subject to their rule (Heath 1972; Heath and Laprade 1982). In a stunningly short time both empires were brought low by their encounter with the better armed Spanish, who represented an expanding Old World power. Neither imperial language disappeared, but each survived with severely reduced social standing. Today Nahuatl and Quechua are low prestige speech forms within the regions where they are spoken, and each is under some threat from still expanding Spanish.  To be sure, cases exist in which a conquering power has given up its own language and adopted the language of the very people whom it has conquered. The Vikings seem to have been particularly susceptible to this, going over to Romance speech forms in Normandy and Sicily and to a Slavic speech form in Russia. It is not unique to them, however; the western Franks and the Bulgars followed a similar pattern, as did the Normans in England, repeating the pattern of their Viking forebears in Normandy. In such cases the conquering group is usually numerically thin, compared with the size of the conquered population, and it may deliberately intermarry with the indigenous aristocracy (for lack of enough women of its own group or for the sake of adding legitimacy to its seizure of local power and property, or both). Distance from the original homeland probably plays a role in some such cases, as in the Viking kingdoms, all established far from Scandinavia. Military loss of home territories can have the same distancing effect. The anglicization of the Normans in England might have been delayed or even prevented if they had been able to retain control of Normandy; but they lost their Norman territories less than a century and a half after conquering England, and from that time forward their focus was on their English territories.  In any event, these are the unusual cases rather than the norm. In the more usual cases, the group that exercises military or political power over others will establish its own language as the language of governance in its contacts with those others. And when one speech form enjoys a favored position as the language of those who control obvious power positions (as administrators, governors, judicial officers, military officers, religious officials, major landholders, and so forth), it requires no great sagacity, but only common sense, to see that it’s likely to be useful to acquire some knowledge of that language. If members of a subordinate population have the opportunity to learn the language of the dominant group, some or all of them will usually do so. They will not necessarily give up their own ancestral language, however. It seems likely that it’s not so much the tendency to learn a dominant-group language which has increased a great deal in modern times, but rather the opportunity to do so, and,

266

chapter 15

concomitantly and more importantly for linguistic diversity, the tendency to abandon one’s ancestral language entirely in the process. To understand this last phenomenon, it may be necessary to consider what in the last two centuries or so is characteristic of Western (i.e. European-derived) attitudes toward non-standard speech forms, since tendencies towards complete ancestral language abandonment seem to be very strong in the widely distributed areas of European settlement.  Ruling powers have not always expected subordinate peoples to give up their ancestral languages or encouraged them to adopt the language of the dominant group. The Ottoman Empire encompassed an extraordinary variety of subordinate ethnic groups but permitted them to retain a good deal of their ethnic identity, including native religious and linguistic practices, in the various milletler (“nations”) within its domains. Even European states were moderately permissive of ethnic languages until relatively recent times. The rise of nationalism in Western Europe at the beginning of the industrial age coincides to a considerable extent with less tolerant attitudes towards subordinate languages. In the present day, for example, France has shown unusual intolerance of ethnic distinctiveness, even for a Western European country (refusing birth certificates and identity cards to children with Breton given names, for example, as recently as the 1970 [New York Times 1975]). Yet cultural and linguistic diversity was an unproblematic fact of life in France until the 1790s, when in the aftermath of the French Revolution a need for a unifying national identity, expressed in part by a single national language, was rather suddenly perceived (Grillo 1989: 22–42; Kuter 1989: 76).   The fact that powerful pressures for cultural and linguistic unity emerged in France around the time of the Revolution is not accidental from the perspective of some students of nationalism. Rather the pressures emerged at that time because of the particular stage of development the country was reaching and the changes attendant on such a stage. Ernest Gellner identifies a pre-industrial “agro-literate polity” in which the uppermost social strata (e.g. nobility, clergy, merchants) are sharply layered horizontally vis-à-vis one another, with the layers prevailing across the polity as a whole, while a variety of distinct small communities coexist laterally separate from each other, within the polity and beneath the upper strata. In societies with this sort of social organization, Gellner (1983: 10) describes the state as “interested in extracting taxes, maintaining the peace, and not much else, and . . . [with] no interest in promoting lateral communication between its subject communities.” In industrial societies, by contrast, conditions are quite different Industrial means of production require universal literacy and numerical skills such that individuals can communicate immediately and effectively with people previously unknown

western language ideologies and small-language prospects

267

to them. Forms of communication must therefore be standardized and able to operate free of local or personal context. This in turn places great importance on educational institutions, which must produce individuals with certain generic capacities that permit slotting and re-slotting into a variety of economic roles. The state is the only organizational level at which an educational infrastructure of the necessary size and costliness can be mounted (Gellner 1983: 35–38).  France offers a particularly good (and particularly well-studied) example of rising standard-language dominance at the dawn of the industrial age. At the time of the Revolution, France was passing out of the agroliterate stage of development into a pre-industrial stage, and with a new focus on the polity as a totality the fact that a number of sizable subcommunities such as the Bretons, Basques, Alsatians, and Occitanians were incapable of understanding and speaking French became unacceptable. In 1794 the Abbé Gregoire, priest and revolutionary, presented a report to the National Convention in which he detailed this lamentable situation and called for the universalization of the French language. Under the monarchy, as the revolutionaries saw it, linguistic heterogeneity had been useful to the crown as a means of keeping “various feudal constituencies from making common cause with one another (Grillo 1989: 35). In the revolutionary French state there could be no place for such policies. As part of the social and ideological transformation they were engaged in, the citizens would be unified by common use of a single language, namely the French language (Grillo 1989: 30, 34).  Sentiment that could be called nationalistic had grown in France from the mid-fifteenth century onwards, as the French crown increased its geographic domain and its political strength (Joseph 1987: 133). Although France had no actual policy of linguistic unification before the Revolution, the prestige of French had been uniquely high nonetheless. The king and his court spoke French, and from that ultimate milieu of power and status the French language gained unrivaled luster (Grillo 1989: 29). The championing of French after the Revolution was perfectly in keeping with the usual linkage between highprestige people and a favored speech form, despite the Revolution’s abrupt termination of the French monarchy. So, too, was the disfavoring of speech varieties spoken by the relatively low-prestige peoples of the country. The speech forms of “vulgar” classes of people were tainted by the status of their speakers: they, too, were “vulgar.” Already in 1790, when the Abbé Gregoire conducted a survey that included questions about the influence of patois (by which he meant both French dialects and non-French vernaculars such as Basque and Breton) and about the consequences of destroying it in the various regions of France, the letters he received in reply to his questions indicated very low

268

chapter 15

opinions of the various regional speech forms, which were labeled coarse and stupid and were considered to keep the people ignorant and superstitious (Grillo 1989: 31, 174).  These were in fact commonplace attitudes in European polities.1 Grillo (1989: 173–74) states bluntly – and accurately, I think – that “an integral feature of the system of linguistic stratification in Europe is an ideology of contempt: subordinate languages are despised languages.” This has been true both where regional dialects are concerned and where the languages of subordinate ethnicities are concerned. In his study of the rise of language standards and standard languages, Joseph (1987: 31) suggests that language is particularly susceptible to what he calls “prestige transfer”:  Because the intrinsic worth of dialects and of their component elements and processes is well nigh impossible to determine, language is highly susceptible to prestige transfer. Persons who are prestigious for quantifiable reasons, physical or material, are on this account emulated by the rest of the community. These others cannot obtain the physical or material resources which confer the prestige directly (at least they cannot obtain them easily, or else no prestige would be associated with them). But prestige is transferred to attributes of the prestigious persons other than those on which their prestige is founded, and these prestigious-by-transfer attributes include things which others in the community may more easily imitate and acquire, if they so choose. Language is one of these.  He further considers that “the power which prestigious dialects hold over nonprestigious speakers goes beyond what logic and rationality can predict or account for,” and that the prestige-holding segment of a population can use the mechanisms of prestige-language standardization to maintain and increase linguistic differences between themselves and speakers of less prestigious speech forms (ibid.). The histories of several of the national languages of Europe, very conspicuously those of French and English (Grillo 1989), are histories of a growing monopoly on legitimacy and prestige by a single dominant speech 1 They were not uncommon in non-European contexts either, for that matter. The Aztecs used a variety of unflattering terms for the languages of their subject peoples, some of which stuck and became the name by which the language is still known, at least to outsiders. Derogatory language names deriving from Aztec labels include Chontal “foreigner”, Popoloca “unintelligible”, and Totanac “rustic” (Heath 1972: 3).

western language ideologies and small-language prospects

269

form, all others being relegated to inferior status. The standard language is typically considered a rich, precise, rationally organized and rationally organizing instrument; dialects and ethnic minority languages, by contrast, are considered impoverished and rude, most likely inadequate to organize the subordinate world itself and certainly inadequate to organize other worlds.2 European states such as France and Great Britain were unusual perhaps chiefly in their determined allocation of unique prestige and legitimacy to a single carefully cultivated supra-local speech variety as the nation’s official language. They were not unusual, certainly, in their allocation of higher prestige to a speech variety originally used by a materially more favored group or in their assumption of the superiority of their own mother tongues. Social status, whatever its basis, seems very generally to rub off on language, as Joseph indicates, so that the possession of wealth, however that wealth is calculated, will enhance not only the social position of the wealthy people but also the social position of the language that they speak. In East Africa, where the “cattle complex” prevails and wealth is measured by the size of the cattle herd, the languages of cattle-herding pastoralists have frequently displaced the languages of huntergatherers who own no cattle (Dimmendaal 1989:  16–24).  Europeans who came from polities with a history of standardizing and promoting just one high-prestige speech form carried their “ideology of contempt” for subordinate languages with them when they conquered far-flung territories, to the serious detriment of indigenous languages. And in addition to a language ideology favoring a single normalized language, derived from the history 2 One result of this is a tendency among other-language speakers in contact with standardlanguage speakers to consider that any feature in which their own language differs markedly from the standard language must indicate some own-language deficiency. In Scottish Gaelic the adjective normally follows the noun. Unfortunately for Gaelic speakers, the dominant standard language to which they compare their own is English rather than French. In East Sutherland, Gaelic speakers frequently remark that Gaelic “puts the cart before the horse” in this regard, implying a failing on the part of Gaelic (since carts don’t belong before horses). That is, they assume that English, which they were taught in school, represents things as they ought to be. Because the ancestral language is measured against dominant-language norms, it’s difficult for speakers who have no special training – and often no schooling in the ancestral language at all – to see in a positive light any unique or highly developed features of their own language. Gaelic, for example, has a very rich system of emphatic suffixes which can attach to nouns, adjectives, many pronouns, and a few verbal forms. Although the emphatic suffixes lend Gaelic a distinctive flavor and constitute a rich discourse device, I’ve never heard an ordinary Gaelic speaker so much as mention the emphatic suffixes, let alone praise them for the subtle effects they make possible in creating discourse tone and expressing point of view and social distance.

270

chapter 15

of national-language standardization in their homelands, Europeans espoused other ideologies that exacerbated their contempt for whatever unstandardized vernaculars they encountered.3 They seriously confounded technological and linguistic development, for example. Unable to conceive that a people who lacked a rich material culture might possess a highly developed, richly complex language, they wrongly assumed that primitive technological means implied primitive linguistic means.4 This misconception condemned most Europeans to total exclusion from the diverse conceptual worlds and rich oral literatures of many peoples whom they encountered; but much more unfortunately, it misled many Europeans into doubting the very humanity of peoples whose languages they mistakenly took to be primitive and undeveloped. Two other European beliefs about language are also likely to have had an unfavorable impact on the survival of indigenous languages in the very considerable portions of the globe where a standardized European language became the language of the dominant social strata (including previously annexed or conquered regions of the home country itself). Particularly widespread and well established is a belief in a linguistic survival of the fittest, a social Darwinism of language. This belief encourages people of European background to assume a correlation between adaptive and expressive capacity in a language and that language’s survival and spread. Since their own languages are prominent among those which have both survived and spread, this is of course a self-serving belief. For obvious reasons it’s also a belief more widespread among English, French, and Spanish speakers than among Czechs or Icelanders, but even among speakers of the smaller standardized and statepromoted languages of Europe there often lurks a notion that the general Indo-European type of language is exceptionally well suited to clear thinking and precise expression.5 Difficult as it often is to convince non-specialists 3 Language ideologies/linguistic ideologies are defined by Silverstein (1979: 173) as “sets of beliefs about language articulated by the users as a rationalization or justification of perceived structure and use” and by Rumsey (1990: 346) as “shared bodies of commonsense notions about the nature of language in the world” (cited in Kroskrity 1993 and Woolard 1992 respectively).  4 There is no evidence, unfortunately, that much progress has been made on this score. Even today only specialists seem to think otherwise, and when linguists and linguistic anthropologists discuss the language endangerment crisis with non-specialists, it’s nearly always necessary to make clear at the  very outset that the languages threatened with extinction are fully developed instruments capable of great precision and rich elaboration in cognitive terms. 5 There is no general awareness of such problems as the poor fit between spoken language and orthography which makes English and French such unnecessarily difficult languages in which to become literate, and there is also no general awareness of such difficulties as pho-

western language ideologies and small-language prospects

271

of the full grammatical and expressive development of all natural languages, it can be even more difficult (without delivering a lecture on the structural properties of the world’s languages, assuming an audience willing to sit still for such a lecture) to persuade them of the extremely useful features of many non-Indo-European language structures, such as, for example, obligatory evidential markers in the verbal system. Notions about the “natural” ability of certain languages to thrive and the “natural” inability of others to do so can be seriously entertained only by people who are not aware of the sudden reversals of linguistic fortune that have occurred when polities have fallen on hard times. Sumerian and Akkadian, two spectacularly successful Mesopotamian languages in their time (roughly the first two-thirds and the last third of the third millennium BCE respectively), not only waned after the falls of Sumer and Akkad but became entirely extinct in the end. Greek, if it crashed less drastically, is today more often learned by non-Greeks in a form that has not been spoken for 2,000 years than in its contemporary form as the living language of a small and not especially prosperous European country.6 Quechua and Nahuatl, the proud languages of oncethriving New World empires, still have strong representation in sheer numbers of speakers but each has poor social standing. Quechua has official-language status in Bolivia and (since 1975 only) in Peru, but not in Ecuador or Chile. Nahuatl has no official status and is seriously threatened in certain Mexican regions where few if any children are acquiring it. The existence of a writing system and even the existence of a notable literature do not necessarily ensure that a language will survive as a living speech form, much less thrive. Hittite has left us copious written materials, yet it is extinct. Irish was one of the earliest northern European languages to be written, and the literature and learning of early Irish are quite distinguished; but as a naturally acquired mother tongue it had declined to the status of a peasant language before late nineteenth-century Irish nationalism encouraged its cultivation once again in literary and expository forms.  The second of the additional beliefs disadvantageous to indigenous languages in regions dominated by speakers of European languages may actually be more characteristic of Anglophones than of speakers of other European languages. Anglophones however are particularly thickly distributed in nologically non-unique morphological elements (e.g. the same sibilant suffixes to express both possessive and plural in most English nouns, the same final vowel for infinitive, secondperson plural, and past participle in many French verbs).  6 A given language has a distinct market value that can be calculated by various objective measures. See the useful chapter on “The value of a language: factors of an economic profile of  languages” in Coulmas 1992.

272

chapter 15

regions that once had large numbers of indigenous languages, so English single-handedly threatens a disproportionate number of other languages. The belief in question is that bilingualism (and by extension multilingualism, all the more so) is onerous, even on the individual level. This belief is so widespread, in fact, that it can be detected even among linguists. Not long ago in reviewing a collection of linguistic papers about first-language attrition I was startled to find that the editors saw bilingualism above all as “a natural setting for the unraveling of native language abilities” (Seliger and Vago 1991: 1). The model of bilingual capacity that underlay the volume was subtractive: the bilingual’s two languages were said to compete – “metaphorically,” said the editors – “for a finite amount of memory and processing space”; the possibility of a full and richly developed bilingualism was acknowledged only by a single reference to “the so-called balanced bilingual” (p. 2). There were no tempering statements about the enrichment potential of bilingualism, nor was first-language attrition, the official subject of the book, distinguished from convergence phenomena. Attrition in language contact implies loss and incompleteness, while convergence implies mutual influence. The latter can reasonably be considered a normal manifestation of bilingualism, on both the individual and the societal level, but it does not by any means constitute an “unraveling,” at least to my mind. It often represents the regional norm, in fact, deviant only when matched against the standard language, and purely in the abstract at that (since the actual performance of even those who claim to be standard-language speakers usually diverges from book and classroom norms). The cumulative effect of the “ideology of contempt,” of ignorance about the complexity and expressivity of indigenous languages, of a belief in linguistic social Darwinism, and of a belief in the onerousness of bi- or multilingualism converge to bear down most of the languages spoken by populations without wealth or power. They are heavy weights for small populations in particular to cast off, and few have so far been able to do so.  In the most general terms, a linguistically distinctive population which has come to have poor standing needs to discover or develop some basis for increased self-regard in order to withstand pressures for ancestral language abandonment and shift to a dominant-group language. Several possible sources of such self-regard can be identified, at least tentatively. Rising prosperity, as an indicator of increasing economic success, can be an effective counterpoise to the social disfavor that typically accompanies a subsistence-level economy. Provided it does not burst suddenly upon a population with no prior experience of it,7 prosperity can boost social self-confidence while also pro7 The history of the Osage tribe after sudden oil wealth illustrates some of the problems associated with abrupt and unexpected wealth.

western language ideologies and small-language prospects

273

viding the resources for institutional language maintenance efforts that might otherwise seem prohibitively expensive. The Ayas Valley, in the Autonomous Aosta Valley Region within the Italian Republic, moved in recent decades from an economy based on agriculture and stock-rearing to one based on tourism. The resulting prosperity of the region supports a trilingual pre-school program for children from 3 to 5. Despite the small size of the population served, five schools deliver a carefully designed and executed program that provides support for local Franco-provençal as well as for Italian and Standard French (Decime 1994). A similar economic transition has changed the outook for tiny Ladin, spoken in three small and discontinuous districts in the South Tyrol region of Italy (Markey 1988). The development of a booming tourist industry, geared in good part to luxury-level skiing, has been the most conspicuous change in local conditions. The seasonal nature of the tourist industry and the fact that the tourists represent no single dominant language, plus the isolation of the districts at other times of the year, perhaps operate to safeguard the ancestral-language base, which has been strengthening in recent decades.  Where prosperity grows less suddenly and dramatically, it may be that its usefulness lies above all in the fostering of a middle class with the social selfconfidence to insist on traditional identity and heritage. Catalan speakers, in a region with a strong economic base and a self-confident tradition, emerged from the severe suppression of the Franco years and were soon able to begin once again attracting new speakers to their language in Castilian-dominated Spain (Woolard & Gahng 1990).8 Wales is considerably less prosperous overall than Catalonia, the Ayas Valley, or the Ladin districts of the South Tyrol. Within the ranks of an established middle class, however, social self-confidence can seemingly emerge despite economic weakness if educational achievement permits marked social mobility. In nineteenth-century Wales in-migration of large numbers of English coal-industry workers originally posed a severe threat to the survival of Welsh. Yet in the longer run the coal industry helped to produce a middle class which in the twentieth century has provided much of the impetus for revitalization of Welsh. Khleif, studying the successful growth of Welsh-medium 8 Catalan can be considered a “small” language in the context of Spain since it is spoken by a much smaller number of people than Spanish and is found chiefly in a single region of the country.  By comparison with minority languages in many other settings, however, it is extremely well represented. Numbers as such form an uncertain measure of linguistic security, of course; some of the distinctly precarious languages of Central India have over a million speakers, e.g. Kurux (Abbi 1995), and so did beleaguered Breton, in France, as recently as 1926 (Timm 1973: 289, citing Meillet 1928: 380). 

274

chapter 15

education, identified the “new middle class” in Wales as the chief factor in the turn-around (1980: 77–78):  The current leaders of Welsh opinion are overwhelmingly sons and daughters of coal miners, agricultural workers, steel workers, shop keepers, and minor civil servants, but especially of coal miners . . . They are all Welsh-speaking and, in a small country such as Wales, know each other very well. Their Welshness sets them apart, for to have spoken Welsh at home, a generation ago, meant that the person by definition was working-class. They are very proud of their Welshness, of their ability to speak Welsh, of their ability to “live a full Welsh life”. They consider their knowledge of Welsh a badge of achievement, for it differentiates them from other middle-class people as well as working-class people who are English monoglots . . . Sons and daughters of the new Welsh-speaking middle class are more  self-assured, many informants remarked. Welsh-medium schools impart self-confidence to the new generation.  The speakers of Ladin, Catalan, and Welsh are themselves Europeans, of course, and as such they may have been at least conceptually less distant to begin with from envisioning their own linguistic success (and less distant, no less significantly, from outside investment capital and the like). Most smalllanguage communities cannot realistically look to rapidly spurting prosperity to reinforce their standing, unfortunately.  Occasionally more accessible as a socially and psychologically invigorating factor may be progress towards political autonomy, preceding or accompanying the rise of a middle class and of a native intelligentsia. In Greenland, for example, after a period of intense Danicization that accompanied a drive toward modernization, reaction set in and pressure for greater autonomy resulted in Home Rule for Greenland in 1979. Prior to that year Greenlandic had been considered a threatened language, with considerable justification in view of the rise in the number of Danish monolinguals during the preceding quarter-century. Since that year Greenlandic has been supported and promoted to an increasing degree, and bilingual Greenlanders have increasingly replaced monolingual Danes in the top institutional and organizational positions (Langgaard 1992).  It would be difficult, or more likely impossible, to identify a precise causeand-effect sequence in most of the cases mentioned so far, since the factors involved frequently intermesh. Rising prosperity and an emerging middle class often coincide. A native intelligentsia is likely to appear in conjunction with

western language ideologies and small-language prospects

275

an emerging middle class, and any or all of these factors may either precede or accompany movement towards greater political autonomy.  While these factors appear to enhance the chances of ancestral language maintenance, their absence need not doom a small language to rapid disappearance. The case of the Arizona Tewa, still in possession of their ancestral language even though long enclaved among the Hopi, suggests a different sort of counterpoise to the negative effects of European-derived linguistic ideologies. On the basis of long-term work among the Arizona Tewa, Kroskrity proposes that in the theocratic Pueblo societies, where political and religious authority are fused, ceremonial speech has a position analogous to that of the standard language in a nation-state. The highly regarded ceremonial speech variety called te’e hi:li ‘kiva talk’ is of critical importance to the Arizona Tewa, and the rigorous standards applied to its maintenance spill over into attitudes towards Arizona Tewa generally (Kroskrity 1993: 37–39):  Their concern is with maintaining and delimiting a distinctive and appropriate linguistic variety, or vocabulary, for religious expression . . . The strong sanctions against foreign expressions in ceremonial speech – violations of which are physically punished – are motivated not by the linguistic expression of xenophobia or extreme ethnocentrism but by the need for stylistic consistency in a highly conventionalized liturgical speech level. Similarly, the negative evaluation of code-mixing in everyday speech by members of the Arizona Tewa speech community does not reflect attitudes about these other languages but rather the functioning of ceremonial speech as a local model of linguistic prestige . . . Just as ceremonial practitioners can neither mix linguistic codes nor use them outside of their circumscribed contexts, Tewa people should observe comparable compartmentalization of their various languages and linguistic levels in their everyday speech.  The Arizona Tewa have maintained their ancestral language for 300 years, despite enclavement within a Hopi environment, despite considerable intermarriage with the Hopi, and despite a small population base. There may well be a variety of elements in their success: they pride themselves on their skill at languages, for example, and they consider their bilingualism in Hopi, while the Hopi do not control Tewa, a form of cultural victory (Kroskrity 1993: 23, 218).9 9 Compare the Emenyo of New Guinea, who likewise consider bilingualism an accomplishment and feel superior to less frequently bilingual Dene-speaking neighbors in whose language they are commonly bilingual (Salisbury 1962: 4). This is the antithesis of the

276

chapter 15

The centrality of religious ceremony has been a factor in some other cases of unexpectedly sustained language maintenance in seemingly adverse circumstanees, however (cf. Hohenthal & MeCorkie 1955; Barber 1973), and a strong religious base certainly cannot be overlooked as a source of the psychosocial confidence necessary for language maintenance in the face of considerable social pressure for shift.  Despite the fact that peoples speaking a variety of small local languages have followed similar paths to a decrease in speaker numbers and to an eventual language shift, the path in question is neither inevitable nor perfectly predictable. Even the prospect of material well-being, for example, seductively associated with Westernization during the long period over which European (and more recently American) power has expanded, does not invariably lure a population away from its traditional culture and traditional language. Natives of Pulap Island in the Western Island group of the Carolines (Micronesia) have proved unusually resistant, during the half century of US dominance in their region since World War II, to American individualistic and materialistic values, which they consider selfish and greedy. While modern opportunities for wageearning work away from the home island have produced welcome material rewards for some Pulap Islanders, indigenous values have led to the sharing of those rewards among extended kin groups at home. Many traditional practices reflecting persisting traditional values still prevail among the Pulapese, not only on Pulap Island itself but also on Moen, the capital of Chuuk State, where land purchases by some Pulap Islanders in the 1950s have come to serve as a more centrally located and urbanized extension of Pulap Island:  The art of [traditional] navigation, production and exchange of local foods, respect behavior toward kin, and traditional dress are the major traits that Pulapese invoke to conceptualize their culture, and this process appears in its most pronounced form on Moen. Pulapese present these cultural characteristics as evidence of their worth in a context in which others are abandoning tradition. (Flinn 1990: 123) Navigating by traditional means is more difficult than navigation using modern techniques, and traditional food preparation is relatively slow and laborious. Traditional clothing permits body parts now normally covered elsewhere in the region to go uncovered, and respect behavior requires women to stoop in the presence of their older brothers, among other things. Pulapese cultural widespread anglophone notion that bilingualism is damaging to the bilingual because the two languages inevitably compete for limited cognitive space.

western language ideologies and small-language prospects

277

conservatism is overt and obvious, therefore, and is acknowledged by other islanders. By laying claim in clearly identifiable fashion to greater traditional “purity,” the Pulapese present themselves as superior in a cultural sense, creatively compensating for a lesser material well-being. And as might be expected, given the other conservative Pulapese behaviors, Pulap Islanders on Moen, even the secondary school pupils whose education on Moen is entirely in English, maintain the use of their Pulap dialect in the home setting (Flinn 1992: 156).  The existence of resistant groups such as the Arizona Tewa and the Pulap Islanders (and of others as well, such as the mountain Kwaio of the Solomon Islands [Keesing 1992]) indicates that one of our particularly acute needs is more in-depth studies of linguistic and cultural persistence in small communities. Except in cases of great geographical or social isolation, the longterm maintenance of a small language implies not just the persistence of one language but the enduring coexistence of two or more. Currently we understand the motivating factors in language shift far better than we understand the psychosocial underpinnings of long-sustained language maintenance. We need to understand not just the staying power of the Arizona Tewa (illuminated both by Yava 1978 and by Kroskrity 1993), but the tolerance of the Hopi, who have permitted long-term Tewa-Hopi bilingualism in their midst while remaining largely without knowledge of Tewa themselves. In similar fashion, Moen natives apparently exert no pressure on the Pulap Islanders on Moen to abandon their somewhat archaic behaviors or to give up their home-island speech variety. Another case of seemingly unproblematic coexistence is to be found in the Circassians who fled to the Middle East in the nineteenth century, with 120 years of persistence in what is now Israel, and Israeli acceptance of them. They are reported to cultivate bi- and trilingualism as a matter of course and to show no signs of incipient language shift (Stern 1990).10 In an ironic turn of events, the excesses of nationalism itself may have begun to effect a change in thinking that could conceivably, if it were to catch hold, lead to an improved outlook for small-language communities submerged in, or under the control of, contemporary nation-states. Recent prime ministers of Ireland and the United Kingdom, recognizing that irreconcilable nationalist aspirations will never offer a basis for peaceful solutions to the problems of 10

There are other cases of great potential interest. English is said to be the language of all monetary rewards for the Koasati of Texas, who number under 200. The children attend English-language schools, furthermore, and yet the ancestral language is successfully transmitted with “monolingual Alabama-Koasati speakers still present in each generation of children” (Saville-Troike 1989: 215).

278

chapter 15

Northern Ireland, came forward with the proposal that what they term “multiple allegiances” be recognized. The Irish prime minister, in a talk before the National Press Club in Washington, DC (Bruton 1995), pointed out that land ownership is no longer the basis of real wealth in modern economies, so that discrete assignment of all land to mutually exclusive nations does not have the urgency it once had. So long as no flags are run up the pole (i.e. so long as certain traditional emblems of purely political allegiance are avoided), individuals with different sets of multiple allegiances should reasonably be able to co-exist in one region. He is so much persuaded of the power of this concept that he recommends it as a potential solution in other ethnically and linguistically complex regions of Europe such as Latvia and Catalonia. Multiple allegiances in this sense might be seen as an extension into the sphere of political organization of the sociologist’s status sets, the totality of all the statuses one occupies (not always entirely congruently) in one’s social life. In the political sphere this suggestion is to some extent “post-nationalist” and to that degree perhaps an escape hatch from the demands of mutually exclusive nationalisms.11 The fact that recognition of multiple allegiances is being recommended as a solution for otherwise irresolvable nationalist conflicts precisely in Europe could be especially helpful, since it is the concept of the nation-state coupled with its official standard language, developed in modern Europe and extended to the many once-colonial territories of European states, that has in modern times posed the keenest threat to both the identities and the languages of small communities. Outside the modern European sphere of interest the same problem of insistent single-language dominance coupled with hostility to minority languages has not necessarily arisen.12 The Ottoman Empire was largely free of it over most of its long history, and the 300year coexistence of the Arizona Tewa with the First Mesa Hopi indicates that both ethnic and linguistic persistence are feasible over long time spans without fatal ethnic hostility on either side. Thailand, with a stable hierarchical social structure until very recently, is said to have had minimal ethnolinguistic

11

12

“What we are trying to do in Ireland is to redefine the concept of nationality, so that it suits the realities or the 21st century, and isn’t mired in the concepts that were the cause of so much war in Europe in the 19th and 20th centuries” (Bruton 1995: 6). This is not to suggest that enforced use of a dominant-group language and intolerance of subordinate languages were unknown outside Europe, since that is clearly not the case. Heath and Laprade describe Inka policies designed to erase both the histories and the languages of conquered tribes, including “a program to spread their language, Quechua, and to prohibit use of the languages of subjugated tribes” (1982: 123).

western language ideologies and small-language prospects

279

conflict over the six and a half centuries of its monarchy despite the presence of a variety of ethnicities within the national borders (Smalley 1994).  In yet another ironic development, there has been a good deal of  consciousness-raising within the European Community recently in connection with small languages and minority languages. The EC member states have not been willing to yield to economic pressures and permit the use of some one or two languages as the Community’s official working languages (English and French being the chief candidates). Instead they have insisted on the nationallanguage principle and have accepted the enormous costs of mounting interpretation services and document-translation services for each of the individual national languages. This unyielding adherence to the national-language ideology has given rise to unprecedented European support for multilingualism, and in an overspill of protective enthusiasm for smaller languages, even minority languages within the EC countries have gained a certain increased recognition and at least a few economic benefits (Coulmas 1992: 116–117). This recent development shows a language ideology which has previously worked against small languages beginning to work for them instead: if all nations, no matter how small, have a right to the use of their own language, then by extension other small-language populations, with or without a nation-state of their own, can with some justice claim the right to the use of their own languages as well. In time some of this change of attitude could conceivably be generalized into wider European spheres of influence.  Popular opinion in the United States and other European-settled parts of the world is unfortunately still largely infected with earlier European language ideologies of the types discussed above, all unfavorable to the survival of smaller indigenous languages. The emergence of government level initiatives to counter some of the negative aspects of nationalism (in the form of the new “multiple allegiances” discussion), the stirrings of a new legitimacy for small languages, and perhaps also the growing acknowledgement in recent decades (in the United States at any rate) of the value for the health of individual and planet of at least some non-Western, small society forms of religious or spiritual world-view, conceivably offer a small window of opportunity to make the case for the wisdom of preserving linguistic and cultural diversity. Still, recent concerns about loss of linguistic and cultural diversity, together with new recognition of the possibilities of multiple sociopolitical allegiances and of the legitimacy of ethnic languages and of multilingualism, come very late in the day for most small languages. Material well-being has been intimately linked to the adoption of dominant languages for a very long time, and the reality of that linkage is undeniable. It requires enormous social and psychological self-confidence for any small group to insist on the importance

280

chapter 15

of ancestral-language retention. (Consider, for example, the case of the English-monolingual speakers who can claim Tlingit ethnicity, as discussed in Dauenhauer and Dauenhauer, this volume.) Precisely that sort of selfconfidence is hard to come by in communities which have suffered the penalties of an ideology of contempt over a long period.  Special problems can arise, furthermore, if language shift is already well underway. Even in settings where remaining fluent speakers of the ancestral language may sense that their culture is deeply bound up with their language (and it is surely germane to the durability of Arizona Tewa that its speakers frequently state “Our language is our history” [Kroskrity 1993: 44]), it becomes impossible to insist on that linkage if a large part of the social group that identifies itself by the ancestral-language label no longer speaks that language. In such cases, defining identity in terms of language would define out of membership most of the younger people whose retention is vital to continued existence as a group. And those without the language will resist the linkage, if my experience in the Scottish Highlands is any indication. I found that when I asked speakers of Scottish Gaelic whether a knowledge of Gaelic was necessary to being a “true Highlander,” they said it was; when I asked people of Highland birth and ancestry who did not speak Gaelic the same question, they said it wasn’t. This is not a surprising division of opinion, but it does greatly complicate the situation for small communities where ancestral-language loss is already well advanced. The question of a linkage between a language and the culture it’s associated with becomes so delicate a matter that it’s almost easier to insist on the importance of language to heritage and identity in settings where the ancestral language is entirely lost than in settings where it’s retained by a relatively small number. Among the Echota Cherokee of Alabama, for example, strong sentiment attaches across various age groups to the Cherokee language and great longing for a lost heritage is expressed in connection with the possibility of introducing it into selected Alabama schools (Sabino 1994: 5); but this outpouring of fervor is for a vanished language that none of the Echota currently speaks. Joshua Fishman points out that language “always exists in a cultural matrix” and that the matrix rather than the language is the point at which support is most needed (1989: 399). He calls attention to the power of “Zeitgeist trends that can contribute as much or even more to [language] spread than language policy per se,” and to a momentum generated by “mobility aspirations” and “the apparent stylishness of the pursuit of modernity itself” (1989: 390). He recognizes (1989: 399) that the staying power of endangered languages “must be intimately tied to a thousand intimate or small-scale network processes, processes too gratifying and rewarding to surrender even if they do not quite

western language ideologies and small-language prospects

281

amount to the pursuit of the higher reaches of power and modernity.” Such rewards cannot be supplied from the outside. They are to be had from within the social web of the community itself or not at all. For this reason it is extraordinarily difficult for even the most sympathetic outsiders to provide useful support for endangered small languages, most especially for non-European small languages within a Euro-American sphere of influence. Moral support and technical expertise, including linguistic expertise, can and should be offered, certainly, but acceptance or rejection will necessarily lie with individual communities. Even in the event of acceptance, effective leadership can only come from inside the community.  One role that knowledgeable outsiders have sometimes usefully played is that of information-disseminator and consciousness-raiser, helping to make a wider public aware of the looming threats to a local language’s survival. This process has only recently begun on a scale more appropriate to the size of the problem, however, and time has grown desperately short for many local languages. Having waited too long before undertaking to rally support for threatened languages, we may find ourselves eulogizing extinct languages whose living uniqueness we had hoped instead to celebrate.   References Abbi, Anvita. 1995. Language contact and language restructuring: A case study of tribal languages in Central India. International Journal of the Sociology of Language 116: 175–85.  Barber, Carroll G. 1973. Trilingualism in an Arizona Yacqui village. In Paul R. Turner, ed., Bilingualism in the Southwest, 295–318. Tucson: University of Arizona Press.  Bruton, John. 1995. Address to the National Press Club, Washington, DC, March 17, 1995. (Federal News Service, Inc., broadcast transcription.)  Coulmas, Florian. 1992. Language and economy. Oxford: Blackwell.  Decime, Rita. 1994. Un projet de trilinguisme intégré pour les enfants des écoles maternelles de la Valée d’Ayas. International Journal of the Sociology of Language 109: 129–37.  Fishman, Joshua A. 1989. Language and ethnicity in minority sociolinguistic perspec­ tive. Philadelphia: Multilingual Matters.  Flinn, Juliana. 1990. We still have our customs: Being Pulapese in Truk. In Joyce Linnekin and Lin Poyer, eds, Cultural identity and ethnicity in the Pacific, 103–26. Honolulu: University of Hawaii Press.  ———. 1992. Diplomas and thatch houses: Asserting tradition in a changing Micro­ nesia. Ann Arbor: The University of Michigan Press. 

282

chapter 15

Gellner, Ernest. 1983. Nations and nationalism. Ithaca, NY: Cornell University Press.  Grillo, Ralph D. 1989. Dominant languages: Language and hierarchy in Britain and France. Cambridge: Cambridge University Press.  Heath, Shirley Brice. 1972. Telling tongues: Language policy in Mexico, colony to nation. New York: Teachers’ College Press.  Heath, Shirley Brice, and Richard Laprade. 1982. Castilian colonialization and indigenous languages: The cases of Quechua and Aymara. In Robert L. Cooper, ed., Language spread, 118–47. Bloomington, IN: Indiana University Press.  Hohenthal, W. D., and Thomas McCorkle. 1955. The problem of aboriginal persistence. Southwestern Journal of Anthropology 11: 288–300.  Joseph, John Earl. 1987. Eloquence and power: The rise of language standards and stan­ dard languages. Oxford: Blackwell.  Khleif, Bud B. 1980. Language, ethnicity, and education in Wales. The Hague: Mouton.  Kroskrity, Paul V. 1993. Language, history, and identity: Ethnolinguistic studies of the Arizona Tewa. Tucson: University of Arizona Press.  Kuter, Lois. 1989. Breton vs. French: Language and the opposition of political, economic, social, and cultural values. In Nancy C. Dorian, ed., Investigating obsoles­ cence: Studies in language contraction and death, 75–89. Cambridge: Cambridge University Press.  Langgaard, Per. 1992. Greenlandic is not an ideology, it is a language. In Nelson H. H. Graburn and Roy Iutzi-Mitchell, eds, Language and educational policy in the north, 167–78. Berkeley: Working Papers of the Canadian Studies Program, International and Area Studies, University of California at Berkeley.  Meillet, Antoine. 1928. Les langues dans l’Europe nouvelle. Paris: Payot.  New York Times. 1975. France refuses to recognize six children because of their Celtic names. January 12.  Rumsey, Alan. 1990. Wording, meaning and linguistic ideology. American Anthropologist 92: 346–61.  Sabino, Robin. 1994. Establishing the necessary database for the restoration of the Cherokee language to the Alabama Cherokee population: A proposal. Unpublished ms.  Salisbury, Richard F. 1962. Notes on bilingualism and linguistic change in New Guinea. Anthropological Linguistcs 4: 1–13.  Saville-Troike, Muriel. 1989. The ethnography of communication. 2nd ed. Oxford: Basil Blackwell.  Seliger, Herbert W., and Robert A. Vago. 1991. The study of first language attrition: An overview. In Herbert W. Seliger and Robert A. Vago, eds, First language attrition, 3–15. Cambridge: Cambridge University Press. 

western language ideologies and small-language prospects

283

Silverstein, Michael. 1979. Language structure and linguistic ideology. In P. R. Clyne, ed., The elements: A parasession on linguistic units and levels, 193–247. Chicago: Chicago Linguistic Society.  Smalley, William A. 1994. Linguistic diversity and national unity: Language ecology in Thailand. Chicago and London: University of Chicago Press.  Stern, Asher. 1990. Educational policy towards the Circassian minority in Israel. In Koen Jaspaert and Sjaak Kroon, eds, Ethnic minority languages and education, 175–84. Amsterdam and Lisse: Swets and Zeitlinger.  Timm, Lenora A. 1973. Modernization and language shift: The case of Brittany. Anthro­ pological Linguistics 15: 281–98.  Woolard, Kathryn A. 1992. Language ideology: Issues and approaches. Pragmatics 2: 235–49.  Woolard, Kathryn A., and Tae-Joong Gahng. 1990. Changing language policies and attitudes in autonomous Catalonia. Language in Society 19: 311–30.  

chapter 16

Bi- and Multilingualism in Minority and Endangered Languages 1 Introduction Languages regarded as endangered are in most cases the languages of minority peoples within the state where the population in question lives. While the languages of some minority peoples have official status of some kind in the state or in a particular region of the state, as do French in Canada, Catalan in Spain, and Assamese in India, far more of the world’s minority languages have no official standing of any kind. The consequences, for speakers of languages without official status, amount to a sharp power differential, both linguistic and social. Government, the legal system, education, print media and broadcasting are conducted largely or wholly in other languages. Insofar as mother-tongue speakers of a minority language without official status hope to be served by those activities or to participate in them, they must either speak an additional language or have the assistance of bilingual intermediaries. How a minority population and its language stand in relation to the official language(s) varies greatly from region to region, reflecting a particular local history of contact and of national policy. Many groups become minorities through migration into an area where another group and their language are dominant. If enough people make the migration and they are able to achieve some settlement density, they may be able to maintain their original language in the new location for a shorter or longer time. Some minority peoples represent an established population which was always small and localized, around which a modern nation-state came into being without much direct contact being effected between state institutions and the original inhabitants, as in the Amazon basin, say, or in parts of Papua New Guinea. But a great many presentday minority peoples have had higher populations, a wider territorial base, and/or more autonomy at some previous time and have become minorities through a process of subordination involving conquest or political reorganization. Movement in the opposite direction also occurs, of course, even when, as in the present day, the ideological climate deeply favors the nation-state and its officially promoted language(s): some peoples long relegated to minoritygroup status achieve more autonomy or even full independence, with a corresponding rise in the status of their language, as has happened in recent times

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�18

bi- and multilingualism in minority and endangered languages

285

with the people of Greenland (greater autonomy) and their language and with the people of Slovenia (independence) and their language. With only something over two hundred nations presently in existence, however, and with most of them intent on maintaining the territorial and political status quo, it seems clear that most of the world’s 6,000 or so languages, insofar as they survive at all, will remain the languages of peoples who are minorities within a relatively small number of nation-states. Under conditions of steadily expanding communication and transport networks and of ongoing economic and cultural globalization, genuine geographic and social isolation can only grow rarer in coming decades. The outlook for endangered minority languages is linked to the willingness of minority peoples to sustain the bi- or multilingualism that characterizes many of them now or to cultivate it deliberately in those cases where the ancestral language has become the valued possession only of the elderly. Such willingness is as much a matter of ideology and cultural values as of objective factors like population size, a viable economy, or political autonomy. Because minority peoples are subordinate peoples, furthermore, the ideologies of the dominant group are fully as important to the outcome as those of the minorities themselves. 2

Monolingualism vs. Bi- or Multilingualism in the Absence of a Nation-State

Monolingualism, now usually considered the unmarked condition by members of the dominant linguistic group in modern nation-states, was in all likelihood less prevalent before the rise of the nation-state gave special sanction to it (see next section). But it was not unknown in pre-state settings where, for example, geographical isolation and/or a reasonably large population buffered a group from frequent contact with other peoples. Such groups would typically have had some members, even in the most demographically central areas, who became bilingual or multilingual for the purpose of representing the group in contacts with other populations as negotiators of some kind (in trade, in alliance building, in assertion of political claims, and so forth), while other group members residing in demographically central areas might well have remained monolingual. At the borders of a large-group territory, however, where members came into more frequent contact with populations speaking other languages, the number of bilinguals and multilinguals would increase. Population size typically had an effect on which group had the greater number of borderarea bilinguals. If one of two neighboring peoples was smaller in size, members of the smaller population were more likely to acquire the large-group language

286

chapter 16

than vice versa, though each group would have included bilingual members. This is the general linguistic situation reconstructed, for example, for the larger and smaller peoples of the Papua New Guinea Highlands in contact with each other in the period before contact with people of European descent (Sankoff 1980: 108–09). Where interaction among a variety of peoples was very frequent, bi- or multilingualism was also more widespread. This seems to have been the norm, for example, in much of Aboriginal Australia and in parts of sub-Saharan Africa, and it is still the prevailing condition among some groups in both settings. Where a number of languages coexisted interactively within a particular region, hierarchies of power and prestige often existed among the languages, reflecting hierarchies of power and prestige among the various groups who spoke them. One factor in the establishment of such hierarchies was demography, as the Highland Papua New Guinea instance suggests: larger peoples had an obvious advantage, since they were better able to assert their control via warfare or political domination. Abundance of natural resources could also play a role. In one interconnected island complex in Micronesia, the resourcerich high-volcanic island of Yap constituted the center of a tribute system that involved a great many outlying island groups made up of coral atolls. Distance from Yap corresponded to place in the social and political hierarchy of islands, so that the most outlying atoll groups ranked lowest while those closest to Yap ranked highest. The origins of Yap dominance are not now entirely certain, but access to resources available only on Yap (e.g. timber, certain spices and foods) and fear of Yapese sorcery seem to have sustained the system, remembered and resented even now by islanders from the outlying atoll groups, despite the fact that it ceased to operate more than a hundred years ago (Flinn 1992: 21–23). 3

Bilingualism and Multilingualism within Nation-State Settings

The establishment of a nation-state typically confers distinct advantages on a select language or set of languages, namely any language(s) adopted as official by the state or acknowledged as the official language(s) of a particular province within the state. Until quite recent years (and indeed in a good many areas still) the best that speakers of other languages could hope for was a benign neglect of their language, while non-benign treatment ran the gamut from denial of the existence of a particular language and its speakers, through the labeling of a particular language as inferior and of its speakers as socially or intellectually limited, to the active suppression – including killing, in the worst cases – of speakers of some minority languages.

bi- and multilingualism in minority and endangered languages

287

Disparities in factors such as population size, access to resources, and control of trade are fully as important in nation-state settings as in non-state settings, since advantages in these spheres are often important to establishing a viable state in the first place. A uniquely significant disparity in the nationstate context, however, is precisely that primary institutional and social status is reserved exclusively for whatever language(s) the state espouses officially. Where the state promotes a particular language as the sole legitimate linguistic medium of national identity and state authority, that language typically moves to an unchallenged place at the pinnacle of a hierarchy of utility and prestige among all the languages that may be spoken within state boundaries. In most cases this concentration on a single favored language works to the disadvantage of all other languages, for which neither comparably high regard nor institutional support will be available. If minority populations in such settings either find by experience or come to believe that the official state language confers major advantages in terms of access to schooling, employment, political participation, and state services, the psychological ground for a language shift may be prepared. Parents who suffered social penalties or educational and occupational disadvantage through limited knowledge of the official language during their own youth may reach a decision not to transmit the ancestral language to their children. Reports of conscious blocking of home-language transmission are commonplace in the literature of language endangerment and shift, e.g. for Scottish Gaelic, for Pennsylvania German, and for Tlingit (Dorian 1981: 104–05; Huffines 1989: 225; and Dauenhauer and Dauenhauer 1998: 64–66, respectively). Unconscious blocking also occurs. Parents in one Papua New Guinea language community, for example, expressed disappointment that their children were not speaking Taiap, the traditional language. But when the parents’ actual speech behavior with their young children was closely observed, they turned out to be switching to Tok Pisin, a creole increasingly used as a lingua franca in Papua New Guinea, when they addressed their children, even though they generally used Taiap when speaking to one another (Kulick 1992). Blocked minority-language transmission can produce majoritylanguage monolingualism within as little as two generations, though three is the commoner pattern. In the latter case, the first of the three generations is either monolingual in the ancestral language or ancestral-language dominant, the second is highly bilingual but more likely to be majority-language dominant, and the third is monolingual in the majority-group language (with or without purely passive bilingualism in the ancestral language). Because shift away from a limited-currency minority language to a widercurrency state-promoted language occurs with considerable frequency, it is sometimes asserted that bilingualism is essentially a practical matter, governed

288

chapter 16

by social or economic necessity. On a very elemental level there is a good deal of truth to this. Acquisition and maintenance of more than one language arises from a contact between peoples that is frequent enough to make it useful for members of at least one group to speak the language of the other group(s). But stronger positions, such as the assertion that people “will not indefinitely maintain two languages when one will serve across all domains” (Edwards 1994: 110), are too extreme, since this would suggest that bilingualism is not long sustained in settings where a lingua franca is in wide use, or in settings where all or very nearly all members of a community are fully bilingual in the language of some other group. Sustained bilingualism can be found in both types of setting, but – tellingly – relatively seldom among people of European origin. 4

The Nation-State and European Language Ideologies

Embrace of the nation-state in the modern era has gone hand-in-hand with embrace of a one-language, one-nation ideology. This ideological construct is generally associated with French and German philosophers of the eighteenth century and linked with policies that took shape in France during the closing decade of the eighteenth century, after the Revolution of 1789 (Grillo 1989: 22–42; Woolard 1998: 16–17). At that time the longstanding and previously unremarkable existence within the French polity of substantial subcommunities who neither spoke nor understood French came to be viewed as unacceptable. The unity of the new revolutionary state was henceforth to be expressed via a common language, replacing the linguistic heterogeneity that in the revolutionary view had served the purposes of a discredited monarchy by preventing various segments of the country’s population from making common cause with one another. The Alsatians, the Basques, the Bretons, and the Occitanians would come to feel their national unity and would express it, according to revolutionary tenets, by adopting the use of the French language. Certain characteristically European ideological positions were given expression in the implementation of this policy. A single language variety associated with people of high social position (the king and his court, in this case) was accorded fixed form and unique authority through standardization, and a monopoly of legitimacy and prestige was conferred on that single form. In the resultant linguistic hierarchy, the unstandardized language varieties of politically and socially subordinate peoples within the state underwent a parallel attitudinal subordination and were subjected to what has been termed an “ideology of contempt” (Grillo 1989: 173–74).

bi- and multilingualism in minority and endangered languages

289

A small but highly suggestive study has found an ideological position that similarly favors linguistic homogeneity still prevailing in the expression of European nationalism at the present day (Blommaert and Verschueren 1998). Coverage of the perceived role of language in nationalist ideologies, as reflected both in editorials and in news stories that appeared in mainstream Western European newspapers and magazines published in Germany, France, the Netherlands, and England (with the International Herald Tribune also included, representing the USA) was closely monitored during the first weeks of November 1990, a time when ethnic/nationalist conflicts were particularly prominent and were much discussed, as were associated questions of asylum and immigration. The media data revealed a very considerable degree of popular consensus on the desirability of homogeneity, with the ideal model of society taken to be not only monolingual but also monoethnic, monoreligious, and monoideological. This European ideological bias in favor of monolingualism has been detected not only in popular opinion but even within the canonical texts of sociolinguistics and the sociology of language. Sociologist Glyn Williams (himself a Welsh-English bilingual) identifies a none-too-subtle evolutionary viewpoint in these texts, according to which traditional societies are characterized by linguistic diversity and multilingualism, while modern societies move steadily (“progress”, in the terms of this viewpoint) toward a single official language and monolingualism. Within this framework, Williams points out, “the elimination of minority languages is a natural, evolutionary process” (1992: 100). Williams objects to what he considers an identification of monolingualism with rationality, as in Edwards’ notion that it is in effect irrational to maintain two languages when one will serve for all purposes (“across all domains”). Edwards’ position represents what has generally been the mainstream European viewpoint,1 however, and it is useful when assessing accounts of minority peoples’ multilingualism to recall that nearly all such accounts have been produced by Europeans (or by their heritors in settings colonized by Europeans), most of them speakers of highly standardized national languages of European origin.

1 In the same chapter, expanding on his claim that bilingualism is sustained only so long as it has practical – chiefly economic – value, Edwards quotes directly from my work (Dorian 1982: 47) in a way that makes it appear as though I were in agreement with his position (Edwards, 1994: 116). It is the unfortunate omission of a lead-in sentence and a following sentence, both strongly qualifying the quoted material, that gives rise to that misleading impression.

290

chapter 16

While the study of minority peoples and their language choices has made the pattern of shift from a minority language to a majority-group language familiar, this pattern is more likely to occur under some circumstances than others. Paulston notes, for example, that the opposite pattern, sustained groupwide bilingualism, is unusual only under particular conditions, namely when the modern nation-state is the setting and when both socioeconomic incentives and access to the dominant language are present: Maintained group bilingualism is unusual, if opportunity of access to the dominant language is present and incentives, especially socioeconomic, motivate a shift to the dominant language. If not, as with India’s former caste system and ascribed status, the result is language maintenance. But given access and incentive, the norm for groups in prolonged contact within a modern nation-state is for the subordinate group to shift to the language of the dominant group, either over several hundred years as with Gaelic in Great Britain or over the span of three generations as has been the case of the European immigrants to Australia and the United States in a very rapid shift (Paulston 1994: 12–13; emphasis in original). Within nation-states the frequency of shift to an official national language when access and socioeconomic incentives are present is undeniable and underlies what many see as a language-endangerment crisis. Furthermore, this shift pattern appears to be growing in geographical distribution as well as in frequency, lending a sense of urgency to discussions about threats to linguistic diversity world-wide. In spite of the frequency and seeming ubiquity of shift from a minoritygroup language to an official majority language, however, it should not be considered an inevitable or “natural” pattern. The conditions and ideologies which give rise to it do not universally hold sway, as recurrent reports by fieldworkers in various parts of the world demonstrate. 5

Environments Favorable to Bi- and Multilingualism

Environments that favor the maintenance of multiple languages are not difficult to find, but they are reported largely from non-European cultural contexts where quite different language ideologies prevail. Since as noted the reporters are in most cases members of Western societies, there is sometimes a striking contrast between the expectations of the Western investigator and the linguistic situation encountered and described. This is the case, for example,

bi- and multilingualism in minority and endangered languages

291

with a 1998 field report by English researcher Roger Blench on two languages that he searched out during a field trip in Plateau State, northern Nigeria. He found the Niger-Kordofanian language Horom spoken in one main village by a maximum of perhaps 1,500 people, while the inhabitants of nearly all the other villages in the area spoke varieties of a Chadic language called Kulere. He described the Horom people as “extremely multilingual”, since they reported themselves to be fluent in Kulere as well as Horom and also in Rindre, another Niger-Kordofanian language, and still more significantly in Hausa, an important and very widely spoken Chadic language of northern Nigeria and the Niger Republic. Kulere and Rindre are both spoken by larger populations than is Horom (Rindre by a considerably larger population), while Hausa has many millions of first-language speakers in Nigeria and the Niger Republic and also increasingly serves as a lingua franca in West Africa. Blench was surprised, by his own account, to find that Horom did not appear to be an endangered language: children present during his language elicitation session “were able to produce the required lexical items simultaneously with the adults” (Blench 1998: 10). He suggested that remoteness might have acted as a buffer in the maintenance of Horom, though he acknowledged that remoteness had not sufficed to keep Chadic languages in the Bauchi area of Nigeria from disappearing. He remarked, too, on the cultural vitality of the Horom people, whose traditional religion, pottery, weaving, and music he found likewise to persist strongly. The linguistic tenacity of another small and isolated Niger-Kordofanianspeaking people in Plateau State, the residents of Tapshin village, struck Blench still more strongly. He gives the following account of their vigorous Nsur language: On the face of it, Nsur should be a prime candidate for language loss. All adults appear to be fluent in Ngas and Hausa and Tapshin is an enclave within the Ngas, a numerous population speaking a Chadic language, by whom they are culturally dominated. The number of speakers [of Nsur] cannot be more than 3–4000. . . . However it was apparent during the interviews that even young children are learning the language and there is no evidence of a decline in competence. Even more surprisingly, but no doubt related, the language is by no means full of loanwords from Hausa and Ngas . . . (1998: 11). The explicit surprise expressed by Blench on encountering this situation can serve as a reflection of the Western ideological perspective on what might be called “unnecessary” multilingualism, or, following Lewis, “irrational” multilingualism.

292

chapter 16

The matter-of-factly maintained multilingualism of these two small peoples is not unduly exceptional, however, in that general part of the world. Routine multilingualism is reported also among the peoples of the Mandara Mountains at the border between northeastern Nigeria and Cameroon, for example (MacEachern 2002). Men belonging to one or another montagnard group typically speak three or four different languages, including their own and that of the closest neighboring group, plus at least one language spoken on the plains below the mountains; women usually speak almost as many. MacEachern traces the ethnic and linguistic complexity of the Mandara Mountains peoples to an uncoordinated but steady movement, over the past five centuries, of various smaller peoples from the plains into the mountains as they attempted to avoid conquest by larger peoples and to escape capture by slave-traders. He notes that multilingualism has been a factor in negotiating temporary alliances that made it possible for montagnard communities to take military action against plains-dwelling peoples, and that as recently as the late 1980s fluent bilingualism in a neighboring language made it feasible for one montagnard people to avoid coming under the political control of the Wandala, the locally dominant plains-dwelling people, by claiming close affiliation with another montagnard group that was successfully contesting a Wandala claim to hegemony in a Cameroon court case. MacEachern makes the point that a people’s use of a particular language is not necessarily a given – that is, a purely passive matter of the group one is born into. Rather, in areas of ethnic complexity and multilingualism linguistic relationships and the identities that they signal can be consciously manipulated to particular sociopolitical effect. In response to centuries of attempts at controlling them, “montagnards have used their language abilities and the cultural complexity that goes along with those abilities to erect alliances and maintain their own independence in a dangerous political environment” (2002: 38). Occasions for deploying multilingualism as a social and political resource have appeared with enough frequency in the Mandara Mountains setting to sustain a pattern of acquisition and maintenance of multiple languages over a period of five hundred years. Africa, of course, is a part of the world where the territories which later became nations were brought into being by European colonial powers concerned with establishing the geographical bounds of their own control. No consideration was given to ethnolinguistic distributions within or across the boundaries created, and in consequence most African states south of the Sahara were multilingual and multicultural at their creation. In the absence of any single language community with numerical or social dominance great enough to support a claim to official-language status, many remain so today. Continuing use of the former colonial language, often as a lingua franca

bi- and multilingualism in minority and endangered languages

293

for urban populations or the educated elite, then adds another layer to the multilingualism. 6

Obligatory Exogamy and Multilingualism

Extreme multilingualism is a well-recognized phenomenon where obligatory exogamous marriage practices prevail: members of any one of a number of language communities in a particular region can contract marriages only with members of a different language community. Among the Hua, a people of the Eastern Highlands of Papua New Guinea, Haiman reported for example that the immemorial practice of women going to live in the village of a husband who must come from a group speaking a different language has made the community phenomenally multilingual. In a survey of 359 adult speakers in 1974, it was found that [in addition to speaking Hua] 305 were fluent in Gimi, 287 in Siane, and 103 in Chimbu. A smaller number of people spoke at least half a dozen other languages. Only two respondents claimed to be totally monolingual, and only eleven knew only one other language besides Hua. All the others spoke at least two, and many were fluent and at ease in four or five (Haiman 1987: 36). The differences among the languages in question are not by any means minor dialectal features. According to Haiman, Gimi, Siane, and Chimbu are impressionistically as different from Hua as French, German, and Russian are from English. Similar patterns of exogamous marriage and resultant multilingualism among all or nearly all group members have been reported from the Vaupés region of the Amazon basin and from Burma. A linguist working in Burma tells of meeting exogamous Kachin in Burma “who can converse happily in at least half a dozen languages, with native knowledge of three or more”, though he relates their profound multilingualism more particularly to a low place in the regional linguistic hierarchy, noting that speakers of another low-position language (Lisu) can be equally multilingual (Bradley 2001: 155). 7

Profound Multilingualism and the Native-Speaker Concept

Among males in the Mandara Moutains, MacEachern speaks of variable command of three or four languages. But another account of routine multilingual-

294

chapter 16

ism from West Africa indicates that the multilingualism in question can be profound indeed, even to the point of making the concept “native language” moot. In an unusual dual presentation of the fieldwork experience, linguist Fiona Mc Laughlin offers her own account of work on noun classes in three Niger-Congo languages of Senegal – Wolof, Pulaar, and Seereer – side by side with a parallel account from her principal Pulaar teacher, Thierno Seydou Sall (Mc Laughlin and Sall 2001). At one point Mc Laughlin declined an offer from Sall, whom she had found to be a natural linguist with remarkable ability to analyze the structure of Pulaar, to work on the noun classes of Wolof with her, in addition to their work on the Pulaar noun classes, because she categorized him as a native speaker of Pulaar. Sall gives this account of the interaction: I grew up in a Seereer village in a Haalpulaar family, so when I was a small child I spoke Pulaar and Seereer better than Wolof, but even then I cannot remember ever not having known Wolof. When I was fourteen I went to Dakar where my Wolof improved, and then I spent five years in Kayor, the heart of Wolof country, where pure Wolof is spoken. By pure Wolof I mean Wolof with very little French in it. Fiona thought that I could not give her the noun classes in Wolof, but for me, it would be the same thing as giving them to her in Pulaar (2001: 207). Sall notes that because he speaks the “deep Wolof” of the heartland he knows the noun classes of Wolof better than a good many Wolofs who live in towns and cities, where they mix with non-Wolofs a great deal, and that he therefore sometimes corrects Wolof people’s Wolof, including their noun classes. Mc Laughlin, reflecting on her experience with multilingual Senegalese like Sall in the Sahelian town of Fatick, where she lived and worked for a year and where Sall lived as part of a large traditional family, revised her notion of what it meant to be a native speaker of a language: [G]iven my experience with native Pulaar and Seereer speakers who spoke fluent Wolof, and learning that many of them could not remember a time when they did not speak Wolof, the very notion of a ‘native speaker of Wolof’ was thrown into question. I had rejected grammatical judgments on Wolof from Thierno because he was a native Pulaar speaker, but could not he, or others like him, also be native speakers of Wolof? In this context, could it not be possible to have more than one native language? Although at the time I did not hold these views, I now think that the urban-rural distinction in Wolof is a much more salient variable in

bi- and multilingualism in minority and endangered languages

295

distinguishing between varieties of the language than whether the Wolof speaker has another mother tongue, such as Seereer or Pulaar (2001: 202). Mc Laughlin is not the only linguist of European descent to have allowed fieldwork practices to be constrained by the notion that an individual has one, and only one, native language. A similar case is reported from Australia, where a researcher realized belatedly that he had probably let precious chances to gather data on an endangered Aboriginal language slip away because he did not take the opportunity to work with profoundly fluent speakers who were officially native speakers of some other Aboriginal language (cited in Evans 2001: 255–56). The passage from Bradley quoted above is notable in attributing “native knowledge” of three or more languages to multilingual Kachin speakers. Bradley does not say whether he did or would accept such speakers as fullfledged data sources for more than one language, but just such work needs to be done by way of plumbing the full capacities of profoundly multilingual individuals. 8

Ideological Aspects of Sustained Bi- and Multilingualism

Apart from the considerable practical benefits bi- or multilingualism may confer (with exogamous marriages and manipulable identities included among such practical benefits), additional benefits that qualify at least in part as ideological are recognized among peoples who have traditionally cultivated knowledge of more than one language. Among some such peoples, multilingual skills are regarded as a sign of intellectual or cultural superiority. In the same general area of the New Guinea Highlands where Hua is spoken, a variety of Siane known as Komunku was spoken by people whose neighbors spoke a variety of Dene. While members of the Emenyo tribe spoke Komunku and often spoke Dene as well, Dene speakers less frequently spoke Komunku. An anthropologist who worked with Emenyo in the 1950s and ’60s summed up Emenyo attitudes as follows (Salisbury 1972 [1962]: 56): The fact that there are more Emenyo bilingual in Dene, than Denespeakers who are bilingual in Komunku is not associated with any feeling among the Emenyo that they are politically less important or that their language is inferior to Dene. Bilingualism is treated as a desirable accomplishment and their command of Dene makes them, if anything, superior to the Dene.

296

chapter 16

Salisbury reported that Emenyo “actively cultivate bilingualism”, noting that when a group of laborers who had acquired pidgin during indentured service at the coast returned to Emenyo village, they immediately began giving the rest of the village males pidgin lessons. Pidgin was of course already a language of potentially unique utility, in the 1960s, because of its role as a lingua franca; but Salisbury reported further that two Emenyo youths took advantage of the availability of a Gahuku speaker who had come to the village in Salisbury’s employ to set about learning some of this more easterly language from the same family as Komunku. The presence of a catechist who came from a different language group was similarly used as an occasion for acquiring knowledge of religious materials in the catechist’s language, and songs seemed almost universally to be learned and publically sung in a variety of foreign languages (1972: 56–57). Prestige attached to knowledge of other languages generally. A well-informed account of Arizona Tewa attitudes toward the acquisition of other languages indicates that they held views similar to those of the Emenyo. The small Arizona Tewa population “enjoyed a reputation for commanding multiple languages” and attained fluency in English earlier than did the Hopi, among whom the Arizona Tewa have been enclaved since finding refuge there following their flight from the Pueblo area of present-day New Mexico after the second Pueblo revolt against the Spanish in 1696 (Kroskrity 1993: 8–9, 23). Kroskrity’s extensive ethnographic work persuaded him that the Arizona Tewa viewed a knowledge of Hopi and Navajo as instrumentally valuable (Hopi as the language of the society within which they were enclaved, and Navajo as a trade language), and of course knowledge of English (and formerly Spanish) had obvious instrumental value as well. But he noted that there was in addition a notion of cultural superiority involved, especially with regard to the Hopi, most of whom did not acquire Arizona Tewa despite several centuries of immediate proximity: “In the Arizona Tewa case . . . the Tewa view the fact that they speak Hopi but few Hopi speak Tewa as a cultural victory on their part” (Kroskrity 1993: 218). The conservative and puristic model of speech associated with religious practice in Arizona Tewa life also acts to support language maintenance and supports linguistic compartmentalization as well. Ceremonial “kiva talk” is a variety strictly reserved for religious practice. Its norms are zealously upheld, with any introduction of foreign terms during ceremonials physically punished. Kroskrity sees in the near-total absence, even in ordinary non-ceremonial Arizona Tewa, of lexical borrowings from Spanish, English, and especially Hopi the effect of a general linguistic conservatism rooted in native cultural ideals (1993: 38, 220).

bi- and multilingualism in minority and endangered languages

297

Where the exogamous multilingual Indians of the Vaupés region are concerned, one observer (Sorensen 1972) described them as having an “instrumental and practical” orientation to multilingualism. But he noted at the same time that multilingualism was actively cultivated, as with the Emenyo, and cultivated furthermore across an entire lifetime despite universal command of a regional lingua franca. Children, typically fluent speakers from their early years of both their mothers’ and their fathers’ languages and also of Tukano, the regional lingua franca, not only acquired during their adolescence several other languages spoken in the community’s longhouse, but might go on to acquire still more in adulthood. Nor did the learning process stop there: “as he [a longhouse resident] approaches old age, field observation indicates, he will go on to perfect his knowledge of all the languages at his disposal” (Sorensen 1972: 86). The degree to which expert knowledge of additional languages was actively pursued indicates an ideological orientation to multilingualism that ultimately transcends the strictly practical level of which Sorensen spoke. 9

Receding Multilingualism

Bi- and multilingualism are familiar phenomena in a number of settings around the world. Attitudes toward bi- and multilingual peoples have been various, but they have often been considerably more favorable than has been typical in Western Europe since the eighteenth century (and since then also in parts of the world heavily settled by people of European descent). Over the course of the twentieth century, however, conditions conducive to language shift on the part of minority peoples became more widespread. Many features associated with modernization and national development contributed to this process, such as the improvements in transport and communications that reduced isolation. With reduced isolation came stronger links between a central government and outlying regions. Schools that used a colonial language or an official national language could be more widely introduced; police and army presences could be established in far-flung areas, reducing local autonomy. Traditional lifeways followed by minority peoples were often disrupted, not only by these developments but also by the movement (sometimes government-sponsored) of expanding populations from more developed regions into less crowded rural areas, in a search for new agricultural land or pasturage. Food resources essential to minority peoples’ traditional subsistence modes were often reduced or lost by such intrusions, and this was even more true wherever major extractive industries moved into previously isolated regions to exploit such resources as timber, ores, gemstones, or oil. Involuntary

298

chapter 16

relocations of minority peoples (already widely perpetrated in earlier centuries) became easier than ever to carry out. Under all these pressures the ancestral languages of many small peoples either passed out of use altogether or came to be spoken by so few people that they appear likely to pass out of use in the near future. In some cases linguists or anthropologists working with a people living in one of the multilingual settings described above can shed light on the process by which traditional multilingualism has receded or been lost. In the Vaupés region of Brazil, where multilingualism was associated with exogamous marriage practices, Aikhenvald worked with the last generation to be fluent in Tariana, an Arawakan language spoken where most of the many other languages in use belong to the Tukano language family. She links the beginnings of a breakdown in traditional cultural and linguistic patterns to the coming of Salesian missionaries to the area in the early 1920s. The Salesians considered the traditional multilingualism of the region a pagan practice, according to Aikhenvald, and in their effort to make the local Indians monolingual “(‘like other civilized people in the world’)”2 they chose to employ only Tukano, the language spoken by the largest number of people. They also relocated Indian settlements closer to mission centers and substituted structures housing nuclear-family units for the traditional longhouses. Participation in the routine multilingualism of the longhouses was eliminated by these changes, and because able-bodied Indian men also began to go off to take up paid employment in Brazilian rubber plantations or mining operations, children’s exposure to their father’s language in particular was reduced. The children of absent fathers might speak other languages more often than Tariana, especially Tukano; furthermore, some Portuguese began to enter the mix via men returning from jobs. Within a few generations Tukano and a regional form of Portuguese predominated, with only some elderly people retaining a good knowledge of Tariana (Aikhenvald 2002). Among the Arizona Tewa, too, Kroskrity found that the way in which education was being delivered favored the official language and disfavored the minority language. Many of the young people got their secondary education at boarding schools in which English was not only the sole medium of instruction but also the only common language among Native American young people who came from a variety of southwestern tribes. Socialization of young people 2 Aikhenvald (p.c., 7 Feb. 02) heard older Salesian missionaries make remarks to this effect in conversation. She notes that a younger generation of Salesians is now active in promoting indigenous languages, but for small language groups such as the Tariana the change in attitude and policy most likely comes too late.

bi- and multilingualism in minority and endangered languages

299

by older kinsfolk and other community members was curtailed by this educational experience, and in addition an influential new reference group was formed for the young people, one that consisted of English-speaking age-mates with off-reservation experience. Kroskrity found young people signaling their allegiance to this new reference group by replying in English when their elders spoke to them in Arizona Tewa. The young people had become aware of the economic disadvantages of reservation life and of the possibilities of material advantage elsewhere, through exposure to mass media and also through their time in urban boarding schools. In Kroskrity’s view socioeconomic factors were primary in motivating young people’s increased use of English, but because the actual economic possibilities for young Arizona Tewa in off-reservation settings are extremely limited, he considered it possible that an unrewarding venture into off-reservation employment might ultimately provide an incentive for return to a more traditional reservation life. The future for Arizona Tewa – maintenance or loss – could depend, consequently, on whether the preference for English among young people proved to be a life-cycle phase, with a later reaffirmation of Arizona Tewa traditional values (including the language), or whether it proved to be permanent (Kroskrity 1993: 103–05). Post-adolescent decisions to stay in the home area after all, and therefore also to maintain the ancestral language, have in fact been known to change the apparent outlook for a small group and their language. In the 1970s a linguist working in an enclaved village in southern Italy, where a Francoprovençal language had been spoken for more than 500 years, predicted on the basis of his census of bilingual speakers of various ages, and also by his observations of language use, that Faetar, the local Francoprovençal language, would be dead by the year 2000. The villagers themselves expressed the same opinion. Another linguist, arriving to work in the same village in the 1990s, found the villagers as bilingual as ever, but she heard them give much the same prediction: Faetar would be dead in twenty years. Because the earlier linguist had done his census by age groups, the second linguist was able to establish that adolescents who were using Italian almost to the exclusion of Faetar in the 1970s were adults using Faetar among themselves and with their children in the 1990s. The adolescents of the 1990s, however, were again using mostly Italian by preference, and their language preference was again giving rise to fears that the language would soon cease to be spoken (Nagy 2000: 128–29). The most serious threat to Faetar appeared to lie in a gradual shrinkage of village population as more people left in pursuit of greater economic opportunity, a depopulation common to agricultural regions of Italy. Lack of sufficient population to keep the village viable remains a future possibility, therefore, but lack of adolescent speakers as such is to some extent an age-graded phenomenon in Faeto. Awareness of such

300

chapter 16

temporary age-related shift patterns explains Kroskrity’s reluctance to predict the future for Arizona Tewa. 10

Shift as the Norm, Maintenance as the Exception

In cases such as those of Tariana and Arizona Tewa, in competition with Tukano and Portuguese and with English respectively, major disparities in the currency of competing languages are evident. Tariana had currency only in the traditional longhouses of Indians in the Vaupés area where some individuals had fathers with a Tariana identity; both Tukano and Portuguese by contrast are used by larger numbers of people and across a wider range of geographical and social settings. Arizona Tewa, too, has obviously limited currency by comparison with English. Yet previous generations of Indians in the Vaupés region spoke the lingua franca Tukano without relinquishing their local-currency languages, and the Arizona Tewa were competent speakers of Spanish and of English relatively early, without the much wider currency of those languages estranging them from Arizona Tewa. The new and unfavorable development in such settings is not the acquisition of other languages, including languages of wider currency, but the abandonment of an ancestral language in the process. That is, at the present day extended encounters between minority peoples and dominant-group members are more likely than not to produce subtractive bilingualism, whereas in the past there was more chance that additive bilingualism would be the result. It is not difficult to understand why members of groups speaking smaller local-currency languages would wish to acquire the wider-currency language of a larger, more prosperous, and relatively dominant people. But a fundamental question that is frequently left unasked is this: why, in view of both the contemporary frequency and the historical frequency of bi- and multilingualism, should speakers of smaller local-currency languages stop speaking their own ancestral languages when they acquire a wider-currency language? Giving up a limited-currency minority language altogether when taking up a wider-currency language may seem so ordinary a phenomenon as to be unremarkable, but this is in all likelihood an ahistorical notion. Even in Europe it was not expected before the end of the eighteenth century, and outside the European sphere of influence dominant groups were known to tolerate the maintenance of subordinate peoples’ languages readily. The Ottoman Turks made no attempt to turn the diverse peoples whom they ruled into Ottoman Turkish speakers, and the Thai kingdom is said to have been equally tolerant

bi- and multilingualism in minority and endangered languages

301

of nearly all the many minority languages spoken under its rule (Smalley 1999: 341–49). Acquisition of a second (or third or fourth) language need not imply loss of a first language. The ideologies and behaviors of two populations are relevant in this matter. Not only do the villagers of Tapshin continue to speak Nsur, even though they also speak Ngas and Hausa, but the numerically and culturally dominant Ngas speakers who surround them apparently have no objection to the continuing use of Nsur in their midst. The relevance of dominant-group attitudes becomes more obvious, perhaps, when the Arizona Tewa case is considered. The Hopi, a somewhat larger population and a well-established one by comparison with the Tewa who took refuge among them, found it acceptable for the Tewa not only to learn Hopi but also to continue to use their original language during the roughly 300 years of their residence on the Hopis’ First Mesa. Paradoxically, anglophone America, with an overwhelming numerical dominance and a uniquely secure official language, exerts serious pressure on the Arizona Tewa (as on all other minority groups) for an assimilation that includes abandonment of the ancestral language as well as adoption of the official language. Hundreds of years of enclavement among a numerically superior group, and even a degree of Tewa-Hopi intermarriage, did not threaten the survival of the Tewa language during its earlier history in Arizona; yet it is possible that boarding-school education in English is now doing just that. Greatly increased pressure for a shift to dominant-language monolingualism in many parts of the contemporary world seems likely to rest in part on continuing dissemination, despite he decline of colonialism, of European language ideologies: above all the one-language, one-nation ideology associated with nationalism and the “ideology of contempt” for subordinate peoples, but also European notions that the languages of peoples who exhibit low technological development must necessarily be equally limited, while the languages of peoples who have achieved politically and technologically dominant positions must necessarily be superior linguistic instruments (Dorian 1998). This selfserving view justifies the established ascendancy of a few “superior” European languages in a post-colonial world still linguistically shaped by European colonialism, while simultaneously rendering the displacement of innumerable “lesser” languages unimportant. Still other elements of European linguistic ideology have been identified in the pressure for monolingualism, for example the notion that acquisition and use of an ancestral minority language early in life is deleterious to full competence in the official language. As one researcher memorably put it, in connection with the pressure exerted on children from minority groups in far northern Norway, Sweden, and Finland to make as rapid

302

chapter 16

a shift to the national language as possible, “it was believed that a child’s head (especially a minority child’s head) would not have space for two languages” (Huss 1999: 129). Huss follows other scholars in tracing such attitudes to the influence of seriously flawed bilingualism research that supposedly detected lower intelligence and inferior linguistic and cognitive development among bilingual children, especially the children of late nineteenth-century and early twentieth-century immigrants to the U.S.3 Certainly such notions have been communicated to minority-group parents, both among immigrant populations and among indigenous peoples, with some of those who have gone over to raising their children with the majority language explaining their decision in such terms. 11

Ethnolinguistic Vitality Assessment

Since the 1970s ethnolinguistic vitality studies have attempted to account for observed cases of shift or maintenance and to predict how likely a minority population will be to maintain its language or to give it up. Three sets of objective factors were originally the chief focus of evaluative efforts: status factors (economic, social, and sociohistorical factors, and status within and without the group), demographic factors (proportion of the overall population, concentration of the minority population, birthrate, etc.), and institutional support and control factors (use in mass media, education, government services, industry, religion, culture, politics). If the “surprisingly” vigorous Nsur language is taken as a test case, these factors are clearly inadequate for predictive purposes. The Ngas language was said to be spoken by many more people than Nsur, and the Ngas were also described as culturally dominant. Whether Ngas had any institutional support is not evident in Blench’s report, but Nsur almost certainly did not. By these measures Nsur speakers ought to be in the process of shift to Ngas, just as Blench expected them to be. Among the various efforts to refine ethnolinguistic vitality assessment, the approach with perhaps the greatest potential for uncovering the roots of maintenance or shift in a non-Western setting like Tapshin village is the development of a Beliefs on Ethnolinguistic Vitality Questionnaire (Allard and Landry 1986). This is not to say that an oral version of the questionnaire as such would be a suitable instrument for research in Tapshin and its neighboring villages (an unlikely prospect, since questionnairing is a research procedure that does not always travel well, culturally speaking), but rather that examination 3 Huss (1999: 129) cites Peal and Lambert, 1962: 1–2 and Skutnabb-Kangas, 1981: 222 in this connection.

bi- and multilingualism in minority and endangered languages

303

of beliefs as a fundamental part of the environment that produces language behaviors might best be able to tap into the underpinnings of Nsur linguistic persistence and of Ngas linguistic tolerance. 12

Reactive Language Maintenance and Assertive Language Maintenance

Despite the weight of the many factors that favor language shift, some countercurrents are observable. One pattern of maintenance involves early shift with later reversion to traditional cultural and linguistic behaviors, in the reactive pattern that Kroskrity considered a possibility for Arizona Tewa young people. Initially, contact with the dominant society, especially educational contact, appears to open a route to economic advance. But where minority peoples are visibly distinguishable from members of the dominant group (and have often acquired a distinctive local version of the dominant language besides), job scarcity combined with a lingering racism can limit advancement and assimilation. This is said to have happened already among the Kwak’wala Indians of Vancouver Island: the Kwak’wala young people have shifted entirely to English, but without reaping the material benefits that the shift seemed to promise (Anonby 1999: 35). In a study of the outlook for the survival of New Caledonian vernaculars, Schooling found that Melanesian New Caledonians educated in French with an expectation of job opportunities that did not materialize settled back into traditional life patterns in their villages of origin, where kin networks remained strong and they had kin-based claims on land. Schooling reported disillusionment among young New Caledonians whose parents had emphasized the acquisition of French more than that of the local vernacular, to the point where some young adults were reversing the pattern and consciously speaking the vernacular with their own children even though husband and wife might frequently use French between themselves (Schooling 1990: 51–52). Another pattern favoring maintenance can appear when a group succeeds in achieving some measure of hoped-for economic advance, whether through education or through development of resources (e.g. scenic resources for tourism or sport), before shift to a dominant language is complete. Economic success can produce enough psychosocial confidence among still-bilingual speakers to encourage assertion of their ethnolinguistic identity and of the right to use their heritage language more widely, including in education, as has happened in Wales, for example. Also favorable to ethnic self-assertion are wider recognition of linguistic human rights in some parts of the world and increased communication (especially electronic communication) among widely separated small peoples.

304

chapter 16

Local, regional, or national governments inclined to ignore or mistreat an indigenous group find themselves in the glare of unwelcome publicity when small and seemingly isolated indigenous peoples succeed in attracting support from human rights observers or from a vocal coalition of indigenous peoples. 13

Revitalization Efforts on Behalf of Endangered Languages

Languages are sometimes viewed as endangered even when nearly all group members are still speakers. If dominant-group assimilative pressures are seen to be rising while resistance on the part of a minority group would seem to require resources that are not in evidence (such as a large population, a strong ethnolinguistic identity, a viable and locally-based means of subsistence), longterm survival of the minority-group language can not be taken for granted. If a shift has already begun – if young children prefer to use a school-acquired dominant language with one another, for example – the future of the language begins to look questionable. If lack of ancestral-language knowledge reaches into the ranks of young parents who are raising children, the outlook becomes a little darker still. In a number of shift-prone settings some young parents can be found going against the tide and taking pains to raise their children in and with the minoritygroup language, as has occurred for example in Scotland and Nova Scotia (Scottish Gaelic), in Finland (Sami), and in Hawai’i (Hawaiian). Among these parents accomplished learners are often to be found, either individuals of the relevant ethnic group who were not raised with the ancestral language themselves or outsiders who have married into the ethnic group and learned the traditional ethnic language. Speakers of both sorts value the ethnic language the more for having had to acquire it effortfully, and with infants of their own they progress from being dedicated learners to being dedicated transmitters. It would take a good many fluent parents deeply devoted to home transmission to produce a numerically significant number of fluent new minoritylanguage speakers, however, and many endangered-language communities turn to schooling, in particular to immersion schooling, for the relatively rapid multiplicative effect it can produce: a handful of dedicated and well-trained teachers, using only the minority-group language in the classroom, can produce scores of new minority-language speakers over a period of several years. Such immersion programs have by now amply demonstrated their success, especially in the cases where primary immersion schooling is followed by secondary immersion schooling.

bi- and multilingualism in minority and endangered languages

305

There are nonetheless limitations to the effectiveness of immersion schooling. Unless there is a good sprinkling of native-speaking home-transmission children in the immersion classroom, the pupils have only one fluent-speaker model available, namely the teacher. They must begin to use the target language themselves before they have had enough exposure to it to acquire its grammatical and phonological structure fully, and before very long their teacher’s well-formed utterances make up only a small part of the classroom language model, since they are providing many imperfect models for each other. The resulting school-based version of the minority-group language often differs quite noticeably from the original native-speaker model, partly by showing a good deal of influence from the majority-group language that most of the children speak at home and partly also by the introduction of many newly coined school-register words. Communities differ about the acceptability of this outcome. For some, a reconfigured version of the group’s language is preferable to no version at all. For others, a version of the language that the children’s grandparents can barely recognize as their own tongue is a very dubious “success”. Few minority-language groups expect or even wish to replace the majority language with their ancestral language. Their goal is rather well-developed dual or multiple language capacities which in the one case offer access to the heritage that is available in and through the ancestral language and in the other case offer participation in at least one broader language community besides. Clearly there is nothing inherent in bi- or multilingualism as cognitive or social phenomena that poses an obstacle to such an outcome. The ideological and political obstacles can be considerable, however, as the issues discussed here indicate. References Aikhenvald, Alexandra Y. 2002. Traditional multilingualism and language endangerment. In David Bradley & Maya Bradley (eds.), Language maintenance for endangered languages: An active approach, 24–33. London: Curzon Press. Allard, Réal, & Landry, Rodrigue. 1994. Subjective ethnolinguistic vitality: A comparison of two measures. International Journal of the Sociology of Language 108, 117–44. Anonby, Stan J. 1999. Reversing language shift: Can Kwak’wala be revived? In Jon Reyhner, Gina Cantoni, Robert N. St. Clair, & Evangeline Parsons Yazzie (eds.), Revitalizing indigenous languages (pp. 33–52). Flagstaff, Arizona: Northern Arizona University.

306

chapter 16

Blench, Roger. 1998. Recent fieldwork in Nigeria: Report on Horom and Tapshin. Ogmios: Newsletter of the Foundation for Endangered Languages 9, 10–11. Blommaert, Jan, & Verschueren, Jef. 1998. The role of language in European nationalist ideologies. In Bambi B. Schieffelin, Kathryn A. Woolard, & Paul V. Kroskrity (eds.), Language ideologies: Practice and theory (pp. 189–210). New York and Oxford: Oxford University Press. Bradley, David. 2001. Language attitudes: The key factor in language maintenance. In Osamu Sakiyama & Fubito Endo (eds.), Lectures on endangered languages: 2 (pp. 151–60). Kyoto: Endangered Languages of the Pacific Rim. Dauenhauer, Richard, & Dauenhauer, Nora Marks. 1998. Technical, emotional, and ideological issues in reversing language shift: Examples from Southeast Alaska. In Lenore A. Grenoble & Lindsay J. Whaley (eds.), Endangered languages: Current issues and future prospects (pp. 57–98). Cambridge: Cambridge University Press. Dorian, Nancy C. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: The University of Pennsylvania Press. ———. 1982. Language loss and maintenance in language contact situations. In Richard D. Lambert & Barbara F. Freed (eds.), The loss of language skills (pp. 158– 67). Rowley, Massachusetts: Newbury House. ———. 1998. Western language ideologies and small-language prospects. In Lenore A. Grenoble & Lindsay J. Whaley (eds.), Endangered languages: Current issues and future prospects (pp. 3–21). Cambridge: Cambridge University Press. Edwards, John. 1994. Multilingualism. London and New York: Routledge. Evans, Nicholas. 2001. The last speaker is dead – long live the last speaker! In Paul Newman & Martha Ratliff (eds.), Linguistic fieldwork (pp. 250–81). Cambridge: Cambridge University Press. Flinn, Juliana. 1992. Diplomas and thatch houses: Asserting tradition in a changing Micronesia. Ann Arbor: University of Michigan Press. Grillo, R.D. 1989. Dominant languages: Language and hierarchy in Britain and France. Cambridge: Cambridge University Press. Haiman, John. 1987. Hua: A Papuan language of New Guinea. In Timothy Shopen (ed.), Languages and their status (pp. 35–89). Philadelphia: University of Pennsylvania Press, for the Center for Applied Linguistics. Huffines, Marion Lois. 1989. Case usage among the Pennsylvania German sectarians and nonsectarians. In Nancy C. Dorian (ed.), Investigating obsolescence: Studies in language contraction and death (pp. 211–26). Cambridge: Cambridge University Press. Huss, Leena. 1999. Reversing language shift in the far north: Linguistic revitalization in northern Scandinavia and Finland. Uppsala: University of Uppsala. Kroskrity, Paul V. 1993. Language, history, and identity: Ethnolinguistic studies of the Arizona Tewa. Tucson: University of Arizona Press.

bi- and multilingualism in minority and endangered languages

307

MacEachern, Scott. 2002. Residuals and resistance: Languages and history in the Mandara Mountains. In Brian D. Joseph, Johanna Destefano, Neil Jacobs, & Ilse Lehiste (eds.), When languages collide: Perspectives of language conflict, language competition, and language coexistence (pp. 21–44). Columbus: The Ohio State University Press. Mc Laughlin, Fiona, and Sall, Thierno Seydou. 2001. The give and take of fieldwork: Noun classes and other concerns in Fatick, Senegal. In Paul A. Newman & Martha Ratliff (eds.), Linguistic fieldwork pp. 189–210. Cambridge: Cambridge University Press. Nagy, Naomi. 2000. Fieldwork for the new century: Working in Faeto, an endangered language community. Southern Journal of Linguistics 24, 121–36. Paulston, Christina Bratt. 1994. Linguistic minorities in multilingual settings: Implications for language policies. Amsterdam/Philadelphia: John Benjamins. Peal, Elisabeth & Lambert, Wallace. (1962). The relation of bilingualism to intelligence. Psychological Monographs 76, No. 546. Salisbury, R. F. 1972. Notes on bilingualism and linguistic change in New Guinea. In J. B. Pride & Janet Holmes (eds.), Sociolinguistics: Selected readings (pp. 52–64). Harmondsworth: Penguin Books. [Reprint of Anthropological Linguistics 4, 1962, 1–13.] Sankoff, Gillian. 1980. Multilingualism in Papua New Guinea. The social life of language (pp. 95–132). Philadelphia: University of Pennsylvania Press. Schooling, Stephen J. 1990. Language maintenance in Melanesia: Sociolinguistics and social networks in New Caledonia. Arlington: The Summer Institute of Linguistics & the University of Texas at Arlington. Skutnabb-Kangas, Tove. 1981. Tvåspråkighet. Lund: Liber Läromedel. Smalley, William A. 1999. Linguistic diversity and national unity: Language ecology in Thailand. Chicago and London: University of Chicago Press. Sorensen, A. P., Jr. 1972. Multilingualism in the northwest Amazon. In J. B. Pride & Janet Holmes (eds.), Sociolinguistics: Selected readings (pp. 78–93). Harmondsworth: Penguin Books. Williams, Glyn. 1992. Sociolinguistics: A sociological critique. London & New York: Routledge. Woolard, Kathryn A. 1998. Language ideology as a field of inquiry. In Bambi B. Schieffelin, Kathryn A. Woolard, & Paul V. Kroskrity (eds.), Language ideologies: Practice and theory (pp. 3–47). New York & Oxford: Oxford University Press.

part four Language Use

∵

chapter 17

Stylistic Variation in a Language Restricted to Private-Sphere Use One of the fundamental principles of sociolinguistic investigation, according to William Labov, is that there “are no single-style speakers” (1970: 19). Dell Hymes makes essentially the same point, asserting that “no normal human being talks the same way all the time” (1984: 44). But dying languages have been included under the rubric “exceptional language” (Obler and Menn 1982), and some researchers have asserted that they are essentially “monostylistic” (Dressler and Wodak-Leodolter 1977: 36–37, Dressler 1982: 326, 1988: 188–89). In invoking the notion of monostyIism in language decay, Dressier specifies “restriction to a very casual style used with very familiar dialogue partners about restricted topics in routine speech situations” (1982: 326). This paper represents an attempt to determine whether, in a language undergoing just that sort of restriction, the fundamental principle articulated by Labov and Hymes still applies, or whether in those unfavorable circumstances a lack of stylistic variation occurs which would warrant the term monostylism. When a language moves close to a foreseeable point of extinction, it is inevitably spoken less and less frequently. Even those speakers who continue to use it regularly have ever fewer fellow-speakers to converse with, because the fluent speaker population is aging and thinning out. Their own kin networks begin to be heavily populated by younger bilinguals who speak the expanding language of the region better than the original ancestral language, and eventually those same kin networks have expanding-language monolinguals among their youngest members. Languages which die out gradually in this fashion, via the progressive failure of intergenerational transmission, usually retreat in the final generations to a few spheres of use: they persist in domestic settings among the older generation, and they are used for casual social intercourse among contemporaries who were schoolmates or workmates in their young years. Occasionally some specialized use of the ancestral languages will remain in more formal spheres: religion, typically, and sometimes a few other more-or-less ritualized types of behavior. This is certainly the profile presented by the East Sutherland dialect of Scottish Gaelic, still spoken in the last half of the twentieth century by a dwindling number of Gaelic-English bilinguals in three villages on the east coast of that very northerly Highland county, all fisherfolk or the offspring of fisherfolk.

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�19

312

chapter 17

For East Sutherland Gaelic (ESG hereafter), the competition comes not from any more vigorous variety of Gaelic, but from English. In the local context the two languages, English and Gaelic, are in competition with one another, and I will therefore speak of ESG as a language approaching extinction, even though it is actually a regional dialect of Gaelic which is fading from the scene; there are certainly other forms of Gaelic which will survive the approaching demise of ESG. As would be expected, the competition between ESG and English has been a very unequal one. English has long had the unwavering support of the national state, so that education has been solely in English since the state took it on, and local people’s experience of military service has also meant the use of English, in modern times, as has any contact with the court system or local administrative services. The one social institution which has favored Gaelic in the 20th century has been the church, with some denominations (all Protestant in this region) persisting longer in the use of Gaelic than others. It was the long-continued use of Gaelic in religious life which prevented the language situation among the ESG-speaking fisherfolk from becoming one of bilingual diglossia during the first half of the twentieth century, in fact. So long as Gaelic remained the language of church services, scripture readings, psalm singing, and praying, it could not be said that English had usurped all H[igh] language functions or that Gaelic had retreated to L[ow] language functions. By the early 1960s, when I first arrived in East Sutherland to begin studying the dialect and its setting, only one village still had Gaelic church services available. All of the Gaelic speakers who were of fisherfolk background (which means all the indigenous Gaelic speakers who remained, by that time) had had their formative religious experiences through the medium of Gaelic, however. Active control of religious language was largely restricted to men, since only males were trained to precent (line out) the psalms during services, and only men had been expected to lead prayer spoken out loud in household settings, at wakes, and so forth. Only men could serve as elders and thus take a limited leadership role in congregational life. (Ministers, too, were exclusively male, but since there were never any ministers of local origin this offered no verbal role for speakers of ESG.) As the number of men who could precent the psalms in the one remaining church with Gaelic services declined, a relatively young woman was sometimes persuaded to precent when no male precentor was available (though she did so from her pew, declining to take the precentor’s seat facing the congregation at the front of the church). She had a strong voice and was self-taught, but her example at least indicates that women could sometimes take a prominent role in using the rather archaic (and very nonlocal) Gaelic of religious usage, at least in the fixed and familiar language of the metrical psalms.

stylistic variation in a language restricted to private-sphere use

313

With the retirement of the last Gaelic-speaking minister in the late 1960s or early 1970s, Gaelic ceased to play any regular formal and public role for Embo villagers; it had ceased to play any such role for members of the other two former fishing communities of East Sutherland long since. Because it seemed neither appropriate nor congenial to explore merely as an observer the use of Gaelic in people’s private devotional lives, I never tried to investigate the active control of religious language among the more than forty fluent speakers with whom I worked. In their work with me a few men demonstrated some control of Biblical vocabulary not in everyday use, and I noted also some occasional use of the language of benediction in formal partings among fluent speakers not in regular contact with one another (including the closing of tape-recorded messages). It seems likely that some individuals used Gaelic in private prayer, and some men were known as good psalm singers. It is possible that some few men read the scriptures in Gaelic. For most ESG speakers in the latter half of the twentieth century, and in particular for most women speakers, Gaelic was the language of hearth and home. There simply were no public spheres in which their native form of Gaelic could appropriately be used. The sole remaining public spheres for any Gaelic at all in these villages, when I began to do research there in the 1960s, were the church (for Embo village only) and the ceilidh. Ceilidhs were evening functions at which admission was charged, with performances of vocal and instrumental music, Highland dance, and sometimes also amusing monologues; they were usually held to benefit some cause or organization. For each ceilidh there was a fear-an-taighe (master-of-ceremonies), and if possible a Gaelic speaker served in this role. It was customary for any Gaelic-speaking fear-an-taighe to use some Gaelic, at least ceremonially (e.g., in welcoming the audience at the outset and in thanking the performers at the close), during the course of the evening. I am not aware of any Gaelic speakers native to the three fishing communities who served as fear-an-taighe, any more than as minister, however; these were roles performed by outsiders to the local communities, and consequently the Gaelic used was not the local variety. All the same, the Gaelic spoken from the platform at ceilidhs exposed local speakers to a relatively formal use of the language in a public function, as church services also did for those who still had them available. These local uses of a more formal Gaelic had some personal reality for ESG speakers, unlike the limited use of Gaelic in broadcasting, but it must be stressed that any variety of Gaelic used by a minister, a fear-an-taighe, or a broadcaster would inevitably be too different from the local East Sutherland variety to serve as a model for local imitation. ESG is radically deviant, from the point of view of the standard language,

314

chapter 17

and the fit between ESG and any variety used in public spheres is too poor to allow any use of the other variety by local speakers beyond the odd adoption of lexical items or turns of phrase. ESG speakers who attempted more than this (and some few occasionally did) quickly came to grief through their inability to sustain the performance over more than a few phrases or sentences.1 When a language has retreated to the private sphere exclusively, as ESG has, there may well be some question about its stylistic range. If it is never used to make a formal speech, to introduce the speaker at a meeting, to give the vote of thanks to a guest performer or speaker at a public function, to debate the agenda of a local organization, to make a motion during a public meeting, to make an announcement of coming events, and so forth, how much capacity will it maintain to express stylistic differences at all? My impression, from participating repeatedly over a fifteen-year period in some of the daily-life activities of fluent ESG speakers who became my friends, was that there remained a useful range of styles available to speakers, though a limited range certainly by comparison with that available to speakers of languages which are used to more public and formal purposes than ESG. In this chapter I try to substantiate my impression of the stylistic flexibility of ESG from tape-recorded but freely spoken data drawn from the single ESG speaker who was most available to me, my landlady over the entire fifteen-year stretch of my recurrent East Sutherland fieldwork. Though it is the best I have available, this body of data is very far from ideal for the purpose of demonstrating the fullest possible range of styles in ESG. For one thing, it all derives from a single kind of verbal event, a very lightly guided variety of interview.2 For another, it derives from a single main participant in 1 I should note that this is just as true of me as of other ESG speakers. Though I’ve often tried, in unavoidable interactions with speakers of more standard varieties, and especially with Gaelic intellectuals, to make some active use of my passive and partial knowledge of standard Gaelic, I’ve been unable to sustain it. Since I have a large advantage over most ESG speakers in passive literacy and in conscious knowledge of the structure of the standard language, my own failures in this respect make me deeply aware of the difficulties of adopting nonlocal models in active Gaelic use. 2 I term these speech events interviews in recognition of the fact that two people were present, one of whom operated a tape recorder and sometimes asked questions or made comments. In actual fact one speaker did the vast majority of the talking, chose the topic a good part of the time, and was relatively seldom interrupted. Linguists sometimes present themselves as power figures in any situation involving their use of a tape recorder (Briggs 1986: 89, 120) and in theory I could of course have terminated the session by turning off the tape recorder; but the reality of these sessions is that the other participant was the dominant party in most respects: she controlled the language being spoken far better than I, she chose most of the

stylistic variation in a language restricted to private-sphere use

315

that verbal event, with one unchanging lesser participant. These are certainly not ideal circumstances for plumbing the far reaches of style change. There are some redeeming factors, nonetheless. My landlady, whom I will call MMK,3 was profoundly illiterate [in Gaelic].4 She was also remarkably uninclined to attempts at shifting in the direction of more standard Gaelic forms, which was striking in view of the fact that she had lived for a few years early in her married life on the west coast of Wester Ross, where she had heard a good deal of the Gaelic dialect (more nearly standard and definitely more prestigious) native to that region. She can be said to represent genuinely local ESG speech norms. She was also a wonderfully uninhibited speaker. That is, she was not only unpretentious in sticking strictly to local Gaelic usage, she was also much more impervious than most local speakers to the scorn of speakers of other dialects and to the disdain of many English monolinguals for speakers of Gaelic generally and for speakers of the local Gaelic particularly. She spoke Gaelic with pleasure and relish, and she was willing to speak it with normal audibility on the street, something a good many local women avoided. She liked to talk, in either of her languages, and had a great appreciation for conversation, gossip, and the recounting of stories. She was, in short, a great talker, harder to persuade to silence than to prompt to exuberant volumes of speech. For a linguistic fieldworker she had another splendid attribute: she had no qualms at all about being tape-recorded. Because of this, and because she and I lived under the same roof so often, she is better represented in my field recordings than most other speakers. And though I have perhaps as much material for two others, no other speaker romped so enthusiastically through as many stories and reminiscences as she. Consequently it was to the material which she spoke on tape for me that I turned when I became interested in stylistic variation, and her freely spoken tape-recorded material forms the basis for this study.5 topics for presentation, she held the floor uninterrupted the majority of the time. Where I affected the topic it was by asking that she repeat for the tape recording a story which she’d already told me. Where I affected the form of the story it was generally by interrupting to ask for more information about something she’d said. 3 She is G2, that is, the second-oldest of my Golspie village sources, in many of my other publications. That seems an excessively impersonal designation for the present paper, however, where the individual aspects of her speech are at issue. 4 The adverb reflects the fact that she could not so much as recognize written Gaelic when she saw it; she used to give me the weekly radio bulletin so that I could tell her which program titles were Gaelic as opposed to French, because both looked equally unfamiliar to her. 5 I was not interested in register variation as such at any point in my original study of ESG and made no effort to gather materials which would be useful for that purpose. By the time I was interested, I was prevented by health problems from returning to East Sutherland to collect

316

chapter 17

During three recording sessions, in different years, MMK had an opportunity to display her verbal skills in Gaelic to good advantage.6 As her audience I certainly felt that she was able to change her style to suit her topic, but I had not looked closely at her texts in order to discover whether my impression was supported by the data. Wishing to do that now, I have concentrated on the first two taping sessions, one of thirteen minutes’ duration dating from 1964,7 and another of twenty-five minutes’ duration dating from 1968. She began the earlier of the two sessions with a particular story she had wanted to tell me, and the story in question had a somber theme. It was a story about a sign of ill luck, followed by drowning deaths among the fishermen in her own kin network, and she wanted to persuade me of the validity of her own belief in premonitions of death. She delivered the story in an unusually serious voice, slowly and with didactic emphasis, so it stood to reason that such a narrative might lean toward the more formal end of her stylistic range. Since most of her stories were far from soberly serious, the presence of this narrative made this first session, in contrast with any other, a good candidate for investigation. And though MMK knew me quite well in 1964, after many months of shared residence, and seemed thoroughly comfortable in talking with me in either of her languages, after much social interaction in both languages and many working sessions on Gaelic as well, she knew me still better in 1968, after intervening visits in 1965 and 1967, so that she could be expected to be still more at ease with me and with tape recording by the time of the second session. The later session could therefore be expected to show a special informality of long acquaintance, by comparison with the earlier one. From the thirty-eight minutes of recordings made with MMK, I selected all of the material which seemed to me to constitute narratives: stories with particular themes, particular actors, and some degree of resolution.8 There were

material specifically to that end. I have not been able to visit East Sutherland since 1978, and most of the fluent speakers with whom I worked (all, in the villages of Brora and Golspie) have died since then. 6 She provided other tape-recorded material on other occasions, including songs, phonologically contrastive words and phrases, and two sections of a letter-tape, but these three sessions contain all of her narrative material. 7 This session takes less time on tape than it took in the event, because I stopped recording each time I myself spoke. The tape was made late in 1963–64, my first year in East Sutherland, and my Gaelic was not yet comfortable. I evidently preferred not to have my own efforts at Gaelic interviewing immortalized on the tape, since I stopped the recording whenever I spoke and restarted it as MMK began to speak. 8 A question from me interrupted narrative 5 before it had reached its natural conclusion.

stylistic variation in a language restricted to private-sphere use

317

eight such narratives, two in the first session and six in the second. Under short titles based on their themes, I offer brief summaries here: 1. The prefigured TRAGEDY (448 words).9 Shattering glass leads to a prediction of tragedy to come. Soon thereafter two local fishermen are drowned. The body of one is recovered immediately, but the body of the other is found and buried only after a considerable interval. 2. Little boy LOST (339 words). While the family is at a herring fishing station MMK’s youngest brother wanders away, but in the direction of the area where his aunt lives; she spots him and retrieves him. 3. The BATHER saved from drowning (974 words). MMK sees an acquaintance in difficulty while bathing and pulls her part of the way out of the water; others help to bring her the rest of the way in. The bather’s sister claims that it was a dog who performed the rescue. 4. Past SINS exposed (144 words). While the bather is being dragged urgently out of the sea, more of her becomes visible than is polite; impolite inferences are drawn about her past history. 5. The Gaelic CLASS (308 words). MMK attends a Gaelic class in preparation for choral singing in a competition, but when she speaks her local Golspie Gaelic the instructor, not a local speaker, rejects it peremptorily. She leaves and refuses to return. 6. The choir COMPETITION (256 words). A strategy for defeating a neighboring village’s choir in the Gaelic choral competition is successful and Golspie wins the challenge cup. 7. The unwelcome Gaelic COACH (860 words). A Gaelic speaker from the Hebrides is rejected as Gaelic choir coach in Golspie but coaches one or two neighboring choirs; the other villages pay her, but not Golspie. Singers from

9 The word counts are approximate. ESG is an unwritten dialect and though I have sometimes contrived to write it in Gaelic orthography for publication purposes, I do not normally, or willingly, do so. All of my ESG tape transcriptions are in quasi-phonemic renderings, and in many cases I followed phonological boundaries rather than those of the traditional orthography. For example, the expression bha aid ‘they were’ has a single long vowel with no rearticulation in ESG: [va:ǰ]. I always wrote it as a unit in my field notes, and I reckoned it as a single word in making the word counts for MMK’s narratives here. There are so many such cases, all involving short words, mostly of high frequency, that I did not trust myself to take these phrases apart with any consistency while counting. Consequently the word counts given here are all lower than they would be if someone more comfortably literate in Gaelic than I were to translate the texts into written ESG, following the word-division conventions of the standard written language.

318

chapter 17

all the villages give her presents when she marries locally, but when the marriage is dissolved within a short time, she has the gifts auctioned off. 8. The defective Gaelic RECORD (214 words). The instructional record which a new boarder is using to help him learn Gaelic seems to MMK’s ears to be teaching him incorrect Gaelic. In referring to these narratives subsequently I will use the number and a single-word rubric: 1, TRAGEDY; 2, LOST; 3, BATHER; 4, SINS; 5, CLASS; 6, COMPETITION; 7, COACH; 8, RECORD. Each but the second (Lost) has a clear narrator’s mood to it, matching the thrust of the story: number 1, serious/ didactic; numbers 3, 5, and 7, indignant; numbers 4 and 8, hilarious; number 6, delighted. The second narrative (Lost) has a certain sober tone which carries over from the first, though it is perfectly cheerful and has a happy ending (the boy is swiftly found and was still alive as an elderly man when the story was told). It will become evident in due course that some ‘soberness’ carryover does appear in the stylistic features of the second story, and this is apparent also in such extralinguistic markers as the absence of any laughter on MMK’s part. The one stylistic marker in ESG of which I was well aware before I investigated MMK’s eight narratives for this study was the handling of obvious loanwords from English.10 I knew from long experience that when ESG speakers were on their best linguistic behavior they tried to avoid both code-switches into English, especially intrasentential ones, and the use of a lot of obvious loanwords from English. With a long history of close contact with English, Scottish Gaelic quite generally shows the effects of the contact in the presence of many English borrowings. Both Gaelic speakers and English speakers notice the more obvious borrowings, and English speakers are often scornful of Gaelic heavily laced with obvious English vocabulary when they hear it spoken. In East Sutherland belittling remarks are sometimes made, for example, “I could speak that Gaelic myself!”, as if to deny the legitimacy of Gaelic as an independent language, and Gaelic speakers are self-conscious about the use of borrowings.11 When I first began to gather tape-recorded texts from each of the East Sutherland fisherfolk villages, it was quickly apparent that those speakers who had the lexical latitude to do so made an effort to minimize their reliance 10

11

Here and subsequently I consider only recognizability, to East Sutherland perceptions, of the English origins of the word. Whether or not the word was borrowed recently or some centuries ago is not relevant to the potential self-consciousness of the ESG speaker using the word. See the discussion in Dorian (1981: 100–2).

stylistic variation in a language restricted to private-sphere use

319

on loanwords from English. One literate male Golspie speaker replaced the nearly universal /phɔlisxən/ ‘policemen’ with /ɫuxkh ən ɫə/ ‘people of the law’, for example, when taping a narrative for me, and a male Embo speaker who needed to use the word /thramph/ ‘tramp’ for his narrative carefully prefaced it with /mər ə xanu aǰ/ ‘as they would say’. In view of all this I expected to see some change in MMK’s handling of loanwords as she moved from the sober, didactic telling of her opening story to more relaxed stories in the later session. In analyzing the use of loanwords in MMK’s eight narratives, I looked at the following features: (1) the sheer number of recognizably English loanwords relative to the total number of words in the story, (2) phrasal switches to English, (3) use of loanwords in markedly Gaelic or markedly English fashion. Within the third category I looked in particular at the application of the initial consonant mutations so characteristic of Gaelic12 to loanwords, in environments where a native Gaelic word would necessarily show mutation, and at the use of inflectional morphology with obviously English loanwords. Where inflectional morphology was concerned, the issue turned out to be whether an obviously English loanword was made plural by means of a Gaelic plural allomorph or an English plural allomorph, since only the plural proved to be variable; all other inflections used with loanwords were uniformly Gaelic. Switches to a wholly English phrase within the Gaelic narratives were also tallied, since it seemed likely that self-consciousness about dependence on English lexicon would extend a fortiori to phrasal switches. Table 17.1 shows MMK’s handling of obviously English material in her eight narratives. There is no indication in MMK’s narratives that she is trying to limit her use of obviously English loanwords as such, since all her narratives have them; nor that a low percentage of such loanwords correlates with the relatively low-key delivery of her first two narratives, since narratives 4 and 5 have a lower percentage than even the first, deeply serious, narrative. The handling of the English material is conservative in the first two narratives overall, however. There are 12

The initial consonants of nouns, verbs, and adjectives can be altered to show grammatical distinctions, as the sole sign of the category in some instances but in conjunction with suffixal morphology in others. In addition, mutations are sometimes an obligatory though grammatically nonsignificant feature of certain constructions. Some ESG examples: /maru a, xə̃ nʹax/ ‘Kill him, Kenneth!’ versus /varu a khə̃ nʹax/ ‘He killed Kenneth’: /hũnig mi ə pra:r/ ‘I saw her brother’ versus /hũnig mi ə vra:r/ ‘I saw his brother.’ The initial mutations are pervasive in all of the Celtic languages, and ESG is no exception. Some consonants are simply not susceptible to mutation, however, and of course some loanwords appear in environments which do not call for mutation, so that this criterion of loanword adaptation does not apply universally to the handling of every loanword which appears in the texts, despite the high frequency of initial mutations.

320

chapter 17

TABLE 17.1 The Handling of Obviously English Loanwords and Phrases in MMK’s Gaelic Narratives, First or Last Instances of Significant Features Indicated by Dividing Lines Loanwords as Percentage of All Words

1964 1 2 1968 3 4 5 6 7 8

3.4 4.7 5.3 2.8 3.2 7.4 3.8 6.5

No Mutation in Mutational Environment

0 of 4 0 of 2 5 of 13 – 0 of 8 0 of 6 2 of 10 0 of 8

English Plurals

Phrasal English Switches

Reversed Phrasal Switches

0 of 3 – 0 of 2 – – 1 of 3 1 of 7 –

0 1 1 0 1 2 4 0

– 1 1 – 0 0 0 –

no failures of mutation in environments where mutation is obligatory in ESG until the third narrative; English plural formations first appear later still, in the sixth and seventh narratives. One phrasal switch to English appears in the second narrative and at least one in all but two of the subsequent narratives, but the early narratives (including 3, in this case) are distinguished from the later narratives by the fact that MMK shows self-consciousness about the phrasal switches, in that she immediately attempts to reformulate the phrasal switch to English as a Gaelic equivalent. In 2 (Lost) she uses the expression Golspie holiday, pauses momentarily, and then says /hɔləde: kəi:špi/. She’s unable to come up with a Gaelic rendering of holiday, but she shifts to the Gaelic version of the place name and she reorders head and modifier to conform to the Gaelic word order rather than the English. In 3 (Bather) she finishes a Gaelic sentence with the phrase a few years, then follows it immediately with a precise Gaelic equivalent /anari vliənəxən/. There are two further signs of lexical self-consciousness in MMK’s two narratives from the earlier recording session. In the first narrative, she prefaces the deliberate use of an English borrowing with the Gaelic phrase /nə pũn tɔ̃ xatən/ ‘or I could say’. She has just described the breaking of glass into “a few pieces” in her narrative, but “a few pieces” was evidently not strong enough to suit her, to describe the fragmentation of the glass, and she adds “or I could say ‘smithereens’,” using the English noun.13 At the opening of the second story 13

The sibilant plural of this noun was not counted as an English plural inflection for purposes of Table 17.1; since the noun has no singular, there is no independent base to be

stylistic variation in a language restricted to private-sphere use

321

she uses two English-based loanwords, /khru/ ‘crew’ and /anti/ ‘aunt’, without self-consciousness, but stops immediately after uttering the third, /wikhεnd/, breaking off her narrative to say to me in English, “What would I say for weekend?” As I murmur that weekend seems fine to me, she carries straight on with her story, but it’s just after this that she uses the phrase Golspie holiday and then immediately offers a Gaelicized version of the phrase. She interrupts herself only once again, in these narratives or the later ones, with doubts about her use of an English word or phrase. In narrative 5, Class, where the acceptability of her Golspie Gaelic is the specific focus of the story, she uses the loanword /khwεsčən/ ‘question’ in starting to describe the teacher’s classroom request for the Gaelic version of an English sentence: /agəs hurd a rəm:əs ə gwεsčən – / ‘and he said to me the question, – ’. She then stops and repeats the offending noun in self-mockery, immediately reformulating the clause with the Gaelic verb ‘ask’. The borrowed noun was well integrated phonologically, with the initial mutation (nasalization) appropriate to a masculine noun of its phonological class after the definite article, but it is evidently not suitable for use in preparing to quote the arrogant Gaelic teacher, even though she uses English loanwords freely and unselfconsciously later in the same story. This instance of corrective self-consciousness in 5 (Class) is spontaneous, while the first narrative remains unique in its deliberate introduction of an English loanword, self-consciously framed with the Gaelic phrase ‘or I could say’. The handling of obviously English loanwords as an index of self-consciously careful style was the first differentiating feature which I looked for in the eight narratives, but it seemed to me that the narratives might also be expected to differ in terms of features which contributed to their relative liveliness as stories, if there was a stylistic continuum of any sort among them. As indicators of liveliness I tallied seven features: the number of simple direct quotes or direct-quote interchanges; the number of those which were in fact interchanges rather than simple direct quotes; the number of interchanges which consisted of more than the minimum two turns; the number of direct quotes which used /(h)ɔrs/, the more vivid of the two quotative past-tense verbs of ESG;14 the number of instances of doubled-up quotatives (akin to English “I sez, ‘_____’, sez I”); the number of strong interjections (i.e., interjections other than the routine ‘oh’, ‘well’, ‘ach’, or ‘och’); and the number of uses of ‘adventurous’

14

inflected. Smithereens is in origin an Irish loanword in English, but it has no Scottish Gaelic counterpart, nor does Scottish Gaelic use the same suffix (anglicized as -een) to form diminutives. MMK is the major user of /(h)ɔrs/ in all freely spoken tape-recorded speech in my corpus. Only one other speaker ever uses it at all in tape-recordings; tellingly, he uses it only in performing jokes.

322

chapter 17

Table 17.2 Incidence of ‘Liveliness Features’ in MMK’s Gaelic Narratives, First Instance or First Major Increase Indicated by Dividing Lines Simple Quotes and Interchanges of Interchanges Interchanges Three or More Turns /(h)ɔrs/ Quotatives # # # # 1964 1968

1 2 3 4 5 6 7 8

1 1 6 1 2 1 1 2

0 0 3 0 2 0 1 0

– – – – 1 – 1 –

0 0 11 1 4 0 7 0



1964 1 2 1968 3 4 5 6 7 8

Doubled-up Quotatives #

Strong Interjections #

‘Adventurous’ Language uses #

0 0­­ 1 0 1 0 0 0

0 0­­ 1 0 0 2 1 3

0 0 0 3 0 1 2 1

language (i.e., profanities and indelicate terms for body parts.)15 Table 17.2 presents the results of this tally. As Table 17.2 indicates, none of the liveliness features of MMK’s narratives except simple direct quotation appears in the first two of her stories (and only 15

MMK is by no means the most profane or indelicate ESG speaker I’ve heard, but she is more likely to use profane or indelicate language on tape than most. This probably reflects both her ease with tape-recording and her particular ease with me as her long-standing boarder. [See Chapter 19 in this volume for an excerpt from narrative 7 that includes two instances of “adventurous” language as that term is used here.]

stylistic variation in a language restricted to private-sphere use 323

one of those in each story), but four liveliness features occur for the first time as of the third story. That third story is certainly a very lively one, but even so it has only one strong interjection and no instances of indelicate or profane language. The fourth story, a very brief afterthought-anecdote connected with narrative 3 (Bather), is wholly indelicate in subject matter, and once she has told that story other ‘adventurous’ usages appear more freely. Of the subsequent stories only 5 (Class) is without them. The combined evidence of Tables 17.1 and 17.2 seems to indicate that MMK is indeed using a relatively restrained style when she begins the earlier taping session, which produces the first two of the eight narratives, and that she opens the later taping session with a more relaxed and casual style. Although MMK is still sufficiently self-conscious about English material, at the opening of the 1968 session, to pause and reformulate a phrasal switch to English back into Gaelic, the third narrative marks the first failures of consonant mutation with an obviously English loanword in environments where mutation would be obligatory in a native Gaelic word. Narrative 3 also provides the first interchanges, in the use of direct quotes, and the vivid quotative /(h)ɔrs/ not only appears, but appears in large numbers. The first instance of doubled-up quotative verbs likewise characterizes this narrative. She uses no ‘adventurous’ language, but she does for the first time use a strong interjection. Though all but two narratives after the third have phrasal switches to English, no narrative after that one shows the reversal of such a switch. Only 5 (Class) among the subsequent narratives has neither a strong interjection nor any ‘adventurous’ language, but 5 does have a number of lively features associated with direct quotation (interchanges, use of /(h)ɔrs/, and the other instance of doubled-up quotatives). Seven of the eight narratives involve MMK personally; only 4 (Sins) does not. She is a particularly central actor in the early part of narrative 3 (Bather), where she is the person who realizes that the bather is in trouble and makes the initial effort at rescue; and in the first half of 5 (Class), where she is the local Gaelic speaker who draws the teacher’s scorn by speaking the Golspie variety of Gaelic. But she is also an actor in five of the six others. (For example, she is present when the glass shatters in 1, and it is she who runs to buy eau de cologne as an antidote to the terrible odor from the second body when it eventually washes up on the shore; she is among the older children who are supposed to be looking after the little boy in 2 when he wanders off.) Narratives 3 and 5, in which MMK plays a very central role, are especially lively stories, but so is 7 (Coach), in which she plays a far less crucial role: she takes a dislike to the Gaelic coach and states her intention to urge the local Gaelic committee not to engage the coach for the Golspie choir’s tuition, but she is not a major figure in most of the story. Personal involvement motivates all of the narratives

324

chapter 17

except 4 (in subject matter an addendum to 3), but degree of involvement does not fully account for degree of liveliness in the narratives. It seems reasonable to look at the third narrative as the first in MMK’s more casual style, which she then maintains to the end of that second taping session (to which the eighth narrative, Record, provides the close). Strictly in terms of Tables 17.1 and 17.2, there is nothing to distinguish the first two narratives from one another except for the absence of any phrasal switches to English in the opening story. For the listener, all the same, there is one characteristic of the first narrative which is uniquely its own, a feature I would term an artistry of repetition which MMK shows nowhere else in this taping session (or for that matter in the second and third). It lends an almost Biblical flavor to her Gaelic here, especially since she keeps her lexicon and her syntax very simple. In setting the scene for the breaking of the glass, for example, she relates that she was sitting with elderly relatives of her mother’s: “Listening to their songs. And their singing.” She repeats the possessive here, making two phrases out of what could easily be one (i.e., “their songs and singing”). She does the same sort of thing several times again in speaking of the rescue attempt and the victims, after the capsizing of the fishing boat: “And my father was on one, and my uncle was on the other one” (instead of “on the other,” i.e., of the boats going out to try to rescue the men in the water); “And they caught the old man, but they didn’t catch his son”; “And the police came, and they all came”; “but one boot was off him and the other boot was on him.” In each case the italicized second instance of a word could have been avoided, and in more ordinary style probably would have been. The example involving the repetition of the verb ‘catch’ is especially notable, because that particular verb is regular, so that repeating it produces identical-sounding verb forms in rapid succession. The more common and less marked verb here would have been ‘get/find’. But ‘get/find’ is suppletive and in the preterite is /h/-initial; because initial /h-/ is unstable in this context and would have disappeared in the second clause, using ‘get/find’ would have produced less similar-sounding verbs in the positive preterite (first clause) and the negative preterite (second clause). One other phrasal oddity appears in this short story, one which resembles the other overly complete repetitions but is actually still more striking, when MMK says, “And the old man was drowned, and his son.” Gaelic is a verb-initial language, and the passive MMK uses to start the sentence would be inflected for third person plural possessive if she had meant “the old man and his son” to act as a compound subject of the passive.16 She inflected it for third person singu16

The passive in question is a partially nominal structure, requiring a possessive pronoun inflected for person and number, and in the third person singular also for gender.

stylistic variation in a language restricted to private-sphere use 325

lar masculine possessive, making it clear (as do the suprasegmentals) that “and his son” is intended to be a second but separate third person singular subject, agreeing with the third person singular masculine passive (or with an elliptical second such passive). The phraseology is unusual, and is the more effective for that reason. In none of the subsequent narratives does MMK use phrasal repetitions in this way. When she uses repetitions in these other stories, they are simply expansions and afterthoughts, as is the case in this example from narrative 7 (Coach), “MacLeod was in the army along with Bob. Along with my husband. He was in India along with my husband.” There are opportunities to use the same sort of repetition effect seen in narrative 1 again, but MMK does not take them. In 3, Bather, for example, a narrative which opens with a neartragedy resembling the tragic scene of the first narrative, she tells of pulling an almost drowned woman from the sea: “And her mouth was full of green foam. And white. And I caught her, and I was pulling her in, but she was too heavy, and I couldn’t get her in.” In this passage MMK does not repeat the noun ‘foam’, to produce “. . . of green foam. And white foam”; and instead of repeating the verb ‘pull’, as she perfectly well could here, she makes a switch from ‘pull’ to ‘get’ when speaking of retrieving the body from the sea. She is using language quite differently from the way she used it in the first narrative of the earlier session, although she is describing a similar scene. The handling of English loanwords and phrasal switches to English and the use of liveliness features constitute two sorts of dimensions along which MMK is able to vary the degree of formality versus casualness in her narratives. By observing the variations along those dimensions, the listener can distinguish between the two opening narratives in her first taping session, which are both relatively formal, and the six subsequent narratives, from the second taping session, which are more casual. By attending to the high degree of purposeful phrasal repetition in the first narrative, the listener can recognize in it a more crafted formality than in the second narrative, even though the second narrative, too, is less casual than those which follow.17 MMK is not the speaker whom I would have chosen if I had been setting out specifically to explore the fullest range of stylistic variation which an ESG speaker in Golspie village could muster. I would have opted instead for one of the three males with whom I worked in Golspie, since I would have expected them to control some of the lexicon and phraseology of religious usage in ESG, giving them an outer limit of formal style more extreme than MMK’s. Neither 17

Tannen (1987: 576) recognizes the repetition of a word, phrase, or longer syntactic unit as part of the “poetics of talk.”

326

chapter 17

would I have chosen to use two recording sessions with myself as interviewer/ audience as my test case, in exploring MMK’s stylistic range or anyone else’s. I would have tried to set up situations in which the individual was speaking to a variety of persons with whom s/he was on more intimate footing or less, in settings of greater or lesser familiarity, and to a variety of purposes. There is one advantage to having less than ideal material to work with, however: if variation in style is identifiable under such conditions, then it becomes that much the more certain that stylistic variation can reliably be claimed for this very “domestic” language which is already close to extinction. It is important to establish the presence of a moderately broad stylistic range in terminal ESG, because it indicates that natural languages, even when restricted by decline in functions and domains to the private sphere, and hence to speech situations which are on the whole informal ones by comparison with those characteristic of languages in vigorous use across an entire society, can still be spoken by the fully fluent in ways appropriate to their various stylistic needs. The evidence of MMK’s narratives, all drawn from two same-setting speech events, does not bear out the “monostylistic” label which Dressler applies to dying languages, for example, in a recent reference work: Terminal language decay seems to show a tendency towards monostylism. That is, recessive languages are more and more used in casual styles only, for example, those which are appropriate for intimate routine interactions at home. This stylistic change is yet another dysfunctional change in so far as the recessive language becomes inadequate for certain speech situations, domains, and functions. (Dressler 1988: 188–89) While it is true that a speaker like MMK uses her Gaelic almost entirely for “routine interactions at home,” it is not true that her Gaelic is “monostylistic.” She can vary it according to her intention in telling a story, so that a serious story which is told for purposes of impressing and convincing shows quite different stylistic features from a hilarious story which is told for its entertainment value above all. No doubt her Gaelic is “inadequate for certain speech situations, domains, and functions.” She probably could not, in Gaelic, pray aloud, welcome a guest speaker to the Women’s Rural Institute, or repeat the multiplication tables. These limitations have a great deal to do with MMK’s range of activities, however; I doubt that she could easily do any of these things in English, either, or would undertake to do so willingly. She would (and did, once I replaced the interviewer whose western dialect she couldn’t understand) do an interview for broadcast on the radio, and I’m certain that she would have had no difficulty whatever in lodging a protest in Gaelic about traffic safety for

stylistic variation in a language restricted to private-sphere use 327

children, or other such daily-life matters, with the village authorities, assuming only that she had a speaker of a Gaelic mutually intelligible with hers to lodge it with. That is, the limits of her Gaelic style range reflect to a considerable degree the limits of her range of activities and not any drastic poverty of her Gaelic, which is rich and fluent by local standards. One of my male sources in Golspie was highly political and quite prepared to engage in serious political discussions in Gaelic. He was not in late adulthood a churchgoer, but though I never heard him pray in Gaelic I suspect he was capable of it because of the religious upbringing which young males of his day experienced. He was an intelligent and resourceful speaker in both Gaelic and English, intellectually inclined despite limited education, and absolutely fearless in verbal interactions regardless of any differences in social status between him and his interlocutor. Listening to Gaelic speakers like this man, and listening to any group of ESG speakers moving from topic to topic and mood to mood, was more than sufficient to convince me that their Gaelic should be termed polystylistic rather than monostylistic: One can locate styles within [both the East Sutherland English and the East Sutherland Gaelic of the Brora, Golspie, and Embo bilinguals] which are appropriate to formality or informality, to vulgarity, to humor, to anger, and the like. Speakers differ in their ability to perform vividly in these styles, but they can certainly shift in the appropriate directions. (Dorian 1981: 85) I believe that it is a mistake to suppose that even a restriction to “routine intimate interactions at home” produces such radical loss of stylistic range in a language as to warrant the label “monostylistic,” at least among speakers who remain fluent. Without leaving the hearthside, fluent last speakers of a fading language are quite likely to speak about subjects as different as their grocery shopping and a recent bereavement, to mention two which I’ve heard women discussing among themselves. The speech style shifts markedly in the course of a casual conversation when a bereavement becomes the topic of discussion. If it does not go over altogether to the religious register of ESG, it certainly moves to the formal end of the range of styles in use among illiterate female speakers of the language. MMK was not literate in Gaelic, was not trained (as the boys of her generation often were) in the use of Gaelic religious language, and did not use either of her languages in the public sphere proper, since she did not take on roles that would have required that of her. Nonetheless her narrative Gaelic speech can be shown to display stylistic differences which are in keeping with

328

chapter 17

topic and purpose in her narrations and also with the increasing ease of the interview situation over two tape recording sessions. She is a testimony to the versatility of speaker and language both, late in the life of each. I am grateful to Edward Finegan for encouraging me, both in general and with specific prompts and questions, to persist in working through my imperfect corpus to locate evidence of stylistic variation, and to Edward Finegan and Douglas Biber for helpful comments on the original draft of this paper and for suggestions on the relevant literature.

References Briggs, Charles L. 1986. Learning how to ask: A sociolinguistic appraisal of the role of the interview in social science research. Cambridge: Cambridge University Press. Dorian, Nancy C. 1978. East Sutherland Gaelic: The dialect of the Brora, Golspie, and Embo fishing communities. Dublin: Dublin Institute for Advanced Studies. ———. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. Dressler, Wolfgang U. 1982. “Acceleration, retardation, and reversal in language decay?” In Robert L. Cooper, ed., Language spread. Bloomington: Indiana University Press. 321–36. ———. 1988. “Language death.” In Frederick J. Newmeyer, ed., Linguistics: The Cambridge survey. Vol. IV. Language: The socio-cultural context. Cambridge: Cambridge University Press. 184–92. Dressler, Wolfgang U., and Ruth Wodak-Leodolter. 1977. “Language preservation and language death in Brittany.” International Journal of the Sociology of Language 12: 33–44. Hymes, Dell. 1984. “Sociolinguistics: Stability and consolidation.” International Journal of the Sociology of Language 45: 39–45. Labov, William. 1970. The study of nonstandard English. Champaign, IL: National Council of Teachers of English. Obler, Loraine K., and Lise Menn, eds. 1982. Exceptional language and linguistics. New York: Academic Press. Tannen, Deborah. 1987. “Repetition in conversation: Toward a poetics of talk.” Language 63: 574–605.

chapter 18

Telling the Monolinguals from the Bilinguals: Unrealistic Code Choices in Direct Quotations within Scottish Gaelic Narratives In the latter half of the twentieth century the sole remaining bilingual population in East Sutherland, on the far northeast coast of mainland Scotland, consisted of former fisherfolk and their descendants in the villages of Brora, Golspie, and Embo. Gaelic-English bilingualism was nearing its end in this region, since no young people were acquiring Gaelic, and productive skills in Gaelic were restricted to the upper portion of the age spectrum. In spite of the recessive state of Gaelic in this locality, older speakers fluent in Gaelic continued to display well developed individual style differences in speech events that they engaged in with some frequency, such as joking, teasing, and telling a story. I propose to consider here one aspect of narrative style displayed by some high proficiency speakers of East Sutherland Gaelic, namely the use of direct quotation in story telling, and in particular code choice in direct quotation. In the speech communities to be looked at here, the first of these two code choices is influenced above all by interlocutor, provided that all parties to the conversation are bilingual (Dorian, 1981: 76–77). Each bilingual has certain conversation partners with whom she or he normally speaks the ancestral language (the East Sutherland variety of Scottish Gaelic, in this case) and others with whom English is the norm. For fully fluent bilinguals, Gaelic is typically used with many or most contemporaries (some once fluent and mostly younger bilinguals, both male and female, decline to use the Gaelic they were raised with) and with all or nearly all bilinguals older than themselves. With speakers younger than themselves, fully fluent bilinguals establish Gaelic or English as the habitual language of conversation on an individual basis. Gaelic is more likely to be chosen as the normal conversational medium with kinfolk not too much younger in age. Gaelic may also be the preferred conversational medium with other younger speakers frequently met with, for example younger members of families who live very near by or younger work-mates. Among less than fully proficient younger bilinguals (semi-speakers), Gaelic is used almost exclusively with certain kinfolk, typically a select few from the first- or secondascending generation. Politeness norms add certain constraints, in terms of fluent bilinguals’ code choices. If a monolingual English speaker is present (other than a young child

© koninklijke brill nv, leiden, ���4 | doi 10.1163/9789004261938_020

330

chapter 18

or a younger person with passive knowledge of Gaelic but little or no active knowledge), politeness norms require that English be used. If conversational overtures are made in one or the other language, politeness norms require that the addressee(s) respond in the same language. And if a clear-cut code switch is made during a conversation by one partner to it, politeness norms require that the other partner(s) make the same switch in responding. Of these three politeness constraints, the first is nearly always honored. The remaining two are honored by most bilinguals on most occasions, but they are more susceptible to being overridden than the first. The second code choice, that of the code in which direct quotations within a narrative are couched, is influenced above all by the first, among East Sutherland bilinguals. If the speaker has embarked on a verbal interaction in Gaelic, then Gaelic will be the ordinary (unmarked) code choice for quoted speech within the narrative as well. The degree to which this general principle holds and the nature of the factors which can override it are the topic of the present study. All of the speakers whose narratives are drawn on here are fully fluent bilinguals from the East Sutherland coastal villages of Golspie and Embo. The direct quotations used to exemplify their code-choice habits are drawn from narratives in which the language of the interaction had been established as East Sutherland Gaelic before the quotations in question appeared. The material considered includes single-speaker or multi-speaker interviews taperecorded in the two villages between 1963 and 1978,1 multi-speaker personal communications added to tapes containing elicited translations requested by me (dating from the 1960s, 1970s, and 1990s), two tapes of extensive reminiscences made for me as gifts by two speakers (a two-speaker tape from the 1960s and a solo tape by the surviving member of that pair in the 1990s), and transatlantic telephone conversations recorded in the 1990s with the permission of the fluent speaker on the other end. One fairly conspicuous difference between Golspie narratives and Embo narratives in respect to quoted speech arises from the large difference between 1 Many such “interviews” were more like story-telling sessions than the rubric suggests, and one or two were more like conversations among the fluent speakers present. In all but one case I was very well known on a long-term basis to all of the other speakers present (and in that one case to two of three others), so that the stiffness that might be expected in a singleencounter interview, the kind conducted by a stranger among strangers, did not appear. I retain the label “interview,” all the same, in recognition of the fact that tape-recording constrains at least potentially all speech events on which it intrudes, and of the fact that I retained the privilege of interposing questions during the course of the recording session and did so with greater or lesser frequency according to the flow of the session or to my need for background information.

telling the monolinguals from the bilinguals

331

the two villages in the number of available bilingual interlocutors. By midtwentieth century Gaelic-English bilingualism was receding swiftly in coastal East Sutherland and the eventual extinction of the local Gaelic dialect was foreseeable. Even in 1963–64, when I first began to record Gaelic narratives in East Sutherland, there were only 54 Gaelic-English bilinguals of local ancestry still remaining in the village of Golspie, in a total population of about 1,167.2 In the course of an ordinary day, a Golspie bilingual would of necessity interact with many English monolinguals, since nearly all shops and offices in the village were staffed by monolinguals and most people encountered on the village streets were likewise monolingual in English. Apart from this, almost none of the Golspie bilinguals, all of them over fifty-five years of age at that time, had children who spoke or understood the local Gaelic. Kin networks consequently included many younger-generation members with whom regular interactions were carried on exclusively in English. In Embo, by contrast, 105 people in a village population of about 275 were bilingual in 1963–64, including the great majority of the adults. The three small shops in the village were staffed by bilinguals, and Gaelic was the language most often heard on the streets. Apart from a rather modest number of incomers, only the youngest one-third or so of the village population spoke no Gaelic and even among some of the young nonspeakers passive skills ranging from fair to excellent were to be encountered. Regular interactions within kin networks were still carried on mostly in Gaelic. When Embo bilinguals produced a Gaelic narrative, contemporary settings as well as past settings offered plenty of incidents in which not only the narrator but all of the other participants as well were Gaelic speakers. For Golspie narrators only past settings offered such incidents, and any Gaelic narrative that reflected recent or current events was likely to involve some participants who were English monolinguals. Because of this difference in the social setting within which Gaelic was spoken, mid-century Gaelic narratives from Embo bilinguals contain considerably more direct quotations which are strictly realistic: Gaelic quotations which represent utterances originally spoken in Gaelic. Golspie narratives from the same period are more likely to include many linguistically unrealistic quotations, with utterances originally spoken in English rendered in Gaelic. Embo narratives from the 1990s however, with the number of remaining bilinguals even in Embo down to a mere handful, resemble the Golspie narratives more closely in that they increasingly contain many quotations of the linguistically unrealistic sort. 2 The census is taken by parish, and neither Golspie village nor Embo village is coterminous with a parish. The whole village population figures given here for Golspie and Embo are estimates given by courtesy of the Scottish Registry.

332

chapter 18

Performance features such as direct quotations, asides, repetitions, sound effects, and gestures, dramatize a story and structure it from the speaker’s point of view (Wolfson, 1978: 216). As Chafe notes in comparing oral and written styles, direct quotation is one expression of the speaker’s involvement with the narrative (1982: 46–48). Among the narratives within my corpus, one striking example of the link between direct quotation and the speaker’s involvement appears in the four pages that a strongly-felt 2,269-word narrative from a Golspie bilingual occupies when phonemically transcribed. The percentage of words-per-page rendered in the form of direct quotation rises from 17.3 on the first to 35.7 on the second and reaches 65.4 and 63.7 respectively on the last two. An interlocutor’s factual status as an English monolingual is by no means a reliable predictor of whether he or she will be quoted in English or not. The quote-heavy narrative just described included within it seven direct-quote conversational interchanges between the bilingual Golspie narrator and three monolingual members of her own kin network, her son, her daughter-in-law, and one of her nephews. All seven of these quoted interchanges appear in Gaelic within the narrative, her younger-generation family members’ utterances no less than her own (G2, 1970).3 Other speakers freely make this same unrealistic choice of codes for direct quotation. An Embo speaker who was acquainted with that same Golspie bilingual’s monolingual son reports a conversational exchange between himself and the son during an Embo narrative, again with the interchange quoted as if it had originally been in Gaelic (E27, 1993). An elderly Embo speaker who reproduced a younger cousin’s conversational remarks in Gaelic, asked later whether the cousin was a Gaelic speaker, replied that she wasn’t (E17, 1996). As logical as it might seem in the abstract for a bilingual to reflect in the conversational quotations embedded in his or her Gaelic narrative any change of code that comes about because of a change of interlocutor, that sort of realism is actually very rare in my corpus. Among copious instances of direct quotation drawn from the Gaelic narratives of Golspie and Embo bilinguals, there is in fact just one instance of a change of codes realistically introduced to represent a change of interlocutors. An octogenarian Embo speaker, telling of a disastrous year when sheep belonging to the local laird destroyed potato and oat crops in the nearby fields of several fisherfolk, quoted his mother’s instructions to him in Gaelic but his subsequent conversation with the laird’s daughter in English. These were fully realistic code choices, since Gaelic was the everyday language of Embo households at that time while by contrast 3 Speakers are coded by village and age. E stands for Embo. G for Golspie, and within each village, 1 represents the oldest speaker among those with whom I worked.

telling the monolinguals from the bilinguals

333

the laird was a well-to-do incomer and neither he nor any of his family spoke Gaelic. (1) hurd mə vã:r “šuɫ an o:rd kəs ə vεkh u said my mother go up until see you ‘My mother said, “Go up till you see

tə xanəs i.” wəl š e: . . . f_____ , š e n ĩrĩn´ what say-FUT she because COP [name] COP the daughter what she’ll say,” because it’s . . . F___ , it’s the daughter



ən uər šə̃ n š ε iš khĩãd as čəi ul – the time that COP she-EMPH looking after all – at that time, it’s she looking after every – everything.



ul – nəh ulə n´ĩçεn agəs all – the.pl all nothing and everything. And I went up

xa went

mĩš ən o:rd. I-EMPH up

kə f____s. hurd mi -- ri f_____ Wεl:, f____ hurd mĩš to [name]-POSSE said I to [name] well [name] said I-EMPH to F__’s. I said to F__ , “Wεl:, f____”, I said, al əur tha:this an ɔ:ts əs ithən wəð ðə šiph. “all our tatties and oats is eaten with the sheep.” al rəith s____ hurd i al gəv yu thu bušəls əf ɔ:ts said she “All right, S____ ,” she said, “I’ll give you two bushels of oats . . .”

(E4, 1970)4

Appropriate as a code switch of this kind may seem to the change of interloculors here, realistic reflection of the original is not a compelling consideration for most East Sutherland narrators when they include direct quotations in their Gaelic stories. The unusual change of codes in this story, rendering the conversational interchange between the narrator and his second (English-speaking) interlocutor, was surely influenced by just who that interlocutor was. In order to hold this particular conversation the narrator had to go up to what is still 4 See the list of abbreviations that follows the body of the paper for grammatical identifications.

334

chapter 18

known in the Highland countryside as an Taigh Mòr, “the Big House,” where the laird and his family resided, and he went essentially as a suppliant, since it was the laird who made available the allotments that the fisherfolk used for their potatoes and oats. Furthermore, Embo householders owned their houses but not the land the houses stood on; for the latter they owed feu duty to the laird, which they had to trek to the Big House to pay. In this region, subject for centuries to anglicized and anglicizing nobility and gentry, there were no Gaelic speakers to be found at the Big House, and both the place and its occupants were thoroughly identified with English in the minds of local bilinguals. Certainly this Embo narrator did not routinely quote all conversations with English monolinguals in English. In another narrative recorded in the same year he quotes the skipper of a Lowland herring fishing boat on which he had hired out as if the skipper had spoken in Gaelic, for example as the Lowlander assuredly had not. Because conversational speech to and from monolinguals is in the usual case freely rendered as if it had been spoken in Gaelic, during the presentation of Gaelic narratives, it is often impossible to tell on the basis of anything in the Gaelic of the narrative itself whether a speaker being quoted is or is not actually a speaker of Gaelic. In favorable cases there may be a clue in the storyline or in the setting, even though the quotation itself gives no clue, but in others there are no grounds for a firm conclusion. The same elderly Embo bilingual quoted above tells in another reminiscence of someone coming to the door of his home with a telegram: (2) hã:nig tə̃ n´ kəs ə dε hurd a a thεligram ə šɔ. came man to the house said he is telegrəm here A man came to the house, he said “There’s a telegram here k iari ǰeǰ u vã:n [ə] wikh wəl ha aǰ . . .  at wanting will.go you down [to] Wick because are they wanting you to go down to Wick. Because they’re širu fairmən stã:n anə wikh. mə heǰ u vã:n hurd a seek-GER fireman down in Wick if will.go you down said he looking for a fireman down in Wick. If you’ll go down,” he said. hãnig fis šɔ er tə hɔn came word this for your sake “This word came for you.” ’

(E4, 1970)

telling the monolinguals from the bilinguals

335

The conversation continues, in the narrative, both parties represented as speaking in Gaelic. The telegram was certain to have been in English, since it had come via official channels and there was on the east coast of Scotland no official use of Gaelic whatsoever at this time (shortly before World War I), nor could the recipient (the bilingual narrator) be assumed to be able to read a message in Gaelic, since many Embo bilinguals were only passively literate in Bible Gaelic, if that. If the telegram was received at the sub-post office in Embo, it’s possible that someone sent to the speaker’s house to deliver the telegram told him its contents in Gaelic. If on the other hand the telegram was received in the much larger and entirely English-speaking village of Dornoch to which Embo is a fishing-village satellite, and delivered by a telegraphic or postal employee from Dornoch, there is virtually no possibility that this discussion of the telegram and of the reply to be sent took place in Gaelic. From the narrative itself there is no way of telling. A good many direct quotations are indeterminate in this fashion, and having learned early on that interrupting the narrative to ask for clarification was deleterious to the narrative flow and might even produce a code switch on the narrator’s part, I seldom did so. In an occasional case there may be indications within the narrative that the quoted speech either is or is not in the language of the original. In one unusual case, for example, a Golspie narrator made an attempt to represent in her quotation the Gaelic dialect of her conversational partner, identified as that woman’s “own” (i.e. not local): (3) hurd mi hĩãn ri pε̃n m____ , te: [ən] danəs a ǰe: šə̃ n said I self to Mrs [name] what the.devil is that.one that ‘I said myself to Mrs. M__, “What the devil is that one širu šɔ agəs ɔrs iš, o an al ə – an al – seek-GER here and said she-EMPH oh not is uh not is wanting here?” And she said, “Oh, [I] don’t – don’t – an al fis akəm,” nε – . . . ə γa:likh εkh hẽn.” not is knowledge at.me or . . . the Gaelic at.her self I don’t know, or – . . . her own Gaelic.’

(G2, 1968)

The quotation’s form indicates that the other woman in fact produced the phrase “I don’t know” in Gaelic, since the final two words of the Gaelic phrase

336

chapter 18

/an al fis akəm/ are rendered in nonlocal dialect forms (though not the first two, which appear in strictly local forms).5 In several other cases where the narrator makes a still more explicit statement about the language used in the original, identifying it as English, the statement constrains the language selected for the quotation in the bilingual’s Gaelic narrative only briefly, or not at all. The Golspie speaker quoted in (3) above told with some indignation on another occasion of meeting two fisherfolk descendants in a shop, neither of whom used Gaelic, and of challenging them about their choice of language: (4) s γ ə̃ i:̃ niç mi riš ə ǰe:əs and PRET ask I to the that.one-EMPH ‘And I asked that one, [ən] də xəi:l´ u γa:likh [INTERROG] PRET lose you the.Gaelic “Did you lose the Gaelic nεr vel a – vel a ad when is it is it at.you when it’s at you [i.e., when you have it]?” agəs hurd i o: hurd i as ə viərl, o s ãĩn´ tɔ̃ and said she oh said she in the English o COP knowledge to.me “Oh,” she said in the English. “Oh, I know phols thiki mi a . . . plenty understand-FUT I it plenty, I’ll understand it.” ’

(G2, 1964)

5 Macaulay (1991: 183–191) distinguishes between mimics and translators, among his Scotsspeaking narrators who use direct quotation. The latter reproduce in their own Scots speech remarks originally produced in other forms of English, where that suits the tone or purpose of the narrative, whereas the former enjoy mimicking other forms of English when quoting non-Scots speakers. In these terms the East Sutherland Gaelic speakers who served as my sources are nearly all translators, when telling a story in Gaelic, and even the efforts of E4 and G2 to reproduce the language of the original are not very successful. The laird’s daughter is quoted in English, to be sure, but her English shows the same strong Gaelic accent with which E4 spoke that language, and Mrs. M.’s Gaelic remark in G2’s rendering appears with only two words out of four in nonlocal form.

telling the monolinguals from the bilinguals

337

An Embo narrator makes a similarly explicit statement about the monolingual status of a particular speaker, in an extensive set of reminiscences she and her husband recorded for my benefit (but without my being present), and she code-switches realistically to English in her first direct quotation from the man in question, a farmer from whose nearby farm some Embo children had been busy stealing apples. But within two sentences of this realistic code switch the same man appears conducting an apparently Gaelic conversation with the narrator’s father: (5) . . . nə va a reǰ mi ax e a____ k____ [a] va when was not believe I that.not COP [name] [REL] was ‘. . . when it was I think A____ G____ [who] was

fɔs khĩn´ ə xru: hã:nig c____ agəs š e piərl above the tree came [name] and COP English up the tree, G____came, and it’s English



a h igəs . . . va a k eax that is at.him . . . was he at shouting that he has, he was shouting



hu:s stilin mai a:pəls, hu:s stilin mai a:pəls!” “Who’s stealing my apples, who’s stealing my apples!”



ha:r šĩn´ εs agəs va a n uər šə̃ n čĩãn ə val fled we off and was he the time that coming to.the village We ran off, and at that time he was coming to the village

l´ε khɔ:rn agəs yax khrεkh mətha:th er či sɔ:rn. with cart and horse selling potatoes on Saturday with a cart and horse, selling potatoes on Saturday.

ãnig a gəs a dε ãn´ agəs hurd a rε m a:r came he to the house at.us and said he to my father He came to our house, and he said to my father [in Gaelic],

va n´ ĩrĩn´ adəs was the daughter at.you-EMPH “Your daughter was

338

chapter 18

kəǰ . . . nə u:ɫən aməs er či s ɔ:rn. . . . stealing the apple-PL at.me-EMPH on Saturday . . . stealing . . . my apples on Saturday. . . .” ’

(E33, 1967)

The conversation between the farmer and the speaker’s father continues, in her narrative, rendered entirely in Gaelic.6 There are a few cases in which the narrator’s conversational partner is quoted so consistently and extensively in Gaelic that I actually began to wonder whether, without my having realized it, an interlocutor known to me purely as an English speaker was just possibly a Gaelic speaker after all, perhaps originally an incomer from a more westerly area where Gaelic was more widely spoken.7 In one such case four direct-quotation conversational interchanges between the bilingual narrator and a woman whom I had always believed to be an English monolingual had already occurred, with both the narrator and her conversation partner represented as speaking in Gaelic, before a direct quotation in English suddenly appeared as a conversational response from the putative monolingual to a direct quotation in Gaelic from the bilingual narrator: (6) hurd mĩš ax khom ɔ____. a – a – a – mi du:l said I-EMPH but indifferent [name] am am am I in.the.hope ‘I said, “But never mind, O____. I -- I -- I hope kə də haxkh aǰ er hulə n´i γ o:ɫ aǰ εs that PRET choke they on every thing PRET drink they from.it that they choked on everything they drank out of it.” 6 The short initial shout of outrage from the English-speaking farmer, with its doubling for narrative effect, may appear here in English because it had become a stereotyped piece of mockery used by the Embo children to plague this particular farmer. There was at any rate a certain repertoire of verbal mockery, regularly shouted out most unkindly by the Embo children after certain individuals, for example, a mocking rhyme called after the man who came to the village selling meat from his cart and a frequently made mistake in her Gaelic called after a female incomer who was trying to acquire Gaelic. 7 Because the number of local Gaelic speakers was so small, and also because local Gaelic speakers are referred to by means of a distinctive set of by-names, it was possible to rule out the woman in question as a previously overlooked member of the local Gaelic speech community. In this case I also checked subsequently with others and made certain that the conversational partner here was, as I had supposed, an English monolingual.

telling the monolinguals from the bilinguals

339

agəs . . . va ɔ____ ka:r “ɔx, wεl:, ə s a da:m še:m,’’ and . . . was [name] laughing And . . . O____ was laughing: “Och, well, it’s a damn shame,” xanu ɔ____. say-COND [name].’ O____ would say.’

(G2, 1970)

The reason for the sudden realism of this particular quotation (followed by two additional English sentences from O. in continuation of the same exchange), within a narrative in which conversation to and from O. appears in Gaelic both before and after this interchange, is straightforward enough in the local context. There is no colloquial equivalent locally for the English word damn, and female bilinguals are in any case more inclined to reproduce English profanity in English than they are to reproduce any other quoted English material in the original. An Embo woman, for example, telling in Gaelic of a drunken supervisor who included a swearword in ordering his workers to leave shelter to perform an outdoor job in vile weather, similarly quoted his profane remark in English (E33, 1964). The same narrator who quotes O.’s profanity in English here was variable in quoting her own profanities, giving them twice in Gaelic (once realistically, since her interlocutor was also a Gaelic speaker [see (3) above]; but once unrealistically, with the interlocutor her own monolingual son). On a third occasion she began realistically in English, representing a dialog with an English monolingual, and then switched in mid-quote to Gaelic, in keeping with the language of the rest of this long story: (7) agəs ɔrš mĩš hwεr ðə dεvəl – khačh an danəs and said I-EMPH – where the devil ‘and I said, “where the devil – where the dεvil ən d uər u bokh: am INTERROG PRET got you the.bag at.me did you get my bag?” ’

(G2, 1970)

Just as profanity can trigger a switch to a more realistic code choice in rendering a conversational original, so can other remarks the negative or positive force

340

chapter 18

of which the narrator especially wishes to convey. This same Golspie Gaelic speaker quoted in English the cutting English remark of a Gaelic teacher who rejected her local Golspie version of a Gaelic phrase he wanted translated (“That’s not in the language!”; G2,1968), and a nonagenarian Embo speaker reproduced in English the callous remark (from the point of view of one who had previously dealt only with doctors who made house calls as a matter of course) of a new doctor who asked instead that she find someone to bring her to his office (“I have a very busy surgery”; E17, 1996). In the same fashion the latter speaker quoted in English, with obvious disgust, the memorably unpleasant abuse hurled at an unfortunate cat by its monolingual owner (E17, 1994). In a happier connection this same elderly Embo speaker switched into English to reproduce a remark that had resonated exceptionally positively with her. She was deeply gratified to find that a journey she had made at some physical cost to herself to visit a younger but terminally ill relative in the Lowlands was as cheering to the ill woman as the speaker had hoped it would be. She quoted from a phone conversation between herself and the relative after the visit, rendering their conversation as if it had been in Gaelic until she came to the remark that had been so particularly gratifying to her, at which point she switched to the English in which the conversation had actually taken place: (8) hurd i ax – tε – te: ǰaxkan ə va am, said she but – what – what the.week REL was at.me ‘She said, “But – what – what a week I had l´ε phĩãn ax hurd i a reǰ mi tənə xuarɫ mi with pain but said she not believe I when heard I with pain. But,” she said, “I believe when I heard kə d rɔ ši čhĩãn rə̃ iñ ´ šə̃ n na šɔ:r mi . . . that PRET were you.pl coming made that COMP better me that you were coming, that made me better.” hurd i ai gɔth ə bu:s[th] hurd i said she said she She said, “I got a boost,” she said.’

(E17, 1994)

telling the monolinguals from the bilinguals

341

Occasionally, even without any code-switched quotations to signal the language of the original, there may be other evidence of an English original that’s being reproduced in Gaelic. The Golspie narrator represented in (3), (4), (6) and (7) above is a conservative older speaker who normally applies the vocative case to personal names used in direct address, whether the personal name is of English or of Celtic origin and whether the individual who bears the name is a bilingual or an English monolingual. In the Gaelic narratives she tape-recorded for me over six or seven years, there is just one case in which she produces a quotation with a personal name in direct address but without the required marking for vocative. This one failure to mark for the vocative in quoted speech is unusual, certainly, but it would have been much more surprising if it had occurred in quoted address to a Gaelic-speaking interlocutor. Instead it appears in a quotation in which she’s addressing her Gaelic to a monolingual English speaker with the distinctively English name of Cecil (G2, 1970). Less subtle evidence of an English original is sometimes present in the form of an unusually high number of short code-switched phrases and prominent loanwords: (9) s γa č____ ə stε əs hurd i wεl š e ə̃ n tha:rštən and went [name] in and said she well COP one tartan ‘And J____went in, and she said, “Well, it’s a tartan one ə v ãũ:n ə brau:n, an ə fɔ:n that was in.it it was. A brown, and a fawn.” agəs – ɔrš nə phɔlisxən wεl an al tharthən ba:g and said the police-PL well not is And – the police said, “Well there’s no tartan bag iǰər anə šɔ at all in here here at all.” ’

(G2, 1970)

A few remarks made originally in English are later quoted realistically in English within a Gaelic narrative, because the narrator is calling attention to

342

chapter 18

the sometimes faulty English typical at one time of East Sutherland bilinguals, whose stronger language was Gaelic: (10) ax e iš ə hurd ən uər šən NEG INTERROG COP she-EMPH REL said the time that” ‘Isn’t it she that said that time, wəl yu go fǝr w____s bɔks nə va aǰ k iari tə̃ n´ when were they at wanting man “Will you go for W___’s box?” When they were wanting a man

xuɫ ə vã:n thor ə çišč ə ɫɔi . . . as ə viərl hurd i going down taking the box of lying in the English said she to go down to get the coffin? . . . In the English, she said.

wəl yu go fər w_____s bɔks a rɔ fis εkh not was knowledge at.her “Will you go for W____’s box?” She didn’t know tε xãnu i riš what say-COND she to.it what she would say for it [i.e., for /khiščh ǝ ɫɔi/ in English].’

(E27, 1967)

The number of unrealistic direct quotations, rendered in Gaelic within Gaelic narratives when they were either originally addressed to, or originally said by English monolinguals, is very large. A few among these quotations are unrealistic to a particularly radical degree, in that the very quotation reproduced in Gaelic bears on the absence of a knowledge of Gaelic in the person or persons being quoted. An Embo bilingual, explaining the force of interlocutor – governed code-choices in his conversational patterns, told of attracting unwanted attention in the office where he worked because he was incapable of speaking English to a particular man with whom Gaelic was his accustomed code choice: (11) . . . tə va a_____ k_____ ha a bru:ra ig ə hai:dro . . . when was [name] is he in.Brora at the Hydro . . . when A_C_ , he’s in Brora, at the Hydro[electric Board].

telling the monolinguals from the bilinguals

343

a hai:dro fonig[u] ə skɔl ha kə mã thrikh is Hydro phone-GER the school is ADV good often The Hydro’s phoning the school, often enough. mə higu a_____ er ə fo:n š e xa:likh ə gõ:ni if come-cond [name] on the phone COP the.Gaelic always If A____ would come on the phone, it’s always the Gaelic. viu hulə bɔdi khĩãd te ha – be-COND every body looking what is – Everybody would be looking, “What’s – te ha fεr šə̃ n ra:? What is the.fellow that saying What’s that fellow saying?” an ũrn tõ:š ə vrĩ:n ri [a____ k____] not able to.me-EMPH COMPL speak to [name] I cant speak to [A__ C__] s ə viərl. a čirax xan ũrn dɔ̃ ye:nu a in the English. is just not able to.me do-GER it in the English. It’s just, I can’t do it.’

(E29, 1974)

The point of this story is that the narrator feels compelled to use a language that no one else in his office knows when he speaks on the phone to someone who is a habitual Gaelic conversational partner of his. But when he quotes the remark made by his fellow-workers about the unintelligibility of his phone conversation, he nonetheless quotes them as if they had made the remark in Gaelic. This same speaker makes a similarly odd-seeming code choice in quoting his own monolingual daughter’s frequent requests, at high-school age, that he teach her Gaelic: (12) ɔx, š ĩmu- š ĩmu uər xãnu i rum och COP many.a COP many.a time say-COND she to.me ‘Och, many – many a time she would say to me,

344

chapter 18

tə va i s ə skɔl yə no: khɔ̃ rn ax when was she in the school NEG INTERROG when she was in school, y’know,” Why don’t yũ:siç u xa:likh tɔ̃ vel is ad? teach you the.Gaelic to.me is knowledge at.you you teach me the Gaelic?,” you know?

(E29, 1974)

Rendering in Gaelic a person’s request to be taught Gaelic represents something of an extreme in unrealistic code choice for direct quotation, it would seem. But the code choice for the recording session was firmly established, not just by the fact that I was asking questions and supplying prompts in Gaelic, but also, and probably much more importantly, by the fact that two other Embo bilinguals were present and had taken an especially active part in the conversation at the beginning of the session. Since they were regular Gaelic conversational partners for the speaker, their participation fixed the language of the interaction as Gaelic for this strongly interlocutor-governed narrator. The force of that original choice apparently kept him speaking Gaelic even when the use of Gaelic in direct quotation made very poor sense in the context of what he was relating. It seems clear that the choice of code for direct quotation within a Gaelic narrative is secondary, among the bilinguals of the East Sutherland fisherfolk communities, to the original choice of code for the narrative itself. Once the language of a particular verbal interaction is established as Gaelic, the narrators are more likely than not to reproduce conversational remarks originally made in English as though they had been made in Gaelic. There may be a hint of the underlying unreality of this sort of counterfactual directquote code choice, either in the form of material that is grammatically out of line with the speaker’s normal Gaelic usage, or in the form of unusually prominent phrasal code switches to English within the quotes. In addition, the choice of Gaelic for quotations within a Gaelic narrative may be reversed temporarily, with the narrative continuing afterwards in Gaelic, in cases where the narrator particularly wishes to convey the negative or positive force of an English original. By and large, however, direct quotation within a Gaelic narrative is likely to be rendered in Gaelic, regardless of whether that was the language used when the quoted remark was originally said and even regardless of whether the person speaking or being spoken to was factually capable of producing or understanding an utterance in Gaelic.

telling the monolinguals from the bilinguals

345

In some bilingual communities the norm clearly differs, while in others pragmatic or stylistic factors may operate selectively, ruling out an established quotational norm. Penelope Gardner-Chloros reports in a study of community usage that “quotations of people’s speech are usually in the original language used,” among French-Alsatian bilinguals in Strasbourg (1991: 176). Other researchers offer examples of similar adherence to the language of the original (Romaine 1989: 148), examples both of adherence to the language of the original and of translation into another community language (Gumperz 1982: 76, 82), and even an example of exact reversal of the language of the original in the quotation, though this last instance comes from a child bilingual (Boeschoten & Verhoeven, quoted in Romaine 1989: 148). In East Sutherland the general preference for quotation in Gaelic within a Gaelic narrative probably reflects the primacy of the interlocutor’s identity when Gaelic is selected as the language of the narrative in the first place. Gaelic is a stigmatized code choice in East Sutherland, in part because of the centuries-long social and economic dominance of English within the country and the region, and in more recent times because of this Gaelic dialect’s association with the fisherfolk as a locally stigmatized social group (Dorian 1981: 61–68). When East Sutherland Gaelic is selected as the conversational medium, a degree of social solidarity is immediately established by the speakers’ shared use of the stigmatized code. Possibly the parties to such a conversation are more likely to maintain their solidary language throughout (allowing for a certain number of the phrasal code-switches typical of this dialect) than would be the case in bilingual communities generally. There are a few particular individuals known for an unusual degree of conversational code-switching (not just in quotation but generally) and criticized for it, but otherwise the practices described here represent a narrative norm. As a result of the preference for quoting in Gaelic when speaking in Gaelic, a listener who understood the language but lacked local knowledge would frequently be unable to tell the monolinguals from the bilinguals on the basis of their quoted speech in a bilingual’s Gaelic narrative. Even at the present time a listener who knows Gaelic, hearing a local speaker report an encounter in which any number of people have apparently been addressed in East Sutherland Gaelic and have replied in it, might easily come to the conclusion that bilinguals are still plentiful in East Sutherland close to the end of the twentieth century. Sad to say, this is not the actual case. Only two of the six highly proficient bilingual speakers from whose narratives the quotations discussed here were drawn are alive at this writing, and the dialect will be carried into the twenty-first century chiefly by imperfect speakers. Fond though some of the imperfect speakers are of Gaelic, they will be unlikely to embark

346

chapter 18

on many Gaelic narratives, not only because of the limitations imposed by their own proficiency levels, but also because there will be ever fewer listeners able to appreciate a tale told in the local East Sutherland Gaelic. The loss is ours as well as theirs. Abbreviations ADV adverbial marker COMP comparative marker COMPL complement particle COND conditional suffix COP copula EMPH emphatic suffix FUT future suffix

GER gerund suffix INTERROG interrogative particle NEG INTERROG negative interrogative particle PL plural suffix POSSE English possessive suffix PRET preterite particle REL relative particle

Acknowledgements I’m indebted to Susanne Romaine and Robin Sabino for helpful criticisms of the original draft of this paper. Responsibility for remaining shortcomings is my own. References Boeschoten, H., & Verhoeven, L. (1985). Integration niederländischer lexicalischer Elemente ins Turkische: Sprachmischung bei Immigranten der ersten und zweiten Generation. Linguistische Berichte, 98, 347–364. Chafe, W. (1982). Integration and involvement in speaking, writing, and oral literature. In Deborah Tannen (Ed.), Spoken and written language: Exploring orality and literacy (pp. 35–53). Norwood, New Jersey: Ablex. Dorian, N. C. (1981). Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: The University of Pennsylvania Press. Gardner-Chloros, P. (1991). Language selection and switching in Strasbourg. Oxford: Oxford University Press. Gumperz, J. J. (1982). Discourse strategies. Cambridge U.K.: Cambridge University Press. Macaulay, R. K. S. (1991). Locating dialect in discourse. New York: Oxford University Press. Romaine, S. (1989). Bilingualism. Oxford: Basil Blackwell. Wolfson, N. (1978). A feature of performed narrative: The conversational historical present. Language in Society, 7, 215–237.

chapter 19

Celebrations: In Praise of the Particular Voices of Languages at Risk Introduction Reading an impressive political novel in German years ago, I was struck by how effectively the author, Joseph Breitbach, made use throughout the entire book of a particular grammatical device that the German language offers its speakers and writers, and also by how impossible it would be to create quite the same effect in English, which lacks a comparable grammatical device. The novel, Bericht über Bruno (‘Report on Bruno’, Breitbach 1964), deals with the career of a malevolent politician (the Bruno of the title). Breitbach uses ordinary indicative verb forms to render the point of view and experience of the first-person narrator, but he uses the so-called subjunctive of indirect discourse to report what Bruno and the other figures in the novel have to say. That particular subjunctive indicates that second-hand report or inference forms the basis of whatever is expressed in it, not first-hand knowledge. It has evidential properties that create a distancing effect in discourse (largely in written discourse, since it is scarcely used in spoken German except in registers that intentionally mimic the style of the written language). It can also suggest doubt about the validity or veracity of whatever is expressed in it. An employee who responds to the boss’s comment on a fellow-worker’s absence by saying “Er ist krank” (‘He’s sick’) implies that he knows of a presumably valid reason for the absence. If he should say “Er sei krank” instead, with the subjunctive of indirect discourse, the effect would be more on the order of ‘He’s supposedly/ reportedly sick’. In that case the speaker would take no responsibility for the validity of the reason offered for the absence, and the selection of the subjunctive of indirect discourse could even suggest that the speaker intended to cast doubt on the statement. In the absence of a parallel in English to the German subjunctive of indirect discourse, an author writing in English would have to take a much more circuitous route to achieve an effect at all similar to the one Joseph Breitbach had ready to hand in his skillfully wrought novel. Adverbs such as purportedly, seemingly, apparently, evidently, and supposedly would probably appear with considerable frequency, and the first-person narrator would need to resort to phrases such as I took that to indicate, one could assume, there seemed to be, © koninklijke brill nv, leiden, ���4 | doi 10.1163/9789004261938_021

348

chapter 19

and that seemed to suggest, when reporting on other figures in the novel. Some of the same implications could be introduced in that way, but such turns of phrase would quickly become repetitious in a way that ongoing use of a single verbal tense-and-mood choice does not. A different tone would result, and a stylistic economy that contributes to the novel’s power in German would certainly be lost. Who would read Breitbach in English (or Dostoyevsky, or Ibsen, or Tagore in English) if s/he had the requisite language skills to read the work in the original language? Languages have their individual voices, created equally by the means which they deploy and by the details of the deployment. As George Steiner says: “Each human language maps the world differently. . . . Each tongue . . . construes a set of possible worlds and geographies of remembrance” (1992: xiv). Semantic mappings across the lexicons of any two languages are often so obviously different that the rankest beginner confronts them immediately and sharply; more subtle differences continue to emerge and confront the second-language learner for years, as familiarity increases. Linguists have the preoccupations of their special field, naturally enough. The properties of a language that fascinate and please them are those that are unusual from a specialist’s point of view: a prominent role for one of the grammatical devices less widely encountered among the world’s languages, or a region’s (infixes, say); the appearance of a syntactic property that had been thought not to exist (languages with both classifiers and gender as separate categories; see Aikhenvald 2003); the occurrence of one phenomenon without another phenomenon believed to be routinely co-occurrent with it (massive grammatical restructuring despite near-absence of lexical borrowing from the language that provides the model for the restructuring; see Aikhenvald 1996). All of these things are of understandably high interest to anyone who studies linguistic properties, but dear as they are to the linguist they are no more likely than much commoner features to form the basis of the very particular effects that native speakers are able to achieve with their languages. A grammatical feature unfamiliar to speakers of English, Spanish, or other Western European languages, such as the partial-reduplication prefixation that expresses intensification in Turkish (beyaz ‘white’, bembeyaz ‘extremely white’; yalnĭz ‘alone’, yapyalnĭz ‘absolutely alone’; etc.), attracts the instant attention of an English speaker who learns Turkish as a foreign language. But the same English speaker who’s struck by intensification prefixation in Turkish may never have noticed in his or her own speech the subjunctive marked by absence of the present-tense indicative 3rd-person suffix -s in the finite verb of the subordinate clause in sentences such as I suggest she try again, I’d prefer that he not go. Sophisticated native English speakers, much given to lamenting the disappearance of the

in praise of the particular voices of languages at risk

349

subjunctive as a grammatically marked category in English, commonly overlook the faithfulness with which they and their peers produce this subjunctive, despite the fact that some speakers deploy just this device to achieve overtones of sarcasm, dry humor, and so forth (“Smith says he can’t find the folder.” “I’d suggest he look again – in the appropriate filing cabinet this time.”). Many linguists are used to arguing, in talking to non-linguists or in introducing the subject matter of Linguistics to students, that each language is unique and has highly individual features that lend it its inimitable and irreplaceable character. Yet as a profession we’ve done very little to date to demonstrate that this is so or to celebrate the particular voices of individual languages. To make good our claims of individually unique expressive capacities for each human language, we need to follow the examples set by Tony Woodbury and Marianne Mithun, who discuss the way in which special properties give two threatened languages their particular expressive flavor: the affective affixes in Yup’ik (Woodbury 1998), and the evidentials and certain other grammatical devices in Mohawk (Mithun 1998). Discussions such as these, acknowledging and celebrating the unique voices of endangered languages, make their point best when they’re mounted in terms of the language(s) that actually threaten to supplant a language poised on the brink of shift (or already over the brink). So, for example, even if the K’emant language of Ethiopia has various properties that seem interestingly exotic from the point of view of English or Spanish (and of English- or Spanish-speaking linguists), the relevant question is not whether K’emant speakers can create semantic or discourse effects that English or Spanish speakers cannot, but whether Amharic, the Semitic language to which K’emant speakers are rapidly shifting, does or does not have properties that offer expressive parallels for the semantic distinctions and the discourse effects that K’emant (a Cushitic language) makes possible (Leyew 1997; see also Leyew 2003). That is, the focus needs to be on what K’emantAmharic bilinguals lose in expressive capacity if they stop speaking K’emant, and on what expressive capacity their descendants never gain if K’emant is not transmitted to them. If speakers of a language that offers an unusually large number of infixes deploys those infixes to express grammatical or semantic categories equally well expressed by prefixes or suffixes in various other languages that they and others around them speak, then they have no particular expressive advantage from the infixing property of their language. If on the other hand the infixes express concepts, make distinctions, or create discourse effects unavailable in any other languages spoken by mother-tongue speakers of the high-infix language, then no matter how multilingual they may be, they lose some expressive capacity if they cease to use their mother tongue.

350

chapter 19

Unless speakers of the high-infix language are lucky enough to be exceptionally well schooled in their ancestral language, or to be unusually observant and reflective about the structural differences between two or more languages they speak fluently, they are not likely to be consciously aware of the losses in expressive capacity that they would face if they stopped using their ancestral language and switched over entirely to an expanding language. This is no discredit to them at all. Most speakers of any language have a hard time becoming aware of some of the most distinctive features of their own language and discovering just how those features work. (See the example of the zero-marked subjunctive in English, above.) Linguists often come to learn about such features and their use in the course of professional training, of course, but relatively few of us were fully aware of them beforehand. All too often, the history of small languages is such that their speakers have been afforded very few chances to recognize, much less to revel in, the special expressive capacities of their ancestral languages. Many local languages have a small population base, and most have much less prestige than some other language spoken in the same region. Schooling is often available only in a higherprestige, wider-currency language, and under those circumstances it can be still harder than in the wide-currency languages themselves to recognize and appreciate the uniquely expressive resources that an ancestral language offers. The material used to create some notable effect may be quite unremarkable in itself (a change in word order, the use of a suffix or prefix), but since the effect created has no match in local speakers’ other language(s), only the resources of their heritage language offer them the opportunity to create the expressive effect in question. Whether native speakers are fully aware of the uniquely expressive features of their ancestral language or not, most of them have the ability to make very effective use of them when they argue, tease, scold, joke, or tell stories, skilled native speakers that they are. Examples for one endangered language follow.

Expressive Bleaching in the Shift from Scottish Gaelic to English in East Sutherland, Scotland

I would like here to look at East Sutherland Gaelic (ESG) from the point of view of expressive uniqueness, highlighting an expressive resource that ESG speakers, all Gaelic-English bilinguals, have available to them in their Gaelic but not in their English. The Gaelic dialect at issue here is a variety spoken in the second half of the twentieth century by a dwindling population of fisherfolk and their descendants in three villages of coastal East Sutherland in Highland Scotland. ESG

in praise of the particular voices of languages at risk

351

speakers represent the last local population segment to shift to English in a regional shift process that began in the twelfth century, when vast lands in the region were granted by a distant monarch to a non-indigenous family without either linguistic or cultural ties to the local population. The upper social strata (not just the ruling aristocratic family, but also their upper-level estate administrators and functionaries, and likewise the clergy whose placements the ruling family controlled) grew more and more exclusively English in speech and culture over the centuries. The large farmers became exclusively Englishspeaking quite abruptly, by contrast, in the process of massive lease transfers in the first half of the 19th century, transfers that simultaneously weakened the position of Scottish Gaelic by displacing great numbers of the original Gaelicspeaking population; the evictees either became fishermen (involuntarily, by estate design) or emigrated. Craftsmen, small tradesmen, sub-subsistence agriculturalists (crofters), and large populations of agricultural wage-laborers and fishers remained exclusively or predominantly Gaelic-speaking throughout most of the 19th century. By the early 20th century most craftsmen and small tradesmen were going over to English, however, and by mid-20th century the crofters and such agricultural laborers as remained had also largely shifted to exclusive use of English. Since only the fisherfolk and their descendants remained proficient speakers of Gaelic as well as English when my work in East Sutherland began in the 1960s, the materials I draw on here represent fisherfolk ESG. The feature of ESG that I will chiefly be discussing, the emphatic marker, takes somewhat different phonological forms in other dialects of Gaelic. Since this suffix has merged phonologically with a deictic suffix in East Sutherland, the details of its use in ESG differ from those in other dialects as well. The general phenomenon of an emphatic marker is common to all dialects of Scottish Gaelic, however, and to Irish Gaelic as well. One notable thing about this feature is its ordinariness in structural terms. It consists of a few rather similar forms of a single suffix, and suffixes could hardly be commoner, in Gaelic or in the world’s languages, as a grammatical device. The grammatical structure of Scottish Gaelic is celebrated among linguists, and among language enthusiasts generally, but not for its suffixes. The celebrated feature is rather its abundant consonant mutations. They occur both at the beginnings of words (very commonly) and at their ends (less frequently). As a grammatical device consonant mutations are much less common in the world’s languages than suffixes, and consequently they have the allure of the unusual. Consonant mutations strike the English speaker as highly unusual, since there’s nothing like them in English, whereas English has a reasonably good supply of suffixes. Some of the consonant mutations carry grammatical information, which makes them functionally important to native speakers and learners alike. Examples (with orthographic forms here and throughout given

352

chapter 19

to suit ESG, rather than as in standard written Scottish Gaelic): phòs a /fo:s a/ ‘he married’, vs. pòs a /pho:s a/ ‘marry him!’; tha cat agham /ha khath am/ ‘I have a cat’, vs. tha cait agham /ha khačh am/ ‘I have cats’. Some of the mutations are obligatory and yet do not carry grammatical information, which doesn’t trouble native speakers in the slightest but can seem an unnecessary and unkind complication to learners. Despite the fact that consonant mutations are relatively unusual as a high-frequency grammatical device, there is relatively little expressed via consonant mutation in ESG that is not either fully matched as a grammatical category by some grammatical element in English, as is true of the past tense and the plural, the two grammatical categories expressed by consonant mutations in the examples given above. (The major exceptions are direct address in the form of the vocative case, which is marked only suprasegmentally – e.g. by pitch- and stress-contours and by timing – in English but by consonant mutation as well as by suprasegmentals in ESG; and grammatical gender, which is marked by consonant mutation for one class of nouns, provided the definite article is present, but can be marked by other devices in various other grammatical environments). By comparison with the attention lavished on the consonant mutations of Scottish Gaelic in most grammars, the emphatic suffix, the chief feature to be discussed here, is only briefly mentioned in most treatments of Gaelic dialects. One reason for the disparity in treatment is natural enough: the consonant mutations affect a large number of different consonants and appear obligatorily in many different environments (and optionally in still others), whereas the emphatic suffix takes a limited number of forms and can be suffixed to only a limited number of elements, while its use is largely optional. Still, there is most likely another reason as well. The emphatic suffix serves above all to create discourse effects, rather than to express grammatical categories, and both traditional grammars and linguistic descriptions show a tendency to concentrate on grammatical elements whose domain is the sentence. Some of the expressive force of the emphatic suffix can be seen within the sentence, or across one or two sentences, but to see its most striking effects it’s necessary to look at longer stretches of discourse.

The Emphatic Suffix, A Focus Marker

Traditional Scottish Gaelic grammar recognizes an emphatic suffix, usually a sibilant or shibilant element plus or minus following vowel, that serves to highlight contrasts, to place emphasis, and generally to mark the speaker’s focus. It also marks changes of focus as the speaker takes conversational turns or moves

in praise of the particular voices of languages at risk

353

along in a narrative. In traditional grammar the emphatic suffix is restricted to occurrence with personal pronouns, prepositional pronouns (the prepositions that conjugate for person in all the Celtic languages), and a limited number of verbal forms. As compared with the forms in the standard language and in many other dialects, the order of vowel and consonant in the emphatic suffix is reversed in East Sutherland Gaelic, taking the basic form -(e)as (rather than -sa, -se). The sibilant element is often (though not always) palatalized to a shibilant if the pronoun to which the emphatic suffix is attached has a front vowel, or had one historically. More nearly mainstream dialects of Scottish Gaelic have a set of so-called emphasizing particles which can appear after nouns, but only if the noun is preceded by the possessive pronoun. These emphasizing particles traditionally take different forms according to person and number, just as the possessive pronouns do. Insofar as these emphasizing particles can be said to exist in ESG, however, analogical leveling eliminates all person-and-number forms except the one common to the 1st and 2nd person singular, -s(a), -s(e) in more westerly dialects (see Oftedal 1956: 212) but -(e)as in ESG: bha h athair-eas na architec’ ann an seo va h a:r-əs nə arkitʰεkʰ an ə šɔ was 3.sg.fem. father-suffixed an architect in here poss.pron. emphatic ‘Her father was an architect here.’ (Golspie speaker, 1968) Scottish Gaelic grammar also recognizes a set of three unstressed and postposed enclitic particles with deictic force, expressing roughly ‘this’, ‘that’, and ‘yon’. The first two of these are recognizably present in ESG, but only one of them is productive, a proximal deictic which takes exactly the same phonological form as the dominant allomorph of the emphatic suffix, namely -(e)as /-əs/ ‘this’. The deictic force of proximal -(e)as is clearly recognizable when it’s applied to nouns with temporal reference, e.g. an t-seachdan-as /ən ǰaxkanəs/ ‘this week’, but otherwise the deictic force is less obvious, often undiscernible. The phonological merger of all these elements – emphatic suffix, emphasizing particle, and deictic enclitics – blurs their distinctiveness in ESG, and it appears that the -(e)as suffix can now combine their semantic force to some extent. In view of the ESG merger of the emphasizing elements (emphatic suffix and emphasizing particle) with the deictic enclitic, it would be more accurate to speak of an emphatic-deictic suffix for the dialect. But since the emphatic function is considerably the more prominent in contemporary usage, the

354

chapter 19

rubric ‘emphatic suffix’ can be used for economy’s sake to cover the single phonological outcome. The emphatic suffix appears at highest frequency with the personal pronouns and with the prepositions that conjugate for person. Among the latter the pronominal forms of the preposition aig ‘at’ are most frequent, since conjugated forms of aig, in conjunction with the verb ‘to be’, serve in the absence of a verb ‘to have’ to express possession (‘I have a brother and a sister’ is expressed as ‘a brother and a sister are at me’). As an example of emphatic suffixation, the ESG forms of the personal pronouns and of the prepositional pronoun aig are given here with their emphatic equivalents (as they might realistically be written if this were a written speech variety, and with the forms from the village of Embo cited wherever there are inter-village differences): personal pronouns

emphatic equivalents

sg.

1 2 3m. f.

mi thu a i

pl.

‘I’ ‘you’ ‘he’ ‘she’

sinn sibh aid

‘we’ ‘you’ ‘they’

prepositional pronouns sg.

1 2 3m. f.

am* ad aig aic

1 2 3m. f.

pl.

mis thus éis is

sinneas sibhs éideas

emphatic equivalents pl.

‘at me’ ‘at you’ ‘at him’ ‘at her’

sg.

aghainn agaibh ac

sg.

‘at us’ ‘at you’ ‘at them’

1 2 3m. f.

pl.

amas ainneas adas agaibhs aigeas ac(a)s** aic(ea)s**

* Although the form normally used in ESG is am, speakers appear to be aware of a fuller underlying form and occasionally someone produces a form such as agam. ** In the case of the two pronouns that end with voiceless final consonant, aic and ac, the emphatic form shows variation, appearing sometimes just as suffixed -s and sometimes as -(e)as.

Apart from its appearance with the personal pronouns and the conjugated prepositional pronouns, the ESG emphatic-deictic suffix can attach to certain other pronominal forms (e.g. an té ‘the one’, used in standard Gaelic for femi-

in praise of the particular voices of languages at risk

355

nine nouns only and likewise in ESG where humans are in question, but used in ESG for both female and male animals and for all inanimate nouns that have purely grammatical gender) and to a noun or an adjective when either is the final element in a noun phrase with a possessive pronoun: mo phiùthaireas /mə fyu:rəs/ ‘MY sister’, an taigh móras /ən dε mo:rəs/ ‘their BIG house’. Note that by contrast to English, where the voice emphasis falls on the possessive pronoun in the case of ‘MY sister’, it’s the noun that takes the emphatic suffix in Gaelic; the possessive pronoun, mo in this case, is always unstressed and cannot combine with other elements. In ESG just one verb form can add the emphatic suffix, namely the first person singular of the conditional: rachainneas air ais /raxĩn´əs er aš/ ‘I WOULD GO back!’ There are occasional occurrences with numerals: bheir mi na dhà-as seachad /ver mi nə γa:əs šaxəd/ ‘I’ll pass THESE TWO on’.

Uses of the Emphatic Suffix

Of particular interest here are uses of the emphatic suffix over more extensive stretches of speech to create discourse effects. I offer below some examples from recordings of ESG interviews, narratives, and taped “letters” to show typical discourse effects. They begin with an instance that could be paralleled quite effectively in English by use of supra-segmentals alone, since emphasizing the equivalent words by stress and pitch in the English translation creates a similar effect. In the later examples, however, the number of emphatic suffixes used goes beyond what could appropriately be matched by voice emphasis in English; the number of sentence elements that can take voice emphasis in English without semantic or affective distortion, over an extended stretch, is limited by comparison with the number of sentence elements with emphatic suffix attached that can comfortably appear in an extended stretch of ESG. adv cond emph fut

Abbreviations in Transcriptions: adverbial marker conditional suffix* emphatic suffix future suffix

imper pret rel fut depend

3 singular imperative suffix preterite particle relative future suffix dependent verb form

* The conditional suffix and the 3 singular imperative (subjunctive) suffix are both /-u/. But in the independent conditional (that is, in the conditionals not governed by certain particles and conjunctions), the suffix /-u/ combines with obligatory verb-initial consonant mutation, whereas no initial consonant mutation appears in the subjunctive.

356 1. 2. 3. 4.

chapter 19

Lines in Transcriptions orthographic rendering of East Sutherland Gaelic; words with emphatic suffixes are capitalized and boldfaced phonemic rendering morpheme-by-morpheme gloss; a period within the gloss signifies a complex morpheme the constituents of which can not be designated by segmenting the surface structure English translation; words that correspond to Gaelic words with emphatic suffixes are capitalized and boldfaced.

1. correction of a misapprehension: ‘You’ve got it wrong!’ ------------------------------------------------ -------------(Source: Embo male, aged 54 at the time of the recording in 1974.) O, bha is – uaosan a’ cheud chuairt o:, va i-š – wəsən ə çiəd xuaršth oh, was she-emph – since the first time ‘Oh, she was – since the first time bha i ’s a’ bhail’, bha i ’g iarraidh va i s ə val va i k iar-i was she in the village was she at want-gerund she was in the village, she wanted thighian ’n a’ bhail. ’S e mis hĩãn n ə val š e mĩ-š to.come to the village. Is it me-emph to come to the village [to live]. It’s me nach d’ robh ’g iarraidh thighian ’n a’ bhail. nax t rɔ k iar-i hĩãn n ə val that.not preterite was at want-gerund to.come to the village that didn’t want to come to the village.” 2. Contrast between two different eras: ‘Those were the days!’ ------------------------------------------------ -------------(Source: Same as in 1.)

in praise of the particular voices of languages at risk

357

Uail, bha Gàidhlig ann an Dòrnach, da’ bha Matty beò. . . . wεl va ka:likh an ə do:rnax tə va mati pyo: . . . well was Gaelic in Dornoch when was Matty alive . . . Well, there was Gaelic [taught] in Dornoch when Matty was alive. . . . Matty M___. Bordaidh ’s am bidh, theidheadh aid mati m___. pçrdi s ə bi hε-u aǰ Matty M___. body in the world would.go-cond they Anybody whatever, they would go uiceas, gheibheadh aid Gàidhlig bhoidh. ik-əs yε-u aǰ ka:likh vɔi to.him-emph would.get-cond they Gaelic from.him to him, they would get Gaelic from him. Geography tidsear bha ann dethas. Ach dh’ fhàg éis, čiɔgrafi thičər v ãũn čε-s ax x a:g e:-š geography teacher was in.it of.him-emph but pret leave he-emph, He was a geography teacher. But he left, dh’ fhalbh a Skye . . . x arlu a skai pret went he [to] Skye he went to Skye. 3. you / your party vs. me / my party: ‘I’ve got a bone to pick with you!’ ------------------------------------------------ -------------(a) a serious matter ------------------------------------------------ -------------(Source: Embo female, aged 42 at the time of the recording in 1967.) (Th)àinig a gus an taigh aghainn, agus thubhard a ãnig a kəs ə dε ãn´ agəs hurd a came he to the house at.us and said he ‘He came to our house, and he said re m’ athair, “Bha ’n irinn aghdas goid rε m a:r va n´ ĩrĩn´ ad-əs kəǰ to my father was the daughter at.you stealing to my father, “Your daughter was stealing

358

chapter 19

an – na h- ùibhlean aghmas ən – nəh u:l-ən am-əs the – the.pl apple-pl at.me-emph the – my apples on Saturday.”

air Di-Sathairn.” er či so:rn on Saturday

’S thubhard m’ athair, “Na saoilinneas s hurd m a:r nə səl-ĩn´-əs and said my father if think-cond.1.sg.-emph And my father said, “If I thought gum bitheadh i goid ùibhlean, kəm bi-u i kəǰ u:l-ən that be-cond she stealing apple-pl, that she would be stealing apples, ghabhainneas am bealt dith.” xa-ĩn´-əs ə bεltʰ či take-cond.1.sg.-emph the belt to.her I would take the belt to her.” (b) joking ------------------------------------------------ -------------(Source: Golspie female, aged 69 at the time of the recording in 1964.) Ach, by golly, ax pai kɔli but by golly ‘But by golly, if you

ma mə if

dus tu-s you-emph

bhithidheas vi-s will.be-rel fut

feuchan nah amhran’n aghmas do sluagh muigh siod, fiax-ən nəh ãũ:ran-n am-əs tə sɫuəu mwĩ šəd show-gerund the.pl song-pl at.me-emph to people out yonder show my songs to people out there, gheobh thus yo u-s will.get you-emph you’ll get it from me!’

bhumas vũm-əs from.me-emph

a! a it

in praise of the particular voices of languages at risk

359

Note that in the case of the oppositional emphatics of 3(a), lines 2 and 3, and 3(b), line 3, both parties are represented by pronouns that show the emphatic suffix, whereas in English one pronoun in each line would certainly receive voice emphasis but probably not the other. This sort of symmetrical emphasizing of the pronoun forms representing both speaker and hearer is frequent in the use of ESG emphatic suffixes, highlighting the interactional dimension of the material. To see the emphatic suffix come into its own most fully, in ESG, the best place to look is a narrative told with plenty of feeling. In the tape-recording from which example 4 is taken, the narrator tells of a rivalry between choirs from the villages of Golspie and Brora. A Gaelic choir from each village will be competing for a cup at the provincial Gaelic music festival, and the singers (none of whom speaks the standard language in which competitive singing is done, and many of whom don’t speak or understand Gaelic at all) will need the services of a Gaelic coach to help them prepare. The narrator of the story, a bilingual Golspie woman, doesn’t take kindly to the woman who has come to coach the East Sutherland choirs, and the force of her opposition to the woman is felt not just in the words she uses but also in the number of emphatic-deictic elements that appear in the narrative. 4. exception is taken to someone or something: ‘Not if I can help it!’ ------------------------------------------------ -------------(Source: Golspie female, aged 73 at the time of the recording in 1968.) ’S bha consairt mòr aghainn s va khɔnsertʰ mo:r ãn´ and was concert big at.us ‘And we had a big concert in the drill hall

a’s an drill hall as ə dril hal in the drill hall

an latha* tha seo, agus bha i – thànaig an té-eas – a staigh. ən ɫɔ: ha šɔ agəs bha i – hã:nig ən ǰe:-əs ə stε the day is this and was she – came the one-emph in this day and she was – that woman came in. ’S dar a dh’ fhalbh i, thubhard mi fhian ** ri Bean M____ , s tər ə γ aɫu i hurd mi hĩãn ri pε̃n m____ and when pret went she said I self to Mrs. M____ And when she went off, I said to Mrs M____,

360

chapter 19

“Dé an danas tha an té sin sireadh seo?” te: ən danəs ha ən ǰe: šən šir-u šɔ what the devil is the one that seek-gerund here “What the devil is that woman looking for here?” Agus ors is, “O, ’an ail – fios acam***”– agəs ɔrs i-š o an al fis ak-əm and said she-emph oh not is knowledge at-me And she said, “Oh, I dont – know”, no – . . . a’ Ghàidhlig aic fhéin** nə ə γa:likh εkh hẽ:n or the Gaelic at.her self or – her own Gaelic. ’S ors mis, “Uail, mar téid i mach á seo, s ɔrs mi-š wεl mər če:j i max ε šɔ and said I-emph well if.not will.go she out of here And I said, “well, if she doesn’t go out of here, bheir mi cic ’s an tòn dith.” ver mi khikh s ə dõ:n či will.give I kick in the rear to.her I’ll give her a kick in the rear.” Agus, ors Bean M___ riumas, “Uail, tha is agəs çrs pε̃n m___ rəm-əs wεl ha i-š and said Mrs. M___ to me-emph well is she-emph And Mrs. M____ said to me, “Well, she’s tidsgeadh a’ Ghàidhlig.” ’S ors mis, thičky-u ə γa:likh s ɔrs mĩ-š teach-gerund the Gaelic and said I-emph teaching the Gaelic.” And I said, “Uail, ’an ail i du’ do thidsgeadh sinneas. wεl an al i tu tə hičky-u šĩn´-əs well not is she going to teach-gerund us-emph “Well, she’s not going to teach us.

in praise of the particular voices of languages at risk

361

Dar ach do ghobh is sinneas, tər ax tə γo i-š šĩn´-əs when not pret took she-emph us-emph When she didnt take us, dar a – thurnaig aid an àrd, tər ə hərnig aǰ ən ɔ:rd when turned they adv high when they turned up, dar a thàinig aid . . . gus a’ hall. tər ə hã:nig aǰ . . . kəs ə hal when came they to the hall when they came to the hall.**** Bheil i smochdadh gum bheil a vel i smɔxk-u kə vel a is.interrog she think-gerund that is.depend it Is she thinking that we aghainneas ri pheidheadh is, an nis, ãn´-əs ri fe:-u i-š ə nĩš at.us-emph to pay-gerund her-emph adv now have to pay her? Now, oir’n thighian an àrd? Ach ’a ruig i fhéin* leas. ɔrn hĩãn ən ɔ:rd ax a rig i hẽ:n l´εs for coming adv high but not reach she self benefit. for coming up? But she needn’t [think so]. Ma tha is sireadh – thidsgeadh cuaoirichean. mə ha i-š šir-u hičky-u kʰwair-içεn if is she-emph seek-gerund teach-gerund choir-pl If she’s looking to – teach choirs. Siubhaileadh is do Bhrùra, agus bithidh i tidsgeadh šul-u i-š tə vru:ra agəs pi-i i thičky-u walk-3sg.imper she-emph to Brora and be-fut she teach-gerund Let her go to Brora, and she’ll be teaching

362

chapter 19

an fheadhann á Brùra. Agus peidheadh éideas i, ən´ yə̃ ũn ε pru:ra agəs phe:-u e:ǰ-əs i the group from Brora and pay-3sg.imper they-emph her the ones from Brora. And let them pay her, uaoil ’an ail sinn du’ wəl an al šĩn´ tu because not are we going because we’re not going to pay her.”

do tə to

pheidheadh i.” fe:-u i pay-gerund her

*A relative particle is omitted here, as often in this structure in ESG. ** Adding the element fhéin ‘self’ (realized as 1st person fhian, otherwise fhéin) to a pronominal form is another way of adding emphasis. If anything, fhéin creates a slightly stronger emphatic effect than suffixed /-əs/, and it’s striking that two instances of emphatic fhéin appear in this passage, in addition to the many emphatic suffixes. (One other instance of fhéin is non-emphatic.) *** Mrs. M___ speaks a different dialect of Gaelic than the narrator does, and in rendering ‘at me’ as acam instead of agham or agam, the latter reproduces one of Mrs. M___’s non-local forms. **** There appears to be a reference here to a grievance, perhaps the rejection of some would-be choir members who tried to join up belatedly.



The Expressive Power of the Emphatic Suffix

Quite generally speaking, the narrator from whose recorded story example 4 is drawn is a highly expressive speaker. In most of her interviews and stories, not just this one, she makes use of a particularly rich array of interjections, and she doesn’t shy away from using mild profanities and other indelicate lexical items, even when she’s being tape-recorded. The pitch- and stress-contours in her stories tend to be greater than average, and she’s inclined to hilarity when there’s the least shade of impropriety or absurdity in whatever matter she relates. In English as well as in Gaelic her stories are lively, then, but in her Gaelic arsenal she has some weapons not available to her in English. One of them is the emphatic suffix, supplemented on occasion by emphatic use of fhéin ‘self’. There are 27 clauses in the narrative stretch offered as example 4 above. Twelve of them, or almost half, include an emphatic suffix; in two instances there are two emphatic suffixes in a single clause. (There are also two instances of emphatic use of fhéin.) The tone is set immediately, when the narrator speaks of seeing the Gaelic coach come into the hall and refers to her as an té-eas ‘that one(-female)’. Two sentences earlier the narrator had referred to the same woman as am boireannach seo ‘this woman’ and had then also used two unemphatic pronominal forms in referring to her; but as she begins the

in praise of the particular voices of languages at risk

363

particular story in which the woman appears as an unwelcome, intrusive presence, the narrator selects an té-eas, creating a distancing effect by using the indefinite pronoun and the emphatic-deictic. An unfriendly tone is set by this change, and it continues throughout the part of the narrative that concerns the offending woman. An unusually high incidence of emphatic suffixes captures the high affect that attaches to this tale of a strongly disliked woman; they pursue the unfortunate woman across many of the clauses that follow her introduction as an té-eas. Registering this trail of emphatic suffixes, the auditor or reader can’t be altogether surprised to find the narrator asking ‘what the devil’ the woman is doing there or threatening to give her ‘a kick in the rear’. Of the 14 emphatic suffixes in example 4, ten are attached to personal or prepositional pronouns that refer directly to the Gaelic coach (as is one of the two emphatic uses of fhéin). The four others (and the other emphatic use of fhéin) appear in conjunction with ors ‘said’, a defective verb used only quotatively. In the full corpus of this speaker’s tape-recorded material, it’s evident that strong affect inclines her to select ors instead of the less marked quotative verb thubhard ‘said’. The additional emphatic pronouns used with ors are therefore in keeping with this additional high-affect word choice. The English translation I chose for an té-eas, ‘that woman ’, with the boldface capital letters used here to indicate strong stress on both words in the English, can serve as an example of an instance in which English offers a good parallel, in the heavy stressing of both words and the choice of a distancing deictic element, to the effect of the emphatic-deictic -eas added to an té in Gaelic. Other features do not correspond so well. Looking first at the resources of English, an English speaker has structural freedom to apply voice emphasis to any noun-phrase element whatever, and to two or more of them together, using voice emphasis to highlight a whole nounphrase or even a whole clause at a time. This not the case for ESG speakers. Neither the pre-nominal nor the post-nominal element of the usual (unmarked) nominal construction ‘that woman’ could take voice stress in Gaelic in a noun phrase such as am boireannach sin ‘that woman’ (lit. ‘the woman that’), and the emphatic suffix, too, would be restricted in its occurrence. It could appear only once within a single noun phrase, and it could be applied only to the subject noun or pronoun, or alternatively to a modifying adjective. Thus far it sounds as though English speakers, with their supra-segmental resources, have more scope for expressing emphasis than Gaelic speakers with their suffixes. But voice emphasis in English is much more intrusive over a long narrative stretch than is the Gaelic emphatic suffix, and narrators who use voice emphasis continually or repeatedly within a limited narrative space risk overdoing the effect and detracting from the story development. Precisely

364

chapter 19

because the emphatic suffix in Gaelic doesn’t require any particular pitch or stress prominence, it can be employed multiply within a single clause and repeated frequently over a series of clauses without making the narrative sound overwrought and without distracting attention from the unfolding of the story line. Gaelic makes available a separation of voice emphasis and focus that English, with only supra-segmental features to indicate emphasis, can not provide. The speaker in example 4 uses emphatic forms of the personal pronoun four times in leading up to direct quotations, as she reconstructs a conversation within her narrative. By keeping the emphatic suffixes coming, she keeps the discourse tone (continuing intense interest in the objectionable Gaelic coach) constant over a long stretch. None of these personal pronouns has voice stress in the Gaelic, and it’s precisely the absence of stress that makes it possible for a series of emphatic personal pronouns to play their role in maintaining discourse tone without diminishing the salience of the quotations that they precede. In addition, symmetrically placed emphatic suffixes can highlight speaker-addressee or subject-object oppositions in ESG (see especially example 3(b) above) in a way that multiple contrastive occurrences of voice emphasis in English can not, at least without distortion. For the written language there is of course also the advantage that the discourse tone of the spoken Gaelic text persists, thanks to the visible presence of the emphatic suffixes, while the discourse tone created by voice emphasis is lost in formal written English. (It can be evoked in casual written English by means of underlinings and exclamation points, liberally resorted to by some people in their private correspondence by way of a substitute for the missing suprasegmentals.) As is evident in example (4), where the use of fhéin ‘self’ is seen to supplement the discourse effect of the emphatic-deictic suffix, languages not only offer distinctive resources but offer the possibility of combining them in distinctive fashion. Gaelic speakers (and Irish speakers, too), besides combining the emphatic use of fhéin with use of the emphatic-deictic suffix -(e)as, are known for their frequent use of “clefting” to allocate emphasis: a “dummy subject”, it, leads off the sentence, linked by a form of be to material highlighted by its postponement. So prevalent is clefting in both Scottish and Irish Gaelic that in its frequent carry-over into Highland and Irish English it’s become a stereotyped feature (e.g. “It’s grand stories he’s telling!”). And once again, speakers can combine this device with the emphatic suffix or fhéin to produce a particularly strong effect. The following two examples drawn from narratives recorded from an octogenarian Embo man in 1970, combine, respectively, clefting with emphatic suffixes and clefting with fhéin:

in praise of the particular voices of languages at risk

365

5.

focus-marking by a combination of clefting and emphatic-deictic suffixation (cf. also the final line of example 1 above). ------------------------------------------------ -------------’S e éis a chuir a cheud taigh riamh an àrd š e e-:š ə xur ə çiəd tʰε riəu ən ɔ:rd cop it he-emph rel put.pret the first house ever adv high It’s he that put up the first house ever a’s a’ bhaileas. as ə val-əs in the village-emph in this village. 6. focus-marking by a combination of clefting and fhéin ------------------------------------------------ -------------’S e mi fhian thug a steach air ais i. š e mĩ hĩãn hug ə stεx ar aš i cop it I self take.pret in back her It’s myself that took her back in [a fog-bound boat]. In Irish, clefting and the emphatic suffixes are used to the exclusion of suprasegmentals to mark focus and emphasis, according to Cotter’s analysis of the Irish of radio broadcasting: The Irish language does not use pitch prominence in the intonation contour in the way that English speakers do, but uses instead syntactic reordering through what could broadly be called clefting, and the so-called ‘emphatic suffixes’ (Cotter 1996: 48). East Sutherland Gaelic is not an especially conservative Gaelic dialect in this respect, and voice emphasis can be used in some environments, in addition to the emphatic-deictic suffix, fhéin, and clefting. But voice emphasis is applicable to far fewer elements in Gaelic than in English, since stress can not be applied to particles and many other functors, and this means that a prominent role falls to alternative devices such as the emphatic-deictic suffix, fhéin, and clefting. By comparison with the practices of English speakers, furthermore, Gaelic speakers quite generally pay a great deal of attention to marking focus, as the very existence of three special devices that can be deployed for the purpose suggests. The frequent use that ordinary speakers make of these elements is evident in the examples given above, but of course gifted creative writers

366

chapter 19

draw on them to powerful effect as well. Here are the opening lines of Gaoir na h-Eorpa (‘The Cry of Europe’) by the late Sorley Maclean, considered by many the finest of modern Scottish Gaelic poets (from MacAulay 1976: 74–75): A nighean a’ chùil bhuidhe, throm-bhuidh, òr-bhuidh, fonn do bheòil-sa ’s gaoir na h-Eòrpa, a nighean gheal chasurlach aighearach bhòidheach, cha bhiodh masladh ar latha-ne searbh ’nad phòig-sa. Girl of the yellow, heavy-yellow, gold-yellow hair, the song of your mouth and Europe’s shivering cry, fair, heavy-haired, spirited, beautiful girl, the disgrace of our day would not be bitter in your kiss. In this first verse of a poem evoking what was for Maclean the dark Europe of the 1930s after Franco’s victory in the Spanish civil war, three emphasizing particles appear in the Gaelic (bheòil-sa ‘your mouth’, latha-ne ‘our day’, and phòigsa ‘your kiss’), setting the lovely girl of spirit and beauty against the darkness of contemporary Europe. The emphasizing particles are an important element in establishing the contrast in Scottish Gaelic, but though the translation is Maclean’s own, nothing is available to him in written English to create a similar effect. If the English version of the poem were to be read aloud, furthermore, and voice stress were applied to your and our, the effect would be distorting, not enhancing. Among the ESG examples offered above, the expressive potential of the Gaelic emphatic suffix is especially evident in the fourth. The deployment of the emphatic suffix seems a pretty straightforward matter in example 1, and perhaps also in examples 3(a) and 3(b). It’s less obvious (at least to me) why some pronominal forms appear with the emphatic suffix in examples 2 and 4 while others don’t. (Why not gheibheadh aid Gàidhlig bhoidheas ‘they would get Gaelic from him’, for example? And why does the high-frequency expression ’s am bidh ‘whatever, at all’ [literally ‘in the world’], never pick up an extra degree of intensification and become ’s am bidh-eas?) Already in example 2, but even more so in example 4, we reach the realm of skilled-native-speaker stylistic choices. It’s beyond me both as linguist and as learner to account fully for the motivating factors behind the native-speaker choices, in texts such as these, where stylistic choices were obviously made not to use the emphatic suffix in some potential environments, as well as to use it in others. At the same time, it’s well within my capacity both as linguist and as learner to recognize and relish certain stylistic effects from the speakers’ deployments

in praise of the particular voices of languages at risk

367

of the suffixes, and to celebrate the liveliness and coherence of tone that the emphatic-deictic suffix brings to their Gaelic discourse. The use made of the emphatic-deictic suffix by the speaker in example 4 is not subtle, and it’s easy to imagine sophisticated story tellers who might introduce suffix-bearing forms less frequently but more slyly and strategically, to more cunningly designed effect. But the particular way the speaker in example 4 uses the emphatic suffix is well suited to her personality and speaking style. The grammatical environment surrounding the central figure, the an té-eas ‘that woman’ of the narrative who then also appears as the direct focus of nine emphatic suffixes, creates a consistent discourse environment in which that high-focus figure becomes an unsurprising target for ‘a kick in the rear’ or an invocation of the devil. The Gaelic passage has a well-sustained narrative tone in which the ESG emphaticdeictic suffix plays a substantial and – for all my morpheme segmentations and glosses – never fully translatable part. Sorley Maclean, practiced translator of his own poetry that he was, fared no better when he came to render the Gaelic of Gaoir na h-Eorpa into English; he had to forego in English the special contrastive effect that the emphasizing particles had given his opening verse in Gaelic. The very distinctiveness of a language’s most particular features creates the insoluble problem any would-be translator (or celebrator) faces, of course: belonging uniquely to that language, they are essentially untranslatable. References Aikhenvald, Alexandra Y. 1996. Areal diffusion in Northwest Amazonia: The case of Tariana. Anthropological Linguistics 38:73–116. —— 2003. Classifiers: A typology of noun categorization devices. Oxford: Oxford University Press. Breitbach, Joseph. 1964. Bericht über Bruno. München/Zürich: Knaur. Cotter, Colleen Marie. 1996. Irish on the air: Media, discourse, and minority-language development. Unpublished University of California, Berkeley, dissertation. Leyew, Zelealem. 1997. Some structural signs of obsolescence in K’emant. Paper given at the Endangered Languages in Africa conference, Leipzig, August, 1997. —— 2003. The Kemantney language: A sociolinguistic and grammatical study of language replacement. Köln: Rüdiger Köppe Verlag. MacAulay, Donald (ed.). 1976. Nua-bhàrdachd Ghàidhlig / Modern Scottish Gaelic Poems. Edinburgh: Southside. Mithun, Marianne. 1998. The significance of diversity in language endangerment and preservation. In Lenore A. Grenoble & Lindsay J. Whaley, eds, Endangered

368

chapter 19

languages: Current issues and future prospects, 163–91. Cambridge: Cambridge University Press. Oftedal, Magne. 1956. The Gaelic of Leurbost Isle of Lewis. (A Linguistic Survey of the Gaelic Dialects of Scotland, Vol. III; Norsk Tidsskrift for Sprogvidenskap, Supplementary Vol. IV) Oslo: H. Aschehoug and Co. Steiner, George. 1992. After Babel: Aspects of language and translation (second edition). Oxford: Oxford University Press. Woodbury, Anthony. 1998. Documenting rhetorical, aesthetic, and expressive loss in language shift. In Lenore A. Grenoble and Lindsay J. Whaley, eds, Endangered languages: Current issues and future prospects, 234–58. Cambridge: Cambridge University Press.

PART five Fieldwork: Methods, Problems, Insights

∵

chapter 20

Gathering Language Data in Terminal Speech Communities The difficulties confronting any linguistic fieldworker are considerable. There are always the problems of choosing an area, locating a useful sample, gaining the cooperation of the speakers, eliciting matter which is both reliable and germane to the investigation, obtaining confirmation of data from others besides the speaker(s) who originally supplied it, and so forth. Most such difficulties are exacerbated by the conditions which obtain in speech communities which are in the process of disappearing as a speech variety passes out of existence; often enough, additional difficulties arise from the special circumstances in speech communities of this sort. In this paper I will draw on my own experience in working with terminal speech forms and on the reports of a number of others to offer a partial “geography” of this difficult terrain and some of the methods which have been adopted by fieldworkers in terminal speech communities to circumvent, or at least reduce, the obstacles. The one problem which is sometimes lessened in working with terminal speech communities is choice of location. Quite often the remaining speech community is enclaved, with only a few small areas left in which it is spoken, while some other language has become increasingly dominant and extended its sway over all other districts. There may be only a single enclaved population, or there may be several. In the former case the site is virtually a given; in the latter, factors such as the familiarity of the surrounding dominant language, the degree of cultural conservatism, the relative health and accessibility or speakers, the number and power of resident speakers of other languages, and matters as mundane as distance, availability of accommodation, climate, the presence or absence or poisonous reptiles and insects or other health hazards, may determine the choice of site. If more than one population is available the question also arises of whether to make the investigation comparative from the start or to work extensively in only one site (the most precarious, say) while planning a second-stage study within one of the other populations. A question easily overlooked, und not always answerable in advance of site choice, is the possibility of a considerable group of exiles or seasonal out-migrants who might either provide a contrastive sample, or, if linguistically conservative, expand the sample of speakers beyond those in regular residence at the site. Equally, if not linguistically conservative, such a group may turn out to serve © koninklijke brill nv, leiden, ���4 | doi 10.1163/9789004261938_022

372

chapter 20

as the means of introducing a flood of loanwords or other interference and render the chosen study population much less conservative than its geographical isolation led the investigator to expect when making the original selection. In extreme situations of impending extinction, Haas (Biloxi; 1968), Salzman (Gros Ventre; 1969), and Swadesh (Chitimacha; 1948) were lucky to find any speakers at all; their options were correspondingly limited. On the other hand, one of the particular attractions for Hill and Hill (1977: 55) in the study of Nahuatl was the quite different states of “health” of the language in various parts of Central Mexico. Mithun and Henry (MS.) were fortunate enough to have available simultaneously two widely separated populations of Cayuga speakers, one of which consisted of only about six speakers showing early signs of decay in their still quite fluent Cayuga while the more northerly community and other northerly Iroquoian languages offered fully proficient speakers and sometimes child acquirers for comparison. Timm (1984: 120–21) chose an interior town in Brittany for her detailed study of the phonology of a Breton dialect partly because the coastal regions of Brittany were thoroughly frenchified and she could expect to find more native speakers available as sources in the interior; but the dialect of her chosen region was also representative of the least studied group, and the town in question was central to a controversy over the history of Breton. My own study of the East Sutherland dialect of Scottish Gaelic was dictated by my offer to coordinate my work with that of the Gaelic Division of the Linguistic Survey of Scotland and go wherever they felt the need for study was greatest. I chose the central one of three residual East Sutherland Gaelic-speaking villages to reside in, although I worked in all three, because it made travelling to the other two less extreme in distance; but my choice was also affected by the fact that it was less isolated than one of the other two and more nucleated than the other less-isolated village. If the population available is vanishingly small, choice of speaker sample may be as obvious as choice of site. In that case the issue is an “uninteresting” one, since nothing is left to the investigator’s discretion. Hence only sampling situations which offer at least some choice will be discussed here. One of the serious problems which faces students of dying speech forms, typically, is lack of sufficient or accurate information about earlier stages of the same speech form. Diachronic study in “real” time, with comparison between the speech form as it was 20, 30, or 50 years earlier and the same speech form at the time of the current investigation, is often impossible because of the rarity with which many of these communities have been studied previously. Where there was an earlier study, it may well have been a study of a related speech form from a nearby area, or have provided a corpus drawn from speakers of dubious reliability (for example, multilinguals only one of whose languages

gathering language data in terminal speech communities

373

is the speech form now under investigation, so that it is difficult to evaluate the amount of cross-language influence – especially if two of the languages spoken were related; the Natchez speakers available to Swadesh [1948: 232] in 1939 were all multilingual, and only one of them apparently spoke Natchez without a Cherokee accent). Sometimes an earlier investigation was executed by a scholar working with different goals, trained in a different school, or of indeterminate reliability him- or herself. Although many investigators bemoan the lack of earlier studies with which to compare their own findings, others may have reason to wish that there had not been an earlier study so that they could start afresh instead of having to reckon with materials difficult to interpret and unassessable as to reliability. If other fieldworkers have done similar or tangentially related work in a given area before, it may on the one hand greatly increase the new investigator’s advance knowledge of conditions and thus his or her likelihood of success. But it will also require that the investigator attempt to assess the accuracy of the other report(s) insofar as these bear on the new work. It can also be the case that earlier fieldworkers have socialized potential sources into a form of “informant role” which is uncongenial or unworkable for subsequent fieldworkers – for example, by paying the sources relatively high wages from grant monies unavailable to a subsequent penurious graduate student, or by discouraging a natural conversational speech style which might show any code mixing actually typical of the community’s normal speech use. The most fortunate investigator is the one who did his or her own study of the speech form some years ago and with or without intending it or planning for it has now managed to return and work again in the same community. Voegelin and Voegelin (1971) offer a good example of this happy development. As often as not the investigator has no earlier materials available. If there are sufficient speakers left, the best that can be done to make up the deficiency is to work in “apparent” time, sorting the remaining speakers into age groups and carefully gathering materials (if possible directly comparable materials) from speakers within each group. One complication which immediately arises is that age and proficiency need not correlate. Dixon (MS.: 6–7) found that the youngest of his three surviving imperfect speakers of Yidiɲ was the most fluent of them; he had been raised by his grandfather, remaining in good contact with Yidiɲ until about age 15, whereas the other two speakers, though older (one by very nearly a decade), had had more contacts with English-speaking Australians and less with Yidiɲ. In such cases grading by proficiency offers an alternative to grading by age. Schmidt (1983) was able to establish progressive deviance from the conservative norm across a group of twelve imperfect speakers of Dyirbal, although her speakers also could not be perfectly ordered along

374

chapter 20

the proficiency continuum by age alone. Neither did the onset of progressively greater deviance appear at exactly the same individual along the continuum for each feature which was tested, yet over the rather large number and type of features which were investigated a quite stable ordering of the twelve imperfect speakers’ relative conservatism proved attainable. The East Sutherland Gaelic semi-speakers with whom I worked also did not themselves fall neatly into line via correlation between age and proficiency, but on the other hand there was perfect age/proficiency grading across the three groups of differentially skilled speakers: older fluent speakers represented a relatively conservative norm for the dialect; younger fluent speakers (who appeared only in the smallest, most cohesive, and least anglicized village where the dialect was dying most slowly) were all younger than the older fluent speakers and showed identifiable deviations from that norm which were not however noticed by the speech community, of which they were fully fluent, skillful members; semi-speakers varied considerably among themselves as to their skills, and greater proficiency did not always correlate with greater age, but within each village where they were studied they were all invariably younger than older fluent speakers and younger fluent speakers (Dorian 1981: 117). As in the case of Schmidt’s study, the large number of features investigated made it possible to group the semi-speakers into stronger and weaker proficiency categories of quite good generality overall. Lack of knowledge about any prior conservative norm for the speech form has occasionally led researchers to compare the dying speech form with a standard language, where one exists. This is usually both diachronically and descriptively precarious and can result in findings which are unprovable at best and seriously misleading at worst. Most often one simply does not know what to make of such findings; certainly no very useful conclusions can be drawn about degree and direction of change, since the differences might be explainable in terms of very early dialect differentiation rather than subsequent deviation from the norm. If we had no independent knowledge about the Alemannic dialect of Old High German, for example, and tried to explain the forms and sounds of Swiss German dialects in terms of changes away from standard German, our discussion would be impossibly flawed. By good fortune we have early attestation of Alemannic Old High German, and we also know a good deal about, the development of a standard German as a rather late compromise form based on East Middle German, with its roots in the Prague and Wittenberg chanceries, serving subsequently as the dominant element in the language of Martin Luther’s influential Bible translation (Waterman 1966: 117). Because of lack of knowledge about earlier local norms, Trudgill (1976–77) was in this sort of unfavorable position where Arvanitika, a form of Albanian

gathering language data in terminal speech communities

375

spoken in Greece, was concerned. Faute de mieux he compared Arvanitika with modern-day Albanian. It is very unlikely that this could give valid results, since the Albanian communities of Greece have been in existence since the fourteenth century (Tsitsipis 1983: 290) and Albanian in the two separate areas can not be expected to have developed completely in tandem. Trudgill notes that despite some difficulties a few of his informants were able to listen to radio broadcasts originating in Albania (1976–77: 49–50); but such generalized intelligibility does not make the kind of direct morphological comparison he undertakes historically or descriptively sound, any more than it would where southwestern Scottish Gaelic and Irish Gaelic are concerned, even though some people in the southernmost Inner Hebrides can follow the broadcasts of Radio Eireann out of Ireland. Despite the foreshortened timescale for change offered by “apparent time” studies, they seem less likely to mislead; and they also better avoid the danger of allowing a dialect with an independent history to be treated as a deviation from (or worse still, a corruption of), some standard language from which it is then erroneously “derived” by a set of ingenious but groundless rewrite rules. If no comparison of any kind is available because of paucity of earlier local materials and serious paucity of speakers as well, then the study will generally have to be descriptive, with only the most common-sensical assumptions of an earlier, fuller form. This had to be Haas’ position with regard to Tunica, for which she had only one surviving speaker. She concluded on the basis of general knowledge of language behavior in more normal circumstances that a language which has survived as a matter of accident via a single speaker and no longer serves any sociological function even for that individual is to all intents and purposes a dead language. Hence it is to be assumed that what Youchigant recalls of Tunica is at best a mere remnant of what the language must have been when many speakers used it as their only means of communication (Haas, n.d., 11). If a degree of comparison internal to the threatened speech form is possible, the difficulties are those of appraising the source-persons and their knowledge, heightened by what are likely to be very negative social circumstances for the target population. Gaining the cooperation of potential source-persons is one of the perennial challenges of fieldwork, but in healthy speech communities where every member is a potential source-person, it can usually be relatively easily achieved, always assuming that the investigator is an individual of some personableness, or entertainment value, or authority, or – sometimes – wealth. In speech communities where the language is dying, apart from the sheer

376

chapter 20

scarcity of available speakers there may be additional factors which complicate the process of finding source-people. Among these are such matters as negative attitudes towards the local speech forms, competition for status among the few remaining speakers, and sometimes the misrepresentation of personal competence in the language or dialect by remaining speakers. It is likely that most fieldworkers in terminal speech communities have to deal with negative attitudes toward the dying language in some form. The mildest form is perhaps a sort of lightly regretful pragmatism which gives rise to general protestations about the regrettable loss or the language unaccompanied by efforts to halt that loss, such as Miller encountered among the Shoshoni. He reported that the Shoshoni “exhibit very little language loyalty”, and that older speakers who sometimes made unfavorable comments about the interlarding of English words and phrases by younger speakers as a “corruption” of Shoshoni themselves “also interlard their Shoshoni with English words and phrases” (Miller 1971: 119, 120). In a case where competing speech forms constitute something more like a continuum, like the Bhojpuri to Standard Hindi scale which Gambhir investigated in Guyana, East Indian speakers applied the label halluk ‘low’ to an utterance in proportion to the number of elements in it which were identifiable as Bhojpuri, a vernacular speech form which has a partially diglossic relationship with Standard Hindi; both are being displaced by Guyanese Creole and Standard English in most domains. Gambhir predicts that while Standard Hindi may have “chances of limited survival”, Guyanese Bhojpuri “is heading for a sure death” (Gambhir 1983: 32–33, 28). Strongly negative attitudes towards enclaved and obsolescing languages are frequently recorded. Tsitsipis found among the remaining speakers of Arvanitika that certain negative attitudes toward the local language are equally shared by fluent and terminal speakers. Surfacing of these attitudes is almost predictable from knowledge of the discourse context such as naturally emerging language focused conversations. The whole set of linguistic attitudes could be summarized as follows: (1) Arvanitika is a bastard language not worth saving; (2) localities other than the speakers’ own speech community always use a deeper or less hellenized variety regardless of actual geographical distances or of objective dialect differences; and (3) children should not be instructed in Arvanitika since its learning interferes with a proper acquisition of competence in Greek (1983: 293). The deliberate non-transmission of the ancestral language to young children is a theme repeated with dreary frequency in communities where a threatened

gathering language data in terminal speech communities

377

minority language is the normal daily speech of the parental and grand parental generations. It is noted for example by Huffines (1980: 52) for Pennsylvania German, by Pulte (1973: 426) for Cherokee, by Denison (1971: 166–67) for German and Friulian in Italy, by Timm (1980: 30) for Breton, and by Dorian (1981: 104) for Scottish Gaelic . The majority population in which the threatened speech form is enclaved may have negative attitudes towards the minority language, shared by some but not all members of the threatened speech community. This was true for the English- and Dyirbal-speaking area studied by Schmidt (1983: 24–26) and for the German- and Hungarian-speaking area studied by Gal in Austria (1979: 106–07). In such areas the range of attitudes and the ambivalences found within each segment of the population can make particularly slippery footing for the investigator. Generally speaking it is certainly true that negative attitudes, whether as wide-spread throughout the bilingual population as was the case with the Arvanitika speakers studied by Tsitsipis or as variably present among both monolingual English speakers and bilingual English-and-Dyirbal speakers as was the case in the population studied by Schmidt, complicate the investigator’s task. In my own work in East Sutherland I met with obstacles ranging from fluent native speakers’ initial recommendations that I should go elsewhere so as to study the “right” Gaelic and attempts on the part of dialect speakers to modify their local Gaelic in the direction of biblical Gaelic or whatever they knew of western dialects, to incredulous questions from English monolinguals about the possibility of doing any serious work in the most isolated and in-bred of the Gaelic-speaking villages, where the natives had (in some circles) the reputation of being backward and slow-witted. Rivalries among remaining speakers, on the other hand, were a very minor problem in East Sutherland. A few people who disliked each other cast aspersions on each other’s language loyalty, suggesting that the other individual was “too proud” to speak the native language, but false allegations of incompetence were non-existent. If anything, a reverse problem occasionally cropped up: people were credited with rather more ability than they actually had. The reasons for overestimation seemed to be two: strong enthusiasm for Gaelic on the part of an imperfect speaker could lead people in relatively little direct contact with that individual to assume more ability than actually existed; and near-perfect receptive control of the language in combination with excellent knowledge of the sociolinguistic norms for the community could make quite limited active use of the language, with heavy reliance on fixed phrases and high-frequency collocations, sufficient for fairly unremarkable interaction with more fluent speakers. In some communities where a language is nearing extinction, familiarity with the ancestral tongue may have special value for the few remaining speakers

378

chapter 20

since it qualifies them, and them alone, to perform certain special rites or services. This in turn entitles them to particular respect as a link with a more intact ethnic past. If, in addition to the high prestige already associated with their rare language skills, linguists or anthropologists appear on the scene and create a new and also prestige-enhancing audience for their abilities, the value of their linguistic performances can rise sharply and kindle a certain amount of competition. In the absence of a speech community large enough and vital enough to permit the investigator either to become a skilled speaker him- or herself or to obtain convincing community consensus on the relative abilities of the remaining speakers, the investigator can find it all but impossible to determine which of the few speakers available are the most reliable and most skillful. The claims of friendship, plus lack of frequent direct interaction, may make the testimony of others among the surviving speakers less than perfectly dependable when they assess the skills of the leading contenders for linguistic laurels. Among the Delaware centered around Dewey, Oklahoma, for example, with a total fairly dispersed speaker-population of only about eight individuals in a community which valued traditional activities with a Delaware-language component, the appearance of investigators from outside the community who also placed a very high premium on language skills complicated the situation and gave rise to evaluation difficulties of this kind (S. Roark-Calnek, pers. comm.). Because of the complex social conditions typical of communities in which languages or dialects are dying, the possibility that potential informants will understate their own abilities is probably as great as the possibility that they will overstate them. It is true that some individuals apparently relish the role of sage and language expert for the prestige it confers and adopt it without the linguistic competence to support it. Swadesh gives a particularly vivid example for Chitimacha, where he describes the bizarre performance of an elderly woman who produced without a minute’s hesitation “a most remarkable mixture of Indian words and invented vocables, for the most part based on French or English with some twist or change” when asked for Indian vocabulary; he felt that she “must have been at least partly conscious of inventing forms in order to maintain her role as one who knew much ancient lore” (1948: 231). On the other hand some individuals, coming to adulthood in a region where the ancestral language is negatively viewed and well aware of the low prestige attached to native-speaker status, may choose to disclaim knowledge of the language. By reason of strong friendships within a particular kin network I was able to test quite extensively one individual who disclaimed speaker status and to establish that her ability to generate sentences in East Sutherland Gaelic was actually slightly superior to that of a near relative almost exactly her age

gathering language data in terminal speech communities

379

who eagerly claimed speaker status and used Gaelic at every opportunity. The disclaimer never voluntarily used Gaelic at all, with the exception of rare occasions when she wanted to keep something secret from her monolingual son, and she had not identified herself as a Gaelic speaker on the national census form (Dorian 1981: 109 and MS.). Altogether untestable of course are the assertions of individuals in terminal speech communities about the “inner speech” of their daily lives. While Melchers cites the experience of the Faroese scholar J. Jakobsen investigating the Norse-derived speech of the Shetlands late in the 19th century, she certainly received one highly contradictory piece of testimony herself: Jakobsen was told that the ‘terminal speaker’ of Norn was a native of Unst who died about 1850, which he found credible, whereas reports of later speakers should be regarded with scepticism. These speakers probably used exceptionally many Norn words but did not actually speak the language. However, during my own visit to Shetland in 1979, a native of Yell told me that he ‘did all his thinking in Norwegian’! (Melchers 1981: 256). Ultimately the most severe difficulties confronting the student of an obsolescing speech form surely lie in obtaining adequate speech samples from whatever source-people may still be available and then in assessing the reliability of the samples both in terms of their reflection of the speaker’s abilities and their reflection of his or her actual usage – and, one might add, in convincing one’s professional colleagues that one has done so at a standard of work which renders one’s conclusions respectable. The investigator’s own ethnicity and linguistic skills clearly play a role here, as do less specifiable personal charaeteristics which facilitate or hinder rapport. Tsitsipis is himself an ethnic Greek and it is precisely in relation to the Greeks and their language that Arvanitika speakers denigrate their language. Gal is a native speaker of standard Hungarian, whereas the form of Hungarian with which she needed to deal in the Austrian border region of Oberwart is a marked regional dialect. Jane and Kenneth Hill, as native English speakers, had to work with two foreign languages in order to study Nahuatl, since the superordinate language of Mexico is of course Spanish. Schmidt, as a white Australian, had the barrier of skin color to contend with in working with the Dyirbal. Working methods can and must vary according to what is sought, of course. In the study which led to their conclusions on vocabulary replacement in Nahuatl, the Hills used a Nahuatl-speaking assistant who conducted taperecorded interviews ranging over conversation, storytelling, translation tasks,

380

chapter 20

questions on language attitudes, and the l00–word lexicostatistical test. The assistant’s cooperation, crucial to the work, nonetheless resulted in some accommodation to his wishes: because he was going to have to transcribe each of the language-attitude interviews for his employers, but not the word-list results (for which they wanted close phonetic detail), he insisted on doing the former before the latter so that he would not have to wind through the wordlist test each time in order to get to the language-attitude section. The original idea had been to have the language-attitude section last, so as to avoid an effect which the Hills suspect did appear: Since the 100-word list was administered as the last item in the interview, after the language-attitude questions, speakers were often particularly concerned to give ‘correct’ . . . responses for the Spanish test items. Thus, many of the ‘missed’ items would doubtless be hispanisms in everyday speech (1981: 216). Gal, given her interest in shift between Hungarian and German and between levels of style within each code, used a variety of strategies to elicit speech variety. Her own identity as a native speaker of standard Hungarian made use of Hungarian to her by bilinguals a matter of courtesy, but she was able to counter this effect by taking advantage or norms of local accommodation to monolingual German speakers and reinterviewing many of the same people with a local monolingual German student along (1979: 66). She was also able, during the second half-year of her residence in Oberwart, to obtain through the close involvement of her participant observation, permission to tape ongoing daily activities such as card-playing, dinner conversation, and housework within eight households – sometimes even in her own absence, with the taperecorder left running. Whatever the effect of the tape-recorder, the relative freedom of this style of recording (as opposed to direct interviewing) did produce the hoped-for wider range of style variation (op. cit.: 67). Subsequent visits to the same community resulted in still further removal of the constraints of formal interviewing as her status as what might be called an “inside outsider” developed; young people were even willing to stand in for her, in a sense, by taking her tape-recorder along in visits to relatives and friends, and she was increasingly able simply to be present during conversations which did not focus on her or even always direetly include her (MS.: 13). My own greatest assets in East Sutherland were firstly that I was not British, so that in spite of having English as my mother tongue I could not be placed anywhere within the British class system; and secondly that I had never studied Scottish Gaelic at all, but rather learned it in East Sutherland by linguistic

gathering language data in terminal speech communities

381

field techniques, so that my Gaelic was the same as that of my sources. I had learned a bit of Inner Hebridean Gaelic before my arrival in East Sutherland, but it was quickly and permanently displaced by the local variety. And since the local Gaelic was extremely distinctive, truly a badge of local identity, I became something of an honorary East Sutherlander simply by becoming a speaker. Hill and Hill (1981), Gal (1979 and MS.), Schmidt (1983), Ralph Cooley with the Delaware of Andarko, Oklahoma (Yoder and Cooley MS.), and I, among others, all worked in settings where more than one speaker was often present, yet where the degree of “naturalism” might still vary markedly and the suitability of the multiple-source technique might also be more or less great according to the particular purpose of the session(s). Hill and Hill found interviews with groups of people highly desirable for the parts of their interview which were aimed at eliciting casual speech, but distracting in the lexicostatistical word-list section, where as they report: Slower respondents (particularly elderly men) often found their role usurped by quicker members of the group, often by impatient wives! (217) Gal’s ability just to “hang around”, after her identity and welcome in the community were well established, was a crucial counterbalance to her nativespeaker use of standard Hungarian. Schmidt was able to overcome the skin color barrier after about two months and gain acceptance within two groups of female young people; but it proved impossible for her as a young white woman to record male speakers of Young People’s Dyirbal in a natural context because her very presence would prompt a switch to English. Among the young women she found that the popularity and general use of cassette recorders made recording much less problematic than she had anticipated, and she carried her recorder in a shoulder bag on group activities such as camping and fishing trips; during story-telling sessions around the camp-fire speakers were not even necessarily aware of whether the recorder was on or off (1983: 185–86). Like Hill and Hill, I found that some types of elicitation work were rendered useless where one of a pair (or group) of source people was particularly eager and was consequently unable to refrain from breaking in with a response to a question directed to someone else. In one especially disappointing case I had to omit a conservative older fluent speaker from a select sample in which I would very much have liked to include him because of the impossibility of arranging elicitation sessions without a quicker and more eager member of his household present. At the same time, like Gal, I found that the many occasions and settings in which I was simply present and able to listen, or sometimes even make notes, without affecting the Gaelic conversation going on

382

chapter 20

in lively fashion around me were a major factor in my ability to speak with some authority about spontaneous as opposed to elicited speech and of course to develop a feel for style-shifting, range of vocabulary, the relative rarity of certain morphological or syntactic structures, and sociolinguistic norms for a variety of verbal behaviors. Cooley was able to develop a highly fruitful field technique by building the interactions – including possible interruptions or contradictions – of his six Delaware speakers into his elicitation sessions. He introduced a story told onto tape by a single one of the six speakers as materials for group discussion, playing it over for the assembled group, all of whom were related to each other and all of whom were fluent, but with considerable variation in motivation and sophistication when it came to the task in hand. He alone had a transcription of the text being played. Working with the tape (for the Delaware speakers) and the transcription-plus-tape (for himself as nominal leader), he prompted the group through a line-by-line review of pronunciation and of wordtranslation with discussion of meanings. The sessions were favorably regarded as important social occasions by the six speakers, and they were paid for their work, in addition. Cooley reports that participation was eager, and he gained material of several important types: close textual analysis and detailed commentary on pronunciation, grammar, and semantics, of course, but also a good deal of general conversation both in English and in Delaware, increasing in spontaneity over time (Yoder and Cooley MS.: 13–15). How valuable this must have been as material for study I can gauge by my own frustration over inability to make use of a one-time similar opportunity among the secular Pennsylvania Dutch a good many years ago. A middle-aged speaker of fairly good language skills and especially intrepid and gregarious personality chose to respond to my request for a translation of a story deliberately larded with a variety of potentially interesting constructions by convening a group of friends to join in the activity. Because the translation was intended for comparison with other German dialect versions of the same story which I was using for instructional purposes in a course on the history of the German language, the finished product was supposed to be a fairly smooth and fluent rendering. Knowing the purpose of the tape, the speaker in charge of the assembly required me to stop the tape-recorder between each line of the group’s mutually ratified translation while they discussed the best version of the next sentence. I ground my teeth over the lost opportunity to record the discussion of the translation, including frequent corrections from a particularly fluent and conservative member of the group and a good deal of allowable variability of rendering which they mulled over among themselves. If I had known that the group approach would be taken, I would have tried to

gathering language data in terminal speech communities

383

evolve a strategy by which to capture both the discussion and the translation; but my chosen informant was in effect in charge, with my own role reduced to that of recording engineer, whereas Cooley continued to co-preside over his group sessions by starting with a single-speaker text and recruiting the group of friendly relatives as discussants. He was justifiably pleased with his results, and indeed his method might very profitably be adopted by other fieldworkers, assuming that a similarly friendly and cooperative group of “discussants” could be assembled, each of whom would be agreeable to minute dissection of a text which he or she had provided in a prior one-to-one setting with the investigator. In direct elicitation, whether with single or multiple source-people, special problems are likely to arise. The most obvious are anxiety on the part of the source people and artificiality in the responses. Much of the investigator’s success in overcoming these obstacles will depend on his or her personality and manner, in the first case, and on his or her ability to gain enough exposure to spontaneous speech so as to spot artificiality in elicitation and reject it, in the second. Since it is difficult to speak for others in these respects, and the problems are seldom fully and explicilty discussed in the professional literature, I will deal here with my own field experiences. They were rather different among the East Sutherland Gaels and among the Berks County Pennsylvania Dutch, which at least makes one direct comparison possible. In my East Sutherland work I was fairly young when I began, still a graduate student and a mere youngster to most of the rather elderly population with which I was to work. For my own part I had grown up in a household with resident grandparents and felt at ease in dealing with the elderly. I tended to like them, and they me; the tradition of Highland hospitality helped, most definitely, but the acceptance went beyond it. My “target population” was strongly stigmatized, but I had not been socialized into such attitudes, of course; my own personal bias was an anti-aristocracy and anti-gentry one, which only endeared me to the ordinary people, who by and large held the same opinions. Serendipity, so rightly invoked in the title of Miller’s 1971 paper, played its usual prominent role. My place in one or two segments of the community at large was aided by two accidental factors. A local family (the parents incomers but long-time residents), one of whose daughters was widely acquainted in Gaelic-promoting circles because of her profession, had lost another daughter to disease only a few years before, and by their account I resembled the deceased daughter. My friendship with this family became close and longstanding, almost certainly aided by the accident of physical resemblance to the much mourned daughter. (See Yoors [1967: 76–82] for an account of his sudden total acceptance by a Romani family when his close companion, their

384

chapter 20

son, was killed in a traffic accident.) In another incomer family, west coast Gaelic speakers even more widely acquainted in Gaelic circles, my acceptance was probably hastened by the fact that I was very musical: they took responsibility for training me, as they did so many others, to take an active role in the Gaelic musical life of the district. Music was also an initial passport to acceptance and eventual sponsorship in another key family, one of whom was an East Sutherland Gaelic speaker, though of crofting rather than fisherfolk background; he held a position of authority in the village richest in local Gaelic speakers and paved the way for my work there through his personal influence. He was himself a Gaelic singer, and music formed the basis of our friendship initially. In their book People Studying People, Georges and Jones (1980: 50–51) discuss in somewhat unfavorable terms scholars who pass themselves off as “students” instead of stating outright that they are fieldworkers, researchers, anthropologists, linguists, folklorists, or what-have-you. I found that in stating that I wanted to learn the local language and write a study of its sound system I was automatically assigned that label by my potential sources; although I never myself used the word, they would speak of my coming for my “lessons” long after my sophistication in the local language had reached a point where I was able to posit the probable existence of complex structures I had never spontaneously heard and actively plan strategies to elicit them. Serendipity again came to my aid in carrying out direct e1icitation tests. All of my sources had television sets, and quiz shows were both popular and frequent in the programming. These quiz shows were low-key and decorous affairs by the standards of American television: contestants pushed buzzers to signal their preparedness to tackle a question, and the contestants as well as the audiences were quietly excited and eager, with delight at correct answers expressed in a restrained if readily apparent way. As my rapidly increasing number of source people grew steadily more accustomed to the task of providing on demand translations in Gaelic for English sentences, I found that person after person likened him- or herself to one of these quiz show contestants, often expressing an actual wish for a buzzer to push because the answers to my questions (as opposed to those on the TV shows) were easily come by and would have made the respondent feel like “the brain of Britain” (after the title of one of the shows). Translation from English to Gaelic (or vice versa) was a common activity in any case, because kinship and friendship networks typically included both bilinguals and monolinguals and a remark made or heard in one language would very often be repeated later in the day in the other language for an interlocutor with whom the original language of the remark was not the one regularly used. It took very little effort on my part to make expert

gathering language data in terminal speech communities

385

informants of the bilinguals with whom I worked, since they were already skilled translators and the role of the TV quiz show contestant was one with which they so readily identified. My skills as an elicitor were put to the test primarily in gauging which informants could be pushed for more conservative or elaborate alternative structures without becoming confused or discouraged, and in finding ways to couch my test sentences so that they would be most likely to produce the desired results. Among the secular (non-Anabaptist) Pennsylvania Dutch the situation was quite different. Clannishness was just as great, but hospitality to outsiders was a less prominent ethnic value. I had entree to the community in question through the family of a former student, and my network of speakers was developed by exploiting that connection. I wanted to work with multiple-generation families rather than with geographically-bounded or occupationally-defined groups, and therefore I needed the cooperation of entire large farnilies. The assent of at least one or two strong-willed central figures in the kin network was the key to success. Since my original entree was through a former student and I was more than a decade older than when I began my East Sutherland work, I had my status as a “professor” to reckon with. Before actually meeting me, one or two people were quite alarmed at the prospect of being interviewed by me, and in one of these cases the individual was an elderly woman with a heart condition who had to be carefully prepared for my appearance. She was too anxious to use her faulty English, and I knew that my standard German would be off-putting, since it represented the language of church services. I finally interviewed her in her farmhouse kitchen, a daughter-in-law and a granddaughter-in-law comfortingly present and preparations for cooking going on as a partial diversion, and our interview was “macaronic”: I asked my questions in English and she answered them in Pennsylvania Dutch. She never spoke any English at all during my visit. By contrast, an equally elderly woman of strong character and high intelligence marshalled her entire vast family for me and virtually orchestrated their cooperation. They became so reconciled to my activities among them that I was encouraged to attend a huge wedding anniversary celebration at which the entire family would be assembled and commandeer one by one any strays whom I had not yet had a chance to interview. Only one grandson notorious for exceptional shyness managed to resist the family pressure for participation in the undertaking. Again I found that direct elicitation und translation tasks were no obstacle (although this is most definitely not the experience of all investigators of obsolescing speech forms: Mohan [MS: 54–55] found translation-style elicitation utterly counterproductive). I did not, however, attempt, as I did in East Sutherland, to acquire control of the local dialect. Knowledge of standard

386

chapter 20

German, together with plenty of experience of the local dialect, was perfectly adequate for fieldwork, and outsiders who had acquired Pennsylvania Dutch were a much comrnoner phenomenon, if not in this day, then in well remembered times, so that becoming a speaker of Berks County Pennsylvania Dutch would not have gained me the sort of instant quasi-local identity that the acquisition of East Sutherland Gaelic guaranteed. Nor is Berks County Pennsylvania Dutch as different, relative to other varieties of Pennsylvania Dutch, as East Sutherland Gaelic is, relative to other varieties of Scottish Gaelic; this again reduces its value as a marker of distinctive “insider” identity. I did, on the other hand, do fairly extensive pilot studies with family members and family friends of my former student, so that I had quite a good understanding of the features of the local dialect before I began my more focussed study with the two large multi-generational networks. I also spent much more time and effort on explaining what it was that I was doing, since a number of the middle-aged members of these networks were well educated and in a good position to understand the purpose of the study. No one ever referred to my interviews or elicitation sessions as “lessons”, under these circumstances, and of course I was not actively acquiring the dialect in any ease. I also never moved into residence in Berks County, but rather made day trips to the area, although I was welcome to stay overnight with family or relatives of my former student and did so on one occasion. The networks I was concentrating on were somewhat dispersed, in any case, and it would not have been a simple matter, as it was in East Sutherland, to choose a location central to the study. The particularly strong-willed matriarch in fact effectively summoned home the more outlying members of her vast family for special times, so that I could interview them at her house. Given the size of her family, I could not possibly have completed the study without her active interest and cooperation, although several of her children also took an active role in recruiting siblings or their own children and in providing settings for interviewing. Although the published literature on language death and language obsolescence deals relatively little with the nitty-gritty of field settings and field methods, some striking differences in working conditions can be discerned. One is the ability or inability to work in a community without sponsorship. I went to East Sutherland because the Linguistic Survey or Scotland asked me to, but I had no sponsors whatever in the area itself and was warned that I might not find enough speakers to work with, or gain acceptance among any I did find and might consequently have to change locations. I found it exceedingly difficult psychologically to tackle a strange community without any advance source of entree, but it nonetheless proved perfectly possible in fairly short order to gain entree without the sponsorship of someone of standing among

gathering language data in terminal speech communities

387

the local population, or, alternatively, of someone with blood connection to them. In both areas where I worked it was possible to recruit local people as “language coaches” who would teach me their language, although I actually undertook this in one of the areas and not in the other. (Literacy in the mother tongue is in no way necessary to the role of “language coach”; very few of my sources in either community were literate in the home language.) There are certainly communities where admitting an outsider to knowledge of the local language may be fiercely resisted, and there are also languages of such difficulty that the investigator’s limited time is better spent in training a gifted bilingual to act as an aide than in trying to master the language for him- or herself. Both direct elicitation and simple “involved presence” in the participantobserver style which permits naturalistic observation were possible among the two populations I dealt with. In some areas one or the other may be possible, but not both. The fieldworker may – sometimes for a fixed wage – be permitted official interviews, but rigidly excluded from informal settings, which are considered out of bounds for outsiders. Alternatively the fieldworker may be relatively welcome as a participant in community activities, especially if he or she makes him- or herself useful in laborious group activities such as work in gardens and fields or work in the preparation of large meals: yet one-to-one formal elicitation sessions may be precluded by shyness, anxiety, inability to cope with the strangeness of the activity, lack of leisure time on the part of overworked potential sources, embarrassment at imperfect command of the language, failure of memory because of the unreality of the situation, and the like. In the settings in which I worked, especially East Sutherland, published accounts which identify individuals are carefully to be avoided. The communities are small, gossip is rife, horror of becoming the subject of public comment strong. One woman, though a loyal friend of 20 years, steadfastly refused to allow me to donate to the local-history collection at one of the area libraries a copy of one of my papers which celebrated the special trove of local lore shared with me by two of her siblings: she feared the community would think the family boastful. I was told by a colleague that one of my books would have been more satisfying if I had allowed the individuals quoted in it to come to life more fully and had identified them for the reader. Not if I ever wanted to work in the area again, I replied. Obviously if a speech forn is dying, some of the terminal speakers will have limited skills and it will be necessary to specify the limitations. Given the usual negative local attitudes towards the minority language, and often towards its speakers as well, the anonymity of the sources had best be guaranteed as well as possible. The community will guess at

388

chapter 20

identities anyway, and experts that they are in local matters, they will usually he right; but at least the author will have respected the native reluctance to name names. On the whole anthropologists have been quicker than linguists to recognize the importance of presenting information on the conditions of their work, the obstacles they encountered, and the reasons (voluntary or involuntary) for the choices they made among sites, sources, and methods. In an era when even the “hard” sciences are increasingly acknowledged to be strongly affected by the human element (see for example Broad and Wade’s volume Betrayers of the Truth: Fraud and Deceit in the Halls of Science [1982) for some of the negative aspects of the situation, and Thomas’ charming chapter on “Endotoxin” in his book The Youngest Science: Notes of a Medicine Watcher [1983] for some of the positive aspects), linguists would do well to follow this lead for the light it is certain to shed on the craft and skill with which they ply their profession and equally on the obstacles and limitations they face in their efforts to advance the boundaries of their discipline. In the sensitive realm of language death studies this may be more difficult to accomplish than usual, but the attempt must be made all the same. This paper is intended among other things as a contribution to such an undertaking, however restricted in scope. References Broad, William, and Nicholas Wade. 1982. Betrayers of the truth: Fraud and deceit in the halls of science. New York: Simon and Schuster. Denison, Norman. 1971. Some observations on language variety and plurilingualism. In Social anthropology and language, ed. by Edwin Ardener, 157–83. London: Tavistock Publications. Dixon, R. M. W. MS. Reassigning underlying forms in Yidiɲ – a change during language death. Dorian, Nancy C. 1981. Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. Gal, Susan. 1979. Language shift: Social determinants of linguistic change in bilingual Austria. New York: Academic Press. ——. MS. Phonological style in bilingualism: The interaction of structure and use. To appear in Meaning, form, and use in context, ed. by. D. Schiffrin. Georgetown: Georgetown University Press. Gambhir, Surendra. 1983. Diglossia in dying languages: A case study of Guyanese Bhojpuri and standard Hindi. Anthropological linguistics 25: 28–38. Georges, Robert A., and Michael O. Jones. 1980. People studying people: The human element in fieldwork. Berkeley: University of California Press.

gathering language data in terminal speech communities

389

Haas, Mary R. n.d. Tunica. Handbook of American Indian languages 4. New York: J. J. Augustin. ——. 1968. The last words of Biloxi. International Journal of American Linguistics 34: 77–84. Hill, Jane, and Kenneth Hill. 1977. Language death and relexification in Tlaxcalan Nahuatl. International Journal of the Sociology of Language 12: 55–69. ——. 1981. Regularities in vocabulary replacement in Modern Nahuatl. International Journal of American Linguistics 47: 215–26. Huffines, Marion Lois. 1980. Pennsylvania German: Maintenance and shift. International Journal of the Sociology of Language 25: 43–57. Melchers, Gunnel. 1981. The Norn element in Shetland dialect today: A case of ‘neveraccepted’ language death. In Tvåspråkighet, ed. by Eva Ejerhed and Inger Henrysson, 254–61. Umeå: Umeå University. Miller, Wick R. 1971. The death of language or serendipity among the Shoshoni. Anthropological linguistics 13: 114–20. Mithun, Marianne, and Reginald Henry. MS. Notes on moribundity: Incipient obsolescence of Oklahoma Iroquois. Mohan, Peggy. MS. A language implodes: The death of Trinidad Bhojpuri. Pulte, William. 1973. Cherokee: A flourishing or obsolescing language? In Language in many ways, ed. by William B. McCormack and Sol Wurm, 423–32. The Hague: Mouton. Salzman, Zdeněk. 1969. Salvage phonology of Gros Ventre (Atsina). International Journal of American Linguistics 35: 307–14. Schmidt, Annette. 1983. Young people’s Dyirbal: An example of language death from Australia. Australian National University Master’s thesis. Swadesh, Morris. 1948. Sociologic notes on obsolescent languages. International Journal of American Linguistics 14: 226–35. Thomas, Lewis. 1983. The youngest science: Notes of a medicine-watcher. New York: The Viking Press. Timm, Lenora A. 1980. Bilingualism, diglossia and language shift in Brittany. International Journal of the Sociology of Language 25: 29–41. ——. 1984. The segmental phonology of Carhaisien Breton. Zeitschrift für celtische Philologie 40: 118–92. Trudgill, Peter. 1976–77. Creolization in reverse: Reduction and simplification in the Albanian dialects of Greece. Transactions of the Philological Society 1976–77: 32–50. Ttsitsipis, Lukas D. 1983. Language shift among the Albanian speakers of Greece. Anthropological linguistics 25: 288–308. Voegelin, C. F. and F. M. Voegelin. 1977. Is Tübatulabal de-acquisition relevant to theories of language acquisition? International Journal of American Linguistics 43: 333–38.

390

chapter 20

Waterman, John T. 1966. A history of the German language. Seattle: University of Washington Press. Yoder, Cecelia K., and Ralph E. Cooley. MS. Ethnomethodology: A psycholinguistic perspective. Yoors, Jan. 1967. The Gypsies. New York: Simon and Schuster.

chapter 21

Surprises in Sutherland: Linguistic Variability amidst Social Uniformity The last thing I would have expected to find in populations of exceptional social uniformity is marked individual linguistic variability, but that’s exactly what I did find in long-term fieldwork with Scottish Gaelic in Sutherland, in the far north of mainland Scotland. I didn’t originally set out to go to Sutherland, and I wasn’t in search of linguistic variability when I arrived there. En route to explaining how Sutherland became my research site and what I found in it, I propose to look at some general issues in field research: What entices a student linguist into the field? How usefully can a research project be focused before the researcher is personally familiar with the field site? When is a fieldwork project “finished”? And finally, how do the professional and the personal experiences of fieldwork conflict or balance? 1

The Library or the Field

A student making routine progress through an academic program volunteers for some discomfort in leaving the familiar academic environment for a fieldwork setting. Entering an unfamiliar social world is guaranteed to plunge the novice researcher into something like a second adolescence: a constant succession of uncomfortable situations in which he or she has no clear idea how to behave and is very likely to behave inappropriately. There must be some substantial inducements to coax the student forth, as of course there are: the excitements of novelty and discovery, and the satisfactions of making a first real trial of professional skills. I don’t recall any explicit discussions during my graduate years, either among students or between graduate students and faculty, about the importance or advisability of undertaking fieldwork as opposed to library research for a dissertation project. Students decided for themselves whether their interests and values made fieldwork attractive, and if so, whether their personal circumstances allowed them to go off to a field site for a year or more. Although the Department of Linguistics at the University of Michigan, where I was studying, didn’t deliberately cultivate a sense of professional mission about undertaking fieldwork, several factors kept the possibility always before us. A linguistic field © koninklijke brill nv, leiden, ���4 | doi 10.1163/9789004261938_023

392

chapter 21

methods course was offered during the summer sessions, and a number of faculty members were themselves either experienced fieldworkers or well-trained Middle English lexicographers who also needed to exercise careful patience in amassing and accounting for primary data. Among the senior faculty, Kenneth Pike was a major presence, and his prowess as a fieldworker was legendary. His occasional “monolingual demonstrations” made a vivid impression on all of us. A speaker of some language unknown to Pike was produced, sworn to speaking his or her mother tongue exclusively, while Pike, armed with a few props such as two sticks and a leaf, asked questions using only Mixteco (a Mexican Indian language he spoke fluently) and miming. Pike wrote everything the speaker of language X said on a blackboard and after half an hour performed an instant grammatical analysis on the material. This was awe-inspiring to watch and no doubt created a certain fieldwork mystique among linguistics graduate students. During my graduate studies Old English had been a delight to me. Here was English as I thought it ought to be, a fully Germanic tongue without the overlay of Romance and Latinate vocabulary that seemed chiefly to serve the causes of euphemism and hypocrisy (“prevaricate” indeed, if the lady had lied!). But somehow a dissertation on Old English struck me as an improper use of my training. I’d been given to understand that a linguist could use the field method techniques we had been taught anywhere, with any language. And since the techniques could be applied anywhere at all, why not go where my interest was highest and try them out on a language I’d wanted to learn more about since childhood, namely Scottish Gaelic? 2

Somewhere Ho!

Good advice is a boon when you’re contemplating fieldwork and the problem of funding it. Mine came from Eric Hamp, famed Celtic scholar at the University of Chicago, who suggested that I link my fieldwork to the needs of the Gaelic Division of the Linguistic Survey of Scotland by offering to write a dissertation on whatever dialect the Survey director considered most in need of study. Hamp predicted that funding sources would see a study linked to an established project as well-focused and worthy of support, as one soon did. I was content to go wherever the Survey directed me and work on whatever project they proposed, since I was off to the country of my choice to work on my top-choice language. In correspondence the Survey director had indicated that a phonological study of the Gaelic spoken in any one of three different Highland locations

surprises in sutherland

393

would be highly suitable from the Survey’s point of view. He seemed at the time to be leaving the final choice to me, but soon after my arrival in Edinburgh he handed me a list of names and wished me well for my work in eastern Sutherland – not the location I had mentally picked for myself among the three, and a daunting distance away on the map, almost as far to the north as one could go without dropping off the mainland altogether. At the time I wasn’t familiar enough with Gaelic dialectology to understand the motivation for the director’s choice. He, however, knew that the Gaelic spoken by the fisherfolk of eastern Sutherland was a dialect of the extreme periphery, and that in classical fashion it differed notably from more central dialects. Furthermore, it was certainly understudied. The speakers whose names appeared on the director’s list had provided answers to a vast questionnaire made up almost entirely of isolated lexical items used by the Survey to track historical phonological development across the whole of Gaelic-speaking Scotland. The questionnaire was well designed for its limited purpose, and the fieldworkers were skilled at their jobs, but most of Scotland’s local Gaelic dialects were otherwise poorly known, especially those of the northern and eastern mainland. The director’s final words to me reflected the extreme scarcity of solid information about the Gaelic of East Sutherland in the early 1960s. He warned me that I might find no speakers left in the three coastal fishing communities he was sending me to, in which case I was to come back and he would give me another assignment; and he urged me to find out, if I did locate speakers, whether it was really true that the Gaelic of eastern Sutherland lacked preaspiration of voiceless stops and affricates (preaspiration being a striking phonological feature of most Scottish Gaelic dialects). Far from finding no one to work with, I soon had an informal census of local Gaelic speakers running to more than 200 people, and the absence of preaspiration, so difficult for the director to credit, proved to be one of the most obvious general features of the whole dialect area, with implications for other parts of the phonological system. The 200 or so local Gaelic speakers still available in East Sutherland did not include many of the people whose names were on the Survey’s list. Survey fieldworkers had moved through eastern Sutherland in 1953 and 1957, and most of the elderly speakers who had served as their sources, or had been mentioned to them as possible additional sources, had died before I reached the area in 1963. After one man who had survived turned out to be lively enough at 86 to make it advisable to keep a table between us at all times, I abandoned the Survey list and searched out my own sources. Whether I was relying on Survey sources or not, my work was still necessarily tied to the Survey’s interests. My funding had been granted on the

394

chapter 21

understanding that I would target my research to their needs, and beyond that they had given me the use of a Survey van for the year. Sutherland has been thinly populated since early in the nineteenth century, when most of the tenantry of the great Highland estates were summarily evicted, often with conspicuous brutality, in order to “clear” the land for sheep farms. There was some distance between villages, and because of the low population public transport was much scantier than is typical of most of Britain. It was impossible to get from one village to another and back again on the same day without private transport, and since the Survey wanted me to cover three villages, the van was quite simply a necessity. Probably the three-village assignment should have alerted me to complexities lying in wait for me in Sutherland, but I only supposed that the Survey director thought I might have to comb through three villages to find enough people to work with. Once I was on location the inconvenient truth of the matter broke over me very quickly: my work was not the relatively simple job of describing a uniform fisherfolk variety of eastern Sutherlandshire Gaelic, but the very much more complicated job of describing each of three slightly different local varieties of fisherfolk Gaelic, one for each village. In the standard field methods fashion that I had been taught, I began my work by eliciting commonplace vocabulary that was likely to be monosyllabic, or at least short. “What do you say for ‘garden’?” I asked. “/yεs/,” said an elderly lady in Brora, the northernmost village. “/l´εs/,” said her counterpart in Golspie, seven road-miles to the south. “/l´es/,” said a woman in Embo, ten road-miles south of Golspie. These were small enough differences, but there was worse to come. ‘Bone’ proved to be /khrẽ:ũ/ in Brora and Golspie, with plural /khrã:vən/; in Embo it was /khrãĩ:/ with plural /khrã:n/. Even when it came to a word as central to the lives of all these fisherfolk descendants as ‘sea’, they didn’t agree: the word was /mur/ in Brora and Embo, but /mwir/ in Golspie. Things were no better when I moved from single words to connected material. My original Brora and Golspie sources gave ‘if you don’t plant oats’ as /mər khur u khɔrkh/, but the equivalent in Embo was /mə khur u khɔrkh/ or /mə gur u khɔrkh/, using a different form of the conjunction ‘if . . . not’ (the first word in each example) and one with variable effects on the initial consonant of the following word. The fact that the Gaelic of these three fishing communities – so similar in their historical origins, so close to one another (especially by sea, once the chief communicative link), and so nearly identical in all economic and social aspects – differed in each locality had immediate consequences for my work. Every word or sentence I gathered had to be checked across all three villages, lest there prove to be local differences. And since there often were such differences, which then had to be checked for possible individual idiosyncrasies, it wasn’t good enough to have

surprises in sutherland

395

a single excellent source in each village. Three converging sources struck me as the minimum needed to confirm a form in any one village, meaning that I needed to consult nine people regularly. If any disagreement turned up, I would have to check with still more speakers. This was a formidable prospect. I had the usual year, more or less, for my field research, conceivably adequate for detailing the phonology of one local dialect, but I now found myself faced with three speech forms, clearly related (the Brora and Golspie forms particularly) but still distinct. And the phonologies, my special assignment, were indeed slightly different, not just in terms of about 300 lexical items which took a different phonological form in at least one of the three villages, but also in terms of phonological inventory and distribution of phones. In Brora and Embo, I enlarged my speaker sample by drawing in relatives of the sources I’d first located (with start-up help from speakers of nonlocal Gaelic dialects in Golspie and Embo, and from an English monolingual supportive of Gaelic causes in Brora). In Golspie, it was the non-local Gaelic speakers again who suggested potential additions to my speaker sample, and the new people happened not to be closely connected to my original pair of speakers. The difference in the way my Golspie speaker sample was enlarged, compared with the Brora and Embo samples, proved instructive. The value to community language studies of following out the natural lines of social networks is well recognized, thanks to the Belfast work of James and Lesley Milroy (Milroy 1980), but in Golspie I found that there were sometimes insights to be gained by working across the grain of social networks as well. My interconnected speaker-networks in Brora and Embo that first year were friendly enough to be largely uncritical of one another, whereas certain tensions within the cobbled-together Golspie sample were more revealing of local language attitudes. It was in Golspie, for example, that I first heard one Gaelic speaker criticize another for being “too proud” to speak Gaelic. The notion that “pride” could keep someone from speaking Gaelic suggested that Gaelic was a social liability in the local context, and so it was. “Gaelic-speaking” and “fisherfolk” had become synonymous, as the rest of the coastal population went over to English, and since fisherfolk origins implied poverty and bottom-rung social standing, some people of fisherfolk descent signaled a wish to distance themselves from their origins by declining to speak Gaelic. During my original fieldwork year, I occasionally encountered people said to be of fisherfolk descent and Gaelic-speaking who turned out not to be fully proficient speakers after all. Regretfully I crossed these interesting people off my list of potential sources. The Survey, like all dialect geography undertakings, was particular about its information sources. Speakers had to be strictly local, preferably elderly, and not too geographically or socially mobile, since

396

chapter 21

people of that description were the ones least likely to have been influenced by any non-local usages they might have been exposed to. Luckily for my future work, it proved difficult in East Sutherland to isolate the speakers who best met the Survey’s criteria from their usual well-peopled social contexts. I was working with them in their own homes, and in several households there were Gaelic-speaking spouses or siblings who were younger than the speaker I had specifically come to work with, plus occasionally a grown son or daughter who spoke some Gaelic. Answers of their own popped out eagerly from some of these others when I put questions to the older speaker. Being young, polite, and deeply grateful to all the families who let me into their homes and tolerated my interminable questions, I considered it proper to write down whatever was offered. So I recorded these extraneous responses, too, and found myself confronted yet again by uncomfortably diverse data. I wasn’t getting reliably identical responses, even though my sources in these cases were not just from the same village but from the same household. The material from younger family members didn’t find much place in my dissertation, since that document was also in effect my report to the Survey, and Survey standards excluded material from such sources. But it was in my notebooks, as was a small amount of material from the very few elderly Gaelicspeaking crofters (sub-subsistence agriculturalists) whom I unearthed in the rural districts round about the three villages when the Survey director handed me another assignment: gathering Gaelic place names for the place-name specialists of the School of Scottish Studies in Edinburgh. Supplied with bundles of oversized map segments, I quickly covered the areas that were well known to my fisherfolk friends, after which I dutifully headed off into the countryside to follow up on uncertain reports of occasional elderly crofters who still spoke Gaelic. I found three, each one the lone surviving Gaelic speaker of his district, and with their help dotted the highly detailed maps with Gaelic names for cleft, knoll, hillside, rivulet, and so forth: ancient indigenous place names certain to be lost all too soon. I worked on a bit with one crofter after the placename task was complete, enjoying the visits to his particularly pleasant family and intrigued by the obvious small differences between his Gaelic and that of the nearest fisherfolk. I was interested, for example, in a number of initial consonant clusters with a prominent bilabial second element (as in the Golspie word /mwir/ ‘sea’) that were typical of the Gaelic of the fisherfolk communities. I knew these to be unusual in terms of western Gaelic dialects, and now I found that they were absent even in the crofter Gaelic once spoken very near at hand. Tying my fieldwork to the Linguistic Survey of Scotland’s interests had a good many consequences. Fieldwork funding and the loan of a car were obvi-

surprises in sutherland

397

ous advantages, as was the access I was given to spectrographic equipment at the University of Edinburgh between academic semesters. For a long while, however, I considered having been set to work on the Gaelic of three different villages a disadvantage, leading me to devote too much time to cross-checking material and not enough to exploring any one variety in real depth. Seemingly unrelated extra assignments, like the highly detailed place-name work, coming my way because I was an available fieldworker in a little known region, had also taken time and attention away from my work with fisherfolk Gaelic, however interesting my brushes with crofter Gaelic had been. 3

The Dissertation Is Done, but Am I?

It doesn’t seem to be necessary to like the people one is studying very much in order to do productive fieldwork. When Malinowski’s diaries were published posthumously (1967), it appeared that he had not had a great liking or respect for the Trobriand Islanders (Van Maanen 1988: 36), and Erving Goffman told me, when I had a chance once to ask him directly, that he had not particularly liked the Shetland Islanders about whom he wrote so illuminatingly in The Presentation of Self in Everyday Life (1959). But liking the people you work with, as I did, certainly makes the fieldwork experience more enjoyable, and for some personalities and casts of mind it may be an important factor in determining the course of future research, since it enhances the appeal of returning to work in depth in a particular fieldwork site. Like many another sojourner in the Highlands, I was astonished by the generosity of people in whose midst I appeared as an unannounced stranger. People were often slow to believe that I could be interested in their local Gaelic, since they had heard nothing but negative comments about it all their lives, both from English monolinguals and from speakers of more conservative westerly Gaelic dialects. But once convinced of my interest, most people showed an almost unlimited willingness in helping me learn about it. Payment was out of the question, since the very mention of it proved offensive, and the small hostess gifts that I learned were acceptable at each of my visits, seemed completely inadequate thanks to people who were giving up whole afternoons or evenings to answering my questions and were regularly pressing great quantities of tea and baked goods on me besides. Even after I left that first year, five of my sources carried on answering my questions, putting long lists of phonologically relevant lexical items and short sentences onto tape for me so that I could consult this material during the ten months that I had spectrographic equipment at my disposal while working on my dissertation. All of the tapes

398

chapter 21

that arrived proved to be the spoken equivalents of letters as well, with added messages giving me news and good wishes in Gaelic, and sometimes including general Gaelic chats among my friends. These were not people easily forgotten. Furthermore, I genuinely liked the East Sutherland variety of Gaelic from an aesthetic point of view, especially the Scandinavian-sounding tonality of its longest vowels and the unusual sonority of its many uninterrupted multivowel sequences. I was also acutely aware that my year’s fieldwork had been barely adequate even to the single task of describing the phonology of this distinctive and little-known Gaelic variety. I lived frugally while I worked on my dissertation, saved money from my fellowship, and left for Scotland again five days after defending the dissertation. Hard pressed though I’d occasionally felt, as I made my perpetual swings from village to village and fanned out into the countryside with the placename survey maps, my limited connections with the Linguistic Survey and the School of Scottish Studies taught me very quickly how precious and how fragile the store of human knowledge and experience among the dwindling Gaelic speakers of East Sutherland was. The material most coveted by the place-name experts, for example, was not the Gaelic place names of East Sutherland itself, but the far rarer Gaelic place names the fisherfolk knew for ports farther down the east coast of Scotland, where Gaelic had not been spoken for centuries. The uniqueness of such knowledge, and the finality of this chance to capture it while some Gaelic-speaking fisherfolk still remained on the east coast, was impressed on me, and it stirred the incipient cultural conservator in me. Phonologically I had certainly encountered phenomena that were curiosities for a Scottish Gaelic dialect, from those initial consonant clusters with /w/ as second member to word-final geminate consonants in unstressed syllables (these last difficult to hear until I met a few of them before vowel-initial words within the same noun phrase). That is to say, I already had evidence, by the end of one year’s work, that East Sutherland Gaelic (ESG) was unusual in more respects than the absence of preaspiration, and I suspected that more surprises might come my way if I spent more time with the dialect. Another reason for my return trips to East Sutherland, in 1965 and after, was that I’d begun to feel a responsibility to document this unusual variety of Gaelic that clearly had a short life-expectancy; family transmission had ceased in this area, and there were no longer any young speakers. It didn’t hurt, either, that the place was beautiful (even if one could hardly say the same for the climate) and that most of the people I worked with personified a fieldworker’s dreams. Originally I worked mostly by elicitation, which by good fortune my sources found congenial and easy (not by any means always the case in fieldwork). My field methods training had stressed elicitation, but the conversational limi-

surprises in sutherland

399

tations of my Gaelic were a more important reason for relying on it. Asking people to produce stories addressed to a tape recorder felt uncomfortable in purely social terms. The tape recorder provided no social reinforcement, such as smiles at humorous bits or nods in response to rhetorical questions, and until I was comfortable enough in the language to supply these ordinary human responses while the story was in progress, I was reluctant to put a microphone in front of people and ask them to tell stories into it. Knowing I should gather texts to exemplify the Gaelic of the three villages, I had done a few taping sessions at the end of my first year, but with my halting Gaelic I found them extremely awkward. It wasn’t until 1967–68, when I was finally comfortable enough in the local Gaelic to make a reasonable conversational partner, that I did a more significant amount of taping; but by then I was working hard on grammar, which again made elicitation (translation tests) the technique of choice. I needed to cover a lot of grammatical territory, and since my sources had proved to handle elicitation with extraordinary ease and even with pleasure – several said it made them feel like the brainy, rapid-fire responders on a popular TV quiz show – elicitation was an efficient way to go about it. As it turned out, elicitation had an unanticipated benefit. The social context in East Sutherland, and especially in Embo, where there was a larger pool of speakers, continued to favor fluid work sessions with more than one family member present. I tended to ask for a good many examples of any structure I was exploring, and during the course of a session a variety of individuals might give their versions of a particular structure. Over time it became apparent that people closely connected with one another were far from unanimous about how certain grammatical niceties were to be expressed. Because I spent a fair amount of purely social time in some of these households, I also heard spontaneous usages that strengthened an impression of ongoing grammatical change in certain constructions. Eventually it seemed important to check on this, and I embarked on batteries of translation tests designed to elicit key constructions from across the widest age-range of speakers available. It also seemed useful to go back, as I did with pleasure and interest, to some of the imperfect speakers whom I’d been sorry to drop from my speaker sample earlier on. Most of the linguistic variation that was being investigated in the 1960s and 1970s was phonological, as in large part it still is today. In ESG, for whatever reason, there was relatively little phonological change in evidence, but a good deal of grammatical change was underway. I had been much impressed by Labovian studies demonstrating correlations between phonological change and social factors such as age, ethnicity, social class, and sex, so I looked long and hard at one clearly advancing phonological change, substitution of [ᶕ l] for the more traditional velarized lateral [ɫ]. But only age seemed to have any

400

chapter 21

bearing on how general the use of [ᶕ l] for [ɫ] became (some younger speakers in Embo having begun to use [ᶕ l] even in word-initial position). When it came to grammatical change, the same was true: age played a clear role in the extent to which an observably advancing change appeared, but no other social correlates emerged. The age differences led me to wonder whether, in excluding younger people as sources, descriptive linguists, who typically insisted on working only with the most traditional speakers, were missing an opportunity to find out just what sorts of changes might be likely to occur as a small and highly localized speech form went out of use. I made a point of enlarging my speaker sample again, this time in Embo, where Gaelic was still widely used and speakers ranged in age from the eighties to the low forties, or even to the upper thirties, if I included some individuals who spoke Gaelic imperfectly with certain older relatives. The results of translation tests presented to Embo’s broad age-range of speakers showed, among other things, that case distinctions were progressively weakening and that one traditional form of the passive was being abandoned (though it was leaving its trace in changes introduced into the other traditional passive (Dorian 1973)). Certainly it was gratifying to find the sort of age-graded changes I’d anticipated when I started probing for these and other grammatical changes. But I was struck, at the same time, by the moderation of many of the changes I looked at. Gender signaling via pronoun reference, for example, was notably weakening, but Gaelic has a number of gender-signaling devices and one or two of the others weren’t showing comparable weakening. It was true that a particularly conservative passive construction was fading out of use, but the passive itself was still fully expressible in ESG, even among the stronger of the imperfect speakers. The hyperabundance of plural and gerund allomorphs in ESG was diminishing, but it wasn’t anywhere near the logical extreme of one universally applied suffix, either for plural or for gerund. It was very far from it in fact: even the imperfect speakers still showed plenty of variety in each case (Dorian 1978b). The limited nature of grammatical “decay” in ESG, even with the dialect’s ultimate extinction in sight, seemed to me as significant a finding as the presence of age-related grammatical change, and I tried to give it equal attention.1 1 In retrospect this seems even more important than it did at the time, since three linguists working with geographically and structurally very different languages have lately found striking evidence of grammatical elaboration among the final speakers of obsolescent languages: Rob Pensalfini (1999) in Jingulu, an Australian Aboriginal language; Alexandra Aikhenvald (in press) in Tariana, an Arawakan language of the Brazilian Amazon; and Silvia Dal Negro (1998) in Pomattertitsch, a Walser dialect of northern Italy. Obsolescence processes clearly needn’t be an unremitting progression into collapse and decay.

surprises in sutherland

401

When I finally felt more or less prepared to write a descriptive monograph on ESG, the training I’d had at the hands of those meticulous Middle English lexicographers at the University of Michigan came into play. I had depended on an unusually large number of sources in my ESG work, originally because the three-village assignment made it unavoidable, then also because sad losses among my elderly early sources made me seek out new speakers as the work went on, and finally because exploring grammatical change called for comparison across as broad an age-range of speakers as possible. It had become obvious long since that no single entity existed that could be labeled “East Sutherland Gaelic” and described in uniform fashion. Not only were there differences from village to village, and from older speakers to younger, but much more awkwardly there were also differences within a single age-group in a single village, as there were among Embo speakers about whether the initial consonant of a following verb would or would not be voiced after /mə/ ‘if . . . not’, as in ‘if you don’t plant oats’ /mə khur u khɔrkh/ or /mə gur u khɔrkh/). Trained as I was to acknowledge differences, the descriptive monograph I eventually wrote, already laden with details about diverse usages because of geographically distinct variants, sprouted another layer of detail that recorded the dialect’s stubborn resistance to uniformity even within the bounds of any one village (Dorian 1978a). Given what I knew of the dialect by then, it would probably have been more difficult to ignore the untidiness and portray ESG in terms of some sort of ideal normalization, than to do as I did and describe the rampant lack of agreement. In the present half-century, the conventions of writing descriptive grammars have permitted reliance on a very small group of sources, or even, as was true of the last Scottish Gaelic dialect grammar produced before my own (Oftedal 1956), on a single highly intelligent and highly cooperative source. This practice reduces the likelihood that linguists will encounter markedly variable usage, or feel obliged to come to grips with it if they do. Oftedal, my immediate predecessor in Gaelic dialect studies, noted that the Gaelic of his single source and that of the man’s wife differed in a number of respects, despite the fact that the two had grown up as next-door neighbors; but after noting the existence of such differences in an early footnote, he never referred to the wife’s Gaelic again. Theoretical preoccupation with detecting the commonalities of universal grammar has meanwhile made it less likely than ever that descriptivists would be interested in pursuing evidence of individually differentiated usage, even if the differences should be of the rather striking sort that Oftedal encountered in the Hebridean dialect he was describing. In both traditional dialect geography and more recent correlational sociolinguistics, researchers have worked chiefly by multi-person single-interview survey, so that persistent differences in the usage of a single individual who is interacting with familiar

402

chapter 21

interlocutors have little or no opportunity to emerge. The level of individual variability I was describing for speakers in socially homogeneous villages such as Brora, Golspie, and Embo seemed unusual, consequently, and by my own account this variability was turning up in small-village speech varieties on their way to foreseeable extinction. Under these circumstances, then, it wasn’t surprising that even a highly knowledgable Gaelic dialect researcher, when reviewing my monograph, took the myriad details of variable usage noted for ESG as an indication of the dialect’s obsolescence (Ó Dochartaigh 1983). Reasonable though his conclusion seemed, I realized on reading it that obsolescence did not in fact provide an adequate explanation for what I had encountered and that the full range of ESG variability was still unaccounted for. Other large-scale projects intervened, and so unfortunately did severe health problems, but with what I trusted was the sort of dogged insistence on respecting the data that my lexicographer mentors would have smiled on, I turned back eventually to the unresolved issue of excessive variability in ESG. Gaelic was dying above all by transmission failure in East Sutherland, not by disuse among those who had grown up with it. When I began my work, Gaelic was still both the first language and the stronger language among a good many older people, and their ESG could reasonably represent the conservative norm for a number of instances of change in progress. But there was a large amount of variability in the dialect that didn’t seem to correlate particularly with age or proficiency differences, and was found in the Gaelic of older and younger speakers alike. For an investigation of the sort of inter-speaker and intra-speaker variability that I had become interested in, the former fishing communities of East Sutherland had some major advantages. Each bilingual group formed an unusually clearly demarcated population, for example. Despite some crossvillage marriages, the Gaelic speakers in each village had recognizably local ways of speaking and could be identified as producing Brora, Golspie, or Embo Gaelic. Living in small clusters of separate streets, as the fisherfolk had, and speaking in each case their own distinctive Gaelic (plus a somewhat distinctive English), the Gaelic speakers of each village formed as clear and unambiguous a speech community as one could hope to find. Their way of life had been locally unique and highly distinctive. Although the fishing industry had died away, all of the fluent bilinguals in my study (and even a number of the imperfect Gaelic speakers) had been deeply involved as children in the shore work that long-line fishing entails, such as gathering and preparing bait, baiting the hundreds of hooks, gathering fir cones for the fish-smoking process, and in the case of the girls, also doing some door-to-door fish selling. This meant that in the childhood years during which Gaelic emerged as the mother tongue,

surprises in sutherland

403

the speakers I worked with had experienced virtually identical social and economic conditions: all lived in a few densely populated streets, in houses of the same general structure and in households sharing identical labor patterns; all were poor and burdened by the same social stigma; all spoke Gaelic in the home; all came from a highly conservative Protestant religious background. Almost no one lived in the fisherfolk streets who did not fish for a living, and after the school years, finished by age 14 in nearly every case, contacts with non-fisherfolk were limited and almost entirely commercial. Even religious life was socially segregated, since there were separate services in Gaelic and in English, with the former attended chiefly by the fisherfolk. Variationist studies have long since demonstrated that the social features of large urban populations in particular, and even the generally smaller number of social distinctions within rural populations, find expression in significant patterns of similarity and difference in the use of phonological and grammatical features. I had gone to the Highlands expecting to find the same sort of phenomena there as well, yet years had passed and I had had nothing of this sort to report on. The very socioeconomic uniformity just described might play some role, of course, and if asked about my lack of findings that’s certainly what I would have pointed to. Yet there wasn’t any shortage of variability. Just the opposite, in fact – there was rampant variability. Faced with this problem, I realized that at last I stood to reap the rewards of the three-village assignment set me by the Survey. Because I had always worked in all three villages and had regularly documented their distinct usages, I knew the purely geographical dimension of ESG variability intimately. I could therefore subtract that form of variation, as well as the strongly age-related variation I had already looked at, and focus on the intra-village and intra-speaker variation that remained. I had recognized this sort of variation early, because it turned up among my sources in puzzling ways. Among my early sources, an Embo brother and sister were unusual in having no other siblings, and in both having married within the home village and lived there lifelong. They also happened to live in adjoining houses as adults and to have a good deal of daily contact. Yet although they claimed they had never noticed it, their speech habits were mysteriously different: the sister, the elder by four years, used /stε/ by preference for adverbial ‘in’, the brother /sčax/; the sister favored /tə(nə)/ for conjunctional ‘when’, the brother /nə(rə)/; the sister used /mwĩç/ for the locational form of the adverb ‘out’, the brother /mwĩ/; the sister used monosyllabic /hãn/ more often than /hãnig/ for ‘came’ and /hũn/ as well as /hũnig/ for ‘saw’, while the brother used only the disyllabic forms of each. Since they were close in age but of opposite sex, the most obvious hypothesis was that these were sex differentiated usages in Embo. But that simply wasn’t the

404

chapter 21

case, as even the most minimal checking quickly showed. The problem, in fact, was that there were no apparent social explanations for this very prevalent kind of variation: socioeconomic background was uniform; age didn’t play the obvious role here that it did in the identifiable changes in progress (although decreasing age could be shown to correlate with a trend toward the favoring of certain variables in several instances); sex could usually be eliminated as a factor; and there was no clustering of favored variants among people who had lived in the same street. I had variation in plenty; what I didn’t have was an explanation for it in terms that variationist studies would have predicted. This is a fascinating conundrum the full dimensions of which I’m still tracking, in fact, especially since learning how to tape-record from the phone (with the permission of those on the other end, needless to say). The still growing database so far supports certain conclusions to which I was inclined in 1994, when on the (mistaken) assumption that I wouldn’t be able to expand my database much, I wrote about the matter (Dorian 1994). The most fundamental of these was that social homogeneity need not imply linguistic homogeneity. Where the two do not correlate, it seems by the East Sutherland evidence that three conditions may play an important part. First, some circumstance must lead to the emergence of an array of variants. The terrible upheaval of the nineteenth-century evictions, in the fisherfolk case, with some degree of population mixture occurring at that time, may account for some of the variation in East Sutherland, and processes of language change for a bit more (decay of former grammatical distinctions, for example). Second, some circumstance must prevent particular variants from acquiring a link with particular social features among groups within the population of speakers. In the fishing communities, small population size and density of interaction, plus a notably uniform socioeconomic background, presumably play this role. Third, some circumstance must impede local speakers’ access to any standard-language norm that may exist for the language and keep them from developing normative judgments in connection with local variants. In the fishing communities the aberrance of the local dialect (which made importation of church-Gaelic norms or more mainstream-dialect norms unworkable) and Gaelic illiteracy (women) or very limited literacy (most men) have this effect. One critical question that the high degree of intra-village and intra-speaker variability in the fisherfolk communities raises is this: if ESG currently represents the only clear-cut case of such prominent but socially unmarked variability, as it appears to, is that because these former fishing communities are genuinely unusual, or is it because the way fieldwork is normally practiced, and to what ends, has precluded recognition of similar cases? There is evidence

surprises in sutherland

405

in the literature to suggest that a considerable amount of socially unweighted linguistic variability exists in small communities that are buffered in one way or another from the development of normative judgments, but in-depth community-wide studies of other small and relatively isolated speech communities with unwritten vernaculars will be needed in order to find out whether it reaches ESG proportions elsewhere. The question can only be addressed if some linguists can be persuaded – even without a Linguistic Survey to give them a nudge – to depart from the usual practice of working with a talented principal informant, plus or minus a few backup sources, and take up the challenge of whole community fieldwork in small communities where people speak a local vernacular that is sharply different from any written languages in regional or national use. That might seem something of a luxury at a time when we’re just coming to grips with a crisis of underdocumentation in the face of impending large-scale language loss. Yet it seems important to determine whether the expectation of a general consensus on phonological and grammatical norms, deeply inculcated in literate researchers whose professional training (and life experience as well, in most cases) took place in highly normed settings, creates a bias inappropriate to the accurate description of some languages in use in small, preliterate, and socially undifferentiated speech communities. 4

The Scholar and the Sojurner

Fieldwork is simultaneously a professional and a personal experience, which of course is the source of much of the tension it engenders. To my thinking, fieldwork is inherently stressful. Work undertaken in a strange setting depends on the goodwill of people whose traditions you’re not fully familiar with and whose values you’ll probably never completely fathom; and sooner or later (or more likely both) you’re bound to offend against local norms. I don’t think I ever prepared to leave for Sutherland without being visited by a recurrent anxiety dream exquisitely well tuned to the East Sutherland social environment. In the dream I found to my horror that I had omitted calling on some one person during my extended round of obligatory fresh-arrival visits. The omitted person changed each time, but the sweaty anxiety provoked by my sudden awareness of an unforgivable oversight never did. Once in the field environment itself, a consistently difficult personal challenge for me was the fishbowl nature of life in a small village setting. Much as I came to appreciate the vivid drama of village life, where every human folly or unlucky flick of fate’s indifferent hand is soon common knowledge, I never got used to being

406

chapter 21

so utterly conspicuous as I unavoidably was. Even so I was very lucky, since personal privacy is respected in Highland Scotland, and I enjoyed a great deal of it whenever my work didn’t require me to be out and about. The linguist arrives at a fieldwork site with a research agenda, looking for native speakers of a certain language and seeing the local people initially as sources of expert knowledge; she may or may not come to see them also as individual people. Local people see the newcomer as an individual (a pretty eccentric individual by local standards); they may or may not come to see her also as a researcher, depending on whether their culture provides any analogs to such a role or whether they’ve previously encountered people with similar preoccupations. Short-term fieldwork is likely to accentuate the researcherand-sources aspect of fieldwork, since the scholar soon moves on to a new project in a new site. Long-term fieldwork can sometimes be managed on the same basis; the researcher might, for example, fit neatly into the role of employer, i.e., someone who returns at intervals and provides jobs for local people. But the tension between the scholar, whose priority is the gathering of information, and the sojourner, who moves among increasingly familiar people and increasingly connects with them as people, can be acute and painful, and never more so than when professional priorities call for subordination of the more human connection. Anthropologist Barbara Tedlock tells of being taught the Zuni cure for fright, an unexpected token of friendship and trust, after a near-accident en route to what was intended as a farewell visit to longterm Zuni consultants. Her immediate professional impulse was to ask a great many questions about this curing treatment she hadn’t previously known of, but personal circumstances ruled that out (Tedlock 1992: 286–87): I kept quiet. Partly because I couldn’t bring myself to objectify the situation so quickly, and partly because of Hapiya. . . . He had given us some of his sacred medicine knowledge, a bit of his own life, his own breath. . . . I also kept quiet because we had something difficult to tell Hapiya. We were starting up new fieldwork, and this time it was far from the Southwest, in Guatemala. It was hard to find the words to explain to him why we would study elsewhere. People don’t see themselves as objects of study. Finding that others do see them that way produces strong reactions. More often than not the reactions are negative, as some eloquent Native American responses to anthropologists’ studies have demonstrated (Deloria 1969); but occasionally a sense of validation and self-worth is roused instead. Social bias against the people who became the East Sutherland fisherfolk arose shortly after 1800, at the time of their invol-

surprises in sutherland

407

untary resettlement as destitute evictees from inland glens, and solidified in the following century, as severe poverty attended their painful transition from agriculture to fishing. The bias against the fisherfolk population was mirrored in a bias against their variety of Gaelic, so that a scholar who aspired to speak it herself, and who returned repeatedly to study it further and to write books about it, represented a vindication of sorts to some speakers. The scholar was indeed very deeply interested and even admiring, and in this particular setting a sincerely interested scholar made a welcome sojourner, too. As low as fieldwork tensions were for me in the East Sutherland setting, they were always present as I blundered about in an environment and a language not my own. In retrospect I wouldn’t wish the tensions or even the painful blunders away. They belong to the learning process of an immersion experience and are often the engine of discovery, casting linguistic and cultural differences into sharp relief. Some of the special insights of fieldwork may hinge on them. Because they’re uncomfortable and unforgettable, they loom large in the consciousness of fieldworkers (and no doubt also in the memories of the people who live where the researcher worked). Occasionally they surface poignantly in their memoirs (e.g., Briggs 1970), to instruct us nearly as usefully as they did the memoirs’ authors. In learning to do fieldwork, as in learning to drive, the learner knows that some mistakes are inevitable. The learner’s hope in both cases is that the first few mistakes will be of a survivable magnitude so that the learning process can continue. References Aikhenvald, Alexandra Y. 2000. Areal typology and grammaticalization: the emergence of new verbal morphology in an obsolescent language. In The interface between comparative linguistics and grammaticalization: Languages of the Americas, ed. Spike Gildea. Amsterdam: John Benjamins. Briggs, Jean L. 1970. Never in anger: Portrait of an Eskimo family. Cambridge, MA: Harvard University Press. Dal Negro, Silvia. 1998. Spracherhaltung in der Beiz – das Überleben von der Walsersprache zu Pomatt/Formazza. Wir Walser 36: 13–16. Deloria, Vine, Jr. 1969. Custer died for your sins. New York: Macmillan. Dorian, Nancy C. 1973. Grammatical change in a dying dialect. Language 49: 413–38. ———. 1978a. East Sutherland Gaelic: The dialect of the Brora, Golspie, and Embo fishing communities. Dublin: Dublin Institute for Advanced Studies. ———. 1978b. The fate of morphological complexity in Scottish Gaelic language death: Evidence from East Sutherland Gaelic. Language 54: 590–609.

408

chapter 21

———. 1994. Varieties of variation in a very small place: Social homogeneity, prestige norms, and linguistic variation. Language 70: 631–96. Goffman, Erving. 1959. The presentation of self in everyday life. Garden City, NY: Doubleday. Malinowski, Bronislaw. 1967. A Diary in the strict sense of the term. New York: Harcourt, Brace and World. Milroy, Lesley. 1980. Language and social networks. Oxford: Basil Blackwell. Ó Dochartaigh, Cathair. 1983. Review of Dorian 1978a. Scottish Gaelic Studies 14: 120–28. Oftedal, Magne. 1956. The Gaelic of Leurbost, Isle of Lewis. (A Linguistic Survey of the Gaelic Dialects of Scotland, 3: Norsk Tidsskrift for Sprogvidenskap, supplementary vol. 4) Oslo: Aschehoug. Pensalfini, Rob. 1999. The rise of case-suffixes as discourse markers in Jingulu – a case of innovation in an obsolescent language. Australian Journal of Linguistics 19: 225–40. Tedlock, Barbara. 1992. The beautiful and the dangerous: Encounters with the Zuni Indians. New York: Viking. Van Maanen, John. 1988. Tales of the field: On writing ethnography. Chicago: University of Chicago Press.

chapter 22

Documentation and Responsibility 1

Introduction: Fieldwork with Endangered Languages

Any researcher who leaves a record of his or her work assumes a number of responsibilities, chief among them responsibility for the record’s accuracy. But when an endangered speech form is involved, as in the chief case I will discuss here, what might be called the ‘last-chance’ responsibility comes strongly into play: right now may be the one and only chance to create a record of the speech form in question, and right or wrong, what the late-stage fieldworker puts on the record is likely to stand. A number of difficulties, some of them more immediately obvious than others, may lie in the way of a researcher who appears at a late stage in the history of a receding language, hoping and intending to leave a reliable linguistic record. If all of the remaining speakers are elderly, it may be that neither the current researcher nor future researchers will have an opportunity to gather additional material for confirmation or refutation of the original record. This is unfortunately a very common circumstance in research with at-risk languages. If the sampling procedure used by the researcher is inadvertently skewed in some fashion, a source who is unrepresentative of the speech community may come to represent the community in the official record of that community’s speech. This happened, for example, with the record for Golspie village in the five-volume Survey of the Gaelic Dialects of Scotland (Ó Dochartaigh 1997), where an overly literate man served as the sole source for the village, supplying written-language substitutes for some local dialect forms (see Dorian 2010, Chapter 9). If the local language is no longer used in ordinary conversation, the researcher may feel obliged to question the naturalness or completeness of such speech as can be retrieved for the record. Haas raised these questions in her work with the last speaker of Tunica, a man who had had no fluent conversation partners for many years, even though he himself appeared to represent a high degree of fluency (Haas 1941). If there are no longer any speakers who know how the language was used in connection with certain traditional practices, it may be impossible to gain a sense of the full semantic range of certain lexical items or expressions. Jocks (1998) describes the dimensions of the semantic-range problem particularly well, coming to it as an adult learner of Mohawk. If the speech community tolerates, or even embraces, a considerable amount of familial or idiosyncratic variation in the ways that Collins (1998) describes for Tolowa and Kroskrity (2002) for Western Mono, the researcher © koninklijke brill nv, leiden, ���4 | doi 10.1163/9789004269385_024

410

chapter 22

may unknowingly take the forms he has recorded to be much more generally representative of a local speech form than they actually are. It should of course be acknowledged that misapprehending some aspects of the data and failing to appreciate the full semantic content of some of the recorded material are not problems unique to working with receding languages. But these problems are heightened in late-stage fieldwork, because of limited opportunity to supplement or correct the record. Impetus for receding-language fieldwork may come entirely from the outside, with researchers arriving to look for languages reported still to survive, as with David Bradley’s quest for remaining speakers of Ugong in Thailand (Bradley 1989), or it may arise from speakers’ own concern for the future of their speech form, as was true for Faetar in Italy (Nagy 2000) and for Rama in Nicaragua (Grinevald 2006). In the latter case sources are readily identifiable, but in the former case the researcher may have to hunt for elusive speakers. In particularly favorable cases the researcher may find speakers who have developed their own sense of mission about leaving a record of their language and are glad to work with a linguist to achieve that goal. In northeastern Australia the last Warrungu speaker, Alf Palmer, told researcher Tasaku Tsunoda, “When I die, this language will die. I’ll teach you everything I know, so put it down properly” (Tsunoda 2005: 98). But as James Collins found in working with a thin scattering of Tolowa speakers in northern California, the linguist’s narrow focus on contrastive forms and their distributions can be a very long way from what the remaining speakers have in mind when agreeing to a joint effort to record their language. “Simply put”, writes Collins, “they were interested in words, not grammar” (Collins 1998: 260; see also Grinevald 2001: 295). Such discrepancies suggest the potential for conflicting objectives in any joint work involving academic researchers and community members and for discordant notions on the part of the two parties about the responsibilities of the researcher in the wake of that work. 2

The Responsibilities of Late-Stage Fieldwork

As gatherers of increasingly scarce and highly valued information, endangered-language researchers are typically responsible to at least three distinct constituencies: other scholars; individuals like Alf Palmer who serve as their sources; and the ethnic community at large (including, for example, younger Warrungu and Tolowa who were growing up without their ancestral language).1 1 A fourth constituency not discussed here, some sort of funding agency, may or may not be involved. For prolonged fieldwork in distant locations, institutional financial support is a

documentation and responsibility

411

Some results of the scholar’s activities may serve all three constituencies – for example, a clear, user-friendly, and accessible dictionary of the language. But it can also happen that the interests of the various stakeholders are at odds and can not easily be reconciled. In that unfavorable case, serving one set of stakeholders well may mean serving another set poorly or not at all. If fieldwork with a single speech variety extends over some years, the sheer passage of time almost guarantees that the researcher’s position with regard to her responsibilities will undergo some change. In my own case, 45 years of work with an isolated and unusual variety of Scottish Gaelic provided plenty of opportunity to reconsider the focus and scope of those responsibilities, thanks both to blunders that I made as a young researcher and to changing circumstances in the scholarly world and also in the ethnic community. Discussing the blunders in particular offers a way of focusing on potential ethical dilemmas in linguistic fieldwork, in particular fieldwork with a rapidly receding language. My orientation as a young scholar in the early 1960s was typical for the time, I believe, in that I considered my research to be undertaken in the interest of other scholars and my responsibility to be primarily to the scholarly community. While I certainly felt a strong connection to the people I was working with from the very beginning, my orientation at the time conformed to the pattern that Himmelmann has described as typical of twentieth century structuralist linguistics (Himmelmann 2008: 341): I did not take the results of my fieldwork to be of any particular interest to the people whose language I was studying. The variety of Scottish Gaelic that I worked with had about 200 speakers in 1964; it currently has three less-than-perfectly-fluent speakers (four, if I count myself). It was a dialect of the extreme Highland periphery and was as atypical as peripheral dialects often are; it had been recorded up to then only in the form of lexical entries in a Gaelic dialect survey. The coastal East Sutherland Gaelic speakers had been fisherfolk, an occupation that created a separate Gaelic-speaking workforce that for some generations had needed only relatively limited English for commercial transactions outside the community and for part-year occupational involvement in the national herring fishery. Separate residential areas for fisherfolk had permitted community members to maintain their home and neighborhood use of Gaelic well into a period when other population segments in the surrounding region had become monolingual in English. But the fishing had come to an end after World War II, and by the 1960s most local speakers were elderly; only a few were under 40, and no children were acquiring the distinctive local dialect. The end of this speech form necessity, but receding languages may also be encountered much closer by. I worked for example with secular Pennsylvania German speakers who were within manageable driving range of my home institution without requiring funding for that research.

412

chapter 22

was already foreseeable, in large part because it was severely stigmatized. If an effort had been mounted to support Gaelic in this region, it would without a doubt have promoted a mainstream form of Gaelic and not the local variety (as in fact has happened in more recent years). The Gaelic that I recorded in the 1960s and ’70s would not be recoverable from the three speakers surviving today, since in every respect – lexicon, syntax, morphology, and phonology – their Gaelic is less full than that of their predecessors. This means that anyone who wishes to see or hear a full-fluency version of East Sutherland fisherfolk Gaelic will be able to achieve that goal only by consulting what I gathered of it between 1963 and the death of the last locally resident, fully fluent speaker in 2001.2 Very early on in my work with the dialect I made some of the classic mistakes of a young and purely academically oriented scholar. I published a sociolinguistic paper on a locally sensitive subject, for example, using actual examples of the phenomenon in question, in the belief that publication in a scholarly journal far from East Sutherland would be like dropping the material into a deep well of scholarly dispassion and anonymity. I soon learned otherwise when I received a phone call from an émigré East Sutherlander in Michigan whose anthropologist son had come across the paper and shared it with his father. As it happened the father, who had been away from the home community for a great many years, was intrigued by the paper rather than offended and was proposing to send a copy to a local minister in East Sutherland who he was sure would be interested. (Indeed he would have been – I’m still in touch with that minister today.) I threw myself on the émigré East Sutherlander’s mercy, sending him a sanitized paper on the same subject, prepared for publication in Scotland itself, and asking him to send that version to his correspondent instead. Since that time only invented examples of the sensitive subject have appeared in any publications of mine that touch on the same phenomenon.3 My second early mistake was to archive several tapes filled with speakers’ reminiscences at the School of Scottish Studies in Edinburgh, thinking of them as invaluable resources for future scholars who would not have the opportunity to hear fluently spoken East Sutherland Gaelic for themselves. They have indeed served that purpose, but at the time I overlooked the question 2 I specify “locally resident” because it’s possible, though by now highly unlikely, that one or two fluent speakers from the East Sutherland diaspora survive in New Zealand, say, or Canada. 3 The sensitive subject is by-naming, a practice similar to nick-naming but with such prominent elements of mockery that a great many by-names are offensive. As a result, by-names, unlike nicknames, are not used in direct address, even though they are universally used in reference.

documentation and responsibility

413

of content entirely, thinking of the material on the tapes essentially as interesting manifestations of linguistic phenomena. It was only after a colleague who listened to one of the tapes remarked that the School could never use the material in any publication because it was “actionable” that I recalled that the tape included a lengthy narrative in which a voluble and exceptionally uninhibited speaker alleged that another villager had committed a theft. There are two oddities in this case, one in connection with my source’s behavior and one in connection with mine. It’s an oddity of fieldwork that the local speaker typically loses track of the potential publicness of the recorded material, even when recording equipment is visibly in operation before his eyes. It was equally an oddity of my early professional and personal naïveté that I could view my recordings simply as linguistic documents without regard to their content. Belatedly, when I reflected on what was actually on my tapes, I came to feel that nearly all of them included material too personal for general listening, and for the next 30 years I archived nothing at all. In effect I had concluded – a little on the late side – that my responsibility to my sources outweighed my responsibility to future colleagues. In part this conclusion was prompted by the very traditional kind of fieldwork I was doing in the 1960s and ’70s. I went regularly to the homes of local speakers and worked either with single individuals or with a small group of household members (and occasionally one or two neighbors). The setting was private, therefore, and it was also quite generally the case that this unwritten form of Gaelic served as a private-sphere language. It was spoken almost exclusively by and among local community members who had grown up in close proximity to one another, in fisherfolk residential areas where people were deeply linked not just by occupation but also by multiple kinship ties, and as a consequence its use automatically invoked some sense of intimacy and social solidarity. Local fisherfolk Gaelic had become in effect something akin to a private language emblematic of community membership. Since I worked in the same three villages intermittently over a 16-year period, returning repeatedly and acquiring the local form of Gaelic myself, it was more or less inevitable that the material people recorded in our sessions would include personal information, sensitive local topics, and individual perspectives that the speakers might not have cared to air in general company. Some speakers were more likely than others to include relatively personal information in their recordings, but many recordings had content that made them questionable for open access. Late in my on-site fieldwork years I also discovered the hard way how impossible it is for an outsider, no matter how well acquainted with the local community, to predict exactly what will be jarring to local sensibilities. In the second half of the 1970s I had begun working on an oral history of the East

414

chapter 22

Sutherland fisherfolk, and I wanted to give some control over the book to the couple whose tape-recorded reminiscences supplied nearly all of the direct quotations in the book. I sent them the full manuscript before publication, so that they could ask to have anything they objected to taken out, and I removed two items of somewhat personal information at the wife’s request. But even so, when the book came out the wife (by then a widow) was distressed by two other matters that she hadn’t noticed on her first reading. This was extremely painful for me as well as for her, since it was the outcome I had most especially been trying to avoid; it took the kindly intervention of her children to convince her that she was not in fact negatively exposed in a public way, as she feared. It was certainly an instructive experience for me. I had been working in these communities for fully a decade and a half by the time the oral history appeared, yet of the four matters that proved sensitive for one of the principals, only one had actually struck me as likely to be problematic when I was writing the book. 3

The Ultimate Implausibility of ‘Informed Consent’

At bottom the issue is by now a familiar one: whose are the materials the fieldworker has gathered, and for what purposes can they legitimately be used? Different constituencies may have quite different responses. In providing material to an outsider who is a scholar, community members may seem to be agreeing to share their knowledge with the wider world. But since local speakers seldom fully understand what scholars do with the materials they gather, the agreement is more apparent than real. In recent years this problem has surfaced most conspicuously in connection with the development and expansion of the internet. We recognize very easily that as little as 20 years ago no one could have foreseen the ease with which specialized information would circulate via the internet and how broadly access to it would be gained. What we do not as easily recognize is that for most of the twentieth century scholarly publications of any kind were just as unimaginable to most of the locallanguage speakers who provided researchers with their source material as the internet was to all of us until recently: local people usually had no experience of academic books or journals and no idea who used them, so that there was no realistic possibility that they could envisage where their materials would appear and how they would be made use of. The impossibility of making the purposes of our fieldwork understandable to our sources is the bedrock dilemma of researcher responsibilities. If we can’t convey our intentions and our goals to those whose knowledge we propose to tap into, then we can’t obtain truly informed consent for the work we’re doing.

documentation and responsibility

415

In the case of the fisherfolk couple who were the central figures in my oral history, two people fully literate in English, I described the project to them in advance, gave them time to discuss their participation between themselves, negotiated a written statement that gave me permission to use their tape recordings and them a guarantee that I would protect identities,4 sent them the manuscript before publication, and removed at their request two items identified at that stage as overly personal. But because the publication process was unfamiliar to them, it was not obvious to them, as I had assumed it would be, that the manuscript stage was the only chance there would be to have items removed. In just the same way, there would have been no possibility of making intelligible or persuasive to local people a scholar’s experience that the locally sensitive examples used in my early sociolinguistic paper would be intellectually interesting and yet socially neutral to the academic readers that that paper was intended for. A particularly poignant example of the gulf created by incommunicable purposes appears in Barbara Tedlock’s description of her and her husband’s final visit to their long-standing Zuni teacher, Hapiya, in New Mexico. For the first time ever, Hapiya treated them with Zuni medicine, an act of great friendship and trust because the Zuni were well aware that outsiders were usually scornful of their medicine. Much as the anthropologist in her wished to explore this unexpected medical treatment with Hapiya, Tedlock felt unable to do so because of the extremely awkward moment that lay ahead: she and her husband faced the necessity of telling Hapiya that they were leaving to work with another people altogether, in Guatemala. She writes (Tedlock 1992: 287): “It was hard to find the words to explain to him why we would study elsewhere”. Indeed it must have been. “Friend” is a graspable concept in most societies, but “scholar”, with its considerably lesser personal commitment, is much less so. Lesley Milroy describes a dilemma similar to Tedlock’s that arose from her Belfast (Northern Ireland) fieldwork: In one of the Belfast inner-city communities, I built up a strong personal relationship with a very poor family. Most of the recording sessions, which took place in the evening, were pleasant and party-like; eating, drinking, smoking, chatting and card-playing often continued into the early hours of the morning. At the end of the observation period, it was extremely difficult to loosen these ties, which of course involved a considerable time commitment. Much of the conversation had focused on the disastrous and pathetic effect upon the family of the civil unrest in Belfast and its 4 They later agreed to allow their own real names to appear in the published book.

416

chapter 22

function was plainly cathartic; many recordings resembled therapy sessions more closely than sociolinguistic field tapes. I sometimes felt that I was taking unfair advantage of the family’s need to talk through their problems with a sympathetic outsider (Milroy 1987: 90; emphasis added). How are people who are not part of the scholarly world to understand a concept like “the observation period”, and what would they think of it if it could be explained to them? 4

The Inherent Social Unnaturalness of the Fieldwork Enterprise

In all of the cases mentioned here, the researcher is the one who is altogether out of line with normal social expectations. How could an ordinary speaker of Tolowa fathom the researcher’s request for endless repetitions of a single Tolowa speech form in the interests of establishing a particular phonological contrast? What does someone for whom language entails social interaction make of a linguist’s insistence on eliciting complete paradigms? And how could any ordinary person, after the sharing of knowledge and experience that extended fieldwork requires, view the published use of sensitive material or the abrupt cessation of long-continued personal contact as anything but evidence of hypocrisy and exploitation? The scholar’s ability to ‘walk away’ after prolonged and intense connection is simply unnatural in terms of ordinary social expectations. Even under the less intimate field conditions of some current team-based documentation projects, the contrast between a relatively short period of intense attention and interest from affluent and powerful outsiders and the subsequent complete disappearance of the supposedly interested outsiders must be a source of confusion and disappointment to the groups undergoing the experience. Increasingly, in recent years, solutions have been consciously sought for avoiding or mitigating this moral dilemma: for bridging the gulf between a connection that can have an intensity, intimacy, and duration that is typical of some degree of friendship, and a connection that from the point of view of the researcher nevertheless requires at least some degree of the detachment of scholarly observation. Fieldwork ethics have consequently become the subject of much discussion and a considerable literature (Grinevald 2006). One way of bridging the gulf, increasingly adopted today, is for the researcher to enlist community members as co-researchers, offering training and/or co-authorship to any community member who might wish it and co-equal researcher status to those who accept the invitation (England 1992: 34; Grinevald 2003: 60).

documentation and responsibility

417

Many papers in scholarly journals already reflect this development.5 Another way is for the researcher to provide useful professional products to the community: a user-friendly dictionary, language-learning primers and storybooks, locally archived tape- or video-recordings, an accessibly written oral history.6 Researchers may become activists and advocates for at-risk communities and their languages (Rhydwen 1998; Walker 2008), and sometimes expert knowledge can be offered to support the redressing of past injustices.7 Instead of cutting ties with the community at the close of an intense fieldwork period, researchers can stay in touch with the people they have worked with by letters, phone calls, tape recordings, and gifts.8 In rare cases a linguist-becomespeaker may be able to reverse roles and act as a source for young community members who missed out on acquiring the ancestral language in childhood.9 Some researchers will argue (several have argued)10 that scholars are not trained for most sorts of concrete support efforts, have no business trying to interfere in language-transmission decisions that rest with community members, and furthermore have a professional obligation to use their limited time and their professional training for the particularly pressing jobs of documentation and description that they are trained for.11 But many others consider that some kind of service is due to individuals or communities for their generosity in sharing the scarce resource that their language represents, and some communities have come to take the same view, making service on the part of the researcher a condition of access.

5 6

7 8

9 10 11

See for example Penfield et al. 2008. Angela Terrill, for example, has written illuminatingly about her reasons for preparing a speech community’s first dictionary and first storybook and about the community’s reception of the books (Terrill 2002). For example, William Shipley’s long-standing involvement with Maidu has enabled him to provide land-claim support to one band of Maidu Indians (Shipley 2000). Lise Dobrin, responding to the particularities of the Papua New Guinea context, urges ongoing material support, in fact, because in Melanesia an empowerment that might be able to support language maintenance is most likely to arise through established exchange relationships with outsiders (Dobrin 2008: 308, 316). Victor Golla has proposed this and William Shipley has actually carried it out (Golla 2001; Shipley 2000). See Ladefoged 1992, Newman 2004, Matras 2005. Himmelmann 2008 makes a very clear distinction between documentation and description; he makes a strong case for the view that documentation is much the more pressing and the more productive of the two tasks.

418 5

chapter 22

Post-Transmission Conditions and Conflicting Responsibilities

There is of course no way for either fieldworker or sources to foresee what responses will emerge in the ethnic population when all fluent speakers of the distinctive local speech variety are gone. In the second half of the twentieth century, fishing for a living became a very rare individual undertaking in East Sutherland rather than the low-return livelihood of an entire ethnic group. As the Gaelic speakers who had been personally involved in the local line fishing and the national herring fishery disappeared from the scene, village residents began to downplay the fierceness of earlier social prejudice against the fisherfolk. This was true not only among locals who were not members of the fisherfolk group but also among descendants of the fisherfolk themselves. Twice, young monolingual offspring of fisherfolk families gave me their opinions that local prejudice against the fisherfolk had been exaggerated, even though their own parents had supplied very powerful accounts of that prejudice.12 This generational disjuncture where recognition of bias toward the fisherfolk is concerned seems to have heralded a change in the representation – both to self and to others – of local village life, but it surprised me how quickly that change ultimately took place. Perhaps as a reaction to an increasing homogenization of local life within national life, interest in local history seemed to increase greatly among both the descendants of the once disdained fisherfolk and the descendants of the non-fishing part of the population. A regional historical society attracted greater interest and participation, and among its most well-informed and active members was a man of fisherfolk descent, a retiree who was a close relative of the central couple in my oral-history study. He knew that what appeared in my book was only a fraction of the material actually recorded in our sessions, and in his role as a local historian he was very eager to see all of the material made fully accessible on the internet. Though raised locally as a child, he was married to a speaker of a more mainstream Gaelic dialect and now spoke a fluent but non-local Gaelic himself. After an adult life spent away from the home village he had returned to reside in the house where he had spent his childhood, but the curious result of distance and scholarly interests on his part was that his dispassion was now greater than mine. I had had unforgettable conversations with his relatives about the sensitivity of cer12

For example, the younger daughter of the couple who served as my central oral-history sources originally questioned my statement that fights had regularly broken out between the young fishermen and their non-fishing peers at the site of a particular house that marked the beginning of the fisherfolk residential section of her village; but as it happened that information came directly from her own father.

documentation and responsibility

419

tain topics and the potential offensiveness of the material touched on here and there in our sessions, and I was also deeply aware that in recordings made in the privacy of their home the wife in particular had touched on family events and personal feelings that she, a very private person, would not have wished to have open to all comers. This situation raised the most difficult question I had yet faced where conflicting responsibilities were concerned. On what basis can an outsider with unrecoverable material of high scholarly interest withhold that material from a birthright member of the ethnic group in question – in this case also a close relative of the principal sources? The relative and I held very different views of my responsibility. I remembered the distress of the widowed wife when she had felt overly exposed by the publication of the two items she had not asked to have removed in time, and I considered myself honor-bound to protect my sources’ privacy. He believed that there was no longer any basis for such sensitivities and that the important thing was to put all of these invaluable recordings into the historical record. There was validity to both points of view, but while I had no difficulty seeing the force of his point of view, he had some difficulty acknowledging the force of mine. The compromise solution I arrived at was to consult a daughter who was still living in Scotland about the tapes, which were mostly in English, and to send them to her so that she could make the decision about public access to her parents’ recordings. She, too, found a good deal of the material too personal for general access, but in the end she selected one or two recordings with lower personal content and made them available to her relative for wider historical use. This experience made me still more aware than I had been already of the growing historical and linguistic value of the recordings I possessed, as did an inquiry from a Scottish scholar who wanted to listen to Sutherland’s east-coast Gaelic in order to compare it with the Gaelic still spoken by a small number of bilinguals on Sutherland’s north coast. Although at this point health problems had kept me from visiting Scotland for more than two decades, I was still in touch with family members in most cases and was able to write to them or speak to them by phone and ask permission to archive tape recordings I had made with their parents, grandparents, aunts, or uncles, promising careful editing in the cases where I knew the tapes contained material of some sensitivity. Permission was granted in all cases, and I proceeded to edit and copy some of my tapes and send them to two archiving bodies in Scotland. There was still the possibility that as an outsider I might include some objectionable material without realizing it, but the passage of time and the less socially charged climate in present-day East Sutherland reduced the likelihood of serious offense.

420

chapter 22

The solutions I’ve adopted in cases of potentially conflicting responsibilities – returning tapes to family members, asking family permission for archiving – resolve some dilemmas but certainly do not cover all imaginable circumstances in which interests might conflict. The taped materials that I copied for archiving purposes, for example, were exactly the sorts of material that fieldworkers have traditionally collected: narratives, stories, and autobiographical accounts (Foley 2003: 85–86). They offer excellent examples of spoken East Sutherland Gaelic, but they are not conversational enough to provide a good basis for constructing teaching materials, an issue that arose for the first time, quite unexpectedly, in 2008. Improbable as it would have seemed to me in the 1960s and ’70s, when the local Gaelic was universally disparaged, an East Sutherlander not of fisherfolk descent was considering a project to encourage some use of locally authentic Gaelic, with the aim of preserving the character of his village in the face of heavy in-migration from outside the region. I had very extensive conversational material on tape, as it happened, recorded over the telephone from 1993 onwards when I had resumed fieldwork by this longdistance method, but permission to record had been asked for and given with the understanding that I would not make extensive or public use of our conversations, only drawing on them for various purely linguistic projects. I was therefore obliged to rule out any direct use of my telephone tapes, although some of the conversational material might conceivably be drawn on indirectly for instructional purposes if a teaching project were to come into being. It is impossible to predict what attitudes will emerge in the region after the final speakers of a strictly local language are gone. Their descendants may wish to erase all memory of earlier ethnic distinctiveness, to alter the nature of that memory, or to enhance it. In East Sutherland, given how virulent the prejudice against the fisherfolk had previously been (see Dorian 1981: 61–68), erasure might well have been expected after the end of the fishing and the passing of those who had participated in it, but that has not been the outcome. The historical society’s level of activity, the frequency of newspaper articles about earlier times, and the appearance of website reminiscences celebrating both fishing and non-fishing aspects of old days in these villages indicate that some degree of altering and enhancing is underway instead. This development was prefigured in the “improved” memories of some individuals as early as the late 1970s. In an interview recorded in 1978, for example, a very elderly monolingual woman expressed her personal pleasure at having lived to see the end of discrimination against the fisherfolk. She claimed in fact to have been annoyed by that discrimination in her young married days – only to be reminded by one of her daughters that she herself (the mother) had forbidden her children to play with fisherfolk children during their childhood (Dorian 1981: 63).

documentation and responsibility

421

The local mode of dealing with the severe social biases of earlier times seems generally to be denial of the strength of the prejudice and mild romanticization of the very lifeways that were once the focus of serious discrimination. This seems to have happened even in the case of the one social group that had still lower standing than the fisherfolk: the tinkers, “travelling” people who spent much of the year walking the roads, selling small, mostly second-hand household items and acquiring others for sale. In recent years the migratory routes that the tinkers took through Sutherland-shire glens have actually been celebrated by reënactment (The Northern Times, May 18, 2007); by contrast, the strong, earlier bias against the tinkers among local villagers appears very clearly in the recollections provided by my oral-history sources (Dorian 1985: 96). Even though revised memories may have prefigured the direction local attitudes would ultimately take, I was no more able to predict the rapidity of this attitudinal change than I was able to detect exactly which matters would be painfully sensitive to the principal figures in my oral-history book. 6 Conclusions Perhaps the most instructive aspect of linguistic research with the East Sutherland fisherfolk is that this fieldwork presents the same fundamental ethical questions as linguistic fieldwork in much more unfamiliar-seeming settings. By contrast with the Dupaningan Agta people discussed by Laura Robinson (this issue), East Sutherlanders in their first-world setting seem ordinary and culturally familiar.13 Working in the East Sutherland setting presents the fieldworker with relatively few difficulties of a cultural or procedural sort. Neither dramatically different lifeways, such as those of a hunter-gatherer group, nor excruciating procedural difficulties such as those Colette Grinevald faced in obtaining the letter-of-consent required by her funding agency for work with the Rama people of Nicaragua (Grinevald 2006: 361–63), confront the fieldworker in East Sutherland. Yet despite the seeming ordinariness and familiarity of East Sutherlanders and their setting, the challenges of avoiding exploitation or breach of confidence and of achieving informed consent are, as I’ve tried to demonstrate, still very much the same – and just as difficult to overcome.

13

This is actually somewhat less the case than it would seem; see Dorian 2010.

422

chapter 22

References Bradley, David. 1989. The disappearance of the Ugong in Thailand. In: Dorian, Nancy C. (Ed.), Investigating obsolescence: Studies in language contraction and death. Cambridge University Press, Cambridge, pp. 33–40. Collins, James. 1998. Our ideology and theirs. In: Schieffelin, Bambi B., Woolard, Kathryn A., Kroskrity, Paul V., (Eds), Language ideologies: Practice and theory. Oxford University Press, New York, pp. 256–70. Dobrin, Lise. 2008. From linguistic elicitation to eliciting the linguist: Lessons in community empowerment from Melanesia. Language 84: 300–24. Dorian, Nancy C. 1981. Language death: The life cycle of a Scottish Gaelic dialect. University of Pennsylvania Press, Philadelphia. ———. 1985. The tyranny of tide: An oral history of the East Sutherland fisherfolk. Karoma Publishers, Ann Arbor, MI. ———. 2010. Investigating variation: The effects of social organization and social setting. Oxford University Press, New York. England, Nora. 1992. Doing Mayan linguistics in Guatemala. Language 68: 29–35. Foley, William A. 2003. Genre, register and language documentation in literate and preliterate communities. In: Austin, Peter K. (Ed.), Language Documentation and Description, Vol. 1. Hans Rausing Endangered Languages Project, London, pp. 85–98. Golla, Victor. 2001. What does it mean for a language to survive? Some thoughts on the (not-so-simple) future of small languages. In: Sakiyama, Osama, Endo, Fubito (Eds), Lectures on endangered languages 2: From Kyoto Conference 2000. Endangered Languages of the Pacific Rim, Kyoto, pp. 171–77. Grinevald, Colette. 2001. Encounters at the brink: Linguistic fieldwork among speakers of endangered languages. In: Sakiyama, Osamu, Endo, Fubito (Eds), Lectures on endangered languages 2: From Kyoto Conference 2000. Endangered Languages of the Pacific Rim, Kyoto, pp. 285–313. ———. 2003. Speakers and documentation of endangered languages. In: Austin, Peter K. (Ed.), Language documentation and descripton, Vol 1. Hans Rausing Endangered Languages Project, London, pp. 52–72. ———. 2006. A view from the field: An Amerindian view, worrying about ethics and wondering about informed consent. In: Saxena, Anju, and Borin, Lars (Eds), Lesser known languages in South Asia: Status and policies, case studies and applications of information technology. Mouton de Gruyter, Berlin, pp. 339–70. Haas, Mary R. 1941. Tunica: Introduction. Handbook of American Indian languages, Vol. IV. J. J. Augustin, New York, 9–11. Himmelmann, Nikolaus P. 2008. Reproduction and preservation of linguistic knowledge: Linguistics’ response to language endangerment. Annual Review of Anthropology 37: 337–50.

documentation and responsibility

423

Jocks, Christopher. 1998. Living words and cartoon translations: Longhouse “texts” and the limitations of English. In: Grenoble, Lenore A., Whalen, Lindsay J. (Eds), Endangered languages: Language loss and community response. Cambridge University Press, Cambridge, pp. 217–33. Kroskrity, Paul V. 2002. Language renewal and the technologies of literacy and postliteracy: Reflections from Western Mono. In: Frawley, William, Hill, Kenneth C., Munro, Pamela (Eds), Making dictionaries: Preserving indigenous languages of the Americas. University of California Press, Berkeley, pp. 171–92. Ladefoged, Peter. 1992. Another view of endangered languages. Language 68: 809–11. Matras, Yaron. 2005. Language contact, language endangerment, and the role of the ‘salvation linguist’. In: Austin, Peter K. (Ed.), Language Documentation and Description, Vol. 3. Hans Rausing Endangered Languages Project, London, pp. 225–51. Milroy, Lesley. 1987. Observing and analyzing natural language: A critical account of sociolinguistic method. Basil Blackwell, Oxford. Nagy, Naomi. 2000. What I didn’t know about working in an endangered language community: Some fieldwork issues. International Journal of the Sociology of Language 144: 143–60. Newman, Paul. 2003. The endangered languages issue as a hopeless cause. In: Janse, Mark, Tol, Sijmen (Eds), Language death and language maintenance: Theoretical, practical and descriptive approaches. John Benjamins, Amsterdam/Philadelphia, pp. 1–13. The Northern Times, May 18, 2007. Summer walkers in the straths – a Mackay Country experience. The Northern Times Company, Golspie (Sutherland). Ó Dochartaigh, Cathair (Ed). 1997. Survey of the Gaelic dialects of Scotland. Dublin: School of Celtic Studies. Penfield, Susan D., Serrato, Angelina, Tucker, Benjamin V., Flores, Amelia, Harper, Gilford, Hill, Johnny Jr., Vasquez, Nora. 2008. Community collaborations: Best practices for North American indigenous language documentation. International Journal of the Sociology of Language 191: 187–202. Rhydwen, Mari. 1998. Strategies for doing the impossible. In: Nicholas Ostler (Ed.), Endangered languages: What role for the specialist? Proceedings of the Second FEL Conference. Foundation for Endangered Languages, Bath, England, pp. 101–06. Shipley, Bill. 2000. The Maidu story. Jorge Hankamer Webfest, http:ling.ucsc.edu/jorge/ index.html. Tedlock, Barbara. 1992. The beautiful and the dangerous: Encounters with the Zuni Indians. Viking, New York. Terrill, Angela. 2002. Why make books for people who don’t read? A perspective on documentation of an endangered language from Solomon Islands. International Journal of the Sociology of Language 155/156: 205–19.

424

chapter 22

Tsunoda, Tasaku. 2005. Language endangerment and language revitalization. Mouton de Gruyter, Berlin. Walker, Alastair. 2008. How can academic institutions help support an endangered language? The case of North Frisian. In: de Graaf, Tjeerd, Ostler, Nicholas, Salverda, Reinier (Eds), Endangered languages and language learning: Proceedings of FEL XII. Foundation for Endangered Languages, Bath, England, pp. 11–17.

chapter 23

The Private and the Public in Language Documentation and Revitalization 1 Introduction Among those who are strongly supportive of efforts to revitalize receding languages, as well as to document them, it is well recognized that the barriers to success are enormous. Fishman’s 1991 volume discusses many of those barriers, and voluminous contributions to the literature on this topic have expanded the discussion. In what follows here, I focus initially on potential problems in the linguistic fieldwork setting and the language revitalization setting, arising from different positions with regard to relatively private versus relatively public language use, acquisition, and transmission. Efforts to lessen such problems and the prospects for lessening them further are discussed thereafter, with particular focus on the increasing professional involvement in these efforts. 2

The Fieldwork Context

2.1 The Problems of “Private” vs. “Public” Fieldwork and Informed Consent Of late there has been a strong emphasis on teamwork in language documentation, with researchers cooperating to record language use in a fuller range of cultural contexts and also to cover spheres outside the strictly linguistic, such as ethnobotany and traditional song (Wittenburg 2003; Himmelmann 2008). Full documentation is now also assumed to include visual as well as auditory recording (Csató and Nathan 2003; Wittenburg 2003), so that more of the discourse context will be recoverable and analyzable for future researchers (and revitalizers), as well as for those currently making the record. Archiving, too, is typically fuller and more responsible, both as to completeness and as to accessibility (Thieberger and Musgrave 2007). One of several advantages in approaching fieldwork in this way is the relative publicness of the activity. The arrival of a team of documenters carrying recording machinery of various kinds is a conspicuous occurrence in most small-language settings, and if community cultural activities are being filmed, many local participants will be involved. Some funding agencies routinely ask

© koninklijke brill nv, leiden, ���4 | doi 10.1163/9789004269385_025

426

chapter 23

for some sort of written consent from the community in which the documentation is undertaken, so that the project will also be as broadly consensual as possible (Grinevald 2006). Serious questions still arise about the degree to which individuals or communities can provide informed consent for projects that are essentially uninterpretable in terms of the local cultures (Dobrin 2008), but the enterprise is at any rate openly and often publicly undertaken. Researchers are also increasingly prepared to share recordings, films, or some printed outcome of this kind of work with the host community (Terrill 2002); in some cases this is an objective of the project from the outset. Prior to the late 1990s, multiple-researcher documentation projects were probably still few in number by comparison with more traditional fieldwork projects in which a lone researcher went “into the field” to make a record of a language, usually with more descriptive than documentary objectives. Videotaping was also relatively uncommon, undertaken more at the initiative of the individual researcher than in response to professional expectations. Some climates and some field settings favored working out of doors, so that the researcher’s activities were open to a great deal of public scrutiny. In other climates and settings, however, the researcher disappeared within some enclosed structure and conducted the research work in considerable privacy, usually on a more or less dyadic basis. In open-air settings much that passed between sources and researcher was audible to others and self-censorship on the part of the local speaker(s) was probably automatic. Self-censorship was less likely in the privacy of a home or a sequestered workroom, and in addition the non-judgmental ear of someone not connected to the community by blood or marriage could produce a freedom of expression neither party originally anticipated, especially if the work continued over a longer period of time. The very intimacy of such sequestered fieldwork encourages a trust that over time reduces inhibitions. This was certainly the case in some of Milroy’s work in Belfast, where she describes the work she did with one family in the following terms: “Much of the conversation had focused on the disastrous and pathetic effect upon the family of the civil unrest in Belfast and its function was plainly cathartic; many recordings resembled therapy sessions more closely than sociolinguistic field tapes” (Milroy 1987: 90). My own experience in fieldwork done prior to the 1990s was that long-term, sequestered fieldwork produced personal and at times very uninhibited content some of which was quite unsuitable for general-access archiving; it would have been equally unsuitable as the basis for printed materials that might promote revitalization, such as story books and autobiographical sketches. The more spontaneous and lively a speaker’s recordings were, the more unsuitable they were likely to be for such purposes. In effect they represented a sharing

private and public in language documentation

427

of confidences on the part of the speaker. One of my best sources produced relatively little that I could feel free to publish as text or to archive, even with time restrictions on archive access. Her stories were full of life, but they frequently told, with gusto and in rich detail, of alleged misbehavior on the part of fellow villagers or other identifiable figures. They were superb resources for descriptive and sociolinguistic purposes, and as a researcher I derived much benefit from them; but it was only long after her death, with the buffer of many intervening years, that I ventured to ask for (and did receive) permission from her remaining family to archive a few carefully selected portions of some of her stories. I took the same steps in the case of problematic material recorded by two other (likewise deceased) sources. But some material remains in my judgment too potentially offensive to archive at all, even with long-term access restrictions. Memories are very long in small villages, and families tend to remain in place for generations. Where a local reputation is at issue, sensibilities are understandably acute. 2.2 Overcoming Problems Related to Privacy in Fieldwork The confidentiality problems occasioned by fieldwork sessions that take place in private locations between a single fieldworker and a single primary source (or perhaps with some other members of the household present as well) are somewhat less likely to arise in twenty-first century conditions for one reason already noted, the increasing adoption of a team-based fieldwork that lessens the frequency of sequestered, dyadic interaction between researcher and source. In addition, the level of discussion about ethical issues of confidentiality and informed consent has risen steadily in workshops, conferences, and the scholarly literature, so that researchers go into the field with greater awareness of the need to protect sources’ privacy. A growing focus on archiving, prompted by recognition of the large number of languages likely to pass out of regular use and of the limited time available to record them, requires researchers to consider just what ought to be permanently on record and to consult with communities on the matter. Not infrequently, the community’s own interest in having their language on record in enough detail to support revitalization efforts is the moving force behind the documentation in the first place. Even so, the issues of privacy and confidentiality remain difficult ones. Milroy offers an example in the form of a recurrent problem she faced in her Belfast fieldwork: a recording session might be underway, by permission of the participants, when people who were not present at the time permission to record was discussed arrived unexpectedly and joined in the conversation; they might not even be aware that recording was going on when they joined in.

428

chapter 23

Yet stopping the session to re-negotiate permission to record in each such case would have been fatal to any degree of naturalness (Milroy 1987: 89). Because of the elusiveness of a truly informed consent from people to whom the methods, purposes, and products of research are bound to remain to some extent obscure, the researcher’s best efforts may not be enough to ensure adequate consent, as one of my own fieldwork experiences can illustrate. In 1976 I embarked on an oral history project in one of the three fishing communities where I had been doing fieldwork intermittently since 1963. I conceived of this as a way of making a record of what was by then a way of life that lingered only in the memories of the elderly and also as a way of making some return to the community for their unstinting generosity in sharing their language with me over the years. I approached a husband and wife in their seventies with whom I had already done linguistic work and asked them whether they would be willing to act as the central sources for an oral history; they were ideal for the role because each of them had had experience of work connected both with the local line fishing and with the national herring fishery. I gave them time to discuss the project between themselves, and when they agreed to participate I wrote an informal agreement into one of my field notebooks, stating that they would allow me to use the material they recorded while I would be careful to protect identities; this we all signed. We proceeded with this work during the summers of 1976 and 1978, and when I had a manuscript ready I posted it to them, so as to give them an opportunity to have anything they objected to removed. This was to be their story, after all, and I wanted it to reflect their lives in a way that felt both accurate and acceptable to them. I was also aware that as an outsider to the community I might not, even after what was by then 16 years, be able to identify exactly what would or would not be objectionable to local sensibilities. Sadly, the husband died while the manuscript was in the mail; but at his widow’s request I removed two items of somewhat personal information from the text. In spite of all my precautions, however, when the book came out the widow was distressed by two other items she had overlooked at first reading. She now wanted these removed as well, which of course was not possible at that point, and it took the kindly intervention of two of her children to persuade her that she was not, as she feared, overly publicly exposed. The flaw in my earnest efforts to make the finished oral history completely inoffensive was that the publication process itself was unfamiliar to my sources: it was not obvious to the surviving spouse, as I had supposed it would be, that the manuscript phase of the book was the only stage at which anything she objected to could be taken out. This was a woman literate in English whose lifestyle in a familiar-seeming, first-world environment was not at that point exotic or unusual in any obvious way, yet her genuinely informed con-

private and public in language documentation

429

sent turned out to be much more difficult to achieve than I had imagined. It is easy to see how much harder it would be to achieve well-informed consent in a cultural context that differed more radically from the researcher’s own. Clearly a good deal of responsibility falls to the researcher when it comes to the protection of privacy. 3

The Language Revitalization Context

3.1 Minority Languages in Private and Public Spheres Where the vitality of a language is high and its dominance unquestioned, use of the language will normally be broad-based. It is likely to serve equally, for example, for intimate family life and for more public gatherings. If the language is written, it will generally be used both for private notes and letters and for more formal purposes as well. Resort to other languages is typically voluntary under these circumstances. Purely statistically speaking this profile is uncommon, since by far the greatest number of all languages coexist with at least one other more dominant language. The more dominant language routinely shows greater vitality in the sense of enjoying either a larger or a socioeconomically better placed population base, more official support, and much wider public use as the result of adoption by the national or regional government and its educational, administrative, and judicial systems. Some small languages in competition with a more dominant language lose ground in a pattern that results in their being reserved for sacred purposes, in invocations, prayers, and the like. More often, however, the pattern of retreat is to the local neighborhood or to just the home and the kin circle, perhaps with some special use of a more formal register persisting in oratory or religious ritual. A small minority language still well established in private spheres but not much used in public spheres is relatively easily documented, assuming access to the speech community is granted. If people still use the language in the streets or in courtyards and homes, and if they are willing to be recorded on tape or film, then discourse in many rich forms can be documented and preserved, for the community’s own use and for the scientific record. But these same languages are less well placed where revitalization is concerned, in particular if the method adopted for promoting the language is schooling, as is increasingly common. On the one hand speakers of a long disfavored language may be glad of any sociopolitical developments that give their language enough legitimacy to claim a place in the educational system. But on the other hand a considerable problem with fit is likely to arise in the early stages of school introduction. If the

430

chapter 23

language has been used mainly in home and neighborhood settings, transferring it from intimate settings into a formal and public setting like the school can produce an inhibiting sense of inappropriateness, for teachers and students alike. If the schools have a long history of excluding local minority languages, children from minority language homes may be particularly uncomfortable about encountering their home language in the classroom and being asked to use it in that environment. Indigenous community members recruited as language teachers might be expected to have no difficulty in using their languages in the school setting, yet school use of the local indigenous language may be problematic for them as well. In Tlaxcala, Mexico, eight individuals who were candidates for positions as indigenous language teachers in a revitalization initiative were observed during parts of a teacher-training program mounted between August and December of 1999. Although the candidates were all speakers of Mexicano (otherwise known as Nahuatl, an indigenous Uto-Aztecan language), the observer reported that such use as they made of Mexicano during their course participation was symbolic, except for the one occasion when they were being tested on their indigenous language skills (Messing 2003: 82). Despite the avowed revitalization purpose of the training course, the classroom was “a formal context, . . . without sufficient intimacy and solidarity between speakers to warrant more use of Mexicano” (ibid.). It should be noted, however, that the reverse of this development is also known to happen: minority-language speakers who become teachers of their indigenous language may adapt well to school use of the language while failing to use it at home with their own children, promoting public sphere use but neglecting private-sphere use (Hinton 2009). Parent-child transmission, usually the swiftest and most complete route to mastery of the target language, is then replaced by purely school-based transmission, which Fishman in his benchmark study of reversing language shift convincingly depicts as ineffective (Fishman 1991: 368–70). Schooling is very often the chosen locus of revitalization efforts, all the same, precisely because it moves revitalization from the less accessible realm of personal motivation into the public and potentially more maneuverable realm of educational policy-making. Once the classroom becomes the setting for revitalization efforts, the requirements of formal teaching present certain foreseeable problems for speakers of private-sphere forms of the minority language. In the interests of promoting literacy and broader use, coinage and codification make their appearance, and their adoption changes the school-promoted version of the language into one that differs from the variety spoken locally. Coinage of new terms is needed so that speakers can deal with topics that are either seldom discussed in the minority language or are discussed by drawing on many loan-

private and public in language documentation

431

words from a more widely used language. Codification is needed in order to replace limited-currency local dialect forms with forms that will be recognized in all regions where the minority language is spoken. The new written-language forms are necessarily unfamiliar to children who normally speak or hear only the local private-sphere form of the language, and if the classroom teachers are not strictly local the pronunciations favored in school may be unfamiliar as well. More importantly, they will be unfamiliar to the children’s parents and grandparents. If local children are taught to produce the variety promoted by the schools, the result may prove counterproductive, at least initially, for the community. King describes, for example, the generational divide that emerged in two Ecuadorian Quichua-speaking communities when elderly local speakers were confronted with the school variety their grandchildren were being taught: grandmothers did not want to converse in Quichua with grandchildren whose speech was full of unfamiliar lexicon (King 2001: 95). In a southeastern Welsh locality, Jones documents support for school-promoted Standard Welsh successful enough to have prevented even passive recognition of the original local dialect. She found that when local children who had acquired schooltaught Welsh responded to a matched guise test,1 they did not so much as recognize certain long-standing features of their own locality’s Welsh and instead identified them as features of northern or western dialects (Jones 1998: 117). If the formal school version of the indigenous language becomes normative in this fashion, the local speech variety may remain essentially where it was: a private-sphere language confined to use in the home and in certain other intimate and solidary settings, perhaps especially among the elderly, as in the case discussed by King. Under these circumstances the local variety is likely to continue to lose speakers by attrition and transmission failure, the phenomenon that usually prompted the revitalization effort in the first place. Bridging the gap between public and private spheres of minority language use may depend, in long-term revitalization efforts, on the degree to which proficient school-taught speakers prove willing to carry their acquired 1 In the matched guise technique, listeners hear one or more speakers gifted at variable renderings of certain key linguistic features read a text that differs only in the way those features are produced. For any given speaker, rate of delivery, tone of voice, and so forth are kept constant, and listeners are not told that the speaker is the same in more than one case. Listeners are asked to evaluate the speaker of each version of the performed text in terms of various subjective responses. In this case the speaker’s rural or urban origin, general age group, general place of residence within Wales, and likelihood of having received Welsh-medium education were evaluated; respondents also assessed the likelihood that the speaker might hold certain specified jobs.

432

chapter 23

language beyond the classroom, into some level of more general social use, and also on the degree to which native speakers prove willing to accept such second-language speakers into their conversational networks. Neither development is a reliable outcome of revitalization undertakings, but hope of such an outcome motivates many language-support movements, and the demonstrated attainability of more modest goals (younger people with at least some active knowledge of the heritage language and some familiarity with its lexicon and structure, as in the case of Tolowa individuals who have passed through the Tolowa language program in northern California; see Collins 1998a: 264) fuels continuing community support for school-based programs. 3.2 Native Speaker Status as a Private-Group Right Most of us have acquired some additional language or languages through schooling without anyone ever challenging our right to learn those languages. Native speakers may well object to our accents or the way we handle the grammar of their languages, and some may try to avoid speaking with foreigners who speak their languages particularly badly, but it would not occur to them to them to warn us off trying to use their languages at all. Languages such as French, Spanish, Russian, English, and German are learned by legions of schoolchildren and university students in Europe and the Americas, for example, as increasingly are Japanese, Chinese, and Arabic, and the right to acquire them is taken for granted. No such freedom of acquisition is assumable in the case of small minority languages. Some have been spoken for generations in essentially closed communities within which only birthright members have access to the minority language. In some cultural contexts the language is deeply associated with the territory in which it is spoken, and access to both land and language is restricted, at least ideally, to members of the indigenous language group. This is famously the case in Australia, for example, where, as Amery puts it, “languages are owned, in the same way that art designs are owned by particular groups or clans”, and “senior individuals are recognized as the owners or custodians of the language”, so that permission must generally be obtained from them to teach the language in a formal course (Amery 2000: 44). Even without an ownership concept as fully developed and asserted as in the Australian context, it may still happen that the minority language comes to be so closely associated with a particular population that it becomes unusual, and possibly unwelcome, for others to acquire it. At that stage would-be learners can be seen as trying to adopt an identity that belongs by rights only to the native-born.

private and public in language documentation

433

This issue arises somewhat counterintuitively in connection with adult second-language acquisition of Scottish Gaelic. The number of Gaelic speakers in Scotland has been in decline for more than a hundred years and the language has only in the last two or three decades enjoyed any governmental support to speak of. With funding for Gaelic initiatives dependant on the continued existence of a population of Gaelic speakers and users, one might suppose that learners would be welcomed by the native-speaking population, but this is far from universally the case. One problem arises from the fact that the identity “Gael” has been heavily romanticized in a backward-looking way (Chapman 1978), so that second-language learners may seem incongruously distant from that identity if they have no Highland ancestry or only very longago and partial Highland ancestry. A second problem arises at a more practical but not unimportant level: Gaelic learners bring limited second-language skills to interactions with a bilingual native speaker population that is fully competent in English, the primary language of most learners. Only the most patient and sympathetic of native speakers are willing to converse with learners who speak Gaelic poorly when fluent English is available to both parties. A third problem arises from the fact that most learners are either English monolinguals or speakers of two major (non-minority) languages and are accustomed to using the languages they speak for all purposes. In keeping with that model of linguistic behavior, they introduce Gaelic into contexts where local community members do not normally use the language; beyond this, activist learners wish to promote the use of Gaelic outside the traditional Gaelic-speaking parts of Scotland. For some native speakers these are unnatural roles for Gaelic, and because they violate local norms for Gaelic use, they seem artificial and offputting. (All of these problems are lucidly discussed by MacCaluim in a study of the potential value of Gaelic learners to reversing language shift in Scotland; MacCaluim 2007.) Note that in the Scottish context, and no doubt in others, two “privatelanguage” obstacles can be seen to coincide: not only may native speakers consider the language a near-exclusive birthright privilege of their own group, but they may also have grown so accustomed to its exclusively private-sphere use that learners’ attempts to expand its functions are viewed as illegitimate. 3.3 Overcoming Problems Related to Acquisition and Use Outside the Home If schools are the only setting in which use of the local language is promoted among children, problems are hard to avoid. Aside from the difficulties noted above – the absence in school of an intimacy and solidarity otherwise

434

chapter 23

associated with the language, elimination of strictly local dialect features in the process of codification, and the introduction of unfamiliar, newly coined lexicon – there are often some young people in the larger minority-language ethnic community who respond poorly to the school environment in general. Alienation arising from a long history of majority language hostility to the minority group is one common source of such a response, all the more so if the minority language was firmly kept out of the schools before a recent policy reversal. The effectiveness of classroom instruction is often in question, too, as in the school-based efforts to promote Ecuadorian Quichua described by King (2001), where neither materials nor teaching strategies were well enough developed to move the instructional program forward. Even where better instruction is available, however, minority-language advocates have often pointed out that school promotion creates problems (so much so that Flores Farfán recommends avoiding dependence on schooling entirely in the Nahua communities he has worked with in Mexico; Flores Farfán 2001: 191). In general, revitalization seems to proceed more effectively if any school instruction that may be available is supplemented by culturally appropriate activities outside school. More effective revitalization of the Keres language, in New Mexico’s Pueblo de Cochiti, has relied, for example, on the embedding of language learning in traditional community practices such as visiting and community clean-up projects, so that younger community members are brought naturally into contact with older members who are skilled habitual speakers of Keres (Pecos and Blum-Martinez 2001). In the Solomon Islands, according to Wurm (1999), Äy̆iwo, a nonAustronesian language with an elaborate noun-class system, a complex nounphrase concordance, and other morphologically challenging features, had begun to show simplifications and losses in the version of the language spoken by young people. In response islanders undertook a gradual revitalization process in the course of which an alphabet, a dictionary, and a text collection were produced, and Äy̆iwo literacy and features of the traditional forms of the language were introduced into some schools. But by deliberate policy young people were also encouraged to take part in traditional crafts such as carving and canoe-building, so that such linguistic features as the mode-of-action prefixes that attached to verbs could be demonstrated and acquired in a natural context of tool-using (Wurm 1999: 171). In northern California, where since the 1920s Tolowa children have learned English first and Tolowa only if their circumstances were unusually favorable to its acquisition, Collins describes the school-based language program mentioned above, underway since the 1970s, as enjoying considerable support from the indigenous community and some

private and public in language documentation

435

success in producing a widely shared though non-fluent knowledge of the ancestral language (Collins 1998a: 264); but he also notes the vital importance of traditional fishing and dancing activities in supporting a Tolowa identity and a cultural ideal that in turn support the school language program (Collins 1998b: 178–93). As for the inclination of some native speakers to consider use of the ancestral language an exclusive in-group privilege, some will no doubt find themselves unable or unwilling to depart from this position. Others, faced with the choice between accepting second-language speakers whose rendition of the language is “inauthentic” in various ways (including phonological and grammatical deviations from traditional norms) or accepting loss of the language altogether, will make their peace with the deviations (Ó Baoill 1987: 102). How many make this second choice has considerable potential significance for the continued oral survival of some form of the language. In Scotland, for example, a small but growing number of children from non-Gaelic-speaking homes are emerging from immersion schooling as fluent speakers of Gaelic. Demand for Gaelic immersion schooling has continued to rise, and if a shortage of teachers and funding can be overcome the numbers will certainly rise further. Yet at the same time, because of continuing transmission failure in the traditional Gaelic heartland, the number of monolingual English-speaking children from Gaelicspeaking homes is rising as well. Monolingual young people from traditionally Gaelic-speaking areas are particularly inclined to resent learners, seeing them as laying claim to a Gaelic identity that rightfully belongs to themselves, even if they have not acquired the language (MacCaluim 2007: 96). At the moment the tide still seems to be flowing against home transmission in the rural heartland, even while Gaelic-medium education strengthens, especially in the cities. This is likely to exacerbate the tensions, and the longer-term outcome is not clear. 4

The Researcher’s Role

The practices of researchers bear directly on the privacy-related problems that arise in fieldwork, whereas any convictions that academic researchers may hold about the significance of schooling and of second language learners for the survival of a language are unlikely to have any bearing on how willing native speakers are to accept a school-based language program and to welcome second-language speakers of their language. In a more general sense, however, the attitudes of researchers may nonetheless have an impact on native speakers’ attitudes and behaviors, at least in some cases. In 1991, at what proved to be the dawn of an era of rising concern

436

chapter 23

about endangered languages, Dixon pointed to the potential value of documentation for the self-image of peoples whose languages were at risk (Dixon 1991: 254). Alongside such concrete steps as implementing bilingual education programs and developing a stock of written literature, he mentioned “helping people to value and cherish their traditional language” as a possibly useful support measure (Dixon 1991: 253). He was exhorting his fellow-linguists in this case, but he credited missionary work in East Sepik Province in Papua New Guinea with a strong revival of Urat (Dixon 1991: 246) and he stated that “any attention that is paid to a local language, whether by linguist or missionary, is likely to enhance the speaker’s image of that language, and of themselves, and can only have a beneficial effect” (Dixon 1991: 247). Though it is not as hard to think of exceptions to this generalization as one could wish, a good many researchers do subscribe to the notion that a fieldworker who takes the trouble to learn the local language sends a message about the worth of that language (Dorian 2001: 149; Dobrin 2008: 318), and that linguists who produce written materials at the request of a community whose language was previously unwritten, or very rarely written, enhance the standing of that language (Terrill 2002). Since negative messages about the value of the receding language are prominent among the factors that bring transmission to an end, linguists’ affirming actions can help to revalorize a small local language “as an important and viable language which still has an important role to play even in the changing urbanizing world” (Terrill, speaking of Lavukaleve, the language of the Solomon Islands community for which she provided a storybook and a dictionary, the latter at the community’s request, 2002: 210). From this point of view the more traditional dyadic form of fieldwork, at least if it is long-term and encourages the researcher to acquire the local language, may have an advantage over team-based documentation projects, which are both unavoidably intrusive (Thieberger and Musgrave 2007) and also expensive to mount, and for the latter reason are perhaps less likely to be sustained over a long stretch of time. Above all, a resident fieldworker who immediately sets out to learn the local language avoids inadvertently modeling the advantages of language shift, whereas a team of affluent and technologically well-equipped outsiders does the opposite if on entering the community they are heard to use a wider-currency language regularly among themselves. In his 1991 article, Dixon maintained that “the work of documentation and that of language maintenance naturally go hand in hand” (Dixon 1991: 254). So they should, perhaps, but we have plentiful testimony from members of indigenous communities that from their point of view this has by no means

private and public in language documentation

437

always been the case. A 2008 conference on “Native American Languages in Crisis: Exploring the Interface between Academia, Technology and Smaller Native Language Communities” at the University of Pennsylvania pointed to the very different agendas of academic linguists and indigenous peoples as a long-standing problem and brought together members of both groups to discuss that problem. Participants hoped to turn a common interest in smalllanguage survival into a more mutually beneficial partnership by identifying and developing best practices for strengthening indigenous American languages at serious risk and by giving indigenous people more control over research and its results. As with the problem of informed consent, this issue is now very much on the table and will presumably receive increasing attention as we go forward. A good example of movement forward appears in Florey’s recent account of a two-session sequence of workshops in Indonesia, designed to create or expand the capacity of Indonesian researchers to document Indonesian languages (Florey 2008). 5

Looking Both Behind and Ahead at Revitalization Issues

The most fundamental issues in connection with documentation and revitalization relate to will and mobilization; they must be addressed initially within the communities in question (Dauenhauer and Dauenhauer 1998). Whether outside researchers may have anything to offer is among the questions for internal discussion. Quite apart from any community doubts about the advisability of seeking outside expertise, however, there are voices within fieldworker ranks arguing that researchers have no business trying to intervene in contexts of language shift, possess no great expertise in the relatively applied forms of linguistics that intervention calls for, and are in any case badly needed for the jobs of description and analysis for which they are specially trained (Ladefoged 1992; Newman 2003; Matras 2005). Certainly communities do not always welcome researchers into their midst. Callaghan (ms: 9) tells of being denied entry to the home of the last speaker of Marin Miwok and Grinevald (2001: 290) of being expelled from a community in Bolivia; the hostility in both cases arose from the potential sources’ prior negative experiences with non-Indians. Wilson describes anthropological fieldwork with Tsimihety (speakers of a regional dialect of Malagasy) in Madagascar which was effectively resisted, the Tsimihety will to “freedom from outside intrusion” extending to keeping the researcher poorly informed, though without active hostility (Wilson 1992: 162).

438

chapter 23

Dwindling speaker communities do not necessarily see their speech form as candidates for revitalization, furthermore. Matras reports that in the community in which he grew up, where only the elderly spoke Yiddish, nontransmission of Yiddish was not regarded as tragic by the people involved (Matras 2005: 227). He also describes the Domari-speaking community he worked in (in Jerusalem, among the Dom, or “gypsies”, whose ancestral community language is Indo-Iranian) as one in which “ongoing language death is accepted by the speakers and the community” (Matras 2005: 244). Many receding languages are spoken by small and poorly placed populations, are unwritten, and receive no official support. If the speaker community has had a traditional economic base of some particular sort, it has typically come under increasing pressure during at least the last half-century from demographic, social, and economic changes associated with land takeovers, resource depletion, urbanization, nation-state development, and economic globalization. If the community has endured a degree of stigma as the result of lesser technological development, low-income subsistence modes, illiteracy, or perceived minority-group distinctiveness, that stigma is likely to have remained strong over the same period in proportion to any continuing distance from regional or national socioeconomic norms. It is not surprising that some such communities have little inclination to regret the loss of their ancestral language, especially if they appear to have a reasonable chance of eventual assimilation. If their distinctiveness was largely related to place of residence, occupation, and income level, for example, and not to differences in physical appearance, ancestral language loss may well shrink in importance for them when set beside the social and economic rewards of assimilation. Even if the prospect for improvement amounts only to some small degree of economic advancement, an ancestral language may be abandoned with little or no regret. Matras does not perceive any conscious attempt among the Dom to integrate into the surrounding Arab society; but he notes that with their recent transition from a cohesive ethnic group with a long history as nomadic metalworkers to a settled “clan” of poor urban wage-laborers, there was no basis for cleaving to traditions (including their language) that marked a stigmatized common group origin (Matras 2005: 242–43). Acceptance of language loss under conditions of this sort has sometimes been characterized as “language suicide”, since it appears that no resistance is mounted to the loss (Dennison 1977: 16). In view of the lengthy period of stigmatization and discrimination that precedes the apparent acquiescence, however, the “suicide” terminology seems to blame the wrong party, the victims rather than the perpetrators. Such is my view, at any rate, after several decades of work with a speaker population whose members, in the wake of generations of stigma associated first with their subsistence

private and public in language documentation

439

mode and subsequently also with their language, likewise expressed no desire for the revitalization of their local speech form (Dorian 1987). Matras takes strong exception to “salvation linguists”, as he calls them, who see themselves as stepping in to “save” an endangered language. He points out that communities do not take a single attitude or speak with a single voice, and that not every community wants its language rescued by a linguist (Matras 2005: 227). Yet it is not uncommon today for speakers of a receding language to take the initiative themselves, asking to have a linguist study and record their language (Nagy 2000; Grinevald 2005) or hiring a linguist to help in creating or refining a writing system, putting together a dictionary, or preparing teaching materials, as do some indigenous Australian and North American groups (Wilkins 1992; Debenport 2009). Hinton describes vividly the yearning for recovery of ancestral languages that brought leaders of the Native California Network together with linguists in 1992 to find effective ways to preserve or restore their languages (Hinton 1994: 221–22). The success of the Indonesian workshops mounted by Florey and her colleagues likewise indicates that local speakers wish both for outside expertise and for greater expertise of their own in furthering documentation and revitalization. Dixon, in his 1991 paper, pointed out that communities often remain unaware of the risk to their language until it has grown too late to change the situation (Dixon 1991: 231). This seems an odd observation initially, but it is not so counter-intuitive as it seems. Fluent speakers who have reached middle age still have available a generation older than themselves who regularly speak their language. They themselves make ample use of the language, and they may simply fail to register the degree to which young people and children are using some more widely spoken language instead. As Kulick demonstrated for the Papua New Guinean village where he worked, they may also not register the degree to which they themselves are failing to use the local language with their children (Kulick 1992). It is only as they become the older generation themselves that some local-language speakers look around and realize that there are no speakers coming along behind them: they are the last remaining speakers, and unless heroic measures are taken their language will disappear with them. At this point the attention of specialist outsiders may rather suddenly be seen as useful where it was not before. Grinevald (2006) has pointed out that interventionist agendas developed among linguists in very particular contexts, namely those in which patterns of language loss had already reached extreme proportions: North and South America and Australia. While in Europe some regional languages were acknowledged to be used less than others – hence the terminology EBLUL, European

440

chapter 23

Bureau of Lesser Used Languages, for the body that represents them – and in South Asia and Africa some languages were recognized as spoken only by certain tribal groups in a context of widespread multilingualism, in the Americas and Australia the context was massive indigenous language loss, already far advanced and in prospect even more severe. In America and Australia, where dwindling numbers of indigenous-language speakers confronted this prospect, indigenous communities were voicing acute concerns about their languages and beginning to agitate for revitalization; the Americanist and Australianist linguists with whom these matters were raised were among the first to express a sense of professional obligation to intervene on behalf of small and receding languages (Grinevald 2006: 340–41). It was no accident that the linguists and activists who contributed to the 1992 issue of the journal Language that famously raised the issue of language endangerment within the U.S. professional community all worked with American or Australian languages (Hale, Krauss, Watahomigie, Yamamoto, Masayesva Jeanne, and England 1992). Linguists have long had their own purely professional reasons for wishing to document and describe receding languages (more typically to describe than to document, although Himmelmann 2008 argues powerfully for the greater value of documentation). Matras articulates the chief such reason at the end of the article cited above when he speaks of “the urgent task of securing a diverse linguistic sample corpus for the sake of future generations of students of language” (Matras 2005: 248). But Woodbury represents a growing voice in the profession when he writes that “it is becoming less and less viable for linguists to think of the stakeholders in language documentation to be constituted only of a vaguely-conceived scientific posterity” (Woodbury 2003: 39), and the same may be said where language description and analysis are concerned. The unhappy fact is that whatever linguistic professionals do, it will be inadequate. The complexity and richness of language and its cultural context are such that they escape all our efforts to capture them. Our descriptions will prove to be more incomplete, our analyses more imperfect, and our documentations more limited than we imagine. Twenty years from now a new generation of linguistic professionals will wonder how we could have failed to raise the questions that interest them most or document the kinds of linguistic behavior that have come to preoccupy them. Whether we scatter ourselves across the globe in what promises to be a last-minute attempt to record more receding languages, or embed ourselves deeply in particular small-language contexts in an attempt to record and understand them more fully, much will elude us. But that is no reason for not trying. The forces arrayed against the survival of small languages are formidable, and our efforts to provide support for them are likely to prove inadequate. But that is also no reason for not trying.

private and public in language documentation

441

References Amery, Rob. 2000. Warrabarna Kaurna! Reclaiming an Australian language. Lisse: Swets and Zeitlinger. Austin, Peter K. and Grenoble, Lenore A. 2007. Current trends in language documentation. In Language Documentation and Description 4, Peter K. Austin (ed.), 12–25. London: The Hans Rausing Endangered Languages Project. Callaghan, Catherine A. Unpublished manuscript. Why work with the last living speaker of a language? Chapman, Malcolm. 1978. The Gaelic vision in Scottish culture. London: Croom Helm. Collins, James. 1998a. Their ideology and ours. In Language ideologies: Practice and theory, Bambi B. Schieffelin, Kathryn A. Woolard, and Paul V. Kroskrity (eds), 256– 270. New York: Oxford University Press. ———. 1998b. Understanding Tolowa histories: Western hegemonies and Native American responses. New York: Routledge. Csató, Eva A. and Nathan, David. 2003. Multimedia and documentation of endangered languages. In Language Documentation and Description 1, Peter K. Austin (ed.), 73–84. London: Hans Rausing Endangered Languages Project. Dauenhauer, Nora Marks, and Dauenhauer, Richard. 1998. Technical, emotional, and ideological issues in reversing language shift: Examples from Southeast Alaska. In Endangered languages: Language loss and community response, Lenore A. Grenoble and Lindsay J. Whaley (eds), 57–98. Cambridge: Cambridge University Press. Debenport, Erin. 2009. “Listen so you can live life the way it’s supposed to be lived”: Paradoxes of text, secrecy and language at a New Mexico pueblo. PhD dissertation, University of Chicago. Dennison, Norman. 1977. Language death or language suicide? International Journal of the Sociology of Language 12: 13–22. Dixon, R. M. W. 1991. The endangered languages of Australia, Indonesia, and Oceania. In Endangered languages, R. H. Robins and E. M. Uhlenbeck (eds), 229–55. Oxford: Berg. Dobrin, Lise. 2008. From linguistic elicitation to eliciting the linguist: Lessons in community empowerment from Melanesia. Language 84(2): 300–24. Dorian, Nancy C. 1987. The value of language-maintenance efforts which are unlikely to succeed. International Journal of the Sociology of Language 68: 57–67. ———. 2001. Surprises in Sutherland: Linguistic variability amidst social uniformity. In Linguistic fieldwork, Paul Newman and Martha Ratliff (eds), 133–51. Cambridge: Cambridge University Press. Fishman, Joshua A. 1991. Reversing language shift. Clevedon: Multilingual Matters.

442

chapter 23

Flores Farfán, José Antonio. 2001. Culture and language revitalization, maintenance, and development in Mexico: The Nahua Alto Balsas communities. International Journal of the Sociology of Language 152: 185–197. Florey, Margaret. 2008. Language activism and the “new linguistics”: Expanding opportunities for documenting endangered languages in Indonesia. In Language Documentation and Description 5, Peter K. Austin (ed.), 120–135. London: Hans Rausing Endangered Languages Project. Grinevald, Colette. 2001. Encounters at the brink: Linguistic fieldwork among speakers of endangered languages. In Lectures on endangered languages 2, Osamu Sakiyama and Fubito Endo (eds), 285–313. Kyoto: Endangered Languages of the Pacific Rim. ———. 2005. Why the Tiger Language and not Rama Cay Creole? Language revitalization made harder. In Language Documentation and Description 3, Peter K. Austin (ed.), 196–24. London: Hans Rausing Endangered Languages Project. ———. 2006. Worrying about ethics and wondering about “informed consent”: Fieldwork from an Americanist perspective. In Lesser-known languages of South Asia, Anju Saxena and Lars Borin (eds), 339–370. Berlin: Mouton de Gruyter. Hale, Ken, Krauss, Michael, Watahomigie, Lucille J., Yamamoto, Akira Y., Craig, Colette, Masayesva Jeanne, LaVerne, and England, Nora C. 1992. Endangered languages. Language 68(1): 1–42. Himmelmann, Nikolaus. 2008. Reproduction and preservation of Linguistic knowledge: Linguistics’ response to language endangerment. Annual Review of Anthropology 37: 337–350. Hinton, Leanne. 1994. Flutes of fire: Essays on California Indian languages. Berkeley: Heyday Books. ———. 2009. Language revitalization at home. Paper delivered at the First International Conference on Language Documentation and Conservation, Manoa, Hawaii, March 14, 2009. Accessible at http:/scholarspace.manoa.hawaii.edu/handle/10125/5961. Jones, Mari C. 1998. Language obsolescence and revitalization: Linguistic change in two sociolinguistically contrasting Welsh communities. Oxford: Oxford University Press. King, Kendall A. 2001. Language revitalization processes and prospects: Quichua in the Ecuadorian Andes. Clevedon: Multilingual Matters. Kulick, Don. 1992. Language shift and cultural reproduction: Socialization, self, and syncretism in a Papua New Guinean village. Cambridge: Cambridge University Press. Ladefoged, Peter. 1992. Discussion note. Another view of endangered languages. Language 68: 809–811. MacCaluim, Alasdair. 2007. Reversing language shift: The social identity and role of adult learners of Scottish Gaelic. Belfast: Cló Ollscoil na Banríona. Matras, Yaron. 2005. Language contact, language endangerment, and the role of the “salvation linguist”. In Language Documentation and Description 3, Peter K. Austin (ed.), 225–251. London: Hans Rausing Endangered Languages Project.

private and public in language documentation

443

Messing, Jacqueline H. E. 2003. Ideological multiplicity in discourse: Language shift and bilingual schooling in Tlaxcala, Mexico. PhD dissertation, University of Arizona. Milroy, Lesley. 1987. Observing and analyzing natural language. Oxford: Basil Blackwell. Nagy, Naomi. 2000. What I didn’t know about working in an endangered language community: Some fieldwork issues. International Journal of the Sociology of Language 44: 143–60. Newman, Paul. 2003. The endangered languages issue as a hopeless cause. In Language death and language maintenance: Theoretical, practical and descriptive approaches, Mark Janse and Sijmen Tol (eds), 1–13. Amsterdam: John Benjamins. Ó Baoill, Dónall P. 1987. Phonological borrowing in Irish and problems of orthographical representation. In Third International Conference on Minority Languages: Celtic Papers, Gearóid Mac Eoin, Anders Ahlqvist, and Donncha Ó hAodha (eds), 89–103. Clevedon: Multilingual Matters. Pecos, Regis, and Blum-Martinez, Rebecca. 2001. The key to cultural survival: Language planning and revitalization in the Pueblo de Cochiti. In The green book of language revitalization in practice, Leanne Hinton and Ken Hale (eds), 75–82. San Diego: Academic Press. Terrill, Angela. 2002. Why write books for people who don’t read? A perspective on documentation of an endangered language from Solomon Islands. International Journal of the Sociology of Language 155/56: 205–219. Thieberger, Nick and Musgrave, Simon. 2007. Documentary linguistics and ethical issues. In Language Documentation and Description 4, Peter K. Austin (ed.), 26–36. London: The Hans Rausing Endangered Languages Project. Wilkins, David. 1992. Linguistic research under Aboriginal control: A personal account of fieldwork in Central Australia. Australian Journal of Linguistics 12(1): 171–200. Wilson, Peter J. 1992. Freedom by a hair’s breadth: Tsimihety in Madagascar. Ann Arbor: University of Michigan Press. Wittenburg, Peter. 2003. The DOBES model of language documentation. In Language Documentation and Description 1, Peter K. Austin (ed.), 122–139. London: Hans Rausing Endangered Languages Project. Woodbury, Anthony. 2003. Defining documentary linguistics. In Language Documentation and Description 1, Peter K. Austin (ed.), 35–51. London: Hans Rausing Endangered Languages Project. Wurm, Stephen A. 1999. Language revivalism and revitalization in Pacific and Asian areas. International Journal of the Sociology of Language 137: 163–72.

Author Index Abbi, Anvita 273n8 Al-Ali, Mohammed N. 244 Aikhenvald, Alexandra Y. 15, 17, 24–25, 179–185, 189–191, 298, 298n2, 305, 348, 367, 400n1, 407 Al-Khatib, Mahmoud 244 Allard, Réal 302, 305 Amery, Rob 432, 441 Andersen, Roger W. 133 Anonby, Stan J. 303, 305 Austerlitz, Robert 92 Austin, Peter K. 3, 25, 441 Ayoungman, Vivian 244 Barber, Carroll G. 276, 281 Bavin, Edith L. 178, 191 Beniak, Édouard 115, 133–134 Berge, Anna 8, 21, 23, 25 Blench, Roger 291, 302, 306 Blommaert, Jan 289, 306 Bloomfield, Leonard 8, 25, 66, 91, 138, 145, 183–185, 190–191 Blum-Martinez, Rebecca 434, 443 Boeschoten, H. 345–346 Bord na Gaeilge 229, 233 Borgström, Carl Hj. 61, 65 Bourhis, Richard Y. 193, 200–201 Bradley, David 21, 25, 293, 295, 306, 410, 422 Brandt, Elizabeth A. 244 Brandl, Maria M. 19, 25 Breatnach, Risteard B. 62, 65 Breitbach, Joseph 347–348, 367 Briggs, Charles L. 314n2, 328 Briggs, Jean L. 407 Broad, William 388

Bruton, John 278, 278n11, 281 Burridge, Kate 244–245 Callaghan, Catherine A. 437, 441 Campbell, John Lorne 148, 155 Campbell, Lyle 16, 26 Castile, George P. 223, 233 Cawley, M. 233 Chafe, Wallace 332, 346 Chapman, Antony J. 193, 200 Chapman, Malcolm 433, 441 Chatsis, Annabelle 14, 24, 26 Childs, G. Tucker 22, 26 Cole, Debbie 14, 24, 26 Collins, James 17–19, 26, 409–410, 422, 432, 434–435, 441 Committee on Irish Language Attitude Research 229, 233 Cooley, Ralph E. 381–383, 390 Cooper, Robert L. 221, 233, 282, 328 Costello, John 220 Cotter, Colleen Marie 365, 367 Coulmas, Florian 247, 262, 271, 279, 281 Craig, Colette (see also Grinevald, Colette) 27, 440, 442 Csató, Eva. A. 425, 441 Dal Negro, Silvia 400n1, 407 Dauenhauer, Nora Marks 280, 287, 306, 437, 441 Dauenhauer, Richard 280, 287, 306, 437, 441 David, Nathan 441 Debenport, Erin 4, 17, 26, 439, 441 Decime, Rita 273, 281 Deloria, Vine, Jr. 406–407

* Both the Author Index and the General Index were assembled with the invaluable help of four student interns from the Tri-College Linguistics Department: Madeleine Booth, Swarthmore College; Zhiyin Ding, Bryn Mawr College; Gregory Nisbet, Haverford College; and Micah Walter, Haverford College. I’m grateful to the students for their interest and energy and to the Tri-College Linguistics Department for sponsoring their internship, which entailed a trip from Philadelphia to Maine.

446 Denison, Norman 174, 176, 207, 220, 377, 388 De Vos, George 193, 200, 231, 233 Dixon, R. M. W. 20, 26, 373, 388, 436, 439, 441 Dobrin, Lise 417n8, 422, 426, 436, 441 Drapeau, Lynn 261–262 Dressler, Wolfgang 8, 26, 67, 90, 92, 311, 326, 328 Eades, Diana 16, 26 Edwards, John 19, 26, 288–289, 289n1, 306 Ellis, Peter Berresford 210, 221, 233, 256, 262 England, Nora C. 27, 440, 442 Evans, Nicholas 1, 16, 26, 295, 306 Fennell, Desmond 228, 233 Fenyvesi, Anna 115, 134 Fillmore, Lily Wong 114, 162, 165, 218, 221 Fishman, Joshua A. 65, 147, 158, 165, 194, 201, 210, 221, 235, 244–245, 249n1, 251, 255, 258, 261–262, 280–281, 425, 430, 441 Flinn, Juliana 276–277, 281, 286, 306 Flores, Amelia 417n5, 423 Flores Farfán, José Antonio 14, 27, 434, 442 Florey, Margaret 437, 439, 442 Foley, William A. 23, 27, 420, 422 Foot, Hugh C. 193, 200 Gaeliconline 242, 245 Gahng, Tae-Joong 273, 283 Gal, Susan 19, 27, 174, 176, 208, 221, 377, 379–381, 388 Gambhir, Surendra 376, 388 Gardner, Robert C. 212, 218, 221 Gardner-Chloros, Penelope 345–346 Gauchat, Louis 63, 65 Gellner, Ernest 266–267, 282 Georges, Robert A. 384, 388 Giles, Howard 193, 200–201 Glazer, Nathan 201 Goffman, Erving 397, 408 Goddard, Ives 24, 27 Golla, Victor 417n9, 422 Greene, David 65, 174, 176, 229, 233 Grillo, Ralph D. 266–268, 282, 288, 306 Grinevald, Colette 17, 24, 27, 410, 416, 421–422, 426, 437, 439–440, 442 Gumperz, John J. 156, 158, 165, 345–346 Gunn, A. 197, 199–201

author index Haiman, John 293, 306 Haas, Mary R. 1, 12, 27, 137–138, 143, 145, 372, 375, 389, 409, 422 Hale, Ken 20, 27, 440, 442 Hamp, Eric P. 92, 209, 221, 258, 260–262, 392 Hansen, Magnus Pharao 234n1, 244–245 Harper, Gilford 417n5, 423 Haugen, Einar 259, 262 Heath, Shirley Brice 175–176, 265, 268n1, 278n12, 282 Henry, Reginald 372, 389 Heyd, Uriel 255, 262 Hill, Johnny Jr. 417n5, 423 Hill, Jane H. 12–13, 27, 67, 92, 143, 144n7, 145, 149, 155, 168, 254, 262, 372, 379, 381, 389 Hill, Kenneth C. 67, 92, 254, 262, 372, 379, 381, 389 Himmelmann, Nikolaus P. 20–21, 27, 411, 417n11, 422, 425, 440, 442 Hindley, Reg 253, 262 Hinojosa, Maria de la Paz 174–175, 177, 208, 220–221 Hinton, Leanne 241, 244–245, 430, 439, 442 Hohenthal, W. D. 193, 201, 208, 221, 276, 282 Holmer, Nils 39n9, 65 Householder, Fred W. 92 Huffines, Marion Lois 174, 177, 258, 261–262, 287, 306, 377, 389 Huss, Leena 302, 302n3, 306 Hymes, Dell H. 64–66, 90, 92, 160, 162, 165, 193n1, 201, 208–209, 221, 311, 328 Ibrahim, Muhammad 247, 262 Irvine, Judith T, 19, 27 Jackson, Jean 193, 201 Jocks, Christopher 409, 423 Jones, Mari C. 133–134, 178, 191, 431, 442 Jones, Michael O. 384, 388 Joseph, John Earl 268–269, 282 Karttunen, Frances 208, 221 Keane, M. 233 Khleif, Bud B. 282 King, Kendall A. 28, 231, 245, 431, 434, 442 Kiparsky, Paul 38n7 Kloss, Heinz 169, 177

447

author index Krauss, Michael E. 20, 66, 92, 137, 143–145, 149, 155 Kroskrity, Paul V. 17, 19, 27, 178, 191, 275, 270n3, 277, 280, 282, 296, 298–300, 303, 306, 409, 423 Kulick, Don 287, 439, 442 Kuryɫowicz, Jerzy 82, 92 Kushner, Gilbert 223, 233 Kuter, Lois 20, 28, 226, 233, 235, 245, 257, 262, 266, 282 Labov, William 63, 65, 101, 114, 144–145, 158, 163, 166, 311, 328 Ladefoged, Peter 22, 26, 28, 417, 423, 437, 442 Lambert, Wallace E. 91–92, 218–219, 221, 302, 307 Landry, Rodrigue 302, 305 Langgaard, Per 274, 282 Lee, Jennifer 249–251, 262 Lehmann, Winfred 35, 65 Leonard, Wesley Y. 16, 28 Lewis, (E.) Glyn 29, 146, 155, 210, 221, 291 Leyew, Zelealam 349, 367 Long, Michael H. 14, 28 mac a’ Ghobhainn, Seumas 210, 221, 223, 233 Macaulay, R. K. S. 336n5, 346 MacAulay, Donald 366–367 MacCaluim, Alasdair 433, 435, 442 MacEachern, Scott 292–293, 307 MacLaren, James 41n14, 65 MacKay, J. 201 Mackey, William F. 146, 155 Maclean, Sorley 366–367 Macnamara, John 210, 221, 227–228, 233 Maguire, Gabrielle 244–245, 259–260, 262 Malinowski, Bronislaw 397, 408 Margolin, David 224–225, 233 Masayesva Jeanne, LaVerne 27, 440, 442 Matras, Yaron 240, 245, 417n10, 423, 437–440, 442 McCorkle, Thomas 193, 201, 208, 221, 282 McEwan-Fujita, Emily 14, 28 McGill, Stuart 3, 25 Mc Laughlin, Fiona 294–295, 307 McLendon, Sally 17, 28, 209, 221 Meillet, Antoine 273n8, 282

Menn, Lise 176, 311, 328 Mertz, Elizabeth 147, 155, 174, 177 Melchers, Gunnel 379, 389 Messing, Jacqueline H. E. 430, 443 Miller, Elmer 22, 28 Miller, Wick R. 66, 92, 149, 155, 376, 383, 389 Milroy, James (Jim) 18, 28 Milroy, Lesley 163, 395, 408, 415–416, 423, 426–428, 443 Mithun, Marianne 349, 367, 372, 389 Miyashita, Mizuki 14, 24, 26 Mohan, Peggy 385, 389 Moore, Robert E. 20, 28 Mougeon, Raymond 115, 133–134 Muehlmann, Shaylih 20, 28 Mühlhäusler, Peter 67, 92 Muntzel, Martha 16, 26 Musgrave, Simon 425, 436, 443 Nagy, Naomi 299, 307, 410, 423, 439, 443 Nathan, David 425, 441 New York Times 266, 282 Newman, Paul 4, 28, 417n10, 423, 437, 443 The Northern Times 421, 423 Ó Baoill, Dónall P. 253, 260, 263, 435, 443 Obler, Loraine K. 176, 311, 328 O’Callahan, Joseph 256–257, 263 Ó Cinneide, M. S. 229, 233 Ó Ciosáin, Séamus 223n1, 227, 233 Ó Cuív, Brian 239, 245 Ó Dochartaigh, Cathair 2, 28, 402, 408–409, 423 Oftedal, Magne 39n9, 41n14, 65, 353, 368, 401, 408 Ó hIfearnáin, Tadhg 243, 245 Ó Riagáin, Pádraig 238–239, 243, 246, 263 Ó Tuathaigh, Gearóid 239, 246 Paulston, Christina Bratt 264n, 290, 307 Peacock, John Hunt 14, 29 Peal, Elisabeth 302n3, 307 Pecos, Regis 434, 443 Penfield, Susan D. 417n5, 423 Pensalfini, Rob 400n1, 408 Perley, Bernard C. 13–14, 29 Petersen, William 193, 201 Polinsky, Maria 11, 29

448 Pride, J. B. 169, 177, 307 Pulte, William 174, 177, 377, 389 Rankin, Robert L. 220–221 Ratliff, Martha 4, 28 Reinecke, John E. 89n9, 92 Rhydwen, Mari 417, 423 Richards, Jack C. 220–221 Romaine, Suzanne 114, 240, 246, 345–346 Rumsey, Alan 270n3, 282 Sabino, Robin 280, 282, 346 Salisbury, Richard F. 275n9, 282, 295–296, 307 Sall, Thierno Seydou 294, 307 Salzman, Zdeněk 149, 155, 372, 389 Samarin, William J. 88, 92 Sankoff, Gillian 286, 307 Sasse, Hans-Jürgen 10, 15, 29, 115, 134 Saville-Troike, Muriel 277, 282 Schmid, Monika S. 11, 29 Schmidt, Annette 20, 29, 107n6, 114, 178, 192, 252, 263, 373–374, 377, 379, 381, 389 Schooling, Stephen J. 303, 307 Schumann, John 218, 221 Seliger, Herbert W. 19, 29, 272, 282 Serrato, Angelina 417n5, 423 Shipley, William (Bill) 16, 29, 417n7, 417n9, 423 Silva-Corvalán, Carmen 115, 133–134 Silverstein, Michael 87n8, 146, 179, 192–193, 270, 283 Skutnabb-Kangas, Tove 302n3, 307 Smalley, William A. 279, 283, 301, 307 Smith, Jean R. 193, 200 Sorensen, A. P., Jr. 297, 307 Spolsky, Bernard 230–231, 233 Stebbins, Tonya 14, 29 Steiner, George 348, 368 Stenson, Nancy 220, 222 Stern, Asher 277, 283 Stevick, Robert D. 63, 65 Swadesh, Morris 137, 143, 145, 372–373, 378, 389 Taylor, Donald M. 193, 200–201 Tannen, Deborah 325n7, 328 Tedlock, Barbara 406, 408, 415, 423 Terrill, Angela 235, 246, 417n6, 423, 426, 436, 443

author index Thieberger, Nick 425, 436, 443 Thomas, Lewis 388–389 Thomason, Sarah G. 8–9, 29, 115, 134 Thomson, Derick S. 211, 222 Timm, Lenora A. 174, 177, 225, 233, 273n8, 283, 372, 377, 389 Tollefson, James W. 236, 246 Trudgill, Peter 374–375, 389 Tschirch, Fritz 256n3, 263 Tsitsipis, Lukas D. 15, 20, 29, 168, 177, 375–377, 379, 389 Tsunoda, Tasaku 410, 424 Tucker, Benjamin V. 417n5, 423 Vago, Robert A. 19, 29, 272, 282 Valenzuela, Pilar 235, 246 Van Maanen, John 397, 408 Vasquez, Nora 417n5, 423 Verhoeven, L. 345–346 Verschueren, Jef 289, 306 Voegelin, Charles F. 178, 192, 373, 389 Voegelin, Florence M. 178, 192, 373, 389 Wade, Nicholas 388 Walker, Alastair 417, 424 Walsh, Michael 19, 25, 244, 246 Watahomigie, Lucille J. 27, 440, 442 
Waterman, John T. 210, 222, 374, 390 Watson, Joseph 156n1, 166 Will, Vanessa Katharina Angela 14, 29 Williams, Glyn 289, 307 Wilkins, David 439, 443 Wilson, Peter J. 437, 443 Wittenburg, Peter 425, 443 Wolfson, Nessa 332, 346 Woodbury, Anthony (Tony) 21, 29, 349, 368, 440, 443 Woolard, Kathryn A. 14, 23, 29, 237, 283, 288, 307 Wurm, Stephen A. 16, 29, 434, 443 Wylie, Jonathan 224–225, 233 Yamamoto, Akira Y. 27, 440, 447 Yamane, Linda 236, 246 Yoder, Cecelia K. 381–382, 390 Yoors, Jan 383, 390 Zengel, Marjorie S.

63, 65

General Index abandonment 5, 155, 195, 264, 266, 272, 300–301, 329 Abbé Gregoire 267 Aborigine, Aboriginals (see also language) 16, 26, 250, 329 Aboriginal Australia 15, 286 access 17, 22, 153, 235, 286–87, 290, 305, 397, 404, 413–414, 417, 419, 426–427, 429, 432 accuracy 22, 96, 143, 373, 409 accessibility 231, 371, 425 acquisition (dialect/language) 1, 9–10, 13–15, 20, 107, 149n2, 153, 170, 212, 235, 239, 288, 292, 296, 300–301, 303, 376, 386, 425, 432, 433–35 adult 14 child 14, 372 foreign-language 206, 212 second-language 212, 218, 433 acquisition(al) history 10, 107, 213, 213n4, 214, 219 activist, activism 235, 237–238, 241–242, 417, 439–440 adjective 39, 70, 98, 115, 140, 259, 269n2, 319n12, 355, 363 possessive 217 adjective comparison 188 adverb (see also particle, phrase) 23, 39, 49, 98, 118, 118n3, 118n5, 132, 315n4, 347, 403 directional 117, 118n3, 119–121, 132, 188–189 locational 117, 119–121, 123, 129, 132, 188–189 adverbial marker 346, 355 Ælfric 261 affiliation 146, 193, 292 Africa 210, 269, 286, 291–292, 294, 440 age (as a variable; see also variation) age-and-proficiency continuum, age-based proficiency continuum 6, 9, 16, 116, 118n5, 119, 122–123, 124–131

age group 36, 44, 46, 48, 63, 68n2, 94, 96, 178, 189, 213, 217, 219, 230, 299, 373, 401, 431n1 age-based proficiency groupings 186–190 age-related shift pattern 300 age-related variation 403 age and speaker skills 178–192 age mate 119, 230, 299 agriculture 211, 273, 407 agricultural population 33, 36n2 agriculturalist 33, 195, 351, 396 Akkad, Akkadian 271 Alabama 227, 277n10, 280 Alabama Koasati 277n10 Albania, Albanian 168, 209, 258, 260, 374–375 alienation 247, 434 allomorph 82, 82n5, 83, 90, 219, 319, 353, 400 Alsatian, Alsatians 267, 288, 345 alternation (linguistic; see also vowel) 34, 143 Amazonia 180, 185 Amazon Basin 284, 293 America (see also Indian) 208, 301, 439, 440 North America 205, 208, 264 South America 16, 264, 439 the Americas 432, 440 Amharic 349 Amish 148, 244, 258 analog 73, 255, 406 analogy (see also change, leveling, regularization) 38, 43–44 analogical plural formation, analogically formed plurals 85, 214 analysis (as language structure; see also grammar) 219 analytic(al) construction 140–141 analytic structure 143, 190

* Boldface entries represent chapter titles and chapter section or subsection headings that feature the subject of the index entry.

450 ancestor, ancestry (see also identity, language) 17, 280, 331, 433 ancestral community 438 ancestral culture 232 ancestral identity 236 ancestral language 11, 13, 17, 148, 151, 178, 181, 199, 227, 236, 241–242, 248, 264–266, 269n2, 272–273, 275, 277n10, 280, 285, 287, 299–301, 304–305, 311, 329, 331, 350, 376–378, 410, 417, 435, 438, 439 ancestral population 230 Andes (see Peru) anomaly 98, 149 anthropologist 3, 13, 22, 270n4, 295, 298, 378, 384, 388, 406, 412, 415 apparent time 8, 373, 375 aptitude 24, 212, 216 Arabic 207, 247–248, 252, 261–262 Arawakan 298, 400n1 archaism 181, 184–185 archive, archiving 21–22, 419, 420, 425–427 Argentine Chaco 22 Arizona Tewa 19, 27, 178, 275, 280, 296, 298–301 Arvanitika 10, 258, 260, 374–377, 379 aspect (linguistic) 103, 110 progressive 70 Assamese 284 assessment 95, 146, 179–185, 186, 188–190, 218, 239, 302–303 assimilation (see also pressure) 64, 176, 301, 303, 438; (phonological) 115 assumption 17–20, 137–38, 143, 178, 206n1, 228, 269, 375, 404 Assynt 196 attitude (see also change, purism) 146, 169, 205–206, 211–212, 216, 218, 225, 227–230, 254, 268, 275, 279, 295–297, 298n2, 301–302, 380, 383, 387, 395, 420–421, 435, 439 conservative 64, 252–253, 256 negative 20, 168, 175, 226, 229–230, 276–377, 387 positive 10, 216, 218 puristic 247–249, 252, 258 social 158, 163 Western 20, 266 attrition 1, 146, 213, 215, 215n6, 219–220, 272, 431

general index first-language attrition 11, 133, 272 second-language attrition 133 Australia, Australians (see also Aborigine) 15–16, 19, 58n13, 107n6, 210, 215–216, 236, 249, 286, 290, 295, 373, 379, 410, 432, 439–440 Austro-Hungarian Empire 210 authenticity 253, 257 authority (see also religion) 18, 182, 225, 274–275, 287–288, 375, 382, 384 Autonomous Aosta Valley Region 273 autonomy 238, 274–275, 284–285, 297 awareness 5, 14, 18, 34, 98, 110, 154, 164, 180, 212n2, 226, 259, 270n5, 299, 405, 427 “limits of awareness” 179 Ayas Valley 273 Äy̆iwo 434 Aztec 206, 254, 264, 268n1 backwardness 12, 149, 241 Baniwa 181 Bará 193 Basque, Basques 267, 288 Bauchi area (Nigeria) 291 behavior 24, 51, 56n32, 65, 83, 89, 95, 170, 200, 234, 264, 277, 287, 301, 311, 382, 413, 427, 435 language/linguistic behavior 12, 17, 34, 170, 173, 175, 196, 199–200, 303, 318, 375, 433, 440 respect behavior 276 speech behavior 96, 287 Belfast 259, 395, 415, 426–427 beliefs 17, 162, 225, 270–272, 302–303, 316, 412 Berks County, Pennsylvania 10, 148–149, 149n2, 154, 154n5, 161n5, 170, 172, 230, 383, 386 Bhojpuri 376 bias 46n18, 95, 289, 383, 405–407, 418, 421 bilingualism 12, 19–20, 34, 40, 90, 153n4, 169, 198–199, 205, 207, 213n5, 218, 272, 275, 275n9, 277, 284–307 home-school 169 passive, receptive 213n4, 254, 287 terminal 40, 61–64 bilingual 1–3, 5, 9, 33–34, 36n2, 69, 95, 103, 116–117, 138, 148–149, 157, 159, 164–165, 170–171, 174, 185, 195, 197–199, 272, 285, 189,

general index 311, 327, 329–346, 349–350, 380, 384–385, 387, 402, 419 (fully) fluent 39, 64, 159, 239 (near-)passive 15, 90, 107n6, 154, 157–160, 162, 164, 170 Biloxi 143, 372 birth order 10, 149n2 Boeotia 10 Bolivia 271, 437 Bòrd na Gàidhlig (the Gaelic Board) 242 borrowing 38, 109, 180, 250, 254–255, 261, 296, 318, 320, 348 negative borrowing 115–133 Brazil 194, 208, 298 Brazilian Amazon 15, 400 Breton(s) 8–9, 67, 90, 147, 167, 226, 235, 257, 266–267, 273n8, 288, 372, 377 Britain, Great Britain 4, 169, 175, 207, 215, 236, 269, 290, 384, 394 British (language) 169, 238, 242 broadcast, broadcasting, broadcaster 16, 168, 232, 239, 248, 255, 284, 313, 326, 365, 375 Brora 1, 2, 5–7, 9–10, 24, 33–34, 37n3, 37n5, 38n8, 41n13, 42n15, 51n25, 52–53, 64, 68, 88, 139, 142, 151, 187, 194–195, 197, 217, 316n5, 327, 329, 342, 359, 361–362, 394–395, 402 Bugotu 235 Bulgar 265 Bureau of Indian Affairs 231 Burma 293 California 16, 137, 149, 168, 236, 410, 432, 434, 439 Master-Apprentice Program 240, 244 calque 181–183, 226 Cameroon 292 Canada 167, 169, 208, 210, 237, 284, 412n2 Cape Breton 147, 167, 177 Carnegie, Andrew 227 the Carolines (Western Island group, Micronesia) 276 case (grammar; see also dative, genitive, nominative, vocative) case distinction 112, 400 case system 39, 49–61, 64 caste 175, 290 category: see grammar, morphology Cayuga 372

451 ceilidh (lexical item) 52, 55; (institution) 163, 313 Celtic 197, 217, 316n5, 327, 329, 342, 359, 361–362, 394–395, 402 census 117n2, 242, 299, 331n2, 379, 393 Central Asia 235 ceremony, ceremonials 168, 194, 208, 276 Chadic 291 change (see also analogy, economics, mutation, style, vowel) analogical 82, 82n5 attitudinal 421 demographic 438 grammatical 33–65, 67, 138, 187, 399–401 internally motivated 115 morphological 73 morphophonemic 139–140 mutational 35–36, 40, 51, 61–62 ongoing 7, 34, 39, 170, 399 phonological 63, 399 quantity (phonology) 71–74, 77, 84, 88, 90 social 438 change in progress 5, 8, 35, 37, 39, 58, 63, 116–117, 123–124, 142, 186–187, 402 change process, process of change 13, 115, 133, 144, 190 direction of change 21, 63, 190, 374 rate of change 63–64 Charmey 63 Cherokee 280, 373, 377 Echota Cherokee 280 Chiapanec 16 Chicago, University of 392 childhood 10, 34, 107, 119, 131, 149–152, 178, 195, 217, 392, 402, 417–418, 420 child, children (see learner, speaker) 1–2, 4–5, 9, 13, 38, 94, 112n8, 117, 138, 138n2, 139, 149n2, 150, 152–153, 159, 170–171, 173–175, 178, 180–182, 189n5, 195, 207, 209, 229–232, 237, 240, 250–252, 257–258, 258n4, 259–260, 264, 266, 271, 273, 277n10, 287, 291, 294, 297–299, 301–305, 323, 327, 329, 331, 337, 338n6, 376, 386, 302, 411, 414, 418, 420, 428, 430–431, 433–435, 439 Chile 271 Chimbu 293 Chinese 210, 432

452 Chitimacha 372, 378 Cholón 16 Chontal 268 church 36n2, 148, 251, 259, 312–313, 385, 404 Chuuk State 276 Circassians 277 class (see social class) closed class 124, 127, 255 classroom 11, 206, 211, 218, 226, 272, 304–305, 321, 430–432, 434 classifier 182, 348 clause 7, 102–104, 106, 198, 108n7, 321, 324, 362–364 condition-contrary-to-fact 103 (prepositional) relative 100, 171 subordinate 103, 348 Clearances, Highland Clearances 207 clefting 364–365 clergy 266, 351 code 259, 275, 332, 380 code choice 329–346 code-mixing 275, 373 code-shifting 6 code switch, code switching 96, 318, 330, 333, 335, 337, 341, 344–345 codification 430–431, 434 coinage 254–256, 256n3, 261, 430 colonialism 301 colonial language 292, 297 Columbia 14 communication (see also language of wider communication) 87, 119, 137–138, 226, 234, 260, 266–267, 303, 375 communication network 63, 285 communications 255, 297 community (see also fishing, home, speech, tradition) bilingual 3, 10–11, 149, 345 closed 432 ethnic 13–15, 169, 243, 410–411, 434 exile 11, 216 fishing, fisherfolk 3, 6, 12, 151, 197, 199, 313, 344, 393–394, 396, 402, 404, 428 indigenous 13, 430, 434, 436, 440 language 4, 11, 13–14, 20, 137–138, 144, 149, 178–192, 220, 230, 248, 254, 261, 274, 277, 287, 292–293, 304–305, 345, 395, 437 rural 19, 242

general index small(-language) 18, 187, 209, 254, 266, 277–278, 280, 405 speech 9, 12, 18, 29, 33, 35–36, 85, 114, 128, 156–165, 183, 186, 191, 205, 209, 216, 226, 240, 275, 329, 338, 371–390, 402, 405, 409, 417, 429 community language 11, 153, 191, 239, 345, 395, 438 community members 10–11, 14, 191, 205, 299, 411, 413–414, 416–417, 434 compartmentalization (language) 275, 296 compensation (language) compensatory mechanisms 112–13 competence 22, 25, 162, 183, 291, 301, 376, 378 completeness (linguistic) 10, 18, 137, 409, 425 complexity (see also morphology) 12, 93, 106, 272, 292, 440 comprehension 15, 107n6, 213n6, 218 compromise 169, 176, 207, 224, 237, 247–261, 374, 419 conditional (tense; see also inflection, suffix) 38, 104, 106, 110, 140–143, 171–73, 188–189, 214, 346, 355 confidence 276, 303, 421, 427 self-confidence 225, 230, 272–274, 279 confidentiality 427 conjunction 18–23, 49, 101–102, 106, 108, 108n7, 109, 110–112, 132n9, 140, 171, 224, 274, 319, 354–355, 363, 394, 403 subordinating 214 conquest (see also Normandy) 174, 207, 260, 284, 292 consent, informed 21, 414–424, 425–427, 428–429, 437 conservatism (see also grammar, norm, structure) 38–39, 188, 217, 247–248, 251, 253–255, 258, 277, 374 cultural 371 linguistic 257, 296 paradigmatic 174 consonant length 77–78, 86n7 consultants 18, 20, 183, 406 contact 133, 147–148, 151, 156, 229, 241, 261, 265, 269n2, 284–286, 288, 290, 303, 312–313, 373, 377, 403, 416, 434 language contact 1, 19, 66, 91, 93, 115, 154, 272 language contact phenomena 131, 254

general index language contact settings 115, 133 language contact situations 115, 154, 205–222, 252, 254 contact effects 6, 115 contact with English 115, 241, 318 context (see also culture) discourse 179, 376 fieldwork 425–427 revitalization 429–432 social 1, 244, 396, 399 continuum (see also age, proficiency) 82, 84, 224, 250, 376 contraction 94, 113, 137, 167, 206, 228, 252 control (see also grammar, phonology)  38n8, 67, 98, 109, 153–154, 162, 162n6, 163n8, 171–174, 179, 195, 217–218, 265, 277, 286–287, 292, 302, 385, 414, 437 active, productive 100, 113, 161–162, 216, 219, 312–313 political 444 receptive 158, 377 social 14 convergence (see also grammar) 19, 66, 93, 115, 258–260, 272 conversation (see also interaction, speech style) 5–7, 17, 69, 97, 128, 160, 160n3, 161, 164, 171, 180, 185, 242, 298n2, 315, 327, 329–330, 330n1, 332–335, 337–340, 343–345, 364, 376, 379–382, 409, 415, 418, 420, 426–427 (tele)phone 116n1, 187n3, 330, 340, 343 conversation(al) partner 7, 12, 93, 95–96, 105, 112n8, 119, 139, 186n1, 187n2, 190, 315, 329, 335, 338, 338n7, 343–344, 399, 409 Coptic Egyptian 264 Cornish 36, 236, 256–257 corpus 21, 24, 111, 128, 129n8, 145, 321n14, 328, 332, 363, 372, 440 court 210, 267, 288, 292, 312 courtesy 380 courtesy rules 190 creole, creole language 1, 66, 90, 260, 287, 376 crofter (see also Gaelic) 195, 351, 396 culture (see also conservatism, tradition) 21, 209, 212, 230, 234, 242, 250, 280, 302, 351, 406, 426 ancestral 232

453 material 270 traditional 181, 276 cultural content 179, 236 cultural context 290, 425, 429, 432, 440 cultural disinheritance 232 cultural institutions 230 cultural knowledge 181–182 cultural legacy 243 cultural material 17, 189 cultural practices 178 cultural revitalization 244 cultural settings 179 cultural superiority 295–296 cultural tradition 207, 209, 231 cultural values 210, 285 Cupeño 143, 144n7, 168 currency 68, 146–151, 154–55, 247, 249, 287, 300, 431 language of wider currency (LWC), wider currency language 148, 167, 226, 248, 260, 300, 350, 436 curriculum 232, 239 Cushitic 349 Czech, Czechoslovakia 210 Dakar 294 Dalmatian 36 Danes, Danicization 274 Darwinism (social) 270, 272 data 6, 8, 18, 23, 25, 35, 46n18, 58n33, 68, 75, 89–91, 118, 129, 131, 138, 144–145, 171, 187, 220, 289, 295, 314, 316, 371–390, 392, 396, 402, 410 elicited 119, 122, 126, 314 freely spoken (as opposed to elicited) 119, 122, 126, 314 database 6–7, 404 dative 51–52, 53–59, 60–62 death (see language) bottom-to-top 16 decay 17, 35, 67, 90, 105, 311, 326, 372, 400, 400n1, 404 decline 2, 5, 12–13, 34, 76–77, 80, 83–84, 94, 105, 113, 141, 176, 197, 223, 227, 231, 252, 258, 291, 326, 433 deficiency (see also speaker) 96, 269n2, 373 Delaware 378, 381–382

454 delivery 319, 431n1 manner of delivery 137, 142n6, 157 demography (see also change) 286 demographic base 210–211 Dene 275n9, 295 Denmark 211, 224, 238 Dewey 378 descendant 5, 14–15, 33–34, 36n2, 91, 116, 156, 158–159, 186, 329, 336, 349–351, 394, 418, 428 description (linguistic) 1–2, 13, 17, 21–22, 35–36, 64, 185, 352, 405, 417, 417n1, 437, 440 descriptive linguistics 8, 13 descriptive linguists, descriptivists 10, 18, 179, 400–401 deviation (see also grammar) 12, 37, 43, 53, 94, 105, 157, 178, 182, 186n1, 190, 374–375, 435 device see grammar, morphology, syntax dialect 2, 7, 67–68, 74, 81, 85, 93–94, 98–99, 117, 142n6, 154, 168, 198–200, 224, 226, 268–269, 375, 392, 400n1, 411 dying 2, 33–65, 68, 90, 113, 149, 154n5, 196 Breton 90, 372 English 171, 226 French 267 German 148, 297, 230, 382 Gaelic 36, 64, 67–68, 133, 147–148, 156, 186, 226, 255, 315, 331, 335, 345, 350, 352, 365, 393, 395–398, 401–402, 409, 411, 418 local 7, 117, 147–148, 209, 227, 243, 385–386, 409, 411, 431, 434 mainstream 117, 138, 241, 353, 404, 418 minority 225–226 prestigious 154n5, 268 regional 154, 237, 254, 268, 312, 379, 437 unwritten 36n2, 113, 234, 317 dialect geography 395, 401 dialect (in)tolerance 14, 226 dialect variation 226 dialect(al) differences 183, 225, 253, 376 dialectological features 4–6 dialectological survey 4, 411 dialectologist, dialectology 241, 393 dictionary 18, 22, 251, 411, 417, 417n6, 434, 436, 439 diglossia 312 diglossic relationship 376 diminutive (see also suffix) 7, 320n13

general index direct address 118, 122, 128, 129n8, 132, 341, 352, 412n3 direction of change see change disappearance 12, 19, 61, 93–94, 115, 239, 241, 275, 348, 416 discourse (see also particle) 347, 352, 367, 429 context 179, 376, 425 device 269 effects 349, 352, 364 tone 269n2, 364 discrimination 69, 149, 420–421, 438 disfluencies 95, 190 displacement 205, 301 distance, distancing (see also social distance) 36, 63, 252, 438 emotional (self-)distance/distancing 217–218 distancing effect 265, 347, 363 distinctiveness (see also occupation) 2, 200, 209, 353, 367, 438 ethnic 266, 420 linguistic 33, 199 distinction (language-related; see also grammar) 8, 13, 60, 68, 81, 84, 104, 106, 112, 117, 123, 132, 188–89, 253, 294, 319n12, 349, 400, 404, 417n11 social distinction (see also status) 403 disuse 5, 93, 402 diversity 209, 266, 279, 289–290 documentation 13, 16, 18, 20–25, 36, 199, 236, 409–443 Dom, Domari 240, 438 domain (language-related) 19, 146, 288–89, 326, 352, 376 domain-separation 176 dominance (see also language) 19, 68, 90, 95, 265, 267, 276, 278, 286, 292, 301, 345, 429 dominant group 275–276, 272, 278n12, 285, 290, 300–301, 303–304 domination 207, 261, 286 Dornoch 156, 194, 335, 357 Dunrobin 156 Dupaningan Agta 421 Dyirbal 128, 426, 430, 433 Young People’s Dyirbal 435 East Sepik Province 436 Eastern Pomo 17, 28, 209, 221

general index Easter Ross 156 Echota (see Cherokee) 280 economy 272–273, 278, 285 cash 178 farming 148 economics (see also dominance, pressure) economic base 148, 273, 438 economic change 224, 438 economic conditions 207–208, 403 economic development 227, 235, 239 economic factor 175, 235, 302 economic growth 228 economic incentive 174 economic opportunity 175, 299 economic success 272, 303 Ecuador 271 Edinburgh 151, 217, 393, 396, 412 University of 397 education (see also Gaelic, Welsh) 148, 164, 197, 211, 238–239, 242–243, 254, 277, 284, 298, 301–303, 312, 327 bilingual 230–232, 436 primary 241 secondary 298 education(al) system 239, 255, 273, 429 elder (church) 312 elicitation 5–9, 11, 18, 69, 129, 144, 171, 291, 381–383, 385–387, 398–99 elimination 118, 289, 434 elite (see also urban) 148, 210, 238, 247–248, 293 Embo 1, 2, 5–7, 9–11, 24, 26, 33–39, 41–45, 49–54, 56–58, 63–64, 68, 88, 98, 102, 104, 112n8, 116, 116n1, 117, 117n2, 118, 118n5, 119, 123–124, 127, 129, 131–133, 138–139, 142–143, 186–191, 194–195, 197–198, 258, 313, 319, 327, 334–335, 337–340, 342, 344, 354, 356–357, 364, 394–395, 399–403 Emenyo 275n9, 295–297 emigration 146, 238 emphasis see suffix employment 239, 287, 298–299 enclave 209, 257, 261, 291 enclaved community/people/population 9, 223, 371 enclavement 275, 301 endangerment (language) 2–3, 12–17, 235, 243, 270n4, 287, 440 endangerment crisis 20, 290 endangered language see language

455 endearments 118, 132 endogamy 148, 159 England 265, 289, 416, 440 English 3–5 American 123n6, 227, 229 East Sutherland 68, 106, 327 Middle 392, 401 nonstandard 101 Old 392 Oxford 227, 229 Standard 16, 68, 103, 376 erasure 420 errors 38, 173, 220 ethics 416 ethical dilemmas, problems 21, 411 ethical issues, questions 421, 427 Ethiopia 349 ethnicity (see also community, distinctiveness, heritage, identity, language) 193– 196, 199–200, 225, 236, 268, 279–280, 379, 399, 438 ethnic group 16, 22, 156, 164, 200, 215, 235, 237 243, 266, 304, 418–419 ethnic heritage 231 ethnic history 230 ethnic marker 193–202 ethnic past 232, 378 ethnic population 17, 235–37, 242, 418 Europe, Europeans 169, 210, 216, 250, 266, 268–270, 274, 278, 278n11, 278n12, 297, 300, 366, 432, 439 European Community 279 European Union 240 evaluation 137, 143, 158, 163–164, 178–192, 237, 260, 275, 378 self-evaluation 216 eviction 404 evictees 116, 351, 407 evidential 181–183, 349 evidential marker 271 evidential property 347 exile (see also community) 133, 139, 139n4, 144, 149, 149n2, 151, 152n3, 154, 159, 215, 215n6, 216–218, 371 exile/exiled (semi-)speaker 142, 152, 213–214 exploitation 416, 421 exogamy (see also marriage) 293 expansion (see also language) 61, 66, 118n3, 197, 208, 210, 261, 325, 414

456 expert, expertise 13, 21–22, 182, 211, 244, 281, 378, 388, 398, 437, 439 expert knowledge 179, 297, 406, 417 exposure (to language) 6, 197, 154, 170, 213, 261, 298, 305, 383 expressivity, expressiveness 179, 272 expressive bleaching 350–352 expressive capacity 270, 349–350 expressive power 362–367 extinction (see also language) 33, 35–36, 82, 89, 91, 94, 139, 144, 144n7, 154n5, 167, 207, 209, 211, 213, 220, 270n4, 311–312, 326, 331, 372, 377, 400, 402 Eyak 66 factions, factionalism 24, 256–257 factors (see grandmother) demographic 302 economic, socioeconomic 175, 299, 302 motivating 277, 366 Faetar, Faeto 299 family 167–177 language 143, 298 nuclear 4, 149, 151, 154 Faroese 224–225, 379 fate 20, 66–92, 104, 209, 223, 405 Fatick 294 feature see dialect, discourse, grammar, morphology, salience fieldwork, fieldworkers (see also anthropology) 1–6, 8, 23–25, 33, 46n18, 137–138, 142, 144, 159n2, 175, 186–187, 188n4, 240, 290, 294–295, 314–315, 371, 373, 375–376, 383–384, 386–387, 391–393, 395–398, 404–407, 409–410, 415, 416–417, 418, 420–421, 425–429, 435–437 late-stage 499, 410–414 long-term 25, 391, 406 team, team-based 21, 427 telephone 23 Finnish, Finland 210, 238 First Mesa 278, 301 fishing (see also industry) 5, 12, 33, 116, 148, 150, 407, 411, 418, 420, 435 fishing community 3, 6, 12, 151, 197, 199, 313, 393–394, 402, 404, 428 fishing village 1, 4–6, 11–12, 58n33, 116, 152, 156n1, 172, 241, 335

general index herring fishing/fishery 36n1, 317, 334 line fishing 402, 418, 428 fluency 10, 19, 96, 124, 137–138, 157, 162, 171–172, 174, 179, 186, 186n1, 217, 236–237, 234, 296, 409 focus marker, focus marking 352–367 formality (see also context) 168, 247, 325, 327 formulaic expressions 162, 162n6 fossil, fossilization (linguistic) 40n10, 93, 98 fossil(ized) expression, form, phrase 41, 51n24, 248 Franco (Francisco) 273, 366 Franco-Provençal 273 Franks 265 Frederick the Great 210 French 89n9, 169, 210, 237, 260–261, 267–268, 269n2, 270, 270n5, 279, 284, 288, 293–294, 303, 315n4, 345, 378, 432 Cajun 167 Norman 261 Ontario 115 Pidgin 89n9 Standard 89n9, 273 French Canadian 167 frequency (of occurrence, linguistic) 6, 15–16, 39, 58–61, 63, 70, 73–74, 77–78, 85–86, 90, 97n1, 98, 101, 106–108, 111, 118, 128–129, 132–133, 151, 157, 169, 171, 319, 352, 354, 366, 377 friendship 378, 383–384, 406, 415–416 friendship networks  148, 159, 384 Friulian 207, 377 Fulnio 194, 208 function (language) 6, 12, 91, 113, 146, 194, 205, 209, 238, 312–314, 326, 353, 375, 433 functional restriction 167 functional segregation 169 funding 22, 242, 392–393, 396, 410n1, 421, 415, 433, 435 future (tense) 140, 142–143, 172–173, 333, 336, 346, 355, 358, 361 Gael 433 Gaelic (see also Irish) Bible, Biblical 36, 95, 117, 313, 324, 335, 377 Common 2

general index “correct” 34, 225, 230, 321 crofter 199, 396–397 East Sutherland passim; as section header, 66–91, 116–124 (East Sutherland) fisherfolk 1, 4, 6–10, 12–13, 18, 23–24, 116–117, 133, 190, 196, 240, 394, 397, 412–413 fisher Gaelic 199 fisherfolk 116–117 formal 313 local 1–2, 5, 7, 12, 15, 133, 148, 157–158, 172, 213n5, 241–242, 315, 323, 331, 338n7, 377, 381, 384, 393, 397, 399, 420 locally authentic 243, 253, 420 mainstream 117, 133, 241, 353, 404, 412, 418 non-local 241, 312, 395, 418 Scottish 66–92, 93–94, 116, 147–148, 156, 167, 174, 195, 207, 211, 223–224, 226, 234n1, 241, 254–255, 269n2, 280, 287, 304, 311, 318, 320n13, 329–346, 372, 375, 377, 380, 386, 391–393, 398, 401, 411, 433 Gaelic Board 242 Gaelic League 238–239 Gaelic-medium (primary) education  241, 435 Gaelic-medium schools 241–242, 435 Gaeltacht, Gaeltachtaí 228, 232, 239, 243, 253 Gahuku 296 Ge’ez 264 gender (language; see also grammar, noun) 51, 51n25, 52–53, 60–61, 67, 70n3, 85, 106, 112, 115, 259, 324n16, 348, 352, 355, 400 generation 4, 35–36, 53, 87, 91, 146–148, 154, 156, 180, 182–183, 189n5, 206, 209, 229, 232, 238, 248, 252, 274, 277n10, 287, 290, 298, 298n2, 300, 311, 327, 411, 427, 432, 438, 440 bilingual 39, 64, 249 final, last 298, 311 first-ascending 87, 329 first-descending 182 future 60, 146, 440 grandparental 377 older, oldest 50, 87–88, 147, 180, 258, 311, 439 parental 150, 337

457 second-ascending 87, 91, 150, 229 third 189n5, 209 younger 53, 64, 298n2, 331–332 generational seniority 181–183 genitive 49–51, 51n24, 53–54, 56, 60–62, 259 genocide 168 geography (see also variation) 2, 167, 201, 235, 252, 267, 290, 292, 300, 348, 371, 376, 385, 400n1, 403 dialect 395, 401 geographic isolation 33, 36, 93, 261, 277, 205, 372 geolinguistics (see linguistics) German 148, 148n2, 154, 167–169, 207–208, 210, 230, 256n3, 258, 260, 287, 288, 293, 347, 348, 377, 380, 382, 385–386, 410n1, 432 Alemannic Old High 374 East Middle 374 New World 264, 271 Pennsylvania (Pennsylvania Dutch) 10, 148, 148n1, 149, 153–54, 161n5, 167, 170, 172, 230, 258, 260, 287, 377, 382–383, 385–386, 411n1 Old World 206, 265 Swiss 374 standard 148 Germany (see also language) 210, 226, 289 gerund 40, 67–68, 69–74, 78–91, 99, 101, 123, 123n6, 125, 346, 400 formation 9, 71–74, 78, 80–81, 82–88, 142 Gimi 293 Glamorgan County 146 globalization 285, 438 Golspie 1–2, 5–7, 24, 33–34, 37n3, 37n5, 39n8, 41n13, 42n15, 50, 51n25, 52–53, 63–64, 88, 104, 194–198, 315n3, 316n5, 317n5, 317n6, 317n7, 319–321, 323, 325, 327, 329–332, 340–341, 353, 358–359, 394–396, 402, 409 government (see also nation, support) 146, 148, 168, 211–212, 228–229, 235, 237–238, 242, 279, 284, 297, 302, 304, 429 government services 211, 232 grammar (see also change, deviation) analytic 49, 62 synthetic 62 grammatical category 49, 49n20, 50, 62, 67, 105, 132, 142–143, 190, 226, 352

458 grammatical completeness 18 grammatical conservatism 186–187, 189 grammatical control 11, 113, 162 grammatical convergence 19 grammatical device 118n5, 347–349, 351–352 grammatical deviation 157, 186n1, 435 grammatical distinction 8, 13, 319n12, 404 grammatical features 4, 10, 115–116, 188, 241, 259, 348, 403 grammatical form 18, 186, 190 grammatical norm 64, 105, 157–158, 163, 405 grammatical structure 117, 123–124, 128–129, 133, 187–188, 305, 351 grammatical system 35, 38–39, 49, 62 grammatical(ity) judgment 18, 294 grandchild 5, 181, 431 grandfather 151, 181, 373 grandmother, granny 87, 112n8, 150–152, 431 grandmother factor 87 grandparent 34, 87, 209, 234, 305, 383, 419, 431 Greece 10, 209, 258, 375 Greek 168, 206–207, 209, 260, 271, 376, 379 Greenland, Greenlanders 237, 274, 285 Gros Ventres 372 group see age, dominance, ethnicity, indigenous group, kin, membership, minority, peer, private, resistance, social group, stigma, subordination Guatemala 406, 415, 422 Guayqueries 194, 209 Guyana 376 Guyanese Creole 376 Guyanese Bhojpuri 376 Hamburg (Pennsylvania) 148, 153, 170 Hamp, Eric 209, 259–261, 392 Hapiya 406, 415 Hawaii, Hawaiian 237, 304 Hebrew 210, 223, 237, 256 Hebrides 232, 317, 375 hegemony 292 heterogeneity (linguistic) 267, 288 heritage 22, 152, 230–232, 235, 237, 240, 243, 273, 280, 305

general index heritage language 10–11, 14, 21, 303, 350, 432 hierarchy (see also social) 67, 164, 175, 286–288, 293 Highlands (Scottish; see also Scotland) 1, 33, 94, 97, 105, 116–117, 123, 123n6, 152, 156, 167, 169, 186, 194, 200, 207, 215, 224, 231–232, 240–241, 280, 311, 313, 350, 364, 383, 392, 394, 397, 403, 406, 411, 433 Hindi 247–248, 261, 376 historical linguistics 36, 64, 138 historian 210, 232, 418 history (see also acquisition) 10, 17, 19, 64, 156, 167, 174–175, 217, 223, 226, 231, 235–236, 241–242, 250, 252, 268–269, 272n7, 278, 278n12, 280, 284, 301, 317–318, 350, 372, 375, 382, 387, 409, 418, 430, 434, 438 oral 3, 413–415, 417–418, 418n12, 421, 428 Hittite 271 home (see also transmission) 9, 11, 88, 94, 112n8, 139, 139n4, 152–154, 159, 161, 168, 170, 178, 195, 195n2, 209, 214, 215n6, 232, 234, 237, 244, 252, 258n4, 259, 265, 270, 274, 276, 299, 305, 326, 334, 396, 403, 411, 413, 419, 426, 429, 430–431, 433–435 home community 139, 215, 215n6, 412 home language 20, 88, 149, 149n2, 168, 216, 219, 387, 430 home village 9–11, 15, 149n2, 213, 217, 403, 418 homogeneity 119, 289, 404 homogenization 418 Hopi 275, 277–278, 296, 301 Horom 291 hostility 176, 196, 230, 235, 278, 434, 437 household, householder 47, 94, 107, 112n8, 138n2, 151–152, 159, 164, 172–174, 216, 312, 332, 334, 380–381, 383, 396, 403, 412, 421, 427 Hua 293, 295 human rights (linguistic) 303 Hungarian 19, 208, 210, 377, 379–381 American Hungarian 115 Icelanders 270 identity (see also ethnicity, nation) 12, 17, 174, 194, 199–200, 208–209, 218, 223, 231, 234, 240, 242, 259–260, 273, 292, 295, 280, 300, 345, 380–381, 388, 415, 428, 432–433, 435

general index ancestral 236 disfavored 264 ethnic, ethnolinguistic 193, 195, 198–199, 224, 240, 266, 303–304 linguistic 180, 303–304 local, quasi-local 381, 386 national 242, 266, 287 ideology 4, 15, 17–20, 23, 181, 267, 270, 270n3, 279–280, 284–285, 288–290, 295–297, 301, 305 European 288–289, 301 language, linguistic 185, 191, 264–281, 291 local 17, 179–181 puristic 19 “ideology of contempt” 268–269, 272, 280, 288, 301 idiom 197, 215n6, 253–254 idiosyncrasy 37, 51n25, 56n32, 58, 71, 78, 81, 85, 105, 107, 394, 409 idiosyncratic form, formation 73, 81, 85 idiosyncratic variation 409 illiteracy 36, 63, 105, 113, 183, 327, 404 illiterate society/speech community 138, 183 immersion (see also schooling) 304–305, 435 experience 407 programs 304 immigration 208–209 immigrants 133, 290, 302 immigrant language 205 imperative negative 123–124, 124n7, 125, 128–129, 131–132, 140, 188–189, 214, 217 imperative plural 214 inalienables 41, 41n12, 217 inanimates 70n3, 118, 355 inauthenticity 237 incentive (see also economics) 212, 290, 299 socioeconomic 290, 299 income 149 low-income livelihood/occupation/subsistence mode 116, 241, 438 incomers 167, 331, 333, 338n6 Indian (see Bará, Eastern Pomo, Kwak’wala, Maidu, Menomini, Rumsien Ohlone) 22, 193, 209, 231, 235, 297–298, 300, 378 American, North American 149, 176 East Indian 376

459 Mexican Indian 174–175, 392 South American Indian 194, 208 indigenous community/group/people (see also language) 22–23, 302, 304, 430, 434, 440 indigenous-language survival 243 indigenous-language teacher 244, 430 Indo-Iranian 438 Indonesia 437, 439 industrialization 146, 228–229, 266–267 industry 302 coal 237 fishing 1, 5, 33, 148, 402 herring 36n2 publishing, broadcasting 225 tourist 273 infix 80, 348–350 inflection (see also morphology) 8, 60, 66, 89n9, 184–185, 249, 253, 319, 320n13 1st-person-singular 189 conditional 140, 189 plural 320 synthetic 188–189 vocative 139 informality 316, 326–327, 387, 393, 428 informant 35, 41n13, 43n16, 56n31, 58n33, 69, 71, 90, 105n5, 137, 139n3–4, 144n8, 145, 163, 196, 207, 375, 378, 383, 385, 405 in-group marker/membership/privilege 168, 199, 435 Inka, Inka Empire 264–265, 278n12 in-migration 146, 273, 420 innovation 19, 81, 85, 87, 126 innovative forms 21, 181 institution (see also culture) 148, 228, 230, 284, 312, 411 academic, educational 175, 267, 424 regional 228 state, national 284 institutional support 148, 287, 302, 410n1 instruction 234, 239, 241, 251, 318, 332, 382, 420, 434 school instruction 434 medium, materials of instruction 230, 250, 259, 298 intactness 137, 142, 144, 186, 186n1, 258, 261 integrity 10, 258 structural 94 intelligentsia 274

460 interaction (see also social interaction) 11, 36n2, 69, 88, 107–108, 151, 156, 159, 160–161, 161n5, 162, 164, 171, 213n5, 247, 286, 294, 314n1, 316, 326–327, 330–331, 344, 359, 377–378, 382, 404, 427, 433 conversational 128, 332, 338 cross-generational 159 sender/receiver 113 verbal 114n9, 160, 162, 327, 330, 344 interaction networks 159, 161 interference 219–220, 372 covert 115 lexical 219 structural 219 interjection 321–323, 362 interlocutor 7, 95, 132, 327, 329, 331–333, 338–339, 341–342, 344–345, 384, 402 interlocutor factor 95 intermarriage 275, 301 internet 414, 418 intervention 414, 428, 437 interventionist agenda 439 interview, interviewer, interviewing 2, 15, 45, 46n18, 58n33, 69, 149, 161, 161n4, 196, 212n3, 255, 291, 314, 314n2, 316n7, 326, 328, 330n1, 355, 362, 379–381, 385–387, 420 multispeaker interview 330, 401 intimacy 413, 416, 426, 430, 433 intimate setting 431 intolerance 19, 226, 240, 266, 278n12 invariance 10, 18, 25 investigation 17, 34, 51, 66–67, 100, 147, 311, 316, 371–373, 402 Ireland 168, 193, 228, 231, 233, 239–240, 259, 277, 278n11, 375 Irish, Irish Gaelic 16, 62, 210, 223, 227–229, 231–232, 238–240, 242–243, 252–254, 259–260, 271, 278, 321n13, 351, 364–365, 375 irregularity (see also verb) 38n8, 97, 111, 140–143, 172, 187–189, 219 irregular noun plural 62, 140, 142, 214 isogloss 37n5 isolation (see also geography) 68, 96, 147, 176, 243, 297 linguistic 64 social 93, 277, 285 Israel 237, 277, 283 Isthmus of Tehuantepec 172

general index Italian 207, 209, 273, 299 Italy 209, 226, 237, 377, 410 Northern 237, 400n1 Southern 299 Japanese 212, 432 Jerusalem 240, 438 Jingulu 400n1 job 211, 230, 232, 239, 251, 303 joke 159, 213n5, 321n14 Juchitán 174–175 juxtaposition 109–110, 112, 182 Kachin 293, 295 Kayor 294 Kaurna 236 Kazakh 235 K’emant 349 Keres 434 kin, kin(s)folk 88, 107–109, 112, 181, 187n2, 189, 214, 276, 299, 329 kin circle 15, 429 kin group 276 kin(ship) network 7, 10–11, 107, 117, 159, 171–172, 303, 316, 331–332, 378, 384, 413 kin(ship) tie 5, 119, 159, 181, 413 kinship system 182 kinship term 41, 181 kinsperson, kinswoman 154, 159, 171, 218 Koasati 277n10 Komunku 295–296 Kulere 291 Kurux 273n8 Kwaio 277 Kwak’wala 303, 305 Labov, William 8, 63, 101, 163, 311 Labovian correlational sociolinguistics/ studies 4, 399 Ladin 237, 273–274 lag (linguistic) 193–200 language (see also ancestor, attrition, behavior, colonialism, community, contact, currency, endangerment, heritage, home, loss, maintenance, nation, norm, privacy, revitalization, schooling, shift, structure, support, trade, tradition) Aboriginal 16, 194, 236, 295, 400n1

general index colonial 292, 297 creole 66, 90 disfavored 147, 155, 298, 429 dominant, dominant-group 3, 6, 86, 115–133, 167–177, 216, 220, 243, 260, 265, 272–273, 278–279, 290, 303–304, 371, 429 dying 2, 13, 35, 64, 66–68, 91, 137–138, 148–149, 210, 219, 231, 311, 326, 376 endangered 3, 8, 17, 19–21, 23–24, 280, 284–307, 349–350, 409–410, 436, 439 ethnic 237, 240, 260, 266, 279, 304 expanding 1, 115, 178, 206, 234, 238, 265, 311, 350 extinct 167, 281 first 1, 94, 117, 239, 252, 259, 301, 402 foreign 198, 206, 211–212, 216, 296, 348, 379 Germanic 260, 392 habitual 112n8, 329 heritage 10–11, 14, 21, 235, 237, 240, 303, 350, 432 immigrant 167, 205 indigenous 36, 115–134, 175, 210, 238, 243–44, 249–250, 269–272, 279, 298n2, 430–432, 440 local 12, 15–16, 21, 144, 149, 168, 179, 207, 238, 240, 276, 281, 350, 376, 384, 387, 395, 409, 420, 430, 433, 436, 439 low-prestige 149, 154–155 majority(-group) 1, 115, 168, 287, 290, 302, 305, 434 minority 1, 4, 12, 23, 115, 168–69, 176, 205, 220, 226–227, 236–237, 248, 252, 269, 273n8, 278–279, 284–305, 377, 387, 429–435 native 16, 19–20, 95, 128, 220, 258, 272, 294–295, 377 Native American 236, 437 non-standard (see speech) obsolescent, obsolescing 167–176, 376, 400n1 pidgin 66 prestige 210 private(-sphere) 413, 431, 433 receding 8, 12–13, 19, 33, 178–191, 237, 240, 243, 409–411, 410n1, 425, 436, 438– 440

461 second (see also acquisition, learner, skills, speaker, teaching) 220, 251 secret 16 “sleeping” 16, 243 small (see also community, population, setting) 6, 20, 244, 255, 258, 261, 264–283, 350, 429, 440 standard, standardized 18, 25, 34, 36n2, 40, 40n11, 49n20, 50, 56, 60–61, 183, 247, 253, 268–270, 269n2, 272, 275, 278, 289, 313, 314n1, 353, 359, 374 state-promoted 168, 270, 287 stigmatized, stigmatizing 5, 240, 345, 412 subordinate 93, 266, 268–269, 278n12 superordinate 93, 379 threatened 2, 17, 154, 167–168, 210, 220, 223, 225–226, 240, 249n1, 260–261, 264, 270–271, 274, 349, 375–377 traditional 249–252, 276, 287, 304, 436 unwritten 1, 8, 24, 36n2, 113, 179, 224, 317n9, 405, 436, 438 written 17–18, 41n14, 117, 168, 224–226, 241, 252–253, 256, 271, 317n9, 347, 354, 364, 405, 409, 429, 431, 436 language choice 290 language death 12–17, 19, 66–91, 93–114, 137–145, 154, 154n5, 167, 174, 219–220, 386, 388, 438 language endangerment 2, 3, 12–17, 235, 243, 270n4, 287, 290, 440 language loyalty 10, 67, 139, 146–147, 149n2, 175, 190, 206, 206n1, 207, 216, 376–377 language maintenance 4, 10, 14, 154, 205–222, 223–233, 234–246, 273, 275–277, 290–292, 296, 299, 300–02, 303–304, 417n8, 436 language mixing 17, 180, 260 language obsolescence 2, 4–6, 8–17, 91, 93, 131, 257, 386, 400n1, 402 language program 432, 434, 435 language retreat 6, 12, 16, 223, 311–312, 314, 429 language shift 7–8, 14, 115–133, 146–155, 178, 187n2, 234–237, 240, 276–277, 280, 287, 297, 303, 430, 433, 436–437 language suicide 174, 438 language use 1, 3, 66, 158, 169, 299, 367, 425, 431

462 language of wider communication (LWC) 21, 234–235 Latin 264, 392 Latvia 278 Lavukaleve 436 law 82, 148, 231, 319 learner 11, 16, 21, 107, 133, 158, 162, 241–242, 247, 256–257, 259–260, 304, 351–352, 366, 407, 409 second language learner, learning 11, 218–219, 237, 348, 433, 435 foreign language 158, 162 legal system 231, 284 lenition (consonant mutation) 37, 37n3, 39–40, 41n13, 42–44, 45n17, 49n20, 49–50, 52–53, 55–59, 61–62, 100–101, 106, 112, 118, 118n5, 128, 132, 319, 323, 351–352 obligatory-lenition adjuncts 39–40 leveling 62, 88, 260 analogical 34, 62, 140–144, 353 Lewis, Isle of 224 lexicon 15, 34, 38, 60, 63, 67, 110, 180, 207, 215n6, 243, 249, 254–256, 319, 324–325, 348, 412, 431–432, 434 open-class lexicon 110 lexical gaps 142n6 lexical interference 219 lexical items 4, 8, 15, 24, 105, 157, 247, 259, 291, 314, 362, 393, 395, 397, 409 lexical recall 96 lexical verb 70, 99n2 lexicostatistics, lexicostatistical test/(word-) list 69, 380–381 lifeways 230, 297, 421 lingua franca 19, 287–288, 291–292, 296–297, 300 linguist 3, 8, 14, 16–23, 63, 114, 186, 188–190, 178–179, 191, 240, 252, 270n4, 272, 295, 298, 314n2, 348–351, 366, 378, 384, 388, 391–392, 400n1, 405–406, 410, 436–437, 439–440 Americanist 440 Australianist 440 descriptive 10, 13, 400 “parachute” linguist 21 “salvation” linguist 439 Linguistic Survey of Scotland, Survey of the Gaelic Dialects of Scotland, Survey 2–4, 8, 372, 386, 392–396, 398, 403, 405, 409

general index linguistics (see also description, sociolinguistics) 349, 437 descriptive 8 historical 36, 64, 138 structural 411 geolinguistics 146 Lisu 293 literacy 63, 168, 224, 226, 248, 251, 255, 266, 387, 404 430, 434 passive 36n2, 117, 314n1, 335 livelihood 33, 116, 230, 241, 418 loanword 73, 96, 146, 153, 196, 200, 220, 226, 247, 249, 256n3, 258, 291, 318, 319, 319n12, 320, 320n13, 323, 325, 341, 372, 396, 403 loanword avoidance, avoidance of loanwords 23, 260 London 11, 112n8, 151–152, 152n3, 153, 152n3, 159, 215n6 lore (see also proverb) 3, 181, 230, 236, 378, 387 loss 94, 113, 142, 241, 272, 346, 350, 434 of diversity 279 of land 265 of language 2–3, 167, 205–222, 224, 240–241, 244, 280, 291, 299, 301, 376, 405, 435, 438–440 of speakers 3, 401 of structural features 6, 12–13, 19, 38, 49n20, 49–51, 54, 61, 62–64, 67, 77–78, 81, 85, 93–94, 98, 113, 115–116, 129, 133, 137, 142–143, 146, 190, 219, 226, 272, 434 of stylistic options/range 67, 327 Louisiana 167, 169 Lowlands (Scotland) 139, 213, 340 loyalty (see also language loyalty) 156, 169, 215 Luther, Martin 374 Luiseño 143–144, 144n7, 168 LWC see currency Machair-Chat (Sutherland, Scotland) 196 Maclean, Sorley 366–367 Madagasgar 437 Maidu 16, 417n7 Maine 167 maintenance see language Malagasy 437 Malinowski, Bronislaw 397

general index Mandara Mountains 292–293 Mandarin Chinese 210 Maori 237 Marin Miwok 437 markedness 6 marriage (see also intermarriage) 94, 148, 293, 318, 426 exogamous 180, 209, 293, 295, 297–298 Massachusetts 236 Master-Apprentice Program 240, 244 media 20, 146, 148, 257, 284, 289 broadcast media 239 mass media 299, 302 medium of instruction 239, 250, 259, 298 membership 15, 91, 158–160, 162, 183, 193, 199, 218, 280, 413 (in-)group 199, 231 Mennonites 244 Menomini 8, 66, 138, 183–185 merger 93, 105, 110–112, 190, 205, 353 Mexico 12, 16, 67, 175, 207, 254, 372, 379, 430, 434 Mexicano see Nahuatl Miami 16 Michigan 412 University of 391, 401 Micronesia 76, 286 migration 146, 228, 273, 284, 371, 420 in-migration 146, 273, 420 out-migrants 371 military service 312 minister 197–198, 278, 313, 412 minority (see also community, dialect, language, speaker) 1, 22, 240, 242 minority group 217, 284, 290, 301–302 minority peoples 284–285, 289–290, 297–300, 303 minority population 235, 284, 287, 302 missionary 3, 22, 436 mistake 7, 9, 25, 34–35, 98, 102, 104, 119, 147, 157, 160n3, 214–215, 327, 338n6, 407, 412 Mixteco 392 mobility 273, 280 mobilization 236, 248, 437 model 43, 46n18, 58, 73, 272, 275, 289, 305, 313, 314n1, 348, 433 puristic 296 traditional 252, 298

463 uniform-state 18 for analogical change/innovation/ restructuring 38, 82n5, 348 modernity 280–281 modernization 274, 297 Moen 276–277 Mohawk 349, 409 monolingual(s) 9, 12, 69, 137, 138n1, 139, 156, 158–159, 171, 187n2, 196, 198–199, 217, 274, 311, 315, 329–345, 384, 397, 433 monolingualism 12, 19, 197–199, 285–286, 286–288, 289, 301 monostylism 311 Montana Salish 9 Moray 116, 156 morpheme 34, 61, 144n8, 182, 356–357 morphology (see also norm, morphophonology, morphosyntax) 34, 38, 49n20, 50, 61–62, 68, 74, 85, 87–91, 99, 107, 214–215, 319, 319n12, 412 derivational 87, 90 inflectional 87, 319 suffixal 40n10, 85–86, 319n12 morphological category 88 morphological complexity 66–92, 88, 93 morphological device 71–78, 83, 86 morphological elements, non-unique 270n5 morphological feature 85, 180 morphological formation 67, 81, 84, 98 morphological simplification 9, 88–89, 91, 93 morphological structure 88, 96–99, 382 morphological system 87, 253 morphophonology 50, 60, 96–97, 107, 112, 214–215 morphosyntax 180 mother tongue 34, 105, 168–170, 195, 208 motivation 15, 22, 211–213, 219, 382, 393, 430 motivating factors 277, 366 multilingualism 19–20, 205, 272, 279, 284–308, 440 “irrational” 291 receding 297–300 multilinguals 19, 285, 372 multilingual peoples/populations 19–20, 297

464 multilingual region/setting 20, 180, 193, 298 mutation 34, 36–38, 38n8, 39–43, 45n17, 49, 49n20, 51–53, 54n28, 61–62, 64, 101, 106, 108, 112, 132, 253, 319, 319n12, 320, 323, 352 consonant 40, 62, 100–101, 118n5, 323, 351–352, 355 final 99, 62, 71–81, 83–84, 86n9, 88, 90 initial 34–35, 36–40, 43, 49, 62, 112, 122, 132n9, 143, 319n12, 321 change-resistant initial mutations 36–40 mutational change 35–36, 40, 49, 51, 62 mutational grammar 34, 40, 60, 63 mutational system 35–36, 38, 38n8, 39–40, 48–49, 61, 63 mutational variability 49, 51–52 Mutsun 236 Nahuatl (see also Mexicano) 12, 67, 206, 254, 265, 271, 372, 379, 430 names 86, 156, 180, 189, 193, 225, 268, 415 given 132, 189n5, 266 personal 118, 128, 129n8, 341 by-names 189, 193, 338n7, 412n3 surname, family name 189n5, 225 narrative 23, 103, 160, 316, 316n6, 316n8, 317, 318–322, 322n7, 322n15, 323–27, 329–346, 353, 355, 359, 362–364, 367, 413, 420 narrative tone 367 nasalization (consonant mutation) 37, 37n4, 38, 38n7, 41n13, 43–44, 50, 52–53, 55–60, 106, 112, 132n9, 143, 321 Natchez 373 nation, nation state (see also government, setting) 13–14, 19, 22, 168–169, 175, 212, 236, 239, 261, 266, 269, 275, 277–79, 285–290, 301, 438 national government 235, 237–238, 302, 429 national identity 242, 266, 287 national language 210–211, 220, 227, 236, 247, 266, 268, 270, 279, 289–290, 297, 302 nationalism 210, 266, 271, 277–279, 280, 289, 301 nationality 211, 278 Native American (used adjectivally) 168, 230, 298, 406, 437 Native California Network 240, 439

general index native speaker see speaker Navajo 230–231, 296 negation 37–38, 103, 108n7 neighbor 20, 251, 187, 200, 226, 229, 235, 241, 244, 248, 259–260, 265, 275, 286, 292, 302, 317, 401, 411, 413, 429, 430 neighboring (ethnic) group/people 20, 200, 235, 285, 292 neighboring language 248, 260, 265, 292 neighborhood 151, 226, 229, 244, 259, 411, 429–430 the Netherlands 211, 289 networks (see also friendship, kin(ship), language, marriage) 11, 63, 160, 214, 239, 285, 385–386, 395, 432 interaction 107, 159, 161, 171 social 148, 239, 307, 395 network processes 244, 280 New York City 63, 158, 163 New Zealand 237, 412n2 newspaper 16, 148, 289, 420 Ngas 291, 301–303 Nicaragua 17, 410, 421 Niger-Kordofanian 291 Nigeria 291–292, 294 Plateau State 291 nobility 266, 334 nominative 50, 50n21, 50n23, 51–56, 56n30, 57–62 norm, norming (see also grammar, rhetoric) 12, 18, 24, 34, 37, 53, 57–58, 63–64, 71, 73–74, 77–78, 81, 87–88, 94, 105, 105n5, 138, 169, 174, 198–199, 213–214, 252, 255, 262, 265, 272, 286, 292, 300, 329, 345, 373–374, 402, 404 church-Gaelic 404 conservative 53, 58, 68, 71, 73–74, 78, 81, 138, 188, 213–214, 259, 373–374, 402 dominant-language 269n2 external, extra-community, nonlocal, outside 18, 24, 224–225 fluent-speaker, older fluent speaker (o.f.s.) 77, 83, 87–88, 157–158, 258n4 grammatical 64, 157–158, 163, 405 linguistic 188, 199 local 199, 224–225, 315, 374, 405, 433 narrative 345 phonological 64, 158, 163, 405 politeness 329–330

general index quotational 345 socioeconomic 438 sociolinguistic 10, 158, 160, 162, 218–219, 258, 377, 382 spoken 223, 255 standard-language 404 traditional 248, 435 Normandy, Normans, Norman period/ Conquest 265, 238, 259 Norman French 134 Norn 379 North America (see also Indian) 208, 264, 439 North Queensland 107n6 North Sea 224 Northern Ireland 259, 278, 415 Norway, Norwegian 301, 379 noun (see also plural, verb) 7, 38, 40–42, 45, 48–49, 49n20, 50–51, 51n24, 51n25, 53–54, 54n28, 55, 56n30, 57, 60–62, 67, 67n1, 68–75, 77–78, 82–86, 86n7, 87–91, 97, 97n1, 101, 118n4, 140–142, 157, 172, 214, 249, 255, 269n2, 294, 320, 320n13, 321, 325, 353, 355, 363, 367, 398 attributive 49, 60, 62 feminine 51, 51n25 inanimate 355 masculine 7, 49–51, 53, 60–61, 321 noun class, noun classification, noun class system 51n25, 249, 294, 434 noun plural 67–68, 69–78, 82–87, 88–91, 140, 142, 172, 214 nominal gender 7, 11 nominal structure 324n16 nominal system 38–39 Nova Scotia 147, 167, 304 Nsur 291, 301–303 number (grammatical) 43, 48–49, 62, 70, 106 number, cardinal and ordinal 96–97 number of speakers 5, 24, 68, 91, 93–94, 113, 129, 217, 238–239, 242, 252, 276, 291 numbering of speakers/sources 43n16, 118 Oaxaca State 207–208 Oberwart 19, 208, 379, 380 obsolescence 1, 4–6, 8–12, 12–17, 91, 131, 133, 191, 257, 386, 402 obsolescence processes 400 Occitanian 267, 288

465 occupation 116, 164, 194, 241, 411, 413, 438 occupational distinctiveness/ separateness 1, 33 Ofo 64 Ojibwa 184, 185 Oklahoma 378, 381 opportunity (for language use) 9, 96, 139, 190, 213, 216, 265, 379 economic 175, 299 orthography 22, 113, 226–227, 256, 270n5, 317n9 Osage 272n7 Otomanguean 16 Ottoman Empire 266, 278 Turkish, Turks 206–207, 300 outsider 13, 16–17, 168, 179, 216, 224, 228–229, 258, 268n1, 281, 304, 313, 380, 385–387, 413–416, 417n8, 419, 428, 436, 439 Papua New Guinea 284, 286–287, 293, 417n8, 436, 439 paradigm 18, 40–41, 59 paradigmatic conservatism 179, 416 parallel (structural, social, attitudinal) 6, 67, 86–87, 93, 100, 115–117, 123, 127, 288, 294, 347, 363 parent 4, 5, 10, 34, 94, 112n8, 150–154, 170, 174, 178, 232, 234, 240, 244, 252, 259, 264, 287, 302–304, 383, 418–419, 430–431 parental generation 150, 377 participant-observation 95, 159, 380 particle 23, 106, 108, 108n7, 110, 132, 132n9, 346, 353, 355, 365 adverbial 98 complement 346 discourse 23 emphasizing 353, 366–367 enclitic 353 negative 123–124, 132 negative imperative 123 negative interrogative 346 preterite 124, 126, 129, 131–132, 132n9, 346, 355 relative 106, 111–112, 346, 362 vocative 49n20, 118 passive 39, 40–49, 62, 90, 100, 103–105, 140–142, 219, 324, 324n16, 325, 400 bith-passive 42–43

466 dol-passive 43–48 passive structure 46n18, 48, 219 past tense 103–104, 159, 321, 352 independent 118, 118n5 negated, negative 123–124, 126, 129, 131–132 peer 51n25, 56n32, 187, 258n4, 349, 418n12 peer group 87, 170, 215 Pennsylvania 10, 115, 148, 161n5, 170, 230, 437 Berks County see Pennsylvania German McKeesport 115 University of Pennsylvania 437 Pennsylvania German, Dutch 10, 148, 148n1, 149, 153–154, 161n5, 167, 169–170, 172, 230, 258, 260, 287, 377, 382–383, 385–386, 410n1 performance 15–16, 20, 88–89, 93, 139, 141, 143–145, 147, 171, 185, 214, 216, 272, 313–314, 332, 378 periphery (see also dialect) 117, 206, 261, 393, 411 periphrasis 70 Persian 206–207, 247 persistence 113, 132, 261, 277–278, 303 personality 64, 96, 149, 152–54, 217, 367, 382–383 Peru 271 Peruvian Andes 16 Amazonian Peru 235 phone see telephone phonology (see also change, morphophonology) 2, 51, 62, 64, 67, 107, 146, 149, 162n8, 180, 186, 213, 236, 249, 258, 372, 378, 395, 398 phonological class 7, 51, 321 phonological inventory 13, 395 phonological structure 305 phonological system 2, 393, 403 phonological variation 158 phrase (see verb) 42, 50, 54n28, 61, 70, 85, 97–98, 101, 102n4, 106, 218, 314, 319–321, 325n17, 340, 348, 355, 363, 398, 434 adverbial 98 conjunctival 101–102, 106 fixed 107, 377 fossilized 51 noun 70, 85, 355, 363, 398 prepositional 42, 50–51, 54n28, 60–62 prepositional relative 106

general index verb 38, 70, 106, 118n3, 132 phrasal switch 319–320, 323–325 Pictish 156, 242 pidgin, pidgin language 66, 89n9, 91, 250, 250n2, 296 Pidgin French (see French) pidginization 66–67, 89–91 Pike, Kenneth 392 place-names 179, 181, 396, 398 Plateau State see Nigeria pluperfect (verb form) 103–104, 106 plural (see also imperative, inflection, noun, vocative) 9, 40n11, 41–43, 45n17, 50n22, 62, 97, 97n1, 141–142, 219–220, 270n5, 319–320, 320n13, 324, 346, 356, 394, 400 sibilant 90, 220, 320 zero 78, 86–86, 89–90 plural allomorph 82, 319 plural noun 67–68, 71, 73–74, 78, 82–91, 140–142, 214 plural suffix, suffixal plural 82, 85–86, 346 policy, policy-making 10, 148n1, 175, 209, 231–232, 239, 242–243, 247, 251, 267, 280, 284, 288, 289n2, 430, 434 economic 243 language 175, 239, 243, 246, 280 national 284 social 243, 287 politeness (see also norm) 329–330 polysynthetic language 249 Pomattertitsch 400n1 pool (of speakers, sources) 4, 7, 137–138, 146, 157, 399 Popoluca 268 population (see also ancestor, enclave, ethnicity, speaker, subordination, urbanization) 2, 5, 9, 12–13, 17, 19, 24, 33–34, 36, 63, 68, 91, 94–95, 116–117, 117n2, 138, 149, 158, 167, 169, 174, 176, 185, 194–195, 195n2, 197–200, 206, 208, 210, 219, 223–224, 228, 238–239, 242, 244, 254, 260, 265, 268, 272–273, 275–276, 279, 284–285, 288, 291, 293, 296–297, 299, 301–302, 304, 331, 331n2, 350–351, 371–372, 375, 377–378, 383, 387, 394–395, 402–403, 404, 407, 411, 418, 429, 432–433 bilingual 36, 329, 377, 433 fisherfolk 9, 116, 407

general index majority 19, 377 minority 19–20, 235, 284, 287, 302 monolingual 12, 199 multilingual 19–20 rural 403 school 259 small language 244, 279 population base 223, 275, 350, 429 population mixture 404 population size 63, 228, 285, 287, 404 positive reinforcement 105 possession (grammatical; see also preposition, relative) 182, 217, 269, 275, 285, 354 possessive adjective 217 possessive pronoun 40–41, 42–43, 48–49, 62, 64, 324n16, 353, 355 possessive pronoun system 64 posterity 22, 440 Potawatomi 184 poverty 116, 238, 327, 395, 407 power 23, 146, 156, 168, 206–207, 264–265, 267–268, 272, 276–278, 280–281, 284, 286, 314n2, 348, 362–367, 371 pragmatism 17, 207–208, 376 Prague 374 prayer, praying 17, 95, 216, 312–313, 429 precenting, precentor 312 prefix, prefixation 182, 249, 348–350, 434 prejudice 196, 225, 418, 420–421 preposition (see also phrase) 41, 44–45, 50, 53–54, 54n8, 55, 56n30, 56n31, 57– 58, 61, 101, 106, 110, 140, 217, 354 complex 188–189 compound 56 conjugated/conjugating prepositions, prepositional pronouns 140, 353–54, 363 double prepositional structure 101 preservation 53, 61–62, 86, 91, 124, 133, 187, 260 pressure 6, 115, 146, 169, 174, 226, 248–249, 274, 276–277, 301, 385, 438 assimilative 169, 176, 304 economic 279 social 169, 276 system 115, 133 prestige 93, 146–147, 168, 205–206, 210, 241, 253, 264, 267–269, 275, 286–288, 296, 350, 378

467 high 264, 378 higher 154, 269 low 147, 149, 154–155, 264–265, 378 negative 87, 148–149 prestige transfer 268 preterite (see also particle) 324 primers 22, 250, 417 privacy 406, 419, 426, 427–429 private sphere 311–328, 429, 431 private sphere language 413, 431 private-sphere use 311–328, 430, 433 productivity 73 proficiency (see also age) 3, 5, 7, 9–11, 15, 21, 68, 68n2, 69, 84, 88–90, 93–94, 116, 118–119, 122–124, 127, 132, 138–139, 141–142, 157–161, 161n5, 162, 164–165, 178–179, 183–184, 186, 186n1, 188–191, 213n4, 216–219, 239, 243, 329, 346, 373–374, 402 proficiency continuum 9, 88, 94, 116, 118–119, 122–123, 127, 132, 139, 374 proficiency/age-and-proficiency group, groupings 119, 186–188, 189–190 proficiency level 11, 118, 188–190, 346 proficiency scale 157, 184 program (see also Master-Apprentice, revitalization) 231, 240, 244–245, 278, 282, 430, 432, 434–435 bilingual (education) 230, 250–251, 436 immersion 304 language-maintenance 223 school(-based) 251, 432 teacher-training 251, 430 trilingual 273 pronoun 7, 67, 188–189, 269, 353, 359, 363 indefinite 363 personal 353–354, 364 possessive 40–41, 43, 48, 62, 70, 324n16, 353, 355 possessive pronoun system 64 prepositional 353–354, 363 subject 48, 48n19, 49, 62, 217 pronoun object 140, 214, 217 pronoun reference 400 pronoun replacement 7, 70, 70n3, 188 pronominal forms 354, 362, 366 pronominal system 39 pronunciation 50n22, 56n32, 181–182, 184, 226, 241, 256, 382, 431

468 prospects 3, 20, 238, 242–244, 264, 425, 442 small-language 264–283 prosperity 231, 272–274 proverb 54n28 proverbial lore 95 psychologists 114 public sphere 313–314, 327, 429–432 Pueblo 17, 275, 296 Pueblo de Cochiti 434 Pulaar 294–295 Pulap, Pulap Islanders, Pulapese 276–277 purism (see also ideology) 247–263 puristic attitudes 247–249, 252, 258 purity (linguistic) 196, 198, 258–259, 261, 277 Quapaw 220–221 quasi-speaker 14 Quebec 237, 261 Quechua 206, 265, 271, 278n12 Quichua 431, 434 questionnaire 1–2, 14, 216, 302, 393 quotation (see also verb) 129, 218 direct 329–346, 364, 414 race 24 radio 15, 36n2, 105–168, 212n3, 232, 315n4, 326, 365, 375 Radio Eireann 375 Rama 17, 410, 421 rarity 61–62, 93, 105–107, 372, 382 Reay Country 196 reclamation 16, 236, 240 reconstruction (historical) 35, 138, 256 recording (see also taping, video) 21, 315–16, 316n7, 355–359, 380–381, 383, 413, 416, 419, 425–426 recording equipment/machinery 413, 425 recording session 316, 320, 326, 328, 330n1, 344, 415, 429 reduction 13, 63, 66–67, 74, 93–94, 113, 115, 137, 142–144, 181–182, 238 reduced form 63, 137–138, 144 reduced proficiency 178 reduced use 13, 138, 144, 159 redundancy 42, 49, 62, 86 reduplication, partial 348 register (see also religion) 14, 185, 305, 315n5, 347, 429

general index regularization 132n9 analogical 219 regularity 132 relative (grammar) 106, 171, 173 relative future suffix 355 relative particle 106, 111–112, 346, 362 prepositional (possessive) relative  100–101, 106, 166 relative (kinsperson) 158–159, 161, 218, 241, 324, 340, 378, 380, 383, 386, 395, 400, 418–419 reliability 372–373, 379 religion (see also register) 88, 168, 194, 209, 291, 302, 311 religious authority 275, 327 religious ceremony/ceremonials/ritual 168, 194, 208, 276, 429 religious language 312–13, 327 religious life 22, 312, 327, 403 religious practice 264, 266, 296 religious register 327 religious usage 312, 325 relocation 298 remoteness 175, 291 repertory, repertoire 19, 36, 96, 108, 110, 210, 338n6 repetition 111, 189n5, 324–325, 325n17, 332, 416 replacement (see also pronoun) 37n3, 37n4, 31, 56, 137, 256, 379 reported speech 7, 129, 187n2 research community-directed 14 field 391, 395 flawed 302 linguistic 179, 421 minority language 4 research agenda/goals/purposes/ priorities 23, 58n33, 212, 405 researcher 1, 4, 8, 13–14, 17, 19, 23–24, 105, 169, 240, 291, 295, 301, 311, 345, 374, 384, 391, 401–402, 405–407, 409–411, 414, 416–417, 425–429, 435–437 endangered language 23, 410 linguistic 179 outside 14, 437 Western 20 research priorities 23 resettlement 407

general index residence 33n1, 124, 301, 316, 371, 380, 386, 431n1 residence patterns 4 residential areas 5, 195, 411, 413 resistance 5, 37, 58, 60, 91, 97, 147, 154, 174, 219, 235–237, 249, 256, 304, 307, 401, 438 resistant groups 277 resource 21, 107–108, 235, 257, 268, 273, 286–287, 292, 297, 303–304, 350, 363–364, 412, 417, 427 responsibility 20, 228, 347, 398, 409–424 restoration 236, 238–239, 242–243 restriction 252, 311, 327, 427 functional 29 retention 11, 90, 93, 106, 123, 129, 131–133, 140, 147, 188–189, 194, 208, 211–212, 214–218, 229, 244, 280 retreat (language) 12, 16, 223, 429 revalorization 237, 242 revitalization (see also culture) 14, 16–17, 20, 191, 234–246, 247–263, 273, 304–305, 425–444 revitalization program 20 revival 14, 16, 223, 227, 230–231, 247–263, 436 rhetoric 13 rhetorical norms 259 rhymes 95, 189, 339n6 ridicule 230 Rindre 291 rite, ritual 16, 378, 429 ritual use 16 ritualistic words 184 Romance 207, 265, 392 Romani 383 Rumsien Ohlone 236 rural area/district 297, 396 Russia, Russian 206, 210, 265, 293, 432 Sahara 286, 292 Salesians 298, 298n2 salience 144, 364 salient feature 77 Salina Cruz 174 Sami 304 sample 2, 6–8, 35, 41n14, 69, 71, 73, 81, 96, 108, 119, 132n9, 133, 142–143, 161, 186–187, 189, 213n4, 214, 215n6, 217, 219, 371–372, 379, 381, 395, 399–400, 440 sampling procedure 409

469 Sanskrit 247, 264 Sauris 207 Scandinavia 208, 237, 265, 398 scholar 18–19, 22, 174, 196, 231, 239, 302, 373, 379, 384, 392, 405–407, 410–412, 414–417, 419 scholarship 224, 231 school, schooling (see also bilingualism, instruction) 20, 44, 117, 148, 148n2, 168, 170, 174–75, 197, 206–207, 209, 213, 218, 224, 227–228, 230–231, 239, 242–244, 247, 250–252, 259, 269n2, 277n10, 280, 287, 297, 304, 350, 373, 403, 429–431, 433–434 boarding 298–99, 301 English-language 148, 149n2, 277n10 primary 17, 172, 174, 259 secondary 243, 277 immersion schooling 304–305, 435 school subject 230 school(-based) (language) program 146, 251, 273, 305, 432, 435 schoolchild(ren) 17, 225, 231–232, 239, 432 schoolmate 96, 311 School of Scottish Studies 33, 396, 398, 412–13 schoolteacher see teacher Schwyzertütsch 168 Scotland (see also Linguistic Survey) 68, 123n6, 137, 138, 147, 152, 168, 174–175, 197, 213, 226, 231–232, 241, 258n4, 304, 329, 335, 350, 391, 393, 398, 412, 419, 433, 435 Highland Scotland 1, 33, 116, 186, 200, 232, 350, 406 Linguistic Survey of Scotland, Survey of the Gaelic Dialects of Scotland 372, 386, 392, 396, 409 Scots 116, 242, 255, 336n5 Scottish Gaelic see Gaelic Scottish Highlands see Highlands, Highland Scotland Scottish Parliament 242 scripture 312–313 Seereer 294–295 segregation 94, 148, 169 self-confidence see confidence self-evaluation see evaluation semantics 64, 85, 382 semantic collapse 104

470 semantic complexity 93 semantic distinctions 349 semantic extension 255 semantic field 18 semantic range 41n12, 409 semi-speakers 9–11, 13–14, 38n8, 68–69, 87, 94, 118n5, 119, 129, 133, 137–145, 146–155, 157–160, 160n3, 161, 161n4, 161n5, 162, 162n6, 164–165, 186, 186n1, 187, 187n2, 188–190, 195, 213n4, 260, 329, 374 low-proficiency 10, 15, 157–161, 161n5, 162, 164–165, 186n1 high-proficiency 158–159 short-burst 157–158 semi-speaker status 142 strengths of 107–108 Semitic 349 separateness (see also occupation) 260 social 1, 199 setting (see also culture, intimacy) fieldwork 391, 425 home 277, 430 multilingual 193, 298, 307 nation-state 286–289 public 430 school 230, 430 small-language 425 small-village 405 solidary 431 sex (as a variable) 399, 403–404 shame 195, 230, 236, 339 Shetlands, Shetland Islanders 379, 397 shift (see age, code, style, utility) 10, 22, 82, 101, 109, 115–134, 156, 207–208, 234–237, 239, 272, 276, 287, 290, 300–302, 304, 349, 350–352, 380 language shift 7–8, 14, 115–133, 144, 146–155, 174–175, 178, 187n2, 234–237, 240, 276–277, 280, 287, 297, 303, 430, 433, 436–437 reversing language shift (RLS) 235, 430, 433 Shiwilu 235 Shoshoni 67, 376 Siane 293, 295 sibling 4–5, 96, 150–151, 159, 170–173, 187, 215, 386–387, 396, 403 older 5, 149n2, 150, 170, 178

general index younger 119, 171 sibling set 4, 182 Sicily 265 simplification (see also morphology) 9, 66–67, 88–91, 434 simplicity 67, 88, 106, 132 singular 41–42, 50n22, 51n24, 69, 70n3, 72, 78, 97, 97n1, 140–141, 189, 214, 321n13, 324n16, 325, 353, 355 Siouan 64, 220 skills (language-related; see also speaker skills) 3, 11, 13, 21, 95, 105, 107, 107n6, 108, 111, 135–192, 211, 219, 228, 232, 275, 348, 374, 378–379, 382, 385, 387, 430 active 107–108, 161, 217 foreign-language 211 morphological 98 passive 331 productive 10, 98, 154, 159–160, 162, 165, 219, 329 receptive 10, 107, 160–162 multilingual 296 second-language 433 verbal 95, 190, 316 Skye, Isle of Skye 224, 232, 357 Slovenia 236, 285 society (see also change, homogeneity, norm) 148, 227, 279, 296, 303, 326, 418, 420, 438 social context 1, 244, 396, 399 social correlates 400 social distance 269n2 social group 24, 160, 163–164, 280, 345, 421 social interaction 171, 316 social intercourse 94, 311 social liability 395 social mobility 273 social organization 24, 266 social penalties 150, 287 social pressure 169, 276 social standing 199, 265, 271, 395 social strata 266, 270, 351 social stratification 18 social structure 10, 209, 278 social unnaturalness 416–418 socialization 91, 149–150, 298 socioeconomic condition, status 153, 208, 299, 403–404, 429

general index socioeconomic hierarchy/structure 175, 194, 209, 290, 299, 438 sociolinguistics 23, 289 correlational 4, 8, 401 solidarity (see also setting) 52, 215, 345, 413, 430, 433 Solomon Islands 235, 277, 434, 436 Somalia 169 South America see America, Indian South Asia 440 South Pacific 176, 235 South Tyrol 273 sources (linguistic) 4–5, 7–8, 11, 16, 18, 21–25, 101, 116n1, 118–119, 129, 137, 295, 315n3, 327, 336n5, 372–373, 381, 384, 387–388, 393–401, 403, 405–406, 410, 413–414, 418, 418n12, 419, 421, 426–428, 437 Soviet Union 207, 210, 235 Spanish 67, 169, 174–175, 208, 210, 254, 265, 270, 273n8, 296, 300, 348–349, 379–380, 432 Castilian 273 Los Angeles 115 speaker (see also competence) age-atypical 178 child 9–10, 107, 213, 257 deficient 182–183 final 400n1, 420 formerly fluent (FFS) 7, 9, 15, 119, 123–124, 129, 130–131 fully fluent 1, 4, 9–12, 94–96, 99, 101, 105–106, 118n5, 119, 129, 157, 159–160, 170–173, 186–187, 189–190, 213, 218, 239, 242, 326, 329–330, 374, 412 habitual 12, 434 imperfect 7–12, 14–15, 25, 107, 107n6, 139n4, 142, 147, 149, 154, 170–172, 186, 190, 213, 213n4, 213n5, 214, 217–220, 345, 373–374, 377, 399–400 local 2–3, 12–13, 117, 225, 241, 313–315, 317, 345, 350, 404, 411, 413–414, 426, 431, 439 marginal 165 minority-language 11, 226, 304, 430 native 9, 20, 95, 105, 107, 114, 179, 184, 205, 207, 227, 230, 236, 240, 242, 248, 253–255, 259, 293–295, 305, 348, 350–352, 366, 372, 377–381, 406, 432–433, 435

471 non-native 229 older fluent (o.f.s., OFS) 6, 68, 68n2, 69, 71, 73–84, 86–89, 94, 98, 100n3, 102, 104, 119, 121–126, 130–131, 187–190, 213, 258, 374, 381 second-language 432, 435 short-burst 190 terminal 9, 64, 143–144, 376, 379, 387 traditional 181, 400 younger fluent (y.f.s., YFS) 6, 68, 68n2, 69, 74–84, 86–88, 94, 100n3, 104, 119–132, 187–190, 215, 217, 374 semi-speaker (SS) 9–13, 15, 38n8, 68–69, 87, 94, 106, 107–108, 109–113, 114n9, 118n5, 119–132,133, 137–145, 146–155, 157–160, 160n3, 161, 161n4, 161n5, 162, 162n6, 164–165, 186, 186n1, 187, 187n2, 188–190, 195, 213n4, 329, 374 speaker base 238, 248, 261 speaker population 13, 167, 224–225, 236–37, 248, 311, 378, 433, 438 speaker skills 178–192 speaker status 15, 142, 183, 190, 378–379, 432–433 speaker typology 190 speech (see also community) conversational 334, 373 good 17, 181, 183 inner 216, 379 non-standard (see also English) 1, 266, 328 quoted/reported 7, 129, 187n2, 330, 335, 341, 345 spontaneous 96, 99, 383 speech behavior 96, 287 speech form 1, 3, 5, 12–13, 15, 18, 20–21, 23, 68, 93, 113, 133, 154n5, 168, 223, 237, 258, 260, 265–269, 271, 371–377, 379, 385, 395, 400, 409–411, 416, 438–439 speech habit 7, 403 speech island 2, 156, 224 speech performance 185 speech style 133, 327, 373 speech variety 129, 132, 158, 162, 269, 275, 277, 354, 371, 380, 411, 418, 431 spouse 52, 88, 96, 229, 231, 396, 428 Sprachbund 133

472 standing (see also social standing) 168, 237–238, 242, 264, 272, 274, 284, 386, 421, 436 Statistical Account (of Scotland, of Sutherlandshire) 197, 199, 207 status (see also speaker) 15, 61, 119, 149, 153n4, 170, 174, 195, 223, 229, 237, 240–241, 249n1, 267, 269, 271, 278, 284, 290, 302, 332, 337, 376, 380, 385, 416 official (language) 271, 284, 292 social 269, 287, 327 socioeconomic 153n4 status distinctions 164 status factors 302 status sets 278 stereotypes, stereotyping 12, 144, 149, 164, 232, 364 stereotyped feature 364 stigma, stigmatization 116, 149, 236, 241, 403, 438 stigmatized groups 5, 164, 215, 218, 230, 240, 345 stimulus sentences (see also test sentence)  69, 71, 104, 111–112 story (see also tradition) 6, 23, 50, 58n33, 180, 183, 190, 230, 236, 251, 289, 314n2, 316, 318–326, 329, 332–333, 336n5, 339, 343, 350, 359, 362–364, 382, 399, 420, 427–428 story collection 22 storybook 417, 421, 426, 436 storyteller, storytelling 181–183, 185, 190, 329–230, 330n1, 367, 379, 381 stratification 18, 268 structure (see also analysis, deviation, grammar, integrity, interference, lexicon, morphology, noun, passive, phonology, society, surface structure, syntax) conservative 6, 19, 96, 101, 132, 188–189 embedded 171, 173 high-frequency 67 language, linguistic 1, 23, 163, 183, 271 matched 116, 123–124, 129–131 non-traditional 127 socioeconomic 194, 209 synthetic 143, 219 traditional 123–124, 127, 129, 131 troublesome 96–105 unmatched 116, 117–123, 124, 129–132 verb(al) 40, 249 verb complement 100–101, 106

general index style (see also factors, speech) 23, 93, 247, 251–252, 314–316, 324, 326–327, 329, 332, 347, 367, 380 careful 321 casual 311, 323–324, 326 formal 168, 251, 325 narrative 329 personal 96 style-shifting 382 stylistic consistency 275 stylistic option 67, 140, 143 stylistic range 314, 316, 326–27 stylistic variation 24, 311–328, 380 subjunctive 347–350, 355 subordination (language structure; see also clause) 217 subordination (social) 260, 284, 288, 406 subordinate group/people/population 169, 265–266, 268–269, 285, 288, 290, 300–301 subordinate language 93, 266, 268–269, 278n12 subsistence, sub-subsistence (economic mode) 195, 272, 297, 304, 351, 396, 438–439 substitution (of one linguistic element for another) 39, 62, 71, 97, 106, 109–110, 112, 140, 143, 399 success (see also economics) 16, 23, 107, 111, 113, 162, 174, 210, 217–218, 223, 229, 231, 237–238, 244, 247, 255–256, 272, 274–275, 303–305, 373, 383, 385, 425, 435, 439 suffix (see also diminutive, morphology, plural) 271n5, 320n13, 349–351, 400, 408 conditional 355 deictic 351, 353 emphatic 269n2, 352–367 emphatic-deictic 353–354, 359, 363–365, 367 idiosyncratic 78 subjunctive 355 3rd-person 348 vocative 118n4 suffixation 71–73, 75, 77–86, 86n7, 88–90, 98–99, 102, 246, 404, 418 simple 75, 80, 82–83, 86, 88–90 Sumer, Sumerian 271 suppletion 41, 72, 74–75, 78–79, 86, 86n7, 87–88, 324 suppletive past-tense forms 124, 568

general index support 20, 22, 131, 148, 162, 169, 235–239, 241–252, 255, 273, 279–281, 304, 312, 392, 417, 417n7, 417n8, 434, 436, 440 community 230, 432 financial 224, 238, 410n1 government(al), official 93, 148, 205, 231, 238, 429, 433, 438 institutional 148, 287, 302 language 235–239, 240–244, 417, 432 suppression 232, 236, 238, 273, 286 suprasegmentals 62, 325, 352, 364–365 suprasegmental features 188, 132 surface structure 61, 118, 124, 132, 356 survey 4, 229, 267, 293, 398, 411 single-interview 401 Survey see Linguistic Survey of Scotland survival (see also indigenous) 2, 11, 14, 21, 54, 64, 67–68, 169, 176, 223, 241, 243, 254, 257–258, 270, 273, 279, 281, 301, 303–304, 376, 435, 437, 440 Sutherland, East Sutherland passim Sweden, Swedish 210, 255, 301 switch see code, phrase Switzerland 168–169 syncretism 67, 110–12, 254 syntax (see also morphosyntax) 67, 99, 107, 184–85, 214, 249, 324, 412 syntactic construction 85, 93 syntactic device 36 syntactic environment 34–35, 37, 50, 60, 69–71, 105 syntactic structure 12, 99, 382 syntactic variation 163 synthetic form (see also grammar, polysynthetic language, structure) 140, 143, 188–190, 214, 219, 259 system see case, grammar, kinship, morphology, morphophonology, mutation, noun, phonology, possessive pronoun, pressure, pronoun, verb, writing Taiap 287 taping, tape recording 3, 23, 59, 116n1, 164, 314n2, 319, 321n14, 322n15, 330, 359, 415, 417, 419 Tapshin 291, 301–302 Tariana 15, 17, 179–180, 181–183, 187, 189, 191, 298, 298n2, 300, 400n1 younger Tariana speakers 15

473 Tasmania 137 teacher, school teacher 182, 225, 227, 231–232, 244, 247–248, 251, 294, 304–305, 321, 323, 340, 357, 415, 430–431, 435 foreign-language 211 indigenous language 244, 430 second-language 14 teaching materials 256, 420, 439 teacher-training program, teacher trainees 251, 257, 430, 437 teaching 15, 225, 230–231, 251–252, 257, 430 second language 14 teaching strategies 434 telephone (see also conversation) 3, 15, 116n1, 187, 187n3, 255, 330, 340, 343, 404, 412, 417, 419–420 telephone fieldwork 23, 420 television 168, 239, 242, 384 tense 37–38, 100, 103–104, 108, 110, 126, 159, 171, 173, 214, 217, 259, 321, 348, 352 independent past tense 118n5 negated/negative past tense 123–124, 126, 129, 131–132 past tense 103–104, 118n5, 123–124, 126, 129, 131–132, 159, 321, 352 present tense 348 tense formation 215 test, testing 11, 15, 52–53, 55, 55n29, 56, 56n31, 57, 59, 68n2, 69, 82n5, 90, 103–104, 109–111, 115, 138–140, 142, 161–162, 164, 171–173, 212–217, 219, 253, 255, 302, 326, 380, 384 comprehension 107n6 matched guise 431 read-back 11, 59 translation 69, 102, 108, 158, 187, 187n2, 219, 399–400 test phrase, test sentence (see also stimulus sentence) 7, 50, 52, 56, 68–69, 71, 74, 78, 82, 82n5, 83, 86, 88, 90, 102, 104, 109, 171, 385 Texas 277n10 text 16, 41n14, 66, 227, 231, 249–251, 256–257, 289, 316, 317n9, 318, 319n12, 364, 366, 382–383, 399, 427–428, 431n1, 434 textbook 68, 252 Thailand, Thai kingdom 278, 300 threat 2, 168, 254, 256, 265, 273, 278, 281, 290, 299

474 threatened language/speech form/speech community 17, 154, 168, 210, 220, 223, 225–226, 230, 240, 260–261, 270, 274, 281, 349, 375–377 tip (linguistic) 167–177 Tiwi 249–252 Tlaxcala 430 Tlingit 280, 287 Toba 22 tolerance 14, 226, 254, 266, 277, 300, 303 Tolowa 17–18, 409–410, 416, 432, 434–435 tone (see also discourse, narrative) 185, 269n2, 318, 336n5, 348, 362–364, 367, 431n1 topic 8, 13, 66, 95, 311, 314n2, 316, 327–328, 330, 413, 419, 425, 430 tourism 273, 303 trade 285, 287 trade language 296 tradesmen 33, 351 tradition (see also culture, language, lore, norm, society, speaker) 207, 66, 90, 184, 207, 209, 212, 224–225, 228, 231, 241, 252, 256–257, 260, 273, 276, 383, 405, 438 cultural 207, 209, 231 folk 23 written 224, 252 traditional activities, practices 168, 178, 181, 189n5, 276, 378, 409, 434 traditional grammar 43–44, 60, 98, 253–254, 352–353 traditional life, lifeways 230, 248, 297, 303 traditional material 162n6 traditional stories 6 transcription 21, 42n15, 317n9, 355–356, 382 transfer 51n25, 83, 115, 117, 268, 351 translation (see also test) 6–7, 46n18, 53, 68–69, 97n1, 99, 99n2, 100, 100n3, 101–102, 102n4, 103–104, 106–107, 139, 206–207, 139, 159, 171, 187, 213n5, 214, 225, 250, 330, 345, 355–356, 363, 366, 374, 382–384 translation task 6–7, 11, 23, 99, 102–104, 112, 129, 187, 187n2, 379, 385 translator 7, 187n2, 336n5, 367, 385 transmission 4, 13, 180, 229–230, 232, 244, 250–251, 287, 311, 398, 417, 425, 430, 436 home 242, 304–305, 435 non-transmission 376

general index parent-(to-)child 244, 430 transmission bridge 237 transmission failure 167–77, 402, 431, 435 transmission process 241 transport, transportation 88, 297, 394 Trobriand Islanders 397 Tsimihety 437 Tucano/Tukano 180–183, 297–298, 300 Ugong 410 uniformity 18, 67, 158, 254 social uniformity 391–408 United Kingdom 277 United States 207–212, 230, 237, 279, 290 Unst 379 Urat 436 urbanization (see also variation) 148, 276, 436, 438 urban boarding schools 299 urban centers 208 urban communities 239 urban elite 247 urban origin 431n1 urban population 293, 403 urban-rural distinction 294, 431n1 urban wage-laborers 438 usage (see also religion) 5–6, 8, 18, 36, 38n38, 41, 44, 46–47, 53, 56, 58, 58n33, 63–64, 68n2, 73–75, 78, 83, 85–87, 90, 123, 138, 156, 158, 163, 171, 181, 184, 188, 215n6, 247, 250, 254–255, 261, 312, 315, 323, 325, 344–345, 353, 379, 396, 399, 401–403 conservative 53, 58, 86, 88 contemporary, current 58, 353 variable 401–402 use (see also private-sphere, syncretism) active 20, 36, 98, 111, 157–58, 314n1, 377 increased 82, 105, 299 reduced 13, 138, 144 usefulness 207, 235, 254, 273 user 44, 87, 107n6, 162, 162n6, 226, 239–240, 248, 253, 270n3, 321n14, 433 active 240 nonuser 108 utility 190, 205, 208–11, 235, 251, 254, 287, 296 utility-based shift 234–37 Uto-Aztecan 430 Uzbek 235

general index validity 4, 20, 144, 189, 316, 347, 419 value 20–23, 36n2, 61, 66, 152, 179, 181–185, 191, 207–208, 218, 223–233, 234, 236–237, 241, 254, 260, 271n6, 279, 289n1, 296, 304, 377–378, 385–386, 395, 419, 433, 436, 440 entertainment 326, 375 value(s) (see also culture) 20, 148, 163, 183–183, 183–185, 210, 276, 285, 299, 391, 405 Vancouver Island 303 the Vaupés 179, 185, 193, 201, 293, 297–298, 300 variable (linguistic) 24, 294, 404 variability (see also mutation) 18, 24, 37, 39, 45, 49–52, 64, 98, 382, 391–408 variant 24, 37, 58, 60, 111, 115, 144n8, 163–164, 173, 401, 404 variation (see also age, dialect, phonology, register, style, syntax) 4, 24, 48, 64, 158, 163–164, 254, 354, 382, 399, 403–404 geographical 4, 24 idiosyncratic 409 individual 4, 17–18, 24, 180, 216 personal pattern/personally patterned 18, 23–24 regional 163–164 register 315n5 social 163 style, stylistic 24, 311–328, 380 urban 4 Venezuela 194, 209 verb (see also conditional, interaction, performance, phrase, pluperfect, skill) 38–39, 40n10, 41–42, 67, 70, 73, 82–83, 86, 89n9, 99n2, 100, 100n3, 101, 103, 106, 108, 111, 118n5, 123–124, 126, 129, 132, 157, 249, 250n2, 259, 271n5, 319n12, 321, 324–325, 347, 354, 401, 434 defective 363 finite 41–42, 44, 46n18, 140–141, 348 high(est) frequency 39, 111 irregular 41–42, 123–124, 140, 142, 259 quotative 185, 321, 323, 363 dependent verb form 101, 106, 355 verb complement 99–101, 106 verb root, verb stem 83n6, 123–124, 140–142, 172, 180

475 verbal enclitic 180 verbal noun 40, 42, 45, 50, 67n1 verbal system 37–39, 61, 271 Vermont 227 vernacular 223, 231, 248, 267, 270, 303, 376, 405 video taping 22, 24, 417, 426 Vietnam 89n9 Vikings 265 village (passim; see also fishing, home) village life 117, 405, 418 vitality (see also culture) 13, 93, 429 ethnolinguistic 302–303 vocabulary 8, 16, 63, 66–67, 96, 151, 184–85, 219, 247, 275, 313, 318, 378–79, 382, 392, 394 commonplace, high-frequency 107, 314 core 69 vocabulary retention 216 vocative (see also particle) 49n20, 62, 117–118, 118n4, 118n5, 122–123, 128–129, 129n8, 131–133, 139, 188–189, 341, 352 vocative plural 139 Voltaire 210 vowel 24, 37n3, 42, 42n15, 48, 56n30, 60, 75, 77–78, 181–182, 270n5, 317n9, 352–353 initial 98, 398 internal vowel change 60–61 vowel alternation 71–77, 79–84, 86, 86n7, 88, 90 vowel shortening 75, 78–79, 81 wage labor, wage-laborers 178, 351, 438 Wales 146, 237, 273–274, 303, 431n1 Wandala 292 Welsh 237, 273–274, 289, 431 Welsh-medium education 273–274, 431n1 Western Mono 17, 409 Westernization 276 Western language ideologies 264–83, 291 Winnebago 184 Wisconsin 183 Wittenberg 374 Wolof 294–295 workmates 311, 329 word order 38, 62, 140, 190, 320, 350

476 worth 268, 276, 376, 406, 436 writing (see also language) 36n2, 113, 117, 138, 183, 224, 241, 251, 261, 345, 401 writing system 14, 113, 225, 237, 271, 439 writings 19, 22, 113 written materials 271, 436 written record 143 Yahi 137 Yap 245, 286

general index Yiddish 438 Yidiɲ 373 Youchigant 375 Young People’s Dyirbal see Dyirbal Yugoslavia 236 Yup’ik 349 Zapotec, Isthmus Zapotec 174–175, 208 Zuni 406, 415