(9789004261938 - Small-Language Fates and Prospects) Small-Language Fates and Prospects

Small-Language Fates and Prospects
Nancy Dorian - 978-90-04-26193-8

Downloaded from Brill.com02/23/2019 02:45:48PM
via King's College London
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
Nancy Dorian - 978-90-04-26193-8

Small-Language Fates
and Prospects
Lessons of Persistence and Change
from Endangered Languages
Collected Essays
By
Nancy C. Dorian
LEIDEN | BOSTON
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

For the MacRae family of Achnahaird,
Coigach, and Golspie, East Sutherland.
A hundred thousand thanks!
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
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
Nancy Dorian - 978-90-04-26193-8

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
Nancy Dorian - 978-90-04-26193-8

Foreword
Nancy Dorian is a star. She is the founder of at least one burgeoning field within lin-
guistics. 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 lan-
guage 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 com-
pletely 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 work-
ing 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 preser-
vation 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 vol-
ume is it. Here we have twenty three papers, and an introduction, by Nancy, which
Nancy Dorian - 978-90-04-26193-8

x foreword
brings them all together, as the culmination of her life’s work. Each of these is a para-
mount achievement in the areas of endangered languages, language variation, lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
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 lan-
guage 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 circum-
stances in which I began in 1963 (and continued thereafter) to do linguistic field-
work 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 nota-
ble 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 Nancy Dorian - 978-90-04-26193-8

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 dis-
tinctiveness and its fragility made its description important to the Linguistic
Survey, whose fieldworkers had sampled the small population of former fisher-
folk 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 informa-
tion 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 devel-
opment of my own interests, of course, as I continued to work with this small,
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage endangerment. Linguists, sociologists, linguistic anthropologists, and
missionary linguists were by the second half of the twentieth century increas-
ingly 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 col-
lection 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 fish-
erfolk 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 domi-
nant 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 mem-
bers of the bilingual community.
In the 1990s another two focuses emerged. One was the documenting of dis-
tinctive 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
Nancy Dorian - 978-90-04-26193-8

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, fur-
thermore, 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 minority-
language 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 geo-

graphical 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 village-
related 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 dis-
tinctive 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
Nancy Dorian - 978-90-04-26193-8

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 consan-
guinial 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 grow-
ing 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 circum-
stances 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’ fail-
ure 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 par-
ents’ level of proficiency, or in some cases their older siblings’ level of profi-
ciency, 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 ques-
tions might be raised about age-related proficiency differences in the local
Gaelic. For example, how systematic were they? Could an age-based proficiency
Nancy Dorian - 978-90-04-26193-8

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 ele-
ments that had an English parallel? Did conservative features that decreased
in the usage of younger speakers fall out of use altogether, or might they dis-
appear 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 phono-
logical, 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 exception-
ally 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 mate-
rial, 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 compa-
rability was to present the same set of sentences for translation to an age-
differentiated 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
Nancy Dorian - 978-90-04-26193-8

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 transla-
tion battery and the 150 sentences of the 1976 translation battery were designed
to elicit structures that would set elements of ongoing change off against ele-
ments 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 highlight-
ing change. The other advantage was that the speakers I worked with were
highly adept translators. This came about in good part because local prac-
tice in reported speech was to use the language expected by the interlocu-
tor 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 part-
ner, 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 elicita-
tion unproblematic for a good many bilingual speakers.
There was a great advantage in being aware in advance of certain character-
istics of local speech when designing elicitation tasks. Knowing, for example,
that diminutives were exceptionally freely used in the local Gaelic made it pos-
sible 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 dif-
fered 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 low-
proficiency speakers that was particularly important.
Nancy Dorian - 978-90-04-26193-8

8 introduction
4 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 work-
ing with small and endangered languages, linguists sought out elderly repre-
sentatives 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 dili-
gently searched for in the contemporary form of local speech.
In the East Sutherland Gaelic case, very nearly all of the language data avail-
able represents apparent time. Fisherfolk Gaelic in these villages is unwritten,
and there are no surviving materials from earlier periods, apart from the lexi-
cal 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 spec-
trum 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 reced-
ing languages. Researchers looking at communities where marked language
shift was underway began deliberately to include the verbal output of imper-
fect speakers among their study materials. Wolfgang Dressler, a pioneer in lan-
guage 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 imper-
fect 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
Nancy Dorian - 978-90-04-26193-8

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 imper-
fect speakers and determine that while simplification appeared in imperfect-
speaker 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 con-
cluded 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 pop-
ulation, 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 flu-
ent 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 imper-
fect 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 speak-
ers 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 mis-
takes 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
Nancy Dorian - 978-90-04-26193-8

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 occa-
sional 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 expec-
tations 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 mainte-
nance, shift, and linguistic change in East Sutherland fisherfolk Gaelic would
be hard to overstate. Their Gaelic, when set directly beside that of fully flu-
ent 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 imper-
fect 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 fail-
ings 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).
Nancy Dorian - 978-90-04-26193-8

introduction 11
Once information about the language skills of imperfect minority-language

speakers began to be available, the possibility of comparing them with class-
room second-language learners arose: foreign-language instructors were inter-
ested 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 con-
trolled; 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 com-
munity 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 flu-
ent 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 fluent-
speaker sources at a later time, but with the gender-related morphophonologi-
cal 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 integra-
tion, 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.
Nancy Dorian - 978-90-04-26193-8

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 fish-
ing 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 cen-
tury 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 devia-
tion 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 popula-
tion, were very negatively evaluated: first a persistent Gaelic monolingualism,
then a conspicuously imperfect English, and finally, in the last half of the cen-
tury, 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 fisher-
folk identity, adding one more facet to regional perceptions of fisherfolk back-
wardness 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
Nancy Dorian - 978-90-04-26193-8

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, reced-
ing languages have had a set of objectives that privilege Western scientific
modes and topics of inquiry and their objectives and are often – indeed usu-
ally – 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 chil-
dren. 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 mem-
ber of a Canadian First Nation, considers that all of the metaphorical label-
ing 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, pro-
ducing more attention to the artifacts of documentation than to speakers
(op. cit. 134).
One problem inherent in working with receding languages is that such lan-
guages 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 connec-
tion with the term “semi-speaker”, to which negative overtones have become
attached. I was at pains when I introduced the term to celebrate the remark-
able way in which semi-speakers put limited skills to maximum use (ch. 3) and
Nancy Dorian - 978-90-04-26193-8

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 reso-
nances 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 semi-
speakers did.
The perspectives of community members who wish to protect, preserve,
revitalize, or revive their heritage languages are more or less inevitably dis-
tinct 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 produc-
ing 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 acqui-
sition (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 gen-
eral 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 sur-
vival: 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 methodologi-
cally and theoretically problematic. To address this question first among the
East Sutherland fisherfolk descendants, self-identification proved not to be
Nancy Dorian - 978-90-04-26193-8

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 motiva-
tions 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 conversa-
tional capacity. Another formerly fluent speaker suddenly and somewhat mys-
teriously 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 capac-
ity 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 oth-
ers, 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 speak-
ers, 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 birth-
right speakers may inhibit highly competent speakers from claiming skilled-
speaker knowledge of one of their languages. This is a familiar phenomenon
in Aboriginal Australia, where languages are associated with particular terri-
tories and are “owned” by the inhabitants of a given territory; fluent speak-
ers from outside the ownership group are reluctant to claim knowledge, even
Nancy Dorian - 978-90-04-26193-8

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 essen-
tially 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 lan-
guage 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 lan-
guage 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 sur-
prising 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 well-
substantiated 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 fre-
quency 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)?
Nancy Dorian - 978-90-04-26193-8

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 pur-
poses” (Grinevald 2005: 211). No new fluent speakers of Rama have been pro-
duced 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 con-
sider 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 struc-
tural 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 percep-
tions 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 repug-
nance 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 indi-
vidual 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
Nancy Dorian - 978-90-04-26193-8

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 inappropri-
ate uniform-state model of what is actually a highly dynamic phenomenon
(Milroy 1999: 17; 2000: 11–12); although “everyone knows” that language is vari-
able, many people, including scholars of language, nonetheless consider lin-
guistic 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 gram-
matical completeness may be carried into the field by linguists with objectives
that are formed not only by primary personal experience of standardized lan-
guages 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 keep-
ing with our preference for the complete and the invariant, to locating defini-
tive 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 co-
exist 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” lan-
guage 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 unwrit-
ten 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
Nancy Dorian - 978-90-04-26193-8

introduction 19
were altogether different from his: “Simply put, they were interested in words
not grammar” (Collins 1998: 260). Kroskrity, similarly, found that the linguis-
tic 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 unac-
knowledged 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 perspec-
tive 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 speak-
ers, 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 bi-
and 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 reper-
tory 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, partic-
ularly 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
Nancy Dorian - 978-90-04-26193-8

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 particu-
lar 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 communi-
ties to maintain their languages. Not the least of the responsibilities we bear as
researchers is an obligation to scrutinize the ideologies that underlie our schol-
arly 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 com-
munity 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 read-
ing 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 dra-
matic character of warnings about the “endangerment crisis” are now them-
selves the subject of critical commentary (e.g. Moore 2006, Muehlmann 2011).
Since the late 1990s (e.g. Himmelmann 1998), discussions about documen-
tation have come to center on the advantages of (cross-disciplinary) team
Nancy Dorian - 978-90-04-26193-8

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 lin-
guists 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 dif-
ficult ethical problems, first and foremost the simple impossibility of gain-
ing 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 docu-
mented, 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 field-
workers 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 popula-
tion of mixed-proficiency speakers, we call attention to the “innovative” (non-
traditional) 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
Nancy Dorian - 978-90-04-26193-8

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 indis-
criminate 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 contempo-
rary colleagues and scientific posterity. Increasingly, if belatedly, linguists will
be engaged in some kind of support effort – to devise or reform an orthog-
raphy, provide a user-friendly dictionary, help produce primers or story col-
lections, 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 suc-
cessive but very different roles, missionary and anthropologist, among the
Toba Indians of the Argentine Chaco, has lessons for linguists as well as anthro-
pologists: the Toba found him considerably more useful to them as a mission-
ary, 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 oppos-
ing viewpoints of Ladefoged 1992 and Dorian 1993 where the “political” role of
linguistic scholars is concerned).
Nancy Dorian - 978-90-04-26193-8

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 fea-
tures (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 appar-
ent absurdity of nineteenth-century philology’s relentless reading of spiritual
qualities from linguistic structures”, but then wondered whether “the single-
minded reading of power into and out of communicative practices that has
characterized our own late-twentieth-century sociolinguistics will look as ludi-
crously 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 expe-
riences 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, inter-
views, and some folk tradition. Although I asked questions fairly frequently,
these interchanges remained primarily informational. By contrast, the mate-
rial 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
Nancy Dorian - 978-90-04-26193-8

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 forthcom-
ing the sources are (unknowns such as the behavior or even aptitude of previ-
ous 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 record-
ings 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 consider-
able number of speakers could anyone hope to determine what part of the
variability represented geographical variation and what part personally pat-
terned 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 organiza-
tion 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 profu-
sion of variant forms for some sort of late-stage laxity or for the dissolution of
Nancy Dorian - 978-90-04-26193-8

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 linguis-
tic 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 how-
ever 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 retro-
spect 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 seman-
tics. 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.
Nancy Dorian - 978-90-04-26193-8

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 con-
traction and death, 181–96. Cambridge: Cambridge University Press.
Chatsis, Annabelle, Mizuki Miyashita, and Debbie Cole. 2013. A documentary ethnog-
raphy 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 lan-
guages. 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 exer-
cises. 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.
———. 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 set-
ting. 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.
Nancy Dorian - 978-90-04-26193-8

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 pre-
literate 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 biocul-
tural diversity, 175–89. Washington D.C.: Smithsonian Institution Press.
———. 2002. “Expert rhetorics” in advocacy for endangered languages: Who is listen-
ing, 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.
Nancy Dorian - 978-90-04-26193-8

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 minor-
ity languages and language varieties, 23–33. Georgetown: Georgetown University
Press.
Long, Michael H. 1993. Second language acquisition as a function of age: Research find-
ings and methodological issues. In Kenneth Hyltenstam and Åke Viberg, eds,
Progression and regression in language: Sociocultural, neuropsychological and lin-
guistic 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:
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.
Nancy Dorian - 978-90-04-26193-8

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.
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 com-
munities 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 ide-
ologies: 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.
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
Part one
Language Change in an Obsolescent Language
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
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 fish-
ing still existed as a livelihood. At that time the fisherfolk were occupationally
isolated in a district of tradesmen and agriculturalists. They were not linguisti-
cally 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 vil-
lages 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 for-
mer 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.

34 chapter 1
there was nothing to attract new population to the village. As a result, Gaelic-
English 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 vil-
lages. 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 play-
ground 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 bilin-
guals 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 grand-
parents. 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 develop-
ments 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 environ-
ments 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 phe-
nomena 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
Nancy Dorian - 978-90-04-26193-8

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 com-
moner 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 spe-
cifics 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 ques-
tions 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:
Nancy Dorian - 978-90-04-26193-8

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 his-
torical 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 lin-

guistic 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-dis-
tant 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 speak-
ers 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 con-
sonant 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 Gaelic-
English 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 subse-
quently 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 ver-
sion of the Psalms.
Nancy Dorian - 978-90-04-26193-8

initial consonantal choices which includes (to use the traditional terminology)
those of “lenition”,3 those of “nasalization”,4 and those of the unmutated radi-
cal initial. Given this basic system in common, and given a group of older, more
conservative speakers who represent a mutational norm, the younger speak-
ers’ 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.
Nancy Dorian - 978-90-04-26193-8

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, con-
servative 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 innova-
tion 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 lexi-
con is extremely weak and shows borrowing from English on a truly massive
scale. Since there are plenty of verbs whose initial consonants are not suscep-
tible 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 neces-
6 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.
Nancy Dorian - 978-90-04-26193-8

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 ver-
bal 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 obliga-
tory 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).
Nancy Dorian - 978-90-04-26193-8

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 pos-
sessive 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 con-
servative ESG are precisely the ones which produce mutation in the standard
language, and the mutations produced are also the same:
sg pl.
1st person mə L nə11

2nd person tə L nə
3rd person masc. əL ənN
fem. ə
10 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.
11 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.
Nancy Dorian - 978-90-04-26193-8

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 envi-
ronment 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 ‘Where is my sweater?’

{khačh ənN bel təL phεčan} ‘Where is your-sg. sweater?’
{khačh ənN bel ənN phεčan ig} ‘Where is the sweater at-him?’
{khačh ənN bel ənN phεčan εkh} ‘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 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.
13 /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.
14 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.
Nancy Dorian - 978-90-04-26193-8

42 chapter 1
‘go’; this construction will be referred to as the dol-passive. The full construc-
tion 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 sur-
prising to find that younger speakers tend to generalize lenition, the com-
monest 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.
Nancy Dorian - 978-90-04-26193-8

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 muta-
tional 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 léš ənN šεnuar} /va aǰ er ə throgal léš 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 unimport-
ant that many younger speakers use initials which are inappropriate from the
point of view of both traditional Gaelic grammar and conservative ESG gram-
mar. In the dol-passives, on the other hand, any such use of inappropriate ini-
tials 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 muta-
tional 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.
Nancy Dorian - 978-90-04-26193-8

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 sub-
ject 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 con-
struction, 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 decade-
groups, 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 per-
centage 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.
Nancy Dorian - 978-90-04-26193-8

Table 1.1 Embo speakers’ dol-passives in the 3sg. f. and 3pl.
Decade Speaker No Subj. Pron., Subj.Pron. Subj.Pron.

group No Preposition only And Prep.
80’s E1 9 – –
9 (100%) 0 0 Totals
70’s E2 6 – –
E3 3 2 –
E4 2 1 1
E5 4 3 –
15 (68%) 6 (27%) 1 (5%) Totals
60’s E6 – 1 2
E7 3 – 1
E8 7 – 1
E9 1 1 4
11 (52%) 2 (10%) 8 (38%) Totals
50’s E10 1 3 2
E11 – – 5
E12 7 – 5
E13 1 2 3
E14 2 3 7
11 (27%) 8 (19%) 22 (53.5%) Totals
40’s E16 – 4 12
E17 – 1 3
E18 – 3 9
E19 – – 5
0 8 (22%) 29 (78%) 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 appropri-
ate 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.
Nancy Dorian - 978-90-04-26193-8

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 ini-

tials 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.
Nancy Dorian - 978-90-04-26193-8

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. 3 sg.f. and 3 pl.

bith – passive dol – passive
Decade Speaker Appropr. Inappropr. Appropr. Inappropr.
group Initial Initial Initial Initial
80’s E1 – – 9 –
9 0 Totals
70’s E2 2 – 6 –
E3 – – 3 –
E4 5 1 3 1
E5 3 1 5 2
10 2 17 3 Totals
60’s E6 – – 1 2
E7 – – 4 –
E8 – – 8 –
E9 1 – 5 1
1 0 18 3 Totals
50’s E10 8 3 4 2
E11 3 2 2 3
E12 – – 10 2
E13 3 2 4 2
E14 – 1 6 6
14 8 26 15 Totals
40’s E16 1 1 9 7
E17 5 10 1 4
E18 – 1 1 11
E19 – 2 – 5
6 14 11 27 Totals
produces more inappropriate than appropriate initial consonants. But E4, in

the next-older decade-group, differs sharply from her son. Her use of appro-
priate vs. inappropriate initials runs four to one in favor of appropriate ini-
tials, 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.
Nancy Dorian - 978-90-04-26193-8

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 down-
ward 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 pro-
nouns, three of which are homophonous in the presence of a preceding con-
sonant 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 differ-
entiated 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.
Nancy Dorian - 978-90-04-26193-8

mutational redundancy in the dol-passive and the built-in mutational redun-

dancy 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 attrib-
utive 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.
Nancy Dorian - 978-90-04-26193-8

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 innova-
tive 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 per-
suaded 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 substi-
tuted 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 con-
structions 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 gram-
matical 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 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.
22 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.
23 Nominative and accusative are not distinguished. The term nominative will be used to
cover the general nominative/accusative form.
Nancy Dorian - 978-90-04-26193-8

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 particu-
lar 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 defi-
nite 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 An unmutated feminine genitive singular in the definite noun does not survive in ESG,
even in fossilized phrases.
25 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 geographically-
based 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.
26 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 phonologically-
constituted and phonetically coherent sub-group of type 2.
Nancy Dorian - 978-90-04-26193-8

52 chapter 1
Table 1.3 Nominative mutations after the definite article
N after Art. L after Art.
/khe:li/ ‘ceilidh’ 7 speakers 0 speakers

/khrãũ:n/ ‘mast’ 7 0
/khu:/ ‘dog’ 6 0
/khath/ ‘cat’ 6 0
/kho:th/ ‘coat’ 7 0
/khyo:/ ‘fog’ 7 0
/khyɔ:rd/ ‘tinker’ 7 0
/khɔr/ ‘kettle’ 7 0
/khũãn/ ‘ocean’ 6 0
/khɫɔxar/ ‘mason’ 7 0
/khɫu:th/ ‘cloth, rag’ 7 0
/khogu/ ‘war’ 6 0
/khɔrkh/ ‘oats’ 7 1
/khrã:ĩ/ ‘bone’ 6 0
/khalax/ ‘rooster’ 7 0
My early attempt (1968) in the tracking down of the scale and limits of muta-
tional 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
Nancy Dorian - 978-90-04-26193-8

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 respon-
dents, under the pretense of checking to see whether they were recorded cor-
rectly; 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 mascu-
lines 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 stone-
mason?’ 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 pre-
sented 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 depar-
ture of the younger generation from the conservative norm, namely nasaliza-
tion 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.
Nancy Dorian - 978-90-04-26193-8

54 chapter 1
Table 1.4 Embo speakers’ masculine and feminine nominatives
Decade group /khɫɔx/, fem. /khɫɔxar/, masc.

L N L N
80’s E1 E1
70’s E2 E2
E4 E4
E5 E5
60’s E6 E6
E7 E7
E8 E8
E9 E9
50’s E10 < E10 E10
E11 E11
E12 E12
E14 E14
40’s E16 E16 E16
E17 < E17 E17
E18 E18
E19 E19
prepositions.28 As noted above, no use of the genitive with prepositions sur-

vives 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 preposi-
tional phrase with the definite noun, ‘against (to) the wall’, would be regularly lenited by
conservative speakers: {riš əL pɔɫ} /riš ə vɔɫ/.
Nancy Dorian - 978-90-04-26193-8

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 Lenition

Older Younger Older Younger
Speakers Speakers
/khe:li/ ‘ceilidh’ /ɔrn/ ‘for’ 1 – 4 3

/ig/ ‘at’ – 2 6 2
/khrãũ:n/ ‘mast’ /ɔrn/ ‘for’ 3 2 2 2
/yε/ ‘off of’ – 1 5 2
/khu:/ ‘dog’ /er/ ‘on’ 2 1 3 4
/tə/ ‘for’ – 1 3 1
/kho:th/ ‘coat’ /ɔrn/ ‘for’ 4 3 – 1
/yε/ ‘off of’ – 2 4 1
/fo/ ‘under’ 1 1 5 3
/khyo:/ ‘fog’ /thrε̃/’through’ 2 2 3 2
/as/ ‘in’ – - 5 2
/khyɔ:rd/ ‘tinker’ /tə/ ‘to’ 1 1 4 1
/khɔr/ ‘kettle’ /ɔrn/ ‘for’ 5 1 – 1
/as/ ‘in’ – – 4 4
/khũãn/ ‘sea’ /vɔ/ ‘from’ 1 – 3 –
/as/ ‘in’ – – 4 2
/khɫɔxar/ ‘mason’ /ig/ ‘at’ – – 4 3
/léš/ ‘with’ – 3 5 –
/khɫu:th/ ‘cloth’ /kus/ ‘to’ 1 – 5 1
/khogu/ ‘war’ /ɔrn/ ‘for’ 3 – 2 1
/khɔrkh/ ‘oats’ /as/ ‘in’ 2 1 4 1
/yε/ ‘out of’ – – 4 1
/khrã:ĩ/ ‘bone’ /er/ ‘on’ 1 – 3 1
/khalax/ ‘rooster’ /vɔ/ ‘from’ – 1 4 2
Totals 27 22 86 41
29 There are only 14 nouns in the prepositional tests; /khath/ ‘cat’ was inadvertently omitted.
Nancy Dorian - 978-90-04-26193-8

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 nasal-
izations; /léš/ ‘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 sug-
gestive older-speaker results of Table 5, we can establish a conservative ESG
30 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.
31 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.
32 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/
Nancy Dorian - 978-90-04-26193-8

Table 1.6 Embo speakers’ mutations with /orn/ and /as/
Noun /khɔr/ /khrãũ:n/ /kho:th/

Prep. /ɔrn/ /as/ /ɔrn/ /as/ /ɔrn/ /as/
N L N L N L N L N L N L
Speaker
E1 N L N L – – L
E2 N L N L N L
E3 N L N N1 L2 N L
E4 N L N N1 L2 N L
E5 L L L L N L
E6 N N N L N N
E9 N L N L N N
older speakers
younger speakers
E10 N L N N N L
E11 N N N N N N
E12 N N N L N N
E13 N L N L N L
E14 L L N N N L
E15 N L N N N N
E16 N L L L N N
E17 N L L L N N
E18 N N N N N N
E19 N L 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 leni-
tions in Table 5, despite the high number of /ɔrn/-sentences, whereas younger
speakers in the same sentences produced nasalizations in better than half the
Nancy Dorian - 978-90-04-26193-8

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 ‘gov-
erned the nominative’ in conservative ESG usage must have provided a ready
model for nasalization of definite masculines in the prepositional environ-
ment, 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 /léš/ ‘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 letter-
tape 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.
Nancy Dorian - 978-90-04-26193-8

the husband-and-wife team E14 and E16 not only nasalize four-to-one after /
léš/, 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 devel-
opment 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 con-
tender 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 mas-
culine nominative. The briefer recordings of two other younger speakers, E10
and E13, show the same pattern: nasalization without exception in the mas-
culine nominative. Further, even for such young speakers as E18 and E19, who
lenite the masculine nominative /khɫɔxar/, nasalization is the strongly pre-
ferred mutation for masculines elsewhere, as Table 6 indicates. This preference
Nancy Dorian - 978-90-04-26193-8

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 prob-

ably no speakers who use this system exclusively, since at least the highest-
frequency 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. N ~ L L
Gen. N ~ L L
Dat. N ~ L L
The boldface mutational symbols represent current younger-speaker prefer-

ences, and only the prepositional environment which traditional Gaelic gram-
mar 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 mascu-
line 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 morphophonol-
ogy (masculine and feminine palatalization with internal vowel change). The
dative, too, is called for only in one syntactic environment, the prepositional
Nancy Dorian - 978-90-04-26193-8

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 mor-
phological shapes within a case system, and moves toward the status of so
many other morphemes in the language: it becomes a paradigmatically iso-
lated morpheme, much like {kle:L}, {ta:L} etc. Grammatically it is not redun-
dant, 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 sys-
tem, 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 oblig-
atory 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
Nancy Dorian - 978-90-04-26193-8

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 muta-
tion 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 for-
mation 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 sub-
ject 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 muta-
tions 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 sig-
nificant grammatical load, the dol-passives in the 3rd person of both numbers,
analogical introduction of subject pronouns provided the margin of redun-
dancy needed to allow for a leveling through of lenition to all persons.
Nancy Dorian - 978-90-04-26193-8

6.4 The Rate of Change

If the kinds and processes of change which we find within the mutational gram-
mar 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 pho-
nological 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 prog-
ress 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 com-
fortable and more proficient in Gaelic than in English.) But even the 40-year-
plus 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 gram-
mar of ESG are sharply constricted in comparison with relatively healthy west-
ern 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
Nancy Dorian - 978-90-04-26193-8

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 pho-
nological 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 dis-
covered a last, assimilation-resistant Ofo speaker who had ‘practiced the lan-
guage 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, how-
ever, 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 prob-
ably be assumed, wherever a grammar is written or a proto-language recon-
structed on the basis of materials gathered from a few last remaining speakers,
that the stage of the language represented by those speakers is markedly devi-
ant even in terms of the recent history of the language.
Nancy Dorian - 978-90-04-26193-8

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 romani-
schen 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.: Anthropo-
logical Society of Washington. Reprinted in Fishman, 99–138.
Kirparsky, Paul. 1968. Linguistic universals and linguistic change. Universals in linguis-
tic 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.
Nancy Dorian - 978-90-04-26193-8

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 sociolin-
guistic theory. Not least among the interests of recent writers have been issues
of simplification and convergence as linguistic processes; witness the very sub-
stantial 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 vari-
eties of a language, by simplification in some circumstances, expansion in oth-
ers; (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 conflu-
ence 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 characteristi-
cally 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”, “dis-
torted”, “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
Nancy Dorian - 978-90-04-26193-8

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 gen-
der 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 hierar-
chy of morphophonemic decay in terminal East Sutherland Scottish Gaelic,
related apparently to avoidance of grammatical syncretism and to the pres-
ence 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 conse-
quences of the massive lexical importation for language loyalty to Nahuatl.
In the present study, I wish to pursue the issues of simplification and conflu-
ence in language death by examining closely the fate of morphological com-
plexity in a terminal Scottish Gaelic dialect. The structures chosen for
investigation represent the extreme in morphological complexity for this dia-
lect 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 mor-
phophonemically 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 com-
plete 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.
Nancy Dorian - 978-90-04-26193-8

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 isola-
tion from all other Gaelic dialects, and because it is succumbing not in compe-
tition 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 under-
stood 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 speak-
ers whose Gaelic is very good, yet shows certain departures from the conserva-
tive 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 lan-
guage 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 mor-
phology 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 estab-
lished on the basis of differences in grammatical usage between age-groups, as reported in
Nancy Dorian - 978-90-04-26193-8

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 lexico-
statistical 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).
Nancy Dorian - 978-90-04-26193-8

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 agree-
ment 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 ĩnág/

broke the lad wee that the window
‘That wee lad broke the window.’
/vriš nə prɔgiç peg šən nəh ĩnágən/

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 mi/ becomes /pi mi prišu/

will-break I will-be 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.
Nancy Dorian - 978-90-04-26193-8

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 for-
mation of the noun plural and the gerund, as elicited from o.f.s. (taken as rep-
resenting 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 pho-
neme. 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 muta-
tion 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 num-
ber 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 plu-
ral, 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, sub-
traction actually combines only with vowel alternation plus suffixation (ger-
und type IX) and with suffixation (gerund type X).
Nancy Dorian - 978-90-04-26193-8

72 chapter 2
Table 2.1
Noun Plural
Singular Plural
I. suffixation /pre:g/ /pre:gən/ ‘lies’

II. final mutation /phũ:nth/ /phũ:nčh/ ‘pounds’
III. suppletion /thε/ /thro:r/ ‘houses’
IV. quantity change
+ suffixation /phyu:r/ /phyuriçεn/ ‘sisters’
V. final mutation
+ suffixation /se:x/ /se:çεn/ ‘dishes’
VI. vowel alternation /mãkh/ /mĩkh/ ‘sons’
VII. vowel alternation
+ final mutation /thəu:ɫ/ /thwi:l´/ ‘holes’
VIII. vowel alternation
+ suffixation /khu:/ /khɔ̃ n´/ ‘dogs’
IX. vowel alternation
+ final mutation
+ suffixation /yax/ /yəiçu/ ‘horses’
X. vowel alternation
+ syncope
+ suffixation /tarəs/ /tɔrsĩn/ ‘doors’
XI. quantity change /ĩnán/ /ĩnán:/ ‘onions’
Gerund
Root Gerund
I. suffixation /ĩ:š/ /ĩ:šu/ ‘telling’

II. final mutation
vowel alternation /furiç/ /furax/ ‘waiting’
III. suppletion /rax/ /tuɫ/ ‘going’
IV. quantity change
+ suffixation /kyɔ:r/ /kyɔru/ ‘cutting’
V. final mutation
+ suffixation, with
or without vowel /chirəməxu/
alternationa /čhirəmiç/ or /chirəmixu/ ‘drying’
VI. subtraction /ibəriç/ /ibər/ ‘working’
(Continued)
Nancy Dorian - 978-90-04-26193-8

Table 2.1 (Continued)
Gerund
Root Gerund
VII. subtraction
+ quantity change /e:riç/ /eri/b ‘rising’
VIII. subtraction
+ final mutation
+ suffixation /kɫu:əškh/b /kɫu:əsth/ ‘moving’
IX. subtraction
+ vowel alternation /khrəǰ/ or
+ suffixation /khrεǰ/ /khresčən/ ‘believing
X. subtraction
+ suffixation /e:šiç/ /e:šnax/ ‘listening’
XI. zeroc /khrεkh/ /khrεkh/ ‘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 morphologi-
cally 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 speak-
ers 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
Nancy Dorian - 978-90-04-26193-8

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 plu-
rals and the zero formation in gerunds. However, a reduction in vowel alterna-
tion 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
o.f.s. y.f.s. s.s.

Formation Type # % # % # %
/-(ə)xən/ (I)a 48 18 34 12.5 87 19

final mutation (II) 28 10 24 9 23 5
/-ən/ (I) 27 10 24 9 124 27
vowel alternation + final
mutation (VII) 24 9 28 10.5 11 2.5
vowel alternation + /-(V)n(´)/b (VIII) 21 8 33 12 26 5.5
/-čεn/ (I) 20 7.5 25 9 18 4
lengthening of final consonant (XI) 19 7 15 5.5 5 1
/-n(:)/ (I) 16 6 16 6 11 2.5
/-içεn/ (I) 14 5 15 5.5 24 5
(Continued)
Nancy Dorian - 978-90-04-26193-8

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

Formation Type # % # % # %
vowel alternation (VI) 14 5 12 4.5 18 4

vowel shortening + /-içεn/ (IV) 10 4 12 4.5 14 3
vowel alternation + syncope +
/-in/ (X) 8 3 8 3 6 1
suppletion (III) 6 2 8 3 9 2
/-in(:)/ (1) 4 1.5 2 1 7 1.5
vowel alternation + final mutation
+ suffixc (IX) 3 1 4 1.5 7 1.5
/-čax/ (1) 3 1 – –
final mutation + /-εn/ (V) 2 1 3 1 6 1
vowel shortening + /-in(:)/ (IV) – 3 1 –
suffixes unique to a single group of
speakersd 5 2 2 1 21 4.5
zero – 1 .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 for-

mation 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.
Nancy Dorian - 978-90-04-26193-8

76 chapter 2
Table 2.3
SPEAKER GROUP # %
o.f.s. 134 50
y.f.s. 118 44
s.s. 292 63.5
4.2. Among the devices of relative prominence (those constituting 5% or bet-

ter) in the plural formations of fluent speakers, the two which show the great-
est 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 o.f.s. y.f.s. s.s.

# % # % # %
TypeVII 24 9 28 10.5 11 2.5

Type VIII 21 8 33 12 26 5.5
Type VI 14 5 12 4.5 18 4
Type X 8 3 8 3 6 1
Type IX 3 1 4 1.5 7 1.5
TABLE 2.5
o.f.s. y.f.s. s.s.

# % # % # %
Type II 28 10 24 9 23 5
Type VII 24 9 28 10.5 11 2.5
Type IX 3 1 4 1.5 7 1.5
Type V 2 1 3 1 6 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
Nancy Dorian - 978-90-04-26193-8

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 pecu-
liarly 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 forma-
tion 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 lat-
ter) 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 o.f.s y.f.s. s.s.

# % # % # %
long 34 87 29 88 8 35
half-long 4 10 4 12 5 22
short 1 2.5 – – 10 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 consider-
able 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 conso-
nant length become sporadic or disappear. Where quantity change is con-
cerned, 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
Nancy Dorian - 978-90-04-26193-8

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 indepen-
dent 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 for-
mations (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 represent-
ing 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/.
Nancy Dorian - 978-90-04-26193-8

Table 2.7

# % # % # %
FORMATION TYPE
/-u/ (I)a 41 26 38 24 62 27
zero (XI) 25 16 27 17 42 18.5
final mutation + vowel 16 10 20 12.5 4 2
alternation (II)
/-al/ (I) 14 9 16 10 58 25.5
vowel shortening + 9 6 4 2.5 3 1
/-u/ (IV)
vowel shortening + 8 5 9 5.5 11 5
/-al/ (IV)
/-tən/ (I) 8 5 11 7 6 2.5
subtraction (VI) 7 4.5 9 5.5 11 5
/-šən/ (I) 4 2.5 1 .5 – –
suppletion (III) 4 2.5 4 2.5 7 3
/-əm/ (I) 4 2.5 4 2.5 10 4.5
final mutation + 3 2 4 2.5 – –
suffixation (V)
vowel alternation + 3 2 4 2.5 – –
subtraction +
/-sčən/ (IX)
/-i/ (I) 2 1 3 2 3 1
/-ən/ (I) 2 1 – – 2 1
/-d/ (I) 2 1 – – – –
vowel shortening + 1 .5 1 .5 – –
/-tən/ (IV)
subtraction + 1 .5 3 2 1 .5
/-nax/ (X)
vowel shortening + 1 .5 – – – –
subtraction (VII)
subtraction + 1 .5 – – – –
final mutation
+ /-th/ (VIII)
vowel alternation – – – – 3 1
/-kən/ (I) – – – – 1 .5
(Continued)
Nancy Dorian - 978-90-04-26193-8

80 chapter 2

# % # % # %
FORMATION TYPE
/-iç/ (I) – – – – 1 .5
/-in/ (I) – – – – 1 .5
/-x/ (I) – – – – 1 .5
infixation – – – – 1 .5
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. 77 49
y.f.s. 73 46
s.s. 145 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. 20 13
y.f.s. 24 15
s.s. 4 2
Nancy Dorian - 978-90-04-26193-8

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 sub-
traction (type VII; cf. Table I, fn. b).
Table 2.10
SPEAKER GROUP # %
o.f.s. 18 11.5
y.f.s. 14 9
s.s 14 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 non-
suffixal 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 suf-
fixation, type V); the other is the complex formation made up of vowel alterna-
tion 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 for-
mation”. 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%).
Nancy Dorian - 978-90-04-26193-8

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. 50% 49%

y.f.s. 44 46
s.s. 63.5 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 com-
pared 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 contin-
uum 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 suffix-
ation. 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 common-
est 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 sen-
tences, so as to avoid the planting of a built-in model for analogical change in the test itself.
Nancy Dorian - 978-90-04-26193-8

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 combina-
tion 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 57 21 20 13
y.f.s 59 22 24 15
s.s. 47 10 4 2
Again it is conspicuous that the drop-off in the use of final mutation as a mor-
phological device is both characteristic of and peculiar to s.s.; it is not conso-
nant 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 break-
away 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/.
Nancy Dorian - 978-90-04-26193-8

84 chapter 2
Table 2.13a
PLURAL GERUND
o.f.s. 70 26 19 12
y.f.s. 85 31.5 24 15
s.s. 68 15.5 7 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 alterna-
tion (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 quan-
tity 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
o.f.s. 49 18 19 12
y.f.s. 48 18 14 9
s.s. 27 6 14 6
Nancy Dorian - 978-90-04-26193-8

Innovatively idiosyncratic formations not shared by any other group of speak-

ers 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 generaliza-
tion does not always hold). S.s. idiosyncrasy, however, shows rather less stabil-
ity 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 fluent-
speaker 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-estab-
lished in the dialect. Her /-an/ plurals and her zero plurals, however, are idio-
syncratic 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 ger-
unds 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 redun-
dant. The locus of marked plurality may thus change without loss of the gram-
matical 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
Nancy Dorian - 978-90-04-26193-8

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 over-
lap 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 fac-
tors 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 ger-
unds 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 speak-
ers 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 fre-
quency of the noun is not a guarantee of the preservation of a non-productive
plural formation. Some very high-frequency nouns which have vowel-alterna-
tion plurals for fluent speakers developed suffixal plurals for most s.s., or joined
the zero-plural group. Suppletion is undoubtedly reinforced in gerund forma-
tion 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 suffix-
ation and vowel alternation (types II, V, IX), supplied only two instances (separate instances
of the same noun, for two different speakers).
Nancy Dorian - 978-90-04-26193-8

(“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 mor-
phology 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 rela-
tively little with fluent speakers, forming a group apart, both socially and lin-
guistically. 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 genera-
tion of their grandparents, are so conspicuous, in the linguistic autobiogra-
phies 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 linguis-
tic 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, con-
tinue 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 sec-
ond ascending generation – usually a grandmother, but sometimes a great-
aunt 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 main-
tain 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 ‘semi-
speaker’ 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] Wasco-
Wishram Chinookan work as well’.
Nancy Dorian - 978-90-04-26193-8

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-gen-
eration 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 contin-
uum of speakers. This is particularly striking because the morphological com-
plexity of the two structures concerned would seem almost to invite
simplification via leveling, and certainly to provide the means: simple suffix-
ation, 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 morphologi-
cal structures. These y.f.s. not only share, both actively and passively, the mor-
phological 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 func-
tional factor of habitual home use is sufficient to prevent the onset of morpho-
logical 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 complet-
ing 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 perfor-
mance 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
Nancy Dorian - 978-90-04-26193-8

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 simple-
suffix plurals for 42 of her 64 nouns, the most conservative o.f. [speakers] plu-
ralized by simple suffix in 27 of the same cases. Table 2.15 presents a full analysis
of the data.
Table 2.15
I.F. o.s.f. I.H. o.s.f.

# % # % # % # %
total plurals 79 100 79 100 64 100 64 100

plurals by simple suffixation 53 67 38 48 42 65.5 27 42
zero plurals 9 11 0 0 7 11 0 0
plurals by neither simple
suffixation nor zero 17 21.5 41 51.5 15 23.5 37 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 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.”
10 Three of the group are relatively strong speakers, as s.s. go; two are of intermediate profi-
ciency; two are quite weak speakers; one is so weak as to have been barely usable and
should not really be termed a “speaker”.
Nancy Dorian - 978-90-04-26193-8

90 chapter 2
If we return to the comparison of simplification in language death with sim-

plification 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 sen-
tence 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, how-
ever), 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 provid-
ing 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 mor-
phology 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 per-
haps the growing dominance of a single suffix in each case, where the over-
whelming 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 whole-
sale 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 pid-
ginization 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
Nancy Dorian - 978-90-04-26193-8

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 fac-
tors 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 socializa-
tion of the s.s.; and the “integrative” rather than “instrumental” role (see
Lambert 1967: 102) of Gaelic in East Sutherland, where the language is emblem-
atic 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 “typi-
cal” 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 sim-
plification 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.
Nancy Dorian - 978-90-04-26193-8

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 lin-
guistics 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:
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.
Nancy Dorian - 978-90-04-26193-8

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 disap-
pearance 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 perfor-
mance 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 differentia-
tions, 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, partic-
ularly 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

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 cor-
respondingly 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” speak-
ers at the bottom. Even though there is such a pattern (as the very term profi-
ciency 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 phe-
nomena associated with language death and contraction be taken, quite erro-
neously, 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, occupa-
tionally, 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
Nancy Dorian - 978-90-04-26193-8

making do with less 95
when included in my study if the criterion for dominance was taken to be pref-
erence 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 domi-
nance 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 num-
ber 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-than-
ideal 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 opin-
ion that many of the fluent speakers with whom I worked should be consid-
ered Gaelic-dominant, including – very conspicuously – the youngest of them.
A smaller number of individuals could best be considered truly balanced bilin-
guals, almost entirely responsive to the interlocutor factor (or that factor com-
bined 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 mat-
ters; 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;
Nancy Dorian - 978-90-04-26193-8

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 individu-
als 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 dif-
ficulty 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 individu-
als among the ranks of the fully fluent, older or younger, could be pressed for
the most conservative or elaborate structures in their repertoire more success-
fully than others (that is, without becoming confused or hesitant and with rel-
ish for the challenge to their abilities).
4 Troublesome Structures: Morphological
Most of the structures which proved difficult for fully fluent speakers were syn-
tactically somewhat complex and had probably fallen out of common use as
the language receded. One exception was a deficiency in morphological struc-
ture, namely uncertainty about the formation of ordinal numbers as opposed
to cardinal numbers. Gaelic numbers are notoriously difficult compared with
Nancy Dorian - 978-90-04-26193-8

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 con-
versation 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 (cor-
rectly) 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 flu-
ent 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.
Nancy Dorian - 978-90-04-26193-8

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.
4.1 The Subjective Element in the “troublesomeness” of Troublesome

Structures
In the case of these ordinals, speakers were well aware that they were produc-
ing 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 tradi-
tionally 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 adver-
bial 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 accept-
able variability for ESG, then, whereas the anomalies of ordinal formation rep-
resent 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 per-
turb fluent speakers, whereas the anomaly of ordinal modifiers identical with
Nancy Dorian - 978-90-04-26193-8

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 com-
plements 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.
Nancy Dorian - 978-90-04-26193-8

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 differ-
ently: “I’m trying to make that he’ll hear me”; but the structure of the expres-
sion “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; excep-
tional 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 sen-
tence 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 dur-
ing 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 accu-
rate English translations instantly and easily.3
In view of the difficulties with active control of the double verb-comple-
ment 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 conserva-
tive ESG, two identically structured prepositional relative clauses in immedi-
ate succession: literally, “Died the woman at (-whom) was the house in (-which)
happened it” (with the italicized portions representing exactly parallel struc-
tures). 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.”
Nancy Dorian - 978-90-04-26193-8

(as was also true of the double verb-complement structure) but also particles
and dependent as opposed to independent verb forms, they can be consid-
ered 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 muta-
tion (always lenition in this construction, a mutation which generally replaces
stops with spirants and spirants with other spirants or zero: /khɫĩ:nčən/ “hear-
ing,” /fiaxən . . . xɫĩ:nčən/ “making . . . hear,” /fa:gal/ “leaving,” /fiaxən . . . a:gal/
“making . . . leave”).
Fluent speakers asked to produce a Gaelic translation of the English sen-
tence “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 what-
ever. 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 trouble-
some for the fully fluent even before I began to investigate it closely seemed
predictably difficult both because the conjunction involved was of low fre-
quency even in the positive and also because the negative clauses introduced
by that conjunction became semantically very close to a much higher fre-
quency 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”
Nancy Dorian - 978-90-04-26193-8

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 speak-
ers 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 scat-
tered translation-test sentences on a variety of occasions.
Presented with a Gaelic sentence which used /nas ɫə: na/, eight fluent speak-
ers (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 pre-
dicted. 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 flu-
ent 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.
Nancy Dorian - 978-90-04-26193-8

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 sen-
tences 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 spo-
ken 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 mistrans-
late by reversing the negation, where either a negative or a positive subordi-
nate 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 con-
dition 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 typi-
cally 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 “plu-
perfect” did, “Our dinner was eaten before he came.”
In the case of “pluperfect” condition-contrary-to-fact sentences, it is signifi-
cant 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 sit-
uation 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,
Nancy Dorian - 978-90-04-26193-8

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 condi-
tion 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); liter-
ally “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 differ-
entiate. 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-to-
fact 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 identi-
cal 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-
Nancy Dorian - 978-90-04-26193-8

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 confu-
sion 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 dif-
ficulties 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 fac-
tor 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.
Nancy Dorian - 978-90-04-26193-8

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 sen-
tences 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 con-
junctival phrase and not strictly necessary. The most likely explanation for
retention is that /nas ɫə:/ is the comparative of “little” and occurs fairly fre-
quently 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 struc-
ture. 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 sit-
ting 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
Nancy Dorian - 978-90-04-26193-8

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 semi-
speakers (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 flu-
ent speakers. Their acquisition histories are not uniform: some began as flu-
ent 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 vari-
able 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 transla-
tions 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.
Nancy Dorian - 978-90-04-26193-8

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 rela-
tively 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 opportu-
nity 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 ini-
tial 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.
Nancy Dorian - 978-90-04-26193-8

The test sentences were not actually designed to explore the control of con-
junctions, since conjunctions overall seemed a more complex matter than
most SSs could be expected to handle. Nonetheless a number of the test sen-
tences 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 strat-
egies 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 conjunc-
tion 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 mean-
ing permitted, was a useful strategy among SSs for making the complex more
expressable. A strong SS, who knew and used the conjunction “when” on occa-
sion, 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 con-
junctival 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 juxtaposi-
tional 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 nega-
tive form of the conjunction. But since they either did not control that form or
Nancy Dorian - 978-90-04-26193-8

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 conjunc-

tions 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 tradition-
ally 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 sever-
est 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 (incor-
rectly) /tə/. Of the four meanings for which she used /tə/, only “when” was suit-
able; 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 juxtaposi-
tion 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
Nancy Dorian - 978-90-04-26193-8

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 cer-
tainly 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 unex-
pected 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 nega-
tive “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 regu-
lar 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
Nancy Dorian - 978-90-04-26193-8

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 prob-
lem 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 dan-
gerous 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 meth-
ods 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 individu-
als, 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 sem-
blance 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.
Nancy Dorian - 978-90-04-26193-8

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
Nancy Dorian - 978-90-04-26193-8

114 chapter 3
can be conveyed successfully. Linguists and psychologists may want to reexam-

ine 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 lin-
guistic variation. In C. J. Fillmore, D. Kempler, & W. S.-Y. Wang (Eds.), Individual dif-
ferences 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.
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.
Nancy Dorian - 978-90-04-26193-8

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 disappear-
ance 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 reduc-
tion 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 sys-
tem pressure may also be involved in the reduction or loss of a feature or vari-
ant 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, ital-
ics 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 par-
ticularly susceptible to reduction and loss, this paper looks at four grammatical

116 chapter 4
features of the Scottish Gaelic spoken by bilingual fisherfolk and their descen-
dants 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 paral-
lel 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 begin-
ning of the 19th century. At that time great landlords all over the Highlands,
pursuing a new source of cash income, summarily removed their Gaelic-
speaking 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 inter-
mittent 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 stig-
matization. Living in a few densely populated streets beside the sea, all follow-
ing 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.
Nancy Dorian - 978-90-04-26193-8

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 periph-
ery, sharply different in many respects from more mainstream dialects of the
western Highlands. The local dialect, not taught in the schools or used in writ-
ing, was unaffected by any standardization processes. Passive literacy in a stan-
dardized 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 met-
rical 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 direc-
tional ‘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.
Nancy Dorian - 978-90-04-26193-8

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 voca-
tive 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 com-
munity 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 non-
informational 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 speak-
ers and the less than fully fluent speakers (semi-speakers).
Nancy Dorian - 978-90-04-26193-8

of complex kinship ties); they are designated older fluent speakers

(OFSs) here.
b. A somewhat younger group, the individuals aged 41 through 60, were (or
had once been) fully fluent and were not distinguished within the com-
munity 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.
c. 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 vig-
orous 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 English-
speaking 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 rec-
ognized as formerly, rather than currently, fluent.
d. The youngest sources in the sample were aged 40 or below and were con-
sidered 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 reg-
ularly can be recognized across these groupings. Apart from the formerly flu-
ent 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 com-
plete 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 loca-
tion. 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.
Nancy Dorian - 978-90-04-26193-8

120 chapter 4
Table 4.1 Locational-adverb forms used traditionally in environments expressing location vs. directional-
adverb 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 directional adverbs used

to express location innovatively to express location
mwĩ (ç) hurəd stã:n hau:ɫ max nɔ:rd vã:n nũ:ɫ

‘out’ ‘up’ ‘down’ ‘over’ ‘out’ ‘up’ ‘down’ ‘over’
speaker fr el fr el fr el fr el fr el fr el fr el fr el
E3 [85] m. 2 1
E4 82 m. 9 15 28 1 4 1 1
E6 75 m. 3 1
E7 ?74 f. 1
E8 ?71 f. 2
E9 70 f. 2
E10 70 m. 1 1 1
E13 [67] f. 3 1 1 1 1 1
E14 65 f. 2
E15 65 m. 2
E17 64 f. 18 1 4 8 19 10 1 5
E20 58 f. 2
E22 58 f. 9 1 2 1 3 2 1 2 1
E23 57 m. 3
E24 57 f. 5 1 1
E26* 54 f. 16 2 17 16 4 6
E27 54 m. 10 10 2 3 16 3 2 2 3 1
E28 51 m. 1
E29* 50 f. 1 2 3 1 23 3 43 11 77 11 19 2
E30 49 m. 2 1 1
E32* 47 m. 5 1 3
E33 46 m. 1
E34 45 f. 1 1 4 5 2 2
E37 41 m. 1 2 1 6 10 1
E38 40 f. 6 2 6 4 3 2 1
E39 38 f. 2 2 2 3
E40 36 f. [3] 26 6 30 1 39 5 4
E41 31 f. 1
E42 30 f. 1 4 4 8 12 2
Nancy Dorian - 978-90-04-26193-8

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 non-traditional
locational forms directional forms
speaker # % # %
E3 [85] m. 3 100
E4 82 m. 57 96.6 2 3.4
E6 75 m. 4 100
E7 ?74 f. 1
E8 ?71 f. 2 100
E9 70 f. 2 100
E10 70 m. 3 100
E13 [67] f. 7 87.5 1 12.5
E14 65 f. 2 100
E15 65 m. 2 100
E17 64 f. 31 47 35 53
E20 58 f. 2 100
E22 58 f. 21 95.5 1 4.5
E23 57 m. 3 100
E24 57 f. 6 85.7 1 14.3
E26* 54 f. 61 100
E27 54 m. 48 92.3 4 7.7
E28 51 m. 1
E29* 50 f. 7 3.6 189 96.4
E30 49 m. 4 100
E32* 47 m. 6 66.7 3 33.3
E33 46 m. 1
E34 45 f. 2 13.3 13 86.2
E37 41 m. 3 14.3 18 85.7
E38 40 f. 8 33.3 16 66.7
E39 38 f. 4 44.4 5 55.6
E40 36 f. 111 100
E41 31 f. 1
E42 30 f. 1 3.2 30 96.8
Nancy Dorian - 978-90-04-26193-8

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 per-
centage of such forms for each speaker.
Table 4.3 Number and percentage of instances of traditional vocative forms with initial-
mutation 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.
trad. vocative marking no vocative marking

speakers fr el tot % fr el tot %
E4 82 m. 1 1
E6 75 m. 6 6 85.7 1 1 14.3
E7 ?74 f. 1 4 5 100
E9 70 f. 11 11 100
E10 70 m. 4 12 16 100
E13 [67] f. 16 2 18 100
E16 64 m. 1 3 4 100
E17 64 f. 14 14 28 100
E21 58 f. 2 7 9 90 1 1 10
E22 58 f. 13 7 20 95.2 1 1 4.8
E23 57 m. 20 20 100
E24 57 f. 4 1 5 100
E26* 54 f. 40 43 83 92.2 7 7 7.8
E27 54 m. 49 49 81.7 11 11 18.3
E29* 50 f. 7 5 12 100
E30 49 m. 2 2 100
E31 49 m. 3 3 100
E33 46 m. 1 1
E34 45 f. 7 9 16 66.7 5 3 8 33.3
E35 43 f. 2 2 66.7 1 1 33.3
E37 41 m. 1 2 3 23.1 4 6 10 76.9
E38 40 f. 7 7 35 2 11 13 65
E40 36 f. 9 12 21 19.8 57 28 85 80.2
E41 31 f. 3 3 30 7 7 70
E42 30 f. 11 11 100
Nancy Dorian - 978-90-04-26193-8

For both the locational adverbs and the vocative case, it is clear that the tra-
ditional structures are best preserved by the older speakers and less well pre-
served 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 voca-
tive marking at 100%; FFS E26*, who showed no retention of locational adver-
bial 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 lan-
guages, 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 leni-
tion was often the only indication of which particle was being used. Gaining
ground across the age-and-proficiency continuum was an alternative nega-
tive 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 expres-
sions 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).
Nancy Dorian - 978-90-04-26193-8

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 supple-
tive 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 speak-
er’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/.
Nancy Dorian - 978-90-04-26193-8

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 non-trad. negative imperative

/(n)a/ or /(x)a/ with /a čhe:ǰ/
/(n)a/ + /(x)a/ + + lenited + lenited

gerund lenited gerund root
gerund
Speakers fr el fr el tot % fr el fr el tot %
E4 82 m. 2 2 100
E6 75 m. 2 3 5 83.3 1 1 16.7
E9 70 f. 1 4 5 100
E10 70 m. 4 8 2 9 23 100
E13 [67] f. 6 6 100
E17 64 f. 2 9 2 5 18 94.7 1 1 5.3
E21 58 f. 1 1
E22 58 f. 6 11 17 100
E23 57 m. 1 1 2 25 6 6 75
E24 57 f. 1 1 2 100
E26* 54 f. 3 2 6 11 22 78.6 1 5 6 21.4
E27 54 m. 9 8 17 100
E29* 50 f. 3 5 6 14 37.8 5 18 23 62.2
E30 49 m. 1 1
E31 49 m. 2 2 100
E33 46 m. 1 1
E34 45 f. 1 1 9 11 78.6 3 3 21.4
E35 43 f. 1 1 50 1 1 50
E37 41 m. 4 1 5 100
E38 40 f. 1 1 2 4 10.3 22 13 35 89.7
E39 38 f. 2 2 50 2 2 50
E40 36 f. 5 38 2 45 100
E41 31 f. 4 2 6 100
E42 30 f. 1 3 4 66.7 1 1 2 33.3
Nancy Dorian - 978-90-04-26193-8

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.
traditional absence of innovative introduction of

preterite part.: /(x)a rɔ/ preterite part.: /(x)a t rɔ/
speakers fr el tot % fr el tot %
E3 [85] m. 3 3 100
E4 82 m. 49 6 55 75.3 18 18 24.7
E5 77 f. 1
E6 75 m. 2 2 25 2 4 6 75
E7 ?74 f. 4 4 100
E9 70 f. 2 2 100
E10 70 m. 5 10 15 100
E13 [67] f. 13 5 18 94.7 1 1 5.3
E14 65 f. 3 3 100
E15 65 m. 2 2 22.2 7 7 77.8
E17 64 f. 441 23 464 99.1 4 4 .9
E18* 60 f. 4 4 100
E20 58 f. 8 8 100
E21 58 f. 4 4 100
E22 58 f. 12 14 26 100
E23 57 m. 1 1
E24 57 f. 1 9 10 76.9 3 3 23.1
E25 56 m. 1 1
E26* 54 f. 88 36 124 93.9 7 1 8 6.1
E27 54 m. 88 28 116 83.6 18 5 23 16.5
E29* 50 f. 17 3 20 7.3 213 40 253 92.7
E30 50 m. 29 2 31 100
E31 48 m. 1 1
E32* 47 m. 12 12 100
E33 46 m. 1 1 20 4 4 80
E34 45 f. 8 4 12 24 24 14 38 76
E36 41 m. 1 1
E37 41 m. 1 11 12 100
E38 40 f. 22 65 87 100
E39 38 f. 8 14 22 88 4 4 12
E40 36 f. 88 79 167 100
E41 31 f. 3 3 100
E42 30 f. 127 17 144 100
Nancy Dorian - 978-90-04-26193-8
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.
1 2 3 4 5 6 7 8
locatnl voc neg imper neg past locatnl voc neg imper neg past
adv ‘be’ w/o adv ‘be’ w.
part part
trad. trad. trad. trad. non- non- non- non-

trad. trad. trad. trad.
speaker % % % % % % % %
E3 [85] m. 100 100

E4 82 m. 96.6 100 75.3 3.4 24.7
E6 75 m. 100 85.7 83.3 25 14.3 16.7 75
E7 ?74 f. 100 100
E8 ?71 f. 100
E9 70 f. 100 100 100 100
E10 70 m. 100 100 100 100
E13 [67] f. 87.5 100 100 94.7 12.5 5.3
E14 65 f. 100 100
E15 65 m. 100 22.2 77.8
E16 64 m. 100
E17 64 f. 47 100 94.7 99.1 53 5.3 .9
E18* 60 f. 100
E20 58 f. 100 100
E21 58 f. 90 100 10
E22 58 f. 95.5 95.2 100 100 4.5 4.8
E23 57 m. 100 25 100 75
E24 57 f. 85.7 100 100 76.9 14.3 23.1
E26* 54 f. 92.2 78.6 93.9 100 7.8 21.4 6.1
E27 54 m. 92.3 81.7 100 83.6 7.7 18.3 16.5
E29* 50 f. 3.6 37.8 7.3 96.4 100 62.2 92.7
E30 49 m. 100 100 100
E31 49 m. 100 100
E32* 47 m. 66.7 100 33.3
E33 46 m. 20 80
E34 45 f. 13.3 66.7 78.6 24 86.2 33.3 21.4 76
E35 43 f. 66.7 50 33.3 50
(Continued)
Nancy Dorian - 978-90-04-26193-8
128 chapter 4
1 2 3 4 5 6 7 8
locatnl voc neg imper neg past locatnl voc neg imper neg past
adv ‘be’ w/o adv ‘be’ w.
part part
trad. trad. trad. trad. non- non- non- non-

trad. trad. trad. trad.
speaker % % % % % % % %
E37 41 m. 14.3 23.1 100 85.7 76.9 100

E38 40 f. 33.3 35 10.3 100 66.7 65 89.7
E39 38 f. 44.4 50 88 55.6 50 12
E40 36 f. 19.8 100 80.2 100 100
E41 31 f. 30 70 100 100
E42 30 f. 3.2 66.7 96.8 100 33.3 100
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 differ-
ent: 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 endorse-
ment 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* dur-
ing 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 com-
monplace, occurring at high frequency in ordinary conversational interac-
tion. However two of them, the negative imperative and the vocative, did not
occur frequently in my freely spoken corpus, even though that corpus accu-
Nancy Dorian - 978-90-04-26193-8

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 famil-
iar 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 num-
ber of cases in which speakers’ retention of the traditional grammatical struc-
ture 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 struc-
tures); three showed no use of the traditional negative imperative structures
and three showed no negated past tense of the verb ‘to be’ without the pret-
erite particle (matched structures). In nearly all cases the speakers who pro-
duced 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 struc-
ture. 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.
Nancy Dorian - 978-90-04-26193-8

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 per-
centage 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 bold-
faced and italicized):
OFSs unm. m. YFSs unm. m. FFSs unm. m. SSs unm. m.
E3 (1) (1) E20 (1) (1) E26* (2) (2) E38 (2) (2)
100 100 100 100 46.1 86.3 34.2 55.2
E4 (1) (2) E21 (1) (1) E29* (2) (2) E39 (1) (2)
96.6 87.7 90 100 1.8 22.6 44.4 69
E6 (2) (2) E22 (2) (2) E32* (1) (1) E40 (2) (2)
92.9 54.2 95.4 100 66.7 100 9.9 0
E7 (1) (1) E23 (1) (1) E41 (1) (2)
100 100 100 25 30 0
E9 (2) (2) E24 (2) (2) E42 (2) (2)
100 100 92.9 88.5 1.6 33.4
E10 (2) (2) E27 (2) (2)
100 100 87 91.8
E13 (2) (2) E30 (2) (1)
93.8 97.4 100 100
E14 (1) (1) E31 (1) (1)
100 100 100 100
E15 (1) (1) E34 (2) (2)
100 22.2 40 51.3
E17 (2) (2) E35 (1) (1)
73.5 96.9 66.7 50
E37 (2) (2)
18.7 50
Nancy Dorian - 978-90-04-26193-8

The ultimate interest in this comparison lies in how often the average percen-
tage of traditional expression of matched structures is higher than the ave-
rage 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 ave-
rage 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 expres-
sion, 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 speak-
ers 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 resis-
tant 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 sur-
prise 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 borrow-
ing as a language-contact phenomenon in obsolescence. This is noteworthy
Nancy Dorian - 978-90-04-26193-8

132 chapter 4
in view of the fact that three of the changes in progress produce greater regu-
larity 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 impera-
tive 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 envi-
ronment.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 interest-
ing challenge to the negative borrowing principle. In the (matched) nega-
tive 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 construc-
tion, 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 conser-
vative 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 regu-
larized 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.
Nancy Dorian - 978-90-04-26193-8

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) dis-
cuss, 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 unex-
pected 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 bor-
rowing 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 dra-
matically 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.
Nancy Dorian - 978-90-04-26193-8

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 theoreti-
cal 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.:
Nancy Dorian - 978-90-04-26193-8

part two
Speaker Skills and the Speech Community in a
Receding Language Context
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
chapter 5
The Problem of the Semi-Speaker in

Language Death
The fieldworker who is investigating a dying language has by definition a limi-

ted 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 contrac-
tion, 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 halt-
ing”, 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 nei-
ther a comparison available nor a markedly deficient manner of delivery. Faced
with this situation, Haas (n.d.: 10) made a judgment based on the sociolinguis-
tic 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’ assump-
tion 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 com-
pared 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

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 individu-
als 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 imper-
fect 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 semi-
speakers 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 grammati-
cal 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 septu-
agenarian speaker B.R. are all considered fluent speakers, down to and includ-
ing 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.
Nancy Dorian - 978-90-04-26193-8

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 hus-
band 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 commu-
nity 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 plu-
ral 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.
Nancy Dorian - 978-90-04-26193-8

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 miss-
ing 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 nega-
tive 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 coin-
cided 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 regular-
ized 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 per-
son 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) absence of a stylistic option (negative imperative; passive)

2) substitution of an analytic construction for a synthetic one (conjugating
preposition; 1st person conditional)
3) 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.
Nancy Dorian - 978-90-04-26193-8

While I suspect that all of these phenomena are in fact characteristic of lan-
guages 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 perfor-
mances of B.R., A.R., and J.R., we find the following:
1) Neg. Imper. Passive

Option 1 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 per-
son 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
Nancy Dorian - 978-90-04-26193-8

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 pro-
gress 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 charac-
teristic 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 semi-
speakers. The Brora semispeaker J.M., for example, levels 7 irregular noun plu-
rals, 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 field-
worker. 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, repre-
senting 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 dia-
lects or languages) may be helpful.
Nancy Dorian - 978-90-04-26193-8

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 con-
sistent) alternation between the citation form [thε] “house” and its counter-
part [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 infor-
mants 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 reduc-
tion noted for East Sutherland Gaelic may prove to be universally character-
istic 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 mor-
phological 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 pro-
ductions 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:
Nancy Dorian - 978-90-04-26193-8

144 chapter 5
1. Is reduced use of a language accompanied by a reduction in form? (i.e.,

are there semi-speakers?)
2. Can the more proficient members of the language community realisti-
cally pinpoint the onset of that reduction? (i.e., can the community accu-
rately identify the semi-speakers?)
3. 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 investi-
gator 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 (“inappropri-
ate 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).
Nancy Dorian - 978-90-04-26193-8

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 semi-
speaker 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.
Nancy Dorian - 978-90-04-26193-8

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 lan-
guages 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 mem-
bers 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 phonol-
ogy, 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 ver-
sion 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.

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 under-
stand 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 sen-
timents 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 com-
munities 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 profi-
cient 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 speak-
ers whose performances are riddled with what an older, more competent gen-
eration 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 increas-
ingly 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 halt-
ingly. 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 pres-
sure from English. The reasons for the shift are readily grasped. Gaelic has
Nancy Dorian - 978-90-04-26193-8

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 Gaelic-
speaking 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 mini-
mal 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 fish-
ing 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 dia-
lect 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 concomi-
tant 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).
Nancy Dorian - 978-90-04-26193-8

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 discrimina-
tion 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 suffi-
cient 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 characteris-
tic 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 east-
ern Sutherland and a lesser amount in Berks County, I have found three fac-
tors 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
Nancy Dorian - 978-90-04-26193-8

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 trans-
mits 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 dur-
ing 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 semi-
speakers differ, potentially, from the others I shall discuss, in that the conscious or uncon-
scious 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 men-
tioned 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.
Nancy Dorian - 978-90-04-26193-8

Another semi-speaker likewise claimed that she spoke better Gaelic than a sib-
ling 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 sup-
pose 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-genera-
tional ties outside the nuclear family thus play a significant role in maintaining
the local-currency language among individuals whose own immediate house-
holds may have used it relatively little.
Where this feature combines with that of strong community identity fos-
tered 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 intel-
ligible 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
Nancy Dorian - 978-90-04-26193-8

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 personal-
ity. Even her grandmother showed a strong tendency to speak English rather
than Gaelic to her, but eventually she herself demanded Gaelic from her par-
ents 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 house-
hold 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 circum-
stances 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.
Nancy Dorian - 978-90-04-26193-8

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 semi-
speaker 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 con-
trol 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 dia-
lect 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 curi-
ous. Asking questions. And, of course, hearing it.
Like so many other semi-speakers, this young boy was responsive to older peo-
ple 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.
Nancy Dorian - 978-90-04-26193-8

154 chapter 6
A great-grandmother is still alive, in this case, and is the matriarch of a four-

generational 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 language-
contact situation are rather overwhelming, it seems to me that explaining resis-
tance 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.
Nancy Dorian - 978-90-04-26193-8

c ommunity, abandonment of a disfavored, limited-currency, low prestige lan-

guage 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.
Nancy Dorian - 978-90-04-26193-8

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 pro-
ven 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 pro-
bably 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 consti-
tuted 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 fellow-
villagers belong only to one. This is the picture one arrives at by adopting
Gumperz’s definition of the speech community: “any human aggregate char-
acterized 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 recogni-
tion 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).

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 near-
passive 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 devi-
ations 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 (rec-
ognized as “mistakes” by fluent speakers), and by the frequency of such devia-
tions, as well as by the presence of a marked degree of analogical levelling and
a tendency to eliminate syntactic redundancies (Dorian 1977 and 1980). Semi-
speakers differ among themselves in their grammatical and phonological abili-
ties in Gaelic, however, and also in their manner of delivery. Some speak quite
readily, though usually in short bursts; despite their phonological and gram-
matical 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
Nancy Dorian - 978-90-04-26193-8

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 near-
passive 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 com-
munity, their claim to inclusion in an East Sutherland Gaelic speech commu-
nity 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 appro-
priate 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 varia-
tion: “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 cer-
tain important respects they are entirely unlike the English monolinguals who
represent the clear-cut excludable group. The first is their outstanding recep-
tive control of East Sutherland Gaelic, and the second is their knowledge of
the sociolinguistic norms which operate within the Gaelic-speaking commu-
nity. The fisherfolk descendants of East Sutherland number just under 100 at
Nancy Dorian - 978-90-04-26193-8

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 rela-
tives representing a variety of ages interact a good deal. Families also tradition-
ally 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 com-
mon. 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 old-
est 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 eve-
nings 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 profi-
ciency 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 signifi-
cant tidbit of gossip. They occasionally supplied a translation of something dif-
ficult 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.
Nancy Dorian - 978-90-04-26193-8

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 fully-
fluent member of the group could step in, if there was a marked hesitation,
and supply the anticipated conclusion. Low-proficiency members of these net-
works, 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 soci-
olinguistic norms, are provided by Hymes (1974) and Corder (1973). Hymes pro-
poses 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 commu-
nity 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 some-
times 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.
Nancy Dorian - 978-90-04-26193-8

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 well-
integrated 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 con-
versation, 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 neigh-
bour’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 recep-
tive skills and sociolinguistic knowledge. Although it’s clear that she acknowl-
edges 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 pres-
ence 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.
Nancy Dorian - 978-90-04-26193-8

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 pho-
nological control of the speech variety common to the participants. The for-
eign 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, communi-
cative 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 appro-
priately 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 formu-
laic 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 interac-
tion 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.
Nancy Dorian - 978-90-04-26193-8

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 mar-
ginal 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 linguis-
tic structures. A great deal of variation is characteristic of the dialect, morpho-
phonemically, 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 varia-
tion. 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.
Nancy Dorian - 978-90-04-26193-8

164 chapter 7
The chief reason for the absence of social evaluation of linguistic varia-
tion 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 differ-
ences 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 Gaelic-
speaking 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 read-
ily, 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 bilin-
guals, 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 any-
one 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 semi-
speakers of any proficiency, and even from the near-passive bilinguals, who are
Nancy Dorian - 978-90-04-26193-8

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.
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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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 lan-
guages 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 popula-
tions 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 continu-
ing immigration (this is the case with Canadian reinforcement of the French-
speaking 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 num-
ber 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 actu-
ally 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

168 chapter 8
this persuasion usually point to ease of modern travel, the “global village” phe-
nomenon, 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 penal-
izing 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, pedagog-
ically-oriented verbal routines such as Aesopian tales, fairy tales and rhyming
genres (all directed toward children) in Albanian-speaking Greek communi-
ties (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 state-
promoted 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 sur-
prises no one. It’s not the inevitable outcome, however, since people seem to
be capable of quite remarkable segmentation of their lives, including linguis-
tic segmentation. It’s hardly encouraging for a language to be excluded from
the schools, ignored in broadcasting, discouraged in public life, and unpro-
vided for in any officially sponsored activities whatever. But in some societ-
ies 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 hearth-
and-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 minor-
ity-language group and the dominant-language group, as with certain Native
American tribes, the home language may be jealously guarded from mem-
bers 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.
Nancy Dorian - 978-90-04-26193-8

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 ques-
tion 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 tem-
per 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 bilingual-
ism 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 com-
munity 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 sus-
tained 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 accre-
tion 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
Nancy Dorian - 978-90-04-26193-8

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 bilin-
guals 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’ origi-
nal 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-language-
only 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 origi-
nal 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 young-
est 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
Nancy Dorian - 978-90-04-26193-8

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 Gaelic-
speaking 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 elicita-
tion 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 reason-
ably 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 struc-
tures 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
Nancy Dorian - 978-90-04-26193-8

172 chapter 8
Table 1. 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.
Analogically- Analogically- Analogically-levelled 1st pers.

levelled noun plural levelled verb stems sing. conditional verb
(opportunities: 17) (opportunities: 16) (opportunities: 5)
Speaker 1 1a 0 0
Speaker 2 2 0 0
Speaker 3 8 3 5
Analogically- Total of # of forms % of forms

levelled 1st pers. potentially produced that produced
sing. future verb analogically-levelled were analogical: that were
(opportunities: 5) forms produced analogical
Speaker 1 0 43 1 2
Speaker 2 0 43 2 4.5
Speaker 3 5 43 21 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 intrafam-
ily 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 home-
language 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
Nancy Dorian - 978-90-04-26193-8

Table 2. Comparison of control of three tenses and three embedded structures by three
siblings in a single Gaelic-speaking household
Correctly formed Past Future Relative

#/% #/% #/%
Speaker 2 23 of 23 / 100 17 of 18 / 94.5 8 of 8 / 100

Speaker 3 24 of 24a / 100 15 of 18 / 83 3 of 8 / 37.5
Speaker 4 23 of 23 / 100 16 of 18b / 89 6 of 8 / 75
Correctly formed ‘that’ ‘if’ Conditional

#/% #/% #/%
Speaker 2 9 of 9 / 100 5 of 5 / 100 10 of 10 / 100

Speaker 3 8 of 9 / 89 4 of 5 / 80 6 of 11 / 54.5
Speaker 4 1 of 9 / 11 4 of 5 / 80 8 of 10 / 80
# of errors opportunities % incorrect
Speaker 2 1 73 1
Speaker 3 15 75 20
Speaker 4 15 75 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 lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

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 mem-
ber 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 sud-
den 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 monolin-
gual 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 incen-
tive 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
Nancy Dorian - 978-90-04-26193-8

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 dra-
matically 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 loy-
alty 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 bot-
tom 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 contin-
ues, 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 nega-
tive 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 con-
struction 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 them-
selves 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 lan-
guage outside the national linguistic mainstream seemingly turns up its toes
Nancy Dorian - 978-90-04-26193-8

176 chapter 8
so dramatically after persisting with anomalous strength for so long. The fail-
ure 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 sur-
vival potential precisely because they have had a prolonged period in which to
learn to cope with pressures for linguistic assimilation. Metaphorically speak-
ing, 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 mea-
sles virus was long in existence, but slow to reach them: when it did, they suc-
cumbed 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.
———. 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
Nancy Dorian - 978-90-04-26193-8

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.
Tsitsipis, Lukas D. 1983. Narrative performance in a dying language: Evidence from
Albanian in Greece. Word 34:25–36.
Nancy Dorian - 978-90-04-26193-8

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 speak-
ers 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 fea-
tures 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 pat-
terns, 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 child-
hood 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 lan-
guage 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 mem-
bers while others spend more of their working lives among speakers of the
expanding language.

age and speaker skills in receding language communities 179
2 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 sen-
tences, 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 pro-
ficiency, 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 pre-
occupations 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 con-
sidered before the position of individual speakers in the local proficiency spec-
trum 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
Nancy Dorian - 978-90-04-26193-8

180 chapter 9
Amazonia (Aikhenvald 2001). Before she offers characterizations of individual

speakers’ language skills in one of the villages, currently a language-shift set-
ting, she first discusses the community’s own notions of desirably “correct” or
“good” Tariana speech. She examines the strong constraints against borrow-
ing 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 domi-
nant in the area), and that speech considered imperfect or incorrect is openly
ridiculed: those who introduce borrowed elements into their Tariana are sub-
ject 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 old-
est 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 older-
generation model is one of a “critical mass” of speakers who favor particular
forms (2001: 419–20).
Traditionally, in this exogamous region, children derived their linguis-
tic identity by acquiring the language of their fathers. But with transmission
of Tariana falling off and use of Tucano expanding, the children of Tariana-
speaking fathers were no longer routinely becoming proficient Tariana speak-
ers 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 substan-
tial 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
Nancy Dorian - 978-90-04-26193-8

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 pro-
files 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 knowl-
edge 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 sub-
sequent 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
Nancy Dorian - 978-90-04-26193-8

182 chapter 9
fuller set of Tariana evidentials, for which he is faulted, apparently by the com-
munity 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 ungram-
matical” (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 pos-
session 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 cul-
tural 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 ungram-
maticalities” (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 pronuncia-
tion than the others (unreduced vowels and aspirated consonants). Ñu’s older
brothers Saba and Kiri have reputations as “ ‘deficient speakers’ ”, their stories
Nancy Dorian - 978-90-04-26193-8

full of Tucano calques, their speech lacking the traditional person marker
and without the full set of Tariana evidentials: “their lack of language compe-
tence 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).
3.2 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 lan-
guage 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 docu-
ment 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
Nancy Dorian - 978-90-04-26193-8

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 speak-
ers’ 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 spo-
ken 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 con-
structions 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-speak-
ers. 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 White-
Thunder, a man of about 40, who speaks “less English than Menomini”, a seri-
ous 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
Nancy Dorian - 978-90-04-26193-8

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 particu-
larly unclear in Bloomfield’s characterizations, suggesting overall more a nega-
tive 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-Cloud-
Woman, 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 arti-
cle 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 vocabu-
lary, 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 presum-
ably 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.
Nancy Dorian - 978-90-04-26193-8

186 chapter 9
4 Age-Based Proficiency Groupings and Individual Exceptions

in East Sutherland
In ongoing work with the Gaelic dialect spoken by fisherfolk and their descen-
dants 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, reflect-
ing 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 rec-
ognized these imperfect speakers, although they had no special term for them
and even though the group members differed from one another in grammati-
cal intactness and level of fluency. I termed these imperfect speakers semi-
speakers 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, middle-
proficiency, 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.
Nancy Dorian - 978-90-04-26193-8

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 tele-
phone, 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 sis-
ter “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 semi-
speakers 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.
Nancy Dorian - 978-90-04-26193-8

188 chapter 9
usage of some individual fluent speakers embodied conservative linguistic

norms more fully than others.
4.1 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% tradi-
tional (18 of 19), and her feminine pronoun replacements were 93.8% tradi-
tional (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 semi-
speakers but none of the other YFSs he used no traditional negative impera-
tives 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 excep-
tional among younger speakers, and especially among semi-speakers, in pro-
ducing 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 syn-
thetic 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 direc-
tional 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 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.
Nancy Dorian - 978-90-04-26193-8

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 pro-
file. It is also clear that community evaluation of speaker capabilities is not
based on consistent grammatical conservatism. Seònaid demonstrated atypi-
cally low levels of vocative retention, locational vs. directional adverb distinc-
tion, 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 voca-
tive 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-and-
proficiency 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 fam-
ily, 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.
Nancy Dorian - 978-90-04-26193-8

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 con-
servatism of his Gaelic, and he was a relatively shy man, not the conversation-
alist 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 comfort-
ably, 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 rou-
tinely 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 struc-

ture, 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 catego-
ries 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 devi-
ate, sometimes to surprising degrees, from the general profile for their age-
group, 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
Nancy Dorian - 978-90-04-26193-8

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 commu-
nity, 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.
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:
——— (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 set-
ting. 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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

chapter 10
Linguistic Lag as an Ethnic Marker1
It is commonplace in the body of literature on ethnicity to find language iden-

tified 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 impor-
tant 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, lan-
guage is but one of several such markers, others such as physical charac-
teristics, 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 mem-
bership 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 look-
ing 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 possi-

ble to find a people using a language which few of them actually speak as a sym-
bol 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 ear-
lier draft of this paper.

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 essen-
tial) 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 iden-
tity 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 sin-
gled 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 dis-
tinctive 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 segre-
gated. 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 settle-
ment, composed wholly of fisherfolk, to the nonfishing village of Dornoch. All
Nancy Dorian - 978-90-04-26193-8

linguistic lag as an ethnic marker 195
three of these fisherfolk residential areas were known, unofficially and impo-
litely, 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 for-
mer 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 abandon-
ment 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 excep-
tions. 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 syn-
chronically, 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.
Nancy Dorian - 978-90-04-26193-8

196 chapter 10
before World War I, some Gaelic speakers remained even in the villages them-
selves. 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 dia-
chronic 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
Nancy Dorian - 978-90-04-26193-8

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 rel-
ative 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 com-
ment of the Golspie parish minister as typical:
Forty years ago, the Gaelic was the language generally spoken in the par-
ish. 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 communi-
ties, 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 –
Nancy Dorian - 978-90-04-26193-8

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 rela-
tive 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 mono-
lingual, 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 tempo-
rary by-name thanks to an English blunder. One Golspie bilingual, now a sep-
tuagenarian 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 pro-
ficient 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 Gaelic-
English bilingualism which characterizes the fisherfolk when all other seg-
ments of the population are monolingual in English. The pattern seems clear.
For a hundred years or more, the East Sutherland fisherfolk have had a lin-
guistic 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?”
Nancy Dorian - 978-90-04-26193-8

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 bilingual-
ism 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 fish-
erfolk 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 usu-
ally 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 late-
nineteenth-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-differen-
tial, 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 linguis-
tic 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
Nancy Dorian - 978-90-04-26193-8

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 invidi-
ous 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 dur-
ing 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 differ-
ential: a persistent lag in linguistic habits as compared with the linguistic hab-
its 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 lan-
guage 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 rela-
tionship 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 rela-
tions. 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.
Nancy Dorian - 978-90-04-26193-8

———. 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 eth-
nography). 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.
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
part three
Language Shift and Language Maintenance
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
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 exten-
sive 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 lan-
guages 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 com-
plement 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 inter-
est 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 pres-
tige. There are societal attitudes toward specific languages and to their native
speakers. How are these attitudes formed? Are they usually local in origin and

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 manipu-
lable 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 lan-
guages 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 lan-
guage 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 sug-
gest 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 lan-
guages in the world’s linguistic arsenal today, but were at one time the aggres-
sively expanding languages of great political powers. The same was true of
Ottoman Turkish. Aztec (Nahuatl) and Quechua experienced similar expan-
sion 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.
Nancy Dorian - 978-90-04-26193-8

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 Arabic-
and 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 characteris-
tic 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 delib-
erately 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 men-
tioned the general usefulness of Italian as compared with the other lan-
guages (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 dra-
matic 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
Nancy Dorian - 978-90-04-26193-8

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 Hungarian-
speaking segment of the population had its counterpart earlier, when the dis-
trict 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 immi-
grants 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 condu-
cive 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 cer-
emony (to which proper use of the language is essential) as the basis of
their identity. (Hymes 1966: 126)
Nancy Dorian - 978-90-04-26193-8

The other group, the Guayqueries, who live on islands off the coast of
Venezuela, apparently abandoned their language and their pre-Christian reli-
gion 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 indistinguish-
able from other Venezuelans. Hymes comments: “Language, together with reli-
gion, 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 prag-
matic 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 grand-
parent, for example). Characteristic of the Eastern Porno, also, is “a matter-of-
fact 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 prag-
matic 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 integ-
rity 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 dissimi-
lar fates of Albanian· language enclaves in Greece and in Italy. In Italy, where
a tradition of cultivating and valuing localism (including local dialects) pre-
vails, 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 lan-
guage. 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.
Nancy Dorian - 978-90-04-26193-8

210 chapter 11
So very many languages are dying, during our time, in so very many locali-
ties, 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 resis-
tant 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 national-
ism 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 demo-
graphic 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
Nancy Dorian - 978-90-04-26193-8

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 nation-
ality in question who happens to be on the campus. This kind of transforma-
tion 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 atti-
tudes 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 subse-
quent 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
Nancy Dorian - 978-90-04-26193-8

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 anec-
dotal 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 journal-
ism 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 compla-
cency 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 reten-
tion just as it does acquisition (again, see Gardner, this volume). Where reten-
tion 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 aris-
ing 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 train-
ing. 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.
Nancy Dorian - 978-90-04-26193-8

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 his-
tories 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, how-
ever: 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 monolin-
gual 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 oppor-
tunity 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 per-
fect (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 devi-
ant. 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 unusu-
ally 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 c onservative
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 any-
thing 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.
Nancy Dorian - 978-90-04-26193-8

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 appar-
ent 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 mor-
pheme entirely; (4) in the negative imperative, she used none of the most con-
servative 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 plu-
ral 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-
Nancy Dorian - 978-90-04-26193-8

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 dif-
ferences 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, willing-
ness 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 liv-
ing 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.
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage 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 knowl-
edge 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 dif-
ficult 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 self-
evaluation 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 identi-
cal 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.
Nancy Dorian - 978-90-04-26193-8

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 pro-
duction 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% suc-
cess rate). In tests for three subordination structures (relativization of subject-
and 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 conserva-
tive 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 alien-
ables 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
Nancy Dorian - 978-90-04-26193-8

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, inter-
estingly enough, regardless of the level of proficiency they display), I find that
not only the stronger imperfect speakers, but also the weaker imperfect speak-
ers, 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 addi-
tion 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 com-
munity (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
Nancy Dorian - 978-90-04-26193-8

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 acqui-
sitional 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 one-
person 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 admix-
ture 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 weak-
est 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,
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage 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 forget-
ting 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 struc-
ture 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.
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage 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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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 theo-
retically 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 empha-
sizing 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 sec-
tions 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 read-
ing of the prepublication form of this paper and for several corrections and suggestions.

224 chapter 12
Marked political and/or economic change is the scenario most often sug-
gested 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 disengage-
ment from a general British polity in which Scottish and particularly Highland
interests are given disgracefully little attention, create a similarly positive cli-
mate 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 lan-
guage. 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 unconge-
nial 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 intro-
duction of outside norms might be welcomed, or at least tolerated, as pref-
erable 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
Nancy Dorian - 978-90-04-26193-8

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 remind-
ing local speakers of just how deviant their own everyday speech is. Far from
contributing to pride in the native ethnicity and traditions, linguistic or oth-
erwise, 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 sen-

tence 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 sym-
pathy 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
Nancy Dorian - 978-90-04-26193-8

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 exam-
ple), it can be hard to appreciate the aversion among self-conscious and under-
confident 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 lan-
guage is low. Perhaps the awareness that dominant-language speakers typically
already have a negative attitude to the minority language makes the minority-
language 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, like-
wise 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 etymolo-
gies involved, is the contempt sometimes expressed by Scottish Gaelic speak-
ers 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 put-
ting 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 tran-
sit historia verbi.
Nancy Dorian - 978-90-04-26193-8

Still the question remains whether large-scale changes in the political cli-
mate (with attendant changes in attitude toward indigenous minority lan-
guages), 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 pro-
vided by such a benefactor, the wherewithal for revival would not be avail-
able. 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 teach-
ers 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 pos-
sibly 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 dedi-
cated 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 rec-
ognized 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
Nancy Dorian - 978-90-04-26193-8

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 particu-
lar 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
Nancy Dorian - 978-90-04-26193-8

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), indus-
trialization 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 world-
language status, namely English, and the only near neighbour states are offi-
cially 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 atti-
tudes internalized by the speakers or potential speakers themselves – is in
itself a serious reason for attempting to promote the language. A middle-aged
Nancy Dorian - 978-90-04-26193-8

230 chapter 12
native speaker of Berks County Pennsylvania Dutch (a German dialect with

extremely negative loading in that part of Pennsylvania, where dialect speak-
ers 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 com-
munity 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 trans-
mission 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 func-
tioned, 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 bilin-
gual education undertakings.
One of the most important economic effects of a bilingual education pro-

gram 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
Nancy Dorian - 978-90-04-26193-8

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 exper-
iment 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 ter-
ritory 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
Nancy Dorian - 978-90-04-26193-8

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 “con-
sumable” 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 school-
children, 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, cul-
tural 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 genera-
tions 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 main-
tenance efforts nonetheless serve some useful purpose, have some value, one
can still find reasons for answering in the affirmative.
Nancy Dorian - 978-90-04-26193-8

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.
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.
Nancy Dorian - 978-90-04-26193-8

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 com-
plete one.1 If we used language only for communication, we would not main-
tain 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 culture-
carrying 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 acquir-
ing 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 mother-
tongue 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.

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 fol-
low. The breakup of the former Soviet Union created new conditions for lan-
guages 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 mis-
sionaries, 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 loca-

tions 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 spec-
trum: in first-world France, for example, where members of a minority popu-
lation 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 vari-
ety 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-
Nancy Dorian - 978-90-04-26193-8

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 every-
day 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 testify-
ing 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 ances-
tral 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 pho-
nology 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 docu-
mentation readily available, and even if the language still has native speakers,
the potential obstacles to successful language-support efforts can be formida-
ble. 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.
Nancy Dorian - 978-90-04-26193-8

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 replac-
ing clumsy or competing writing systems, and putting forward one particu-
lar speech form (or an engineered compromise form) for adoption without
creating fatal dissension or perceptions of inauthenticity is an equally daunt-
ing 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 lan-

guages, 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 suc-
ceeded 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
Nancy Dorian - 978-90-04-26193-8

238 chapter 13
of language-support movements, and in spite of improved language-rights leg-

islation 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 rel-
atively 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 revi-

talization and restoration is surely Irish, a case which some see as evidence
that language-support movements can achieve little and others see as evi-
dence 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 speak-
ers 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 long-
suppressed 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
Nancy Dorian - 978-90-04-26193-8

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 sug-
gests 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 dispas-
sionate 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 lan-
guage, 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
Nancy Dorian - 978-90-04-26193-8

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 work-
ing 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 minor-
ity, perhaps as many as one in three people in Ireland have some understand-
ing 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 con-
nection 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 dis-
cusses (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 identi-
ties. 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 dam-
aging, 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 lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

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 transmis-
sion 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 writ-
ing 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 vil-
lages and in a neighboring county a relatively short drive away. But the instruc-
tors 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 back-
wardness, 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 seri-
ous about learning Gaelic or showed a strong interest in its history and tradi-
tion; 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
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage; 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 speak-
ers. 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 lan-
guage. 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 pop-
ulation (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 cen-
sus, since the deaths of elderly speakers still far outstrip gains in new speak-
ers 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 offi-
cial 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 pro-

mote ancestral languages will disappear. If anything, they are likely to increase,
Nancy Dorian - 978-90-04-26193-8

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 some-
times spoken of as “sleeping,” that is, viewed as a central part of a cultural leg-
acy 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 sen-
tence 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 poli-
cies that might have contrary effects. In the Irish case, where this has become
particularly apparent, Ó Riagáin points to the counterproductive effect on lan-
guage revitalization of certain policies undertaken in the economic sphere,
in regional planning, and in education. These other policies were rarely con-
sidered 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 neg-
ative 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 language-
support policy itself are counterproductive: in adopting an “official” Irish that
is no one’s actual spoken language for the purposes of country-wide educa-
tion in Irish, schooling in Irish actually weakens the use of authentic local dia-
lects 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 dif-
fer on the threshold of endangerment (in terms of some critical mass of speak-
ers 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
Nancy Dorian - 978-90-04-26193-8

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 suc-
ceed better than others, on whether outside expertise can or should contrib-
ute 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-to-
child 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 transmis-
sion and language continuity are concerned (1991, 368–373) and to the need for
intimate, small-scale network processes “too gratifying and rewarding to sur-
render” 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 (residen-
tial neighborhoods or religious congregations, the latter sometimes also resi-
dentially 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 revitaliza-
tion 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 move-
ments successes, it would also be unwise to declare them all unequivocal fail-
ures. 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 main-
tenance 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.
Nancy Dorian - 978-90-04-26193-8

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.
——— 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, eco-
nomic, 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,
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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 conser-
vatism about language which is consistent with a similar schoolteacher conser-
vatism 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, standard-
izers 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, how-
ever, the problem is not recent norms, but ancient ones. Classical Arabic, codi-
fied 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 rel-
evance 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)

248 chapter 14
In all of these contexts, purism can be seen to represent a form of conserva-

tism, 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 neigh-
boring 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 com-
munities 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 revi-
talization, 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 indi-
viduals 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
Nancy Dorian - 978-90-04-26193-8

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 revi-
talization than in revival, since bringing about alterations in what people are
already saying could be expected to produce more resistance than prescrib-
ing 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 impera-
tive 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).
Nancy Dorian - 978-90-04-26193-8

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 situa-
tion 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.
Nancy Dorian - 978-90-04-26193-8

The older people within the Tiwi community of Bathurst wanted the tra-
ditional 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 sec-
ond 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 actu-
ally 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 prepar-
ing 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 com-
munity, 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 allow-
ing 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:
Nancy Dorian - 978-90-04-26193-8

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 impos-
sible situation as many of the parents of the children, young adults them-
selves, 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: puris-
tic 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 gen-
erations, 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 differ-
ent, 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 docu-
ments 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 revital-
ization 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 con-
traction 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 spo-
ken language was concerned, what remained was rustic in character, surviving
Nancy Dorian - 978-90-04-26193-8

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 stan-
dardizing 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, unnatu-
ral 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 conserva-
tive 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 complexi-
ties 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 com-
petent 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 reduc-
ing the complex system of grammatical rules involving the use of inflec-
tions and the mutations of initial consonants of Irish to express different
shades of meaning. Many of the rules and forms advocated by the stan-
dard 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 speak-
ers, for their part, had found it possible over the centuries to dispense with
Nancy Dorian - 978-90-04-26193-8

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 pur-
ism 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 posi-
tive, 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 popula-
tion 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 self-
consciousness 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 eco-
nomic utility, many people question the instrumental value of the lan-
guage. 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 lexi-
con 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 remain-
ing 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).
Nancy Dorian - 978-90-04-26193-8

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 fel-
low 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 pain-
fully 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 coin-
age 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, seman-
tic 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 effec-
tive, 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 conserva-
tism has usually been profound, and novel usages have been the object of deri-
sion and rejection. These reactions are hardly limited to Gaelic-speaking Scots,
of course; they are responses already familiar from other language-planning
Nancy Dorian - 978-90-04-26193-8

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 atti-
tudes in instances of revitalization. By comparison, revival settings look invit-
ingly 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 dispa-
rate 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 dispa-
rate 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, rec-
ommended 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.
Nancy Dorian - 978-90-04-26193-8

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 cen-
turies. Both factions claim authenticity; the George faction claims the authen-
ticity 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 loyal-
ties 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 lan-
guage 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 produc-
tive 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 lexi-
cal 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 dif-
ferent 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 precari-
ously placed language has survived longer than might have been expected, an
Nancy Dorian - 978-90-04-26193-8

258 chapter 14
absence of puristic attitudes may have characterized some speakers. Hamp

(1989: 198–99) found, for instance, that phonological intactness was no mea-
sure of survival potential among the pockets of Albanian, known as Arvanitika,
scattered through Greece. On the contrary, the youngest speakers with ser-
viceable (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 phonologi-
cal 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 non-
sectarians (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 com-
munity, “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 con-
cludes, regarding the sectarians’ flourishing PG, that “sociolinguistic norms
prescribe its use but not its form” (225). The greater conservatism of non-
sectarian 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 repre-
sent 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 speak-
ers 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 environ-
ment which encouraged their children to speak Gaelic, when most young people of compa-
rable ages were not actively acquiring the language.
Nancy Dorian - 978-90-04-26193-8

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 con-
veyed 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 addi-
tion, English syntactic influence is evident, and English lexical items are plen-
tifully 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 rhetori-
cal norms of the academicians.” But at the same time (191) she contends that
Haugen was not endorsing “allowing the language to landslide into the sys-
tem of the D[ominant] L[anguage],” and she urges “a firm grasp on the real-
ity 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 sur-
vives, as opposed to disappearing, regardless of its form. Others consider a
highly convergent outcome too poor a representative of the original language
Nancy Dorian - 978-90-04-26193-8

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 capa-
ble of some degree of grammatical monitoring and avoidance of English loan-
words 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 sum-
marizing comment (1987: 88) leans to the conservative side: “Although com-
municative 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 pho-
nological 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 tooth-
gritting 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 prolif-
eration 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 neighbor-
ing language of wider currency. Ó Baoill, in a more pragmatic tradition, is ask-
ing 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
Nancy Dorian - 978-90-04-26193-8

possible worlds, one would not need to choose, of course, and Fishman’s char-
acterization 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 lan-
guage 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 isola-
tion 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 prob-
lem 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 find-
ings 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 compro-
mise 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 centu-
ries 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.
Nancy Dorian - 978-90-04-26193-8

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 pre-
sented 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 lan-
guage 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:
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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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 low-
prestige 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 lan-
guage will be lost. The power of the social forces involved is evidently consid-
erable, since under better circumstances attachment to an ancestral mother
tongue is usually strong. The phenomenon of ancestral-language abandon-
ment is worth looking at, then, precisely because a good many people, espe-
cially 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 with-
out ordinary speakers, as sometimes happens in connection with reli-
gious 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.

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 territo-
ries 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 pat-
tern of their Viking forebears in Normandy. In such cases the conquering group
is usually numerically thin, compared with the size of the conquered popula-
tion, 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 lan-
guage of those who control obvious power positions (as administrators, gov-
ernors, 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 mem-
bers 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,
Nancy Dorian - 978-90-04-26193-8

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 ances-
tral language abandonment seem to be very strong in the widely distrib-
uted 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 vari-
ous milletler (“nations”) within its domains. Even European states were mod-
erately permissive of ethnic languages until relatively recent times. The rise
of nationalism in Western Europe at the beginning of the industrial age coin-
cides to a considerable extent with less tolerant attitudes towards subordinate
languages. In the present day, for example, France has shown unusual intol-
erance of ethnic distinctiveness, even for a Western European country (refus-
ing 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 perspec-
tive 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, main-
taining the peace, and not much else, and . . . [with] no interest in promoting
lateral communication between its subject communities.” In industrial soci-
eties, by contrast, conditions are quite different Industrial means of produc-
tion require universal literacy and numerical skills such that individuals can
communicate immediately and effectively with people previously unknown
Nancy Dorian - 978-90-04-26193-8

to them. Forms of communication must therefore be standardized and able

to operate free of local or personal context. This in turn places great impor-
tance on educational institutions, which must produce individuals with cer-
tain generic capacities that permit slotting and re-slotting into a variety of
economic roles. The state is the only organizational level at which an edu-
cational 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 pol-
ity 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 pol-
icies. 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 high-
prestige people and a favored speech form, despite the Revolution’s abrupt ter-
mination of the French monarchy. So, too, was the disfavoring of speech vari-
eties 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
Nancy Dorian - 978-90-04-26193-8

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 fea-
ture of the system of linguistic stratification in Europe is an ideology of con-
tempt: 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 sus-
ceptible to what he calls “prestige transfer”:
Because the intrinsic worth of dialects and of their component ele-

ments 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 presti-
gious-by-transfer attributes include things which others in the commu-
nity 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 non-
prestigious 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 lin-
guistic 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 “unintel-
ligible”, and Totanac “rustic” (Heath 1972: 3).
Nancy Dorian - 978-90-04-26193-8

form, all others being relegated to inferior status. The standard language is

typically considered a rich, precise, rationally organized and rationally orga-
nizing 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, what-
ever 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” pre-
vails and wealth is measured by the size of the cattle herd, the languages of
cattle-herding pastoralists have frequently displaced the languages of hunter-
gatherers who own no cattle (Dimmendaal 1989: 16–24).
Europeans who came from polities with a history of standardizing and pro-
moting just one high-prestige speech form carried their “ideology of contempt”
for subordinate languages with them when they conquered far-flung territo-
ries, to the serious detriment of indigenous languages. And in addition to a lan-
guage 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 standard-
language 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 fea-
tures 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 express-
ing point of view and social distance.
Nancy Dorian - 978-90-04-26193-8

270 chapter 15
of national-language standardization in their homelands, Europeans espoused

other ideologies that exacerbated their contempt for whatever unstandard-
ized vernaculars they encountered.3 They seriously confounded technologi-
cal 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 unfortu-
nately, 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 unfa-
vorable impact on the survival of indigenous languages in the very consider-
able 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 back-
ground 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 wide-
spread among English, French, and Spanish speakers than among Czechs or
Icelanders, but even among speakers of the smaller standardized and state-
promoted languages of Europe there often lurks a notion that the general
Indo-European type of language is exceptionally well suited to clear think-
ing 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 per-
ceived 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 anthropolo-
gists discuss the language endangerment crisis with non-specialists, it’s nearly always neces-
sary 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-
Nancy Dorian - 978-90-04-26193-8

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 evi-
dential 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 peo-
ple 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 pros-
perous European country.6 Quechua and Nahuatl, the proud languages of once-
thriving 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 writ-
ing 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 lan-
guages 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, second-
person plural, and past participle in many French verbs).
6 A given language has a distinct market value that can be calculated by various objective mea-
sures. See the useful chapter on “The value of a language: factors of an economic profile of
languages” in Coulmas 1992.
Nancy Dorian - 978-90-04-26193-8

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 wide-
spread, 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 set-
ting 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 possi-
bility 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 lat-
ter can reasonably be considered a normal manifestation of bilingualism, on
both the individual and the societal level, but it does not by any means con-
stitute 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 popu-
lations 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 lan-
guage abandonment and shift to a dominant-group language. Several possible
sources of such self-regard can be identified, at least tentatively. Rising prosper-
ity, as an indicator of increasing economic success, can be an effective coun-
terpoise 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 pro-
7 The history of the Osage tribe after sudden oil wealth illustrates some of the problems associ-
ated with abrupt and unexpected wealth.
Nancy Dorian - 978-90-04-26193-8

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 pro-
gram 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 indus-
try, geared in good part to luxury-level skiing, has been the most conspicu-
ous change in local conditions. The seasonal nature of the tourist industry and
the fact that the tourists represent no single dominant language, plus the iso-
lation 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 self-
confidence 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 estab-
lished 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 revi-
talization 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 secu-
rity, of course; some of the distinctly precarious languages of Central India have over a mil-
lion speakers, e.g. Kurux (Abbi 1995), and so did beleaguered Breton, in France, as recently as
1926 (Timm 1973: 289, citing Meillet 1928: 380).
Nancy Dorian - 978-90-04-26193-8

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 keep-
ers, 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 small-
language 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 accompa-
nying 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 preced-
ing quarter-century. Since that year Greenlandic has been supported and pro-
moted to an increasing degree, and bilingual Greenlanders have increasingly
replaced monolingual Danes in the top institutional and organizational posi-
tions (Langgaard 1992).
It would be difficult, or more likely impossible, to identify a precise cause-
and-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
Nancy Dorian - 978-90-04-26193-8

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 disap-
pearance. 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 ide-
ologies. 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 appro-

priate 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 every-
day speech by members of the Arizona Tewa speech community does
not reflect attitudes about these other languages but rather the function-
ing 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 lin-
guistic levels in their everyday speech.
The Arizona Tewa have maintained their ancestral language for 300 years,
despite enclavement within a Hopi environment, despite considerable inter-
marriage 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 accomplish-

ment 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
Nancy Dorian - 978-90-04-26193-8

276 chapter 15
The centrality of religious ceremony has been a factor in some other cases of

unexpectedly sustained language maintenance in seemingly adverse circum-
stanees, 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 pre-
dictable. 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 wage-
earning 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 prac-
tices 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 pro-
cess 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 mod-

ern techniques, and traditional food preparation is relatively slow and labori-
ous. 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.
Nancy Dorian - 978-90-04-26193-8

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 com-
munities. Except in cases of great geographical or social isolation, the long-
term maintenance of a small language implies not just the persistence of one
language but the enduring coexistence of two or more. Currently we under-
stand 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 (illumi-
nated 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 cen-
tury, 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 national-
ist 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 genera-
tion of children” (Saville-Troike 1989: 215).
Nancy Dorian - 978-90-04-26193-8

278 chapter 15
Northern Ireland, came forward with the proposal that what they term “mul-
tiple 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), indi-
viduals 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 con-
cept that he recommends it as a potential solution in other ethnically and lin-
guistically 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 sta-
tuses 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 exclu-
sive 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 mod-
ern Europe and extended to the many once-colonial territories of European
states, that has in modern times posed the keenest threat to both the identi-
ties 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 300-
year coexistence of the Arizona Tewa with the First Mesa Hopi indicates that
both ethnic and linguistic persistence are feasible over long time spans with-
out fatal ethnic hostility on either side. Thailand, with a stable hierarchical
social structure until very recently, is said to have had minimal ethnolinguistic
11 “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).
12 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).
Nancy Dorian - 978-90-04-26193-8

conflict over the six and a half centuries of its monarchy despite the pres-
ence 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 connec-
tion 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 national-
language principle and have accepted the enormous costs of mounting inter-
pretation services and document-translation services for each of the individ-
ual national languages. This unyielding adherence to the national-language
ideology has given rise to unprecedented European support for multilingual-
ism, 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 exten-
sion 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 lan-
guage 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 spiri-
tual 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 inti-
mately 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
Nancy Dorian - 978-90-04-26193-8

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 self-
confidence is hard to come by in communities which have suffered the penal-
ties 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 iden-
tifies itself by the ancestral-language label no longer speaks that language. In
such cases, defining identity in terms of language would define out of mem-
bership 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 nec-
essary 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 lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

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 extraor-
dinarily difficult for even the most sympathetic outsiders to provide useful sup-
port 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 com-
munities. 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 lan-
guages. Having waited too long before undertaking to rally support for threat-
ened 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 mater-
nelles 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.
Nancy Dorian - 978-90-04-26193-8

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 indig-
enous 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 persis-
tence. 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, eco-
nomic, 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.
Nancy Dorian - 978-90-04-26193-8

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 atti-
tudes in autonomous Catalonia. Language in Society 19: 311–30.
Nancy Dorian - 978-90-04-26193-8

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 lan-
guages 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 offi-
cial standing of any kind. The consequences, for speakers of languages with-
out 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 repre-
sent 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 present-
day 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 reorganiza-
tion. 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 minority-
group status achieve more autonomy or even full independence, with a corre-
sponding rise in the status of their language, as has happened in recent times

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 some-
thing 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 geo-
graphic and social isolation can only grow rarer in coming decades. The out-
look 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 popu-
lation 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 mem-

bers of the dominant linguistic group in modern nation-states, was in all likeli-
hood 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 alli-
ance 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 mem-
bers came into more frequent contact with populations speaking other lan-
guages, the number of bilinguals and multilinguals would increase. Population
size typically had an effect on which group had the greater number of border-
area 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
Nancy Dorian - 978-90-04-26193-8

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 demogra-
phy, 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 resource-
rich 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 prov-
ince 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.
Nancy Dorian - 978-90-04-26193-8

Disparities in factors such as population size, access to resources, and con-

trol of trade are fully as important in nation-state settings as in non-state set-
tings, since advantages in these spheres are often important to establishing a
viable state in the first place. A uniquely significant disparity in the nation-
state 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 set-
tings either find by experience or come to believe that the official state language
confers major advantages in terms of access to schooling, employment, politi-
cal 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 disappoint-
ment 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 increas-
ingly 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 majority-
language 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 domi-
nant, 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 wider-
currency state-promoted language occurs with considerable frequency, it is
sometimes asserted that bilingualism is essentially a practical matter, governed
Nancy Dorian - 978-90-04-26193-8

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 clos-
ing 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 subcommuni-
ties who neither spoke nor understood French came to be viewed as unaccept-
able. The unity of the new revolutionary state was henceforth to be expressed
via a common language, replacing the linguistic heterogeneity that in the revo-
lutionary view had served the purposes of a discredited monarchy by prevent-
ing 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 peo-
ple 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).
Nancy Dorian - 978-90-04-26193-8

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 popu-
lar consensus on the desirability of homogeneity, with the ideal model of soci-
ety 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 charac-
terized by linguistic diversity and multilingualism, while modern societies
move steadily (“progress”, in the terms of this viewpoint) toward a single offi-
cial 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 mono-
lingualism 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 main-
stream 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.
Nancy Dorian - 978-90-04-26193-8

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 group-
wide bilingualism, is unusual only under particular conditions, namely when
the modern nation-state is the setting and when both socioeconomic incen-
tives 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 con-
tact 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 minority-
group language to an official majority language, however, it should not be con-
sidered 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 diffi-
cult 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 strik-
ing contrast between the expectations of the Western investigator and the
linguistic situation encountered and described. This is the case, for example,
Nancy Dorian - 978-90-04-26193-8

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 impor-
tant 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 lan-
guage: 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 disap-
pearing. 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-Kordofanian-
speaking 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.
Nancy Dorian - 978-90-04-26193-8

292 chapter 16
The matter-of-factly maintained multilingualism of these two small peo-

ples 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 negotiat-
ing 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 fea-
sible 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 con-
testing 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 con-
cerned 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
Nancy Dorian - 978-90-04-26193-8

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 impres-
sionistically 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 lan-
guage (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 com-

mand of three or four languages. But another account of routine multilingual-
Nancy Dorian - 978-90-04-26193-8

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 abil-
ity 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 can-
not 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 judg-
ments 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 lan-
guage? 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
Nancy Dorian - 978-90-04-26193-8

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 field-
work 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 speak-
ers. Bradley does not say whether he did or would accept such speakers as full-
fledged 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 con-
fer (with exogamous marriages and manipulable identities included among
such practical benefits), additional benefits that qualify at least in part as ideo-
logical are recognized among peoples who have traditionally cultivated knowl-
edge 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 anthropolo-
gist who worked with Emenyo in the 1950s and ’60s summed up Emenyo atti-
tudes as follows (Salisbury 1972 [1962]: 56):
The fact that there are more Emenyo bilingual in Dene, than Dene-
speakers 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 accom-
plishment and their command of Dene makes them, if anything, superior
to the Dene.
Nancy Dorian - 978-90-04-26193-8

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 knowl-
edge 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 acquisi-
tion of other languages indicates that they held views similar to those of the
Emenyo. The small Arizona Tewa population “enjoyed a reputation for com-
manding 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 valu-
able (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 centu-
ries 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).
Nancy Dorian - 978-90-04-26193-8

Where the exogamous multilingual Indians of the Vaupés region are con-
cerned, one observer (Sorensen 1972) described them as having an “instru-
mental 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 fea-
tures 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 lan-
guage 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 dis-
rupted, 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 subsis-
tence 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
Nancy Dorian - 978-90-04-26193-8

298 chapter 16
relocations of minority peoples (already widely perpetrated in earlier centu-

ries) became easier than ever to carry out. Under all these pressures the ances-
tral 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 mar-
riage practices, Aikhenvald worked with the last generation to be fluent in
Tariana, an Arawakan language spoken where most of the many other lan-
guages 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 consid-
ered the traditional multilingualism of the region a pagan practice, accord-
ing 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 hous-
ing 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, espe-
cially 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 edu-
cation 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 atti-
tude and policy most likely comes too late.
Nancy Dorian - 978-90-04-26193-8

by older kinsfolk and other community members was curtailed by this edu-
cational 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 set-
tings are extremely limited, he considered it possible that an unrewarding ven-
ture 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 lan-
guage 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 lan-
guage 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 ado-
lescents 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 peo-
ple 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
Nancy Dorian - 978-90-04-26193-8

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 com-
parison 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 bilin-
gualism 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 fundamen-
tal question that is frequently left unasked is this: why, in view of both the con-
temporary 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 unre-
markable, 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
Nancy Dorian - 978-90-04-26193-8

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 con-
sidered. 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 offi-
cial language. Hundreds of years of enclavement among a numerically supe-
rior 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 monolingual-
ism in many parts of the contemporary world seems likely to rest in part on
continuing dissemination, despite he decline of colonialism, of European lan-
guage 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 techno-
logical 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 self-
serving view justifies the established ascendancy of a few “superior” European
languages in a post-colonial world still linguistically shaped by European colo-
nialism, 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
Nancy Dorian - 978-90-04-26193-8

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 popula-
tions 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 objec-
tive 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, con-
centration of the minority population, birthrate, etc.), and institutional sup-
port and control factors (use in mass media, education, government services,
industry, religion, culture, politics). If the “surprisingly” vigorous Nsur lan-
guage 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 main-
tenance or shift in a non-Western setting like Tapshin village is the develop-
ment 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.
Nancy Dorian - 978-90-04-26193-8

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 counter-
currents 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 assimi-
lation. 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 linguis-
tic human rights in some parts of the world and increased communication
(especially electronic communication) among widely separated small peoples.
Nancy Dorian - 978-90-04-26193-8

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), long-
term 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 minority-
group 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 them-
selves 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 trans-
mission to produce a numerically significant number of fluent new minority-
language 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 pro-
duce 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 sec-
ondary immersion schooling.
Nancy Dorian - 978-90-04-26193-8

There are nonetheless limitations to the effectiveness of immersion school-

ing. 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 lan-
guage 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 acceptabil-
ity 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 major-
ity 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 endanger-

ment. In David Bradley & Maya Bradley (eds.), Language maintenance for endan-
gered 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.
Nancy Dorian - 978-90-04-26193-8

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 nation-
alist ideologies. In Bambi B. Schieffelin, Kathryn A. Woolard, & Paul V. Kroskrity
(eds.), Language ideologies: Practice and theory (pp. 189–210). New York and Oxford:
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.
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:
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.
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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
part four
Language Use
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
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 rep-
resents 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 inev-
itably 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 eventu-
ally 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 set-
tings among the older generation, and they are used for casual social inter-
course 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 dwin-
dling 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.

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 func-
tions. By the early 1960s, when I first arrived in East Sutherland to begin study-
ing 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 some-
times 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 met-
rical psalms.
Nancy Dorian - 978-90-04-26193-8

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 for-
mer 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 eve-
ning functions at which admission was charged, with performances of vocal
and instrumental music, Highland dance, and sometimes also amusing mono-
logues; 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 rela-
tively 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 differ-
ent from the local East Sutherland variety to serve as a model for local imita-
tion. ESG is radically deviant, from the point of view of the standard language,
Nancy Dorian - 978-90-04-26193-8

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 lan-
guages 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 stan-
dard 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 lan-
guage, 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
Nancy Dorian - 978-90-04-26193-8

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 mar-
ried 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 audibil-
ity 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 pub-
lications. 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
Nancy Dorian - 978-90-04-26193-8

316 chapter 17
During three recording sessions, in different years, MMK had an opportu-

nity 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 ear-
lier 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 premoni-
tions 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 ses-
sion, 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 par-
ticular 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, phonologi-
cally contrastive words and phrases, and two sections of a letter-tape, but these three ses-
sions 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.
Nancy Dorian - 978-90-04-26193-8

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 infer-
ences 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 neighbor-
ing 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 orthogra-
phy. For example, the expression bha aid ‘they were’ has a single long vowel with no rearticu-
lation 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.
Nancy Dorian - 978-90-04-26193-8

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 car-
ries over from the first, though it is perfectly cheerful and has a happy end-
ing (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 investi-
gated MMK’s eight narratives for this study was the handling of obvious loan-
words 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 pres-
ence 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 spo-
ken. 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 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.
11 See the discussion in Dorian (1981: 100–2).
Nancy Dorian - 978-90-04-26193-8

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 rel-
ative 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 con-
sonant 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 percent-
age 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 grammati-
cal 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 exam-
ples: /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 con-
sonants 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.
Nancy Dorian - 978-90-04-26193-8

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 No Mutation in English Phrasal Reversed

Percentage of All Mutational Plurals English Phrasal
Words Environment Switches Switches
1964 1 3.4 0 of 4 0 of 3 0 –
2 4.7 0 of 2 – 1 1
1968 3 5.3 5 of 13 0 of 2 1 1
4 2.8 – – 0 –
5 3.2 0 of 8 – 1 0
6 7.4 0 of 6 1 of 3 2 0
7 3.8 2 of 10 1 of 7 4 0
8 6.5 0 of 8 – 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 sec-
ond 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 nar-
ratives 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 pur-
poses of Table 17.1; since the noun has no singular, there is no independent base to be
Nancy Dorian - 978-90-04-26193-8

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 week-
end?” 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 her-
self 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 phono-
logical 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 nar-
rative 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 indica-
tors of liveliness I tallied seven features: the number of simple direct quotes
or direct-quote interchanges; the number of those which were in fact inter-
changes 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’
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.
14 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.
Nancy Dorian - 978-90-04-26193-8

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 1 1 0 – 0
2 1 0 – 0
1968 3 6 3 – 11
4 1 0 – 1
5 2 2 1 4
6 1 0 – 0
7 1 1 1 7
8 2 0 – 0

Doubled-up Strong ‘Adventurous’
Quotatives Interjections Language uses
# # #
1964 1 0 0 0
2 0 0 0
1968 3 1 1 0
4 0 0 3
5 1 0 0
6 0 2 1
7 0 1 2
8 0 3 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.]
Nancy Dorian - 978-90-04-26193-8

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 subse-
quent 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 tap-
ing 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 inter-
changes, 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 quota-
tive verbs likewise characterizes this narrative. She uses no ‘adventurous’ lan-
guage, 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
Nancy Dorian - 978-90-04-26193-8

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 set-
ting 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 pro-
duced 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 singu-
16 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.
Nancy Dorian - 978-90-04-26193-8

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 near-
tragedy 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 narra-
tive, 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.”
Nancy Dorian - 978-90-04-26193-8

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 stylis-
tic 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 monostyl-

ism. That is, recessive languages are more and more used in casual styles
only, for example, those which are appropriate for intimate routine inter-
actions 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 “rou-
tine 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 differ-
ent stylistic features from a hilarious story which is told for its entertainment
value above all. No doubt her Gaelic is “inadequate for certain speech situa-
tions, domains, and functions.” She probably could not, in Gaelic, pray aloud,
welcome a guest speaker to the Women’s Rural Institute, or repeat the multi-
plication 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
Nancy Dorian - 978-90-04-26193-8

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 adult-
hood 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 inti-

mate 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 gen-
eration 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
Nancy Dorian - 978-90-04-26193-8

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 help-
ful 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.
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.
Nancy Dorian - 978-90-04-26193-8

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 pop-
ulation 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 con-
tinued 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 second-
ascending 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 Nancy Dorian - 978-90-04-26193-8

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 tape-
recorded 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 single-
encounter 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.
Nancy Dorian - 978-90-04-26193-8

telling the monolinguals from the bilinguals 331
the two villages in the number of available bilingual interlocutors. By mid-

twentieth 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 esti-
mates given by courtesy of the Scottish Registry.
Nancy Dorian - 978-90-04-26193-8

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 vil-
lage, 1 represents the oldest speaker among those with whom I worked.
Nancy Dorian - 978-90-04-26193-8

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 ní̃çεn agəs xa mĩš ən o:rd.

all – the.pl all nothing and went I-EMPH up
everything. And I went 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 interlocu-
lors 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 conver-
sational 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.
Nancy Dorian - 978-90-04-26193-8

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 occu-
pants 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)
Nancy Dorian - 978-90-04-26193-8

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 tele-
gram 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 discus-
sion 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
Nancy Dorian - 978-90-04-26193-8

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 Scots-
speaking 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.
Nancy Dorian - 978-90-04-26193-8

An Embo narrator makes a similarly explicit statement about the monolin-

gual 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
Nancy Dorian - 978-90-04-26193-8

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 ní 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 chil-
dren 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.
Nancy Dorian - 978-90-04-26193-8

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 con-
text. There is no colloquial equivalent locally for the English word damn, and
female bilinguals are in any case more inclined to reproduce English profan-
ity 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 dia-
log 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
Nancy Dorian - 978-90-04-26193-8

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 lat-
ter speaker quoted in English, with obvious disgust, the memorably unpleas-
ant 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)
Nancy Dorian - 978-90-04-26193-8

Occasionally, even without any code-switched quotations to signal the lan-

guage 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 voca-
tive 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 with-
out the required marking for vocative. This one failure to mark for the voca-
tive 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 interlocu-
tor. 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
Nancy Dorian - 978-90-04-26193-8

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 unre-
alistic 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 per-
sons 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].
Nancy Dorian - 978-90-04-26193-8

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,
Nancy Dorian - 978-90-04-26193-8

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 some-

thing 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 direct-
quote 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.
Nancy Dorian - 978-90-04-26193-8

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
Nancy Dorian - 978-90-04-26193-8

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 GER gerund suffix

COMP comparative marker INTERROG interrogative particle
COMPL complement particle NEG INTERROG negative interrogative particle
COND conditional suffix PL plural suffix
COP copula POSSE English possessive suffix
EMPH emphatic suffix PRET preterite particle
FUT future suffix 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 liter-
acy (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:
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.
Nancy Dorian - 978-90-04-26193-8

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 ordi-
nary 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 dis-
course 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 evi-
dential properties that create a distancing effect in discourse (largely in writ-
ten 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 subjunc-
tive 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 indi-
rect 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,

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 cer-
tainly 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 origi-
nal 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 imme-
diately 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 inten-
sification 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
Nancy Dorian - 978-90-04-26193-8

in praise of the particular voices of languages at risk 349
subjunctive as a grammatically marked category in English, commonly over-

look 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 introduc-
ing 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, acknowledg-
ing 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 vari-
ous properties that seem interestingly exotic from the point of view of English
or Spanish (and of English- or Spanish-speaking linguists), the relevant ques-
tion 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’emant-
Amharic 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 lan-
guages 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.
Nancy Dorian - 978-90-04-26193-8

350 chapter 19
Unless speakers of the high-infix language are lucky enough to be excep-

tionally 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 dis-
credit to them at all. Most speakers of any language have a hard time becom-
ing 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 fea-
tures and their use in the course of professional training, of course, but rela-
tively 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 lan-
guage spoken in the same region. Schooling is often available only in a higher-
prestige, 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 speak-
ers, 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 descen-
dants in three villages of coastal East Sutherland in Highland Scotland. ESG
Nancy Dorian - 978-90-04-26193-8

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 admin-
istrators 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 English-
speaking 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 Gaelic-
speaking 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 through-
out 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 gen-
eral 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 cele-
brated 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
Nancy Dorian - 978-90-04-26193-8

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 trou-
ble native speakers in the slightest but can seem an unnecessary and unkind
complication to learners. Despite the fact that consonant mutations are rela-
tively 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 supra-
segmentally – e.g. by pitch- and stress-contours and by timing – in English but
by consonant mutation as well as by suprasegmentals in ESG; and grammati-
cal 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 obligato-
rily 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 con-
centrate 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 high-
light 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
Nancy Dorian - 978-90-04-26193-8

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 suf-
fix 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 tradition-
ally take different forms according to person and number, just as the posses-
sive 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 post-
posed 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 phono-
logical 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, emphasiz-
ing 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
Nancy Dorian - 978-90-04-26193-8

354 chapter 19
rubric ‘emphatic suffix’ can be used for economy’s sake to cover the single pho-
nological outcome.
The emphatic suffix appears at highest frequency with the personal pro-
nouns and with the prepositions that conjugate for person. Among the latter
the pronominal forms of the preposition aig ‘at’ are most frequent, since conju-
gated 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 writ-
ten 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. pl. sg. pl.
1 mi ‘I’ sinn ‘we’ 1 mis sinneas

2 thu ‘you’ sibh ‘you’ 2 thus sibhs
3m. a ‘he’ aid ‘they’ 3m. éis éideas
f. i ‘she’ f. is
prepositional pronouns emphatic equivalents
sg. pl. sg. pl.
1 am* ‘at me’ aghainn ‘at us’ 1 amas ainneas

2 ad ‘at you’ agaibh ‘at you’ 2 adas agaibhs
3m. aig ‘at him’ ac ‘at them’ 3m. aigeas ac(a)s**
f. aic ‘at her’ f. 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-
Nancy Dorian - 978-90-04-26193-8

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: rachain-
neas air ais /raxĩn´əs er aš/ ‘I WOULD GO back!’ There are occasional occur-
rences 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 emphasiz-
ing 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.
Abbreviations in Transcriptions:
adv adverbial marker imper 3 singular imperative suffix

cond conditional suffix* pret preterite particle
emph emphatic suffix rel fut relative future suffix
fut future suffix depend 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.
Nancy Dorian - 978-90-04-26193-8

356 chapter 19
Lines in Transcriptions
1. orthographic rendering of East Sutherland Gaelic; words with emphatic
suffixes are capitalized and boldfaced
2. phonemic rendering
3. morpheme-by-morpheme gloss; a period within the gloss signifies a com-
plex morpheme the constituents of which can not be designated by seg-
menting the surface structure
4. 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.)
Nancy Dorian - 978-90-04-26193-8

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
Nancy Dorian - 978-90-04-26193-8

358 chapter 19
an – na h- ùibhlean aghmas air Di-Sathairn.”

ən – nəh u:l-ən am-əs er či so:rn
the – the.pl apple-pl at.me-emph on Saturday
the – my apples 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, ma bhithidheas dus

ax pai kɔli mə vi-s tu-s
but by golly if will.be-rel fut you-emph
‘But by golly, if you
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 bhumas a!

yo u-s vũm-əs a
will.get you-emph from.me-emph it
you’ll get it from me!’
Nancy Dorian - 978-90-04-26193-8

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 emphasiz-
ing 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 a’s an drill hall

s va khɔnsertʰ mo:r ãn´ as ə dril hal
and was concert big at.us in the drill hall
‘And we had a big concert 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____,
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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
Nancy Dorian - 978-90-04-26193-8

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’ do pheidheadh i.”

wəl an al šĩn´ tu tə fe:-u i
because not are we going to pay-gerund her
because we’re not going to pay 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 sto-
ries 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 arse-
nal 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
Nancy Dorian - 978-90-04-26193-8

particular story in which the woman appears as an unwelcome, intrusive pres-

ence, 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 pur-
sue the unfortunate woman across many of the clauses that follow her intro-
duction 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 quota-
tively. 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 quota-
tive 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 bold-
face 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 noun-
phrase 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
Nancy Dorian - 978-90-04-26193-8

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 objection-
able Gaelic coach) constant over a long stretch. None of these personal pro-
nouns 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 suf-
fixes 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 dis-
course 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 distinc-
tive 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 stereo-
typed 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 particu-
larly 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:
Nancy Dorian - 978-90-04-26193-8

5. focus-marking by a combination of clefting and emphatic-deictic suffix-

ation (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 supra-
segmentals 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 con-
tour in the way that English speakers do, but uses instead syntactic reor-
dering 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 appli-
cable 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 pur-
pose suggests. The frequent use that ordinary speakers make of these elements
is evident in the examples given above, but of course gifted creative writers
Nancy Dorian - 978-90-04-26193-8

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 par-
ticles appear in the Gaelic (bheòil-sa ‘your mouth’, latha-ne ‘our day’, and phòig-
sa ‘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 expres-
sion ’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
Nancy Dorian - 978-90-04-26193-8

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 environ-
ment surrounding the central figure, the an té-eas ‘that woman’ of the narra-
tive 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 emphatic-
deictic suffix plays a substantial and – for all my morpheme segmentations
and glosses – never fully translatable part. Sorley Maclean, practiced transla-
tor 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 lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

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 lan-
guages: Current issues and future prospects, 234–58. Cambridge: Cambridge
University Press.
Nancy Dorian - 978-90-04-26193-8

PART five
Fieldwork: Methods, Problems, Insights
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
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 ger-
mane 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 vari-
ety 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 ques-
tion 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

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 geographi-
cal 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 commu-
nity 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 french-
ified 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 contro-
versy 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
Nancy Dorian - 978-90-04-26193-8

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 inter-
pret 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 “infor-
mant 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 actu-
ally typical of the community’s normal speech use. The most fortunate inves-
tigator 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 flu-
ent 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 speak-
ers of Dyirbal, although her speakers also could not be perfectly ordered along
Nancy Dorian - 978-90-04-26193-8

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 imper-
fect 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 differen-
tially skilled speakers: older fluent speakers represented a relatively conser-
vative 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 subse-
quent 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 com-
promise 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
Nancy Dorian - 978-90-04-26193-8

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 stan-
dard 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 gener-
ally 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
Nancy Dorian - 978-90-04-26193-8

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 nega-
tive 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 unac-
companied 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 “cor-
ruption” 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 inves-
tigated in Guyana, East Indian speakers applied the label halluk ‘low’ to an
utterance in proportion to the number of elements in it which were identifi-
able as Bhojpuri, a vernacular speech form which has a partially diglossic rela-
tionship 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 lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

minority language is the normal daily speech of the parental and grand paren-
tal 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 foot-
ing 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 dia-
lect 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 iso-
lated 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 asper-
sions on each other’s language loyalty, suggesting that the other individual was
“too proud” to speak the native language, but false allegations of incompe-
tence 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
Nancy Dorian - 978-90-04-26193-8

378 chapter 20
since it qualifies them, and them alone, to perform certain special rites or ser-
vices. 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 situa-
tion 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 exam-
ple for Chitimacha, where he describes the bizarre performance of an elderly
woman who produced without a minute’s hesitation “a most remarkable mix-
ture 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
Nancy Dorian - 978-90-04-26193-8

who eagerly claimed speaker status and used Gaelic at every opportunity. The
disclaimer never voluntarily used Gaelic at all, with the exception of rare occa-
sions 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 asser-
tions 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 testi-
mony 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 lan-
guage. 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 obsolesc-

ing speech form surely lie in obtaining adequate speech samples from what-
ever 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 rap-
port. 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 tape-
recorded interviews ranging over conversation, storytelling, translation tasks,
Nancy Dorian - 978-90-04-26193-8

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 word-
list 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 ongo-
ing daily activities such as card-playing, dinner conversation, and housework
within eight households – sometimes even in her own absence, with the tape-
recorder left running. Whatever the effect of the tape-recorder, the relative
freedom of this style of recording (as opposed to direct interviewing) did pro-
duce the hoped-for wider range of style variation (op. cit.: 67). Subsequent vis-
its 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 stud-
ied Scottish Gaelic at all, but rather learned it in East Sutherland by linguistic
Nancy Dorian - 978-90-04-26193-8

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 suitabil-
ity 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 com-
munity were well established, was a crucial counterbalance to her native-
speaker 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 fish-
ing 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 occa-
sions and settings in which I was simply present and able to listen, or some-
times even make notes, without affecting the Gaelic conversation going on
Nancy Dorian - 978-90-04-26193-8

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, play-
ing 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 word-
translation 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 com-
mentary 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 delib-
erately larded with a variety of potentially interesting constructions by con-
vening 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 particu-
larly fluent and conservative member of the group and a good deal of allow-
able 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
Nancy Dorian - 978-90-04-26193-8

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 fieldwork-
ers, 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 suc-
cess 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 litera-
ture, 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 defi-
nitely, 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 daugh-
ter to disease only a few years before, and by their account I resembled the
deceased daughter. My friendship with this family became close and long-
standing, 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
Nancy Dorian - 978-90-04-26193-8

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 responsi-
bility 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 accep-
tance and eventual sponsorship in another key family, one of whom was an
East Sutherland Gaelic speaker, though of crofting rather than fisherfolk back-
ground; 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, anthro-
pologists, 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 sponta-
neously 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 provid-
ing on demand translations in Gaelic for English sentences, I found that per-
son 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 typi-
cally 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 lan-
guage 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
Nancy Dorian - 978-90-04-26193-8

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 devel-
oped 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 meet-
ing me, one or two people were quite alarmed at the prospect of being inter-
viewed 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 ser-
vices. 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 inter-
view. 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 obsta-
cle (although this is most definitely not the experience of all investigators of
obsolescing speech forms: Mohan [MS: 54–55] found translation-style elicita-
tion utterly counterproductive). I did not, however, attempt, as I did in East
Sutherland, to acquire control of the local dialect. Knowledge of standard
Nancy Dorian - 978-90-04-26193-8

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 remem-
bered 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 acqui-
sition 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 mem-
bers and family friends of my former student, so that I had quite a good under-
standing 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 par-
ticularly 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 com-
pleted the study without her active interest and cooperation, although several
of her children also took an active role in recruiting siblings or their own chil-
dren and in providing settings for interviewing.
Although the published literature on language death and language obso-
lescence 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
Nancy Dorian - 978-90-04-26193-8

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 “lan-
guage coaches” who would teach me their language, although I actually under-
took 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 participant-
observer 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 permit-
ted 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 communi-
ties 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
Nancy Dorian - 978-90-04-26193-8

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 nega-
tive 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.
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:
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 ele-
ment in fieldwork. Berkeley: University of California Press.
Nancy Dorian - 978-90-04-26193-8

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
Huffines, Marion Lois. 1980. Pennsylvania German: Maintenance and shift. International
Melchers, Gunnel. 1981. The Norn element in Shetland dialect today: A case of ‘never-
accepted’ 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 obsoles-
cence 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
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
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 theo-
ries of language acquisition? International Journal of American Linguistics 43:
333–38.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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 expe-
riences 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 field-
work setting. Entering an unfamiliar social world is guaranteed to plunge the
novice researcher into something like a second adolescence: a constant suc-
cession 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 dis-
sertation project. Students decided for themselves whether their interests and
values made fieldwork attractive, and if so, whether their personal circum-
stances 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

392 chapter 21
methods course was offered during the summer sessions, and a number of fac-
ulty 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 prob-
lem 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
Nancy Dorian - 978-90-04-26193-8

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 question-
naire 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 dia-
lects were otherwise poorly known, especially those of the northern and east-
ern 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 com-
munities 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 speak-
ers, 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 field-
workers 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 neces-
sarily tied to the Survey’s interests. My funding had been granted on the
Nancy Dorian - 978-90-04-26193-8

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 ten-
antry of the great Highland estates were summarily evicted, often with con-
spicuous 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 complexi-
ties 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 mat-
ter 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 dif-
ferent 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. “/lés/,” 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 vari-
able 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
Nancy Dorian - 978-90-04-26193-8

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 dif-
ferent 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 sup-
portive 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 net-
works 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 inter-
connected 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 them-
selves 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 undertak-
ings, was particular about its information sources. Speakers had to be strictly
local, preferably elderly, and not too geographically or socially mobile, since
Nancy Dorian - 978-90-04-26193-8

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 toler-
ated 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 Gaelic-
speaking 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 spe-
cialists 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 dis-
trict, 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 place-
name 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 con-
sonant clusters with a prominent bilabial second element (as in the Golspie
word /mwir/ ‘sea’) that were typical of the Gaelic of the fisherfolk communi-
ties. 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-
Nancy Dorian - 978-90-04-26193-8

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, com-
ing 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 deter-
mining 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 conserva-
tive 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 quan-
tities 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 phono-
logically 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
Nancy Dorian - 978-90-04-26193-8

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 disserta-
tion, 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 place-
name survey maps, my limited connections with the Linguistic Survey and the
School of Scottish Studies taught me very quickly how precious and how frag-
ile 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 curiosi-
ties 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-
Nancy Dorian - 978-90-04-26193-8

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 work-
ing hard on grammar, which again made elicitation (translation tests) the tech-
nique 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 sponta-
neous 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
Nancy Dorian - 978-90-04-26193-8

400 chapter 21
bearing on how general the use of [ᶕ l] for [ɫ] became (some younger speak-
ers 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 sam-
ple 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 progres-
sively weakening and that one traditional form of the passive was being aban-
doned (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 vari-
ety 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 strik-
ing 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.
Nancy Dorian - 978-90-04-26193-8

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 com-
parison 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 ini-
tial 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 gram-
mars 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 exis-
tence 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 descrip-
tivists 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
Nancy Dorian - 978-90-04-26193-8

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 variabil-
ity 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 cross-
village 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 distinc-
tive English), the Gaelic speakers of each village formed as clear and unambig-
uous 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 imper-
fect 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,
Nancy Dorian - 978-90-04-26193-8

the speakers I worked with had experienced virtually identical social and eco-
nomic 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 grammati-
cal 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 there-
fore subtract that form of variation, as well as the strongly age-related varia-
tion 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
Nancy Dorian - 978-90-04-26193-8

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 track-
ing, 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 data-
base 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 nine-
teenth-century evictions, in the fisherfolk case, with some degree of popula-
tion 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 for-
mer 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 socio-
economic 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 repre-
sents the only clear-cut case of such prominent but socially unmarked vari-
ability, 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
Nancy Dorian - 978-90-04-26193-8

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 commu-
nities 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 impend-
ing 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 environ-
ment. 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 appreci-
ate 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
Nancy Dorian - 978-90-04-26193-8

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 researcher-
and-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 pain-
ful, 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 long-
term 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 situa-
tion 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 neg-
ative, 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-
Nancy Dorian - 978-90-04-26193-8

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 experi-
ence and are often the engine of discovery, casting linguistic and cultural dif-
ferences 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.
Nancy Dorian - 978-90-04-26193-8

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.
Nancy Dorian - 978-90-04-26193-8

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 oth-
ers, may lie in the way of a researcher who appears at a late stage in the his-
tory 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, supply-
ing 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 con-
versation partners for many years, even though he himself appeared to repre-
sent 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 cer-
tain 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

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 sup-
plement or correct the record.
Impetus for receding-language fieldwork may come entirely from the out-
side, 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 devel-
oped 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 nar-
row 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 dis-
crepancies suggest the potential for conflicting objectives in any joint work
involving academic researchers and community members and for discor-
dant 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, endan-

gered-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
Nancy Dorian - 978-90-04-26193-8

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 stake-
holders 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 circum-
stances 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 commu-
nity. 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 atypi-
cal 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 rela-
tively 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 main-
tain 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 chil-
dren 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.
Nancy Dorian - 978-90-04-26193-8

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 sociolinguis-
tic paper on a locally sensitive subject, for example, using actual examples of
the phenomenon in question, in the belief that publication in a scholarly jour-
nal 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 oppor-
tunity 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 promi-
nent 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.
Nancy Dorian - 978-90-04-26193-8

of content entirely, thinking of the material on the tapes essentially as inter-

esting 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 unin-
hibited 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 listen-
ing, 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 field-
work 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 exclu-
sively 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 pri-
vate 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 informa-
tion, 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
Nancy Dorian - 978-90-04-26193-8

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 field-
worker 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 speak-
ers 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 expan-
sion 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 schol-
arly publications of any kind were just as unimaginable to most of the local-
language 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.
Nancy Dorian - 978-90-04-26193-8

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 intellectu-
ally 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 usu-
ally 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 dif-
ficult 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.
Nancy Dorian - 978-90-04-26193-8

416 chapter 22
function was plainly cathartic; many recordings resembled therapy ses-

sions 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 con-
cept 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 phonologi-
cal contrast? What does someone for whom language entails social interac-
tion 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 pro-
longed 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 out-
siders must be a source of confusion and disappointment to the groups under-
going 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 sta-
tus to those who accept the invitation (England 1992: 34; Grinevald 2003: 60).
Nancy Dorian - 978-90-04-26193-8

Many papers in scholarly journals already reflect this development.5 Another

way is for the researcher to provide useful professional products to the com-
munity: 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 knowl-
edge 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 let-
ters, phone calls, tape recordings, and gifts.8 In rare cases a linguist-become-
speaker 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 mem-
bers, and furthermore have a professional obligation to use their limited time
and their professional training for the particularly pressing jobs of documenta-
tion 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 com-
munities have come to take the same view, making service on the part of the
researcher a condition of access.
5 See for example Penfield et al. 2008.

6 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).
7 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).
8 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).
9 Victor Golla has proposed this and William Shipley has actually carried it out (Golla 2001;
Shipley 2000).
10 See Ladefoged 1992, Newman 2004, Matras 2005.
11 Himmelmann 2008 makes a very clear distinction between documentation and descrip-
tion; he makes a strong case for the view that documentation is much the more pressing
and the more productive of the two tasks.
Nancy Dorian - 978-90-04-26193-8

418 chapter 22
5 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 fisher-
folk. This was true not only among locals who were not members of the fish-
erfolk 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 fisher-
folk 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 homog-
enization 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 his-
torical 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 cer-
12 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 hap-
pened that information came directly from her own father.
Nancy Dorian - 978-90-04-26193-8

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 con-
flicting 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 sensi-
tivities and that the important thing was to put all of these invaluable record-
ings 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 prob-
lems 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 sen-
sitivity. 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 objection-
able material without realizing it, but the passage of time and the less socially
charged climate in present-day East Sutherland reduced the likelihood of seri-
ous offense.
Nancy Dorian - 978-90-04-26193-8

420 chapter 22
The solutions I’ve adopted in cases of potentially conflicting responsi-

bilities – 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 autobio-
graphical accounts (Foley 2003: 85–86). They offer excellent examples of spo-
ken 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 long-
distance 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 con-
versations, 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 dis-
crimination 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).
Nancy Dorian - 978-90-04-26193-8

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 roman-
ticization of the very lifeways that were once the focus of serious discrimina-
tion. 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 ordi-
nary 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.
Nancy Dorian - 978-90-04-26193-8

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 com-
munity empowerment from Melanesia. Language 84: 300–24.
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.
Nancy Dorian - 978-90-04-26193-8

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 postlit-
eracy: 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 com-
munity: 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 prac-
tices for North American indigenous language documentation. International
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
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage? 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.
Nancy Dorian - 978-90-04-26193-8

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 lan-
guages, 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 pub-
lic language use, acquisition, and transmission. Efforts to lessen such problems
and the prospects for lessening them further are discussed thereafter, with par-
ticular 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 documen-
tation, 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 dis-
course 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 acces-
sibility (Thieberger and Musgrave 2007).
One of several advantages in approaching fieldwork in this way is the rela-
tive 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

426 chapter 23
for some sort of written consent from the community in which the documen-
tation 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 proj-
ects 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 initia-
tive of the individual researcher than in response to professional expecta-
tions. 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 func-
tion 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 con-
tent some of which was quite unsuitable for general-access archiving; it would
have been equally unsuitable as the basis for printed materials that might pro-
mote 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
Nancy Dorian - 978-90-04-26193-8

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 fre-
quently 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, sensibili-
ties 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 less-
ens the frequency of sequestered, dyadic interaction between researcher and
source. In addition, the level of discussion about ethical issues of confidential-
ity 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 perma-
nently on record and to consult with communities on the matter. Not infre-
quently, 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.
Nancy Dorian - 978-90-04-26193-8

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 ade-
quate 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 will-
ing 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 care-
ful 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 read-
ing. 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 per-
suade 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 any-
thing 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-
Nancy Dorian - 978-90-04-26193-8

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 lan-
guage 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 invoca-
tions, 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 pre-
served, for the community’s own use and for the scientific record. But these
same languages are less well placed where revitalization is concerned, in par-
ticular 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
Nancy Dorian - 978-90-04-26193-8

430 chapter 23
language has been used mainly in home and neighborhood settings, transfer-
ring 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 lan-
guages 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 revi-
talization 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 fail-
ing to use it at home with their own children, promoting public sphere use but
neglecting private-sphere use (Hinton 2009). Parent-child transmission, usu-
ally 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 ineffec-
tive (Fishman 1991: 368–70). Schooling is very often the chosen locus of revi-
talization 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-
Nancy Dorian - 978-90-04-26193-8

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 speak-
ers 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 school-
taught Welsh responded to a matched guise test,1 they did not so much as rec-
ognize 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 inti-
mate and solidary settings, perhaps especially among the elderly, as in the case
discussed by King. Under these circumstances the local variety is likely to con-
tinue 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 lan-
guage 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 ren-
derings 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 con-
stant, 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, gen-
eral place of residence within Wales, and likelihood of having received Welsh-medium edu-
cation were evaluated; respondents also assessed the likelihood that the speaker might hold
certain specified jobs.
Nancy Dorian - 978-90-04-26193-8

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 famil-
iarity 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 gram-
mar 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 birth-
right members have access to the minority language. In some cultural con-
texts 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 lan-
guage 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.
Nancy Dorian - 978-90-04-26193-8

This issue arises somewhat counterintuitively in connection with adult

second-language acquisition of Scottish Gaelic. The number of Gaelic speak-
ers 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 governmen-
tal 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 long-
ago 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 compe-
tent 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 monolin-
guals 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 commu-
nity 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 off-
putting. (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 “private-
language” obstacles can be seen to coincide: not only may native speakers con-
sider 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
Nancy Dorian - 978-90-04-26193-8

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 recom-
mends avoiding dependence on schooling entirely in the Nahua communities
he has worked with in Mexico; Flores Farfán 2001: 191). In general, revitaliza-
tion 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 con-
tact 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 non-
Austronesian language with an elaborate noun-class system, a complex noun-
phrase concordance, and other morphologically challenging features, had
begun to show simplifications and losses in the version of the language spo-
ken 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 pre-
fixes 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 cir-
cumstances 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
Nancy Dorian - 978-90-04-26193-8

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 ances-
tral language an exclusive in-group privilege, some will no doubt find them-
selves 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 gram-
matical 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 Gaelic-
speaking 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 will-
ing native speakers are to accept a school-based language program and to wel-
come second-language speakers of their language.
In a more general sense, however, the attitudes of researchers may none-
theless 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
Nancy Dorian - 978-90-04-26193-8

436 chapter 23
about endangered languages, Dixon pointed to the potential value of docu-

mentation 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 sup-
port 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 previ-
ously unwritten, or very rarely written, enhance the standing of that language
(Terrill 2002).
Since negative messages about the value of the receding language are prom-
inent among the factors that bring transmission to an end, linguists’ affirm-
ing 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 chang-
ing urbanizing world” (Terrill, speaking of Lavukaleve, the language of the
Solomon Islands community for which she provided a storybook and a dic-
tionary, 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 technologi-
cally 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
Nancy Dorian - 978-90-04-26193-8

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 small-
language survival into a more mutually beneficial partnership by identify-
ing 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 revital-

ization 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 advis-
ability of seeking outside expertise, however, there are voices within field-
worker 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 commu-
nity in Bolivia; the hostility in both cases arose from the potential sources’
prior negative experiences with non-Indians. Wilson describes anthropologi-
cal 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).
Nancy Dorian - 978-90-04-26193-8

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, non-
transmission 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 com-
munity 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 popula-
tions, 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, illiter-
acy, or perceived minority-group distinctiveness, that stigma is likely to have
remained strong over the same period in proportion to any continuing dis-
tance from regional or national socioeconomic norms. It is not surprising that
some such communities have little inclination to regret the loss of their ances-
tral 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 transi-
tion 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 condi-
tions 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
Nancy Dorian - 978-90-04-26193-8

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 teach-
ing 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 reg-
ister 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 sud-
denly 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 acknowl-
edged to be used less than others – hence the terminology EBLUL, European
Nancy Dorian - 978-90-04-26193-8

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 cer-
tain 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 reced-
ing 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. profes-
sional 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 lan-
guage” (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 inad-
equate. 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 documen-
tations more limited than we imagine. Twenty years from now a new genera-
tion 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.
Nancy Dorian - 978-90-04-26193-8

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 documenta-
tion. 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 com-
munity 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:
Fishman, Joshua A. 1991. Reversing language shift. Clevedon: Multilingual Matters.
Nancy Dorian - 978-90-04-26193-8

442 chapter 23
Flores Farfán, José Antonio. 2001. Culture and language revitalization, maintenance,
and development in Mexico: The Nahua Alto Balsas communities. International
Florey, Margaret. 2008. Language activism and the “new linguistics”: Expanding oppor-
tunities 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 revitaliza-
tion 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 syn-
cretism 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.
Nancy Dorian - 978-90-04-26193-8

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 com-
munity: 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 orthographi-
cal 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
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.
Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8
Author Index
Abbi, Anvita 273n8 Bruton, John 278, 278n11, 281

Al-Ali, Mohammed N. 244 Burridge, Kate 244–245
Aikhenvald, Alexandra Y. 15, 17, 24–25,
179–185, 189–191, 298, 298n2, 305, 348, 367, Callaghan, Catherine A. 437, 441
400n1, 407 Campbell, John Lorne 148, 155
Al-Khatib, Mahmoud 244 Campbell, Lyle 16, 26
Allard, Réal 302, 305 Castile, George P. 223, 233
Amery, Rob 432, 441 Cawley, M. 233
Andersen, Roger W. 133 Chafe, Wallace 332, 346
Anonby, Stan J. 303, 305 Chapman, Antony J. 193, 200
Austerlitz, Robert 92 Chapman, Malcolm 433, 441
Austin, Peter K. 3, 25, 441 Chatsis, Annabelle 14, 24, 26
Ayoungman, Vivian 244 Childs, G. Tucker 22, 26
Cole, Debbie 14, 24, 26
Barber, Carroll G. 276, 281 Collins, James 17–19, 26, 409–410, 422, 432,
Bavin, Edith L. 178, 191 434–435, 441
Beniak, Édouard 115, 133–134 Committee on Irish Language Attitude
Berge, Anna 8, 21, 23, 25 Research 229, 233
Blench, Roger 291, 302, 306 Cooley, Ralph E. 381–383, 390
Blommaert, Jan 289, 306 Cooper, Robert L. 221, 233, 282, 328
Bloomfield, Leonard 8, 25, 66, 91, 138, 145, Costello, John 220
183–185, 190–191 Cotter, Colleen Marie 365, 367
Blum-Martinez, Rebecca 434, 443 Coulmas, Florian 247, 262, 271, 279, 281
Boeschoten, H. 345–346 Craig, Colette (see also Grinevald,
Bord na Gaeilge 229, 233 Colette) 27, 440, 442
Borgström, Carl Hj. 61, 65 Csató, Eva. A. 425, 441
Bourhis, Richard Y. 193, 200–201
Bradley, David 21, 25, 293, 295, 306, 410, 422 Dal Negro, Silvia 400n1, 407
Brandt, Elizabeth A. 244 Dauenhauer, Nora Marks 280, 287, 306, 437,
Brandl, Maria M. 19, 25 441
Breatnach, Risteard B. 62, 65 Dauenhauer, Richard 280, 287, 306, 437, 441
Breitbach, Joseph 347–348, 367 David, Nathan 441
Briggs, Charles L. 314n2, 328 Debenport, Erin 4, 17, 26, 439, 441
Briggs, Jean L. 407 Decime, Rita 273, 281
Broad, William 388 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.
Nancy Dorian - 978-90-04-26193-8

446 author index
Denison, Norman 174, 176, 207, 220, 377, 388 Haiman, John 293, 306
De Vos, George 193, 200, 231, 233 Haas, Mary R. 1, 12, 27, 137–138, 143, 145, 372,
Dixon, R. M. W. 20, 26, 373, 388, 436, 439, 441 375, 389, 409, 422
Dobrin, Lise 417n8, 422, 426, 436, 441 Hale, Ken 20, 27, 440, 442
Drapeau, Lynn 261–262 Hamp, Eric P. 92, 209, 221, 258, 260–262,
Dressler, Wolfgang 8, 26, 67, 90, 92, 311, 326, 392
328 Hansen, Magnus Pharao 234n1, 244–245
Eades, Diana 16, 26 Harper, Gilford 417n5, 423
Edwards, John 19, 26, 288–289, 289n1, 306 Haugen, Einar 259, 262
Ellis, Peter Berresford 210, 221, 233, 256, Heath, Shirley Brice 175–176, 265, 268n1,
262 278n12, 282
England, Nora C. 27, 440, 442 Henry, Reginald 372, 389
Evans, Nicholas 1, 16, 26, 295, 306 Heyd, Uriel 255, 262
Hill, Johnny Jr. 417n5, 423
Fennell, Desmond 228, 233 Hill, Jane H. 12–13, 27, 67, 92, 143, 144n7, 145,
Fenyvesi, Anna 115, 134 149, 155, 168, 254, 262, 372, 379, 381, 389
Fillmore, Lily Wong 114, 162, 165, 218, 221 Hill, Kenneth C. 67, 92, 254, 262, 372, 379,
Fishman, Joshua A. 65, 147, 158, 165, 194, 201, 381, 389
210, 221, 235, 244–245, 249n1, 251, 255, 258, Himmelmann, Nikolaus P. 20–21, 27, 411,
261–262, 280–281, 425, 430, 441 417n11, 422, 425, 440, 442
Flinn, Juliana 276–277, 281, 286, 306 Hindley, Reg 253, 262
Flores, Amelia 417n5, 423 Hinojosa, Maria de la Paz 174–175, 177, 208,
Flores Farfán, José Antonio 14, 27, 434, 442 220–221
Florey, Margaret 437, 439, 442 Hinton, Leanne 241, 244–245, 430, 439, 442
Foley, William A. 23, 27, 420, 422 Hohenthal, W. D. 193, 201, 208, 221, 276, 282
Foot, Hugh C. 193, 200 Holmer, Nils 39n9, 65
Householder, Fred W. 92
Gaeliconline 242, 245 Huffines, Marion Lois 174, 177, 258, 261–262,
Gahng, Tae-Joong 273, 283 287, 306, 377, 389
Gal, Susan 19, 27, 174, 176, 208, 221, 377, Huss, Leena 302, 302n3, 306
379–381, 388 Hymes, Dell H. 64–66, 90, 92, 160, 162, 165,
Gambhir, Surendra 376, 388 193n1, 201, 208–209, 221, 311, 328
Gardner, Robert C. 212, 218, 221
Gardner-Chloros, Penelope 345–346 Ibrahim, Muhammad 247, 262
Gauchat, Louis 63, 65 Irvine, Judith T, 19, 27
Gellner, Ernest 266–267, 282
Georges, Robert A. 384, 388 Jackson, Jean 193, 201
Giles, Howard 193, 200–201 Jocks, Christopher 409, 423
Glazer, Nathan 201 Jones, Mari C. 133–134, 178, 191, 431, 442
Goffman, Erving 397, 408 Jones, Michael O. 384, 388
Goddard, Ives 24, 27 Joseph, John Earl 268–269, 282
Golla, Victor 417n9, 422
Greene, David 65, 174, 176, 229, 233 Karttunen, Frances 208, 221
Grillo, Ralph D. 266–268, 282, 288, 306 Keane, M. 233
Grinevald, Colette 17, 24, 27, 410, 416, Khleif, Bud B. 282
421–422, 426, 437, 439–440, 442 King, Kendall A. 28, 231, 245, 431, 434, 442
Gumperz, John J. 156, 158, 165, 345–346 Kiparsky, Paul 38n7
Gunn, A. 197, 199–201 Kloss, Heinz 169, 177
Nancy Dorian - 978-90-04-26193-8

author index 447
Krauss, Michael E. 20, 66, 92, 137, 143–145, Menn, Lise 176, 311, 328
149, 155 Mertz, Elizabeth 147, 155, 174, 177
Kroskrity, Paul V. 17, 19, 27, 178, 191, 275, Melchers, Gunnel 379, 389
270n3, 277, 280, 282, 296, 298–300, 303, Messing, Jacqueline H. E. 430, 443
306, 409, 423 Miller, Elmer 22, 28
Kulick, Don 287, 439, 442 Miller, Wick R. 66, 92, 149, 155, 376, 383, 389
Kuryɫowicz, Jerzy 82, 92 Milroy, James (Jim) 18, 28
Kushner, Gilbert 223, 233 Milroy, Lesley 163, 395, 408, 415–416, 423,
Kuter, Lois 20, 28, 226, 233, 235, 245, 257, 426–428, 443
262, 266, 282 Mithun, Marianne 349, 367, 372, 389
Miyashita, Mizuki 14, 24, 26
Labov, William 63, 65, 101, 114, 144–145, 158, Mohan, Peggy 385, 389
163, 166, 311, 328 Moore, Robert E. 20, 28
Ladefoged, Peter 22, 26, 28, 417, 423, 437, Mougeon, Raymond 115, 133–134
442 Muehlmann, Shaylih 20, 28
Lambert, Wallace E. 91–92, 218–219, 221, Mühlhäusler, Peter 67, 92
302, 307 Muntzel, Martha 16, 26
Landry, Rodrigue 302, 305 Musgrave, Simon 425, 436, 443
Langgaard, Per 274, 282
Lee, Jennifer 249–251, 262 Nagy, Naomi 299, 307, 410, 423, 439, 443
Lehmann, Winfred 35, 65 Nathan, David 425, 441
Leonard, Wesley Y. 16, 28 New York Times 266, 282
Lewis, (E.) Glyn 29, 146, 155, 210, 221, 291 Newman, Paul 4, 28, 417n10, 423, 437, 443
Leyew, Zelealam 349, 367 The Northern Times 421, 423
Long, Michael H. 14, 28
Ó Baoill, Dónall P. 253, 260, 263, 435, 443
mac a’ Ghobhainn, Seumas 210, 221, 223, 233 Obler, Loraine K. 176, 311, 328
Macaulay, R. K. S. 336n5, 346 O’Callahan, Joseph 256–257, 263
MacAulay, Donald 366–367 Ó Cinneide, M. S. 229, 233
MacCaluim, Alasdair 433, 435, 442 Ó Ciosáin, Séamus 223n1, 227, 233
MacEachern, Scott 292–293, 307 Ó Cuív, Brian 239, 245
MacLaren, James 41n14, 65 Ó Dochartaigh, Cathair 2, 28, 402, 408–409,
MacKay, J. 201 423
Mackey, William F. 146, 155 Oftedal, Magne 39n9, 41n14, 65, 353, 368,
Maclean, Sorley 366–367 401, 408
Macnamara, John 210, 221, 227–228, 233 Ó hIfearnáin, Tadhg 243, 245
Maguire, Gabrielle 244–245, 259–260, 262 Ó Riagáin, Pádraig 238–239, 243, 246, 263
Malinowski, Bronislaw 397, 408 Ó Tuathaigh, Gearóid 239, 246
Margolin, David 224–225, 233
Masayesva Jeanne, LaVerne 27, 440, 442 Paulston, Christina Bratt 264n, 290, 307
Matras, Yaron 240, 245, 417n10, 423, Peacock, John Hunt 14, 29
437–440, 442 Peal, Elisabeth 302n3, 307
McCorkle, Thomas 193, 201, 208, 221, 282 Pecos, Regis 434, 443
McEwan-Fujita, Emily 14, 28 Penfield, Susan D. 417n5, 423
McGill, Stuart 3, 25 Pensalfini, Rob 400n1, 408
Mc Laughlin, Fiona 294–295, 307 Perley, Bernard C. 13–14, 29
McLendon, Sally 17, 28, 209, 221 Petersen, William 193, 201
Meillet, Antoine 273n8, 282 Polinsky, Maria 11, 29
Nancy Dorian - 978-90-04-26193-8

448 author index
Pride, J. B. 169, 177, 307 Thieberger, Nick 425, 436, 443

Pulte, William 174, 177, 377, 389 Thomas, Lewis 388–389
Thomason, Sarah G. 8–9, 29, 115, 134
Rankin, Robert L. 220–221 Thomson, Derick S. 211, 222
Ratliff, Martha 4, 28 Timm, Lenora A. 174, 177, 225, 233, 273n8,
Reinecke, John E. 89n9, 92 283, 372, 377, 389
Rhydwen, Mari 417, 423 Tollefson, James W. 236, 246
Richards, Jack C. 220–221 Trudgill, Peter 374–375, 389
Romaine, Suzanne 114, 240, 246, 345–346 Tschirch, Fritz 256n3, 263
Rumsey, Alan 270n3, 282 Tsitsipis, Lukas D. 15, 20, 29, 168, 177,
375–377, 379, 389
Sabino, Robin 280, 282, 346 Tsunoda, Tasaku 410, 424
Salisbury, Richard F. 275n9, 282, 295–296, Tucker, Benjamin V. 417n5, 423
307
Sall, Thierno Seydou 294, 307 Vago, Robert A. 19, 29, 272, 282
Salzman, Zdeněk 149, 155, 372, 389 Valenzuela, Pilar 235, 246
Samarin, William J. 88, 92 Van Maanen, John 397, 408
Sankoff, Gillian 286, 307 Vasquez, Nora 417n5, 423
Sasse, Hans-Jürgen 10, 15, 29, 115, 134 Verhoeven, L. 345–346
Saville-Troike, Muriel 277, 282 Verschueren, Jef 289, 306
Schmid, Monika S. 11, 29 Voegelin, Charles F. 178, 192, 373, 389
Schmidt, Annette 20, 29, 107n6, 114, 178, 192, Voegelin, Florence M. 178, 192, 373, 389
252, 263, 373–374, 377, 379, 381, 389
Schooling, Stephen J. 303, 307 Wade, Nicholas 388
Schumann, John 218, 221 Walker, Alastair 417, 424
Seliger, Herbert W. 19, 29, 272, 282 Walsh, Michael 19, 25, 244, 246
Serrato, Angelina 417n5, 423 Watahomigie, Lucille J. 27, 440, 442
Shipley, William (Bill) 16, 29, 417n7, 417n9,  Waterman, John T. 210, 222, 374, 390
423 Watson, Joseph 156n1, 166
Silva-Corvalán, Carmen 115, 133–134 Will, Vanessa Katharina Angela 14, 29
Silverstein, Michael 87n8, 146, 179, 192–193, Williams, Glyn 289, 307
270, 283 Wilkins, David 439, 443
Skutnabb-Kangas, Tove 302n3, 307 Wilson, Peter J. 437, 443
Smalley, William A. 279, 283, 301, 307 Wittenburg, Peter 425, 443
Smith, Jean R. 193, 200 Wolfson, Nessa 332, 346
Sorensen, A. P., Jr. 297, 307 Woodbury, Anthony (Tony) 21, 29, 349, 368,
Spolsky, Bernard 230–231, 233 440, 443
Stebbins, Tonya 14, 29 Woolard, Kathryn A. 14, 23, 29, 237, 283,
Steiner, George 348, 368 288, 307
Stenson, Nancy 220, 222 Wurm, Stephen A. 16, 29, 434, 443
Stern, Asher 277, 283 Wylie, Jonathan 224–225, 233
Stevick, Robert D. 63, 65
Swadesh, Morris 137, 143, 145, 372–373, 378, Yamamoto, Akira Y. 27, 440, 447
389 Yamane, Linda 236, 246
Yoder, Cecelia K. 381–382, 390
Taylor, Donald M. 193, 200–201 Yoors, Jan 383, 390
Tannen, Deborah 325n7, 328
Tedlock, Barbara 406, 408, 415, 423 Zengel, Marjorie S. 63, 65
Terrill, Angela 235, 246, 417n6, 423, 426, 436,
443
Nancy Dorian - 978-90-04-26193-8

General Index
abandonment 5, 155, 195, 264, 266, 272, age group 36, 44, 46, 48, 63, 68n2, 94,
300–301, 329 96, 178, 189, 213, 217, 219, 230, 299, 373,
Abbé Gregoire 267 401, 431n1
Aborigine, Aboriginals (see also language) age-based proficiency
16, 26, 250, 329 groupings 186–190
Aboriginal Australia 15, 286 age-related shift pattern 300
access 17, 22, 153, 235, 286–87, 290, 305, 397, age-related variation 403
404, 413–414, 417, 419, 426–427, 429, 432 age and speaker skills 178–192
accuracy 22, 96, 143, 373, 409 age mate 119, 230, 299
accessibility 231, 371, 425 agriculture 211, 273, 407
acquisition (dialect/language) 1, 9–10, 13–15, agricultural population 33, 36n2
20, 107, 149n2, 153, 170, 212, 235, 239, 288, agriculturalist 33, 195, 351, 396
292, 296, 300–301, 303, 376, 386, 425, 432, Akkad, Akkadian 271
433–35 Alabama 227, 277n10, 280
adult 14 Alabama Koasati 277n10
child 14, 372 Albania, Albanian 168, 209, 258, 260,
foreign-language 206, 212 374–375
second-language 212, 218, 433 alienation 247, 434
acquisition(al) history 10, 107, 213, 213n4, allomorph 82, 82n5, 83, 90, 219, 319, 353,
214, 219 400
activist, activism 235, 237–238, 241–242, 417, Alsatian, Alsatians 267, 288, 345
439–440 alternation (linguistic; see also vowel) 34,
adjective 39, 70, 98, 115, 140, 259, 269n2, 143
319n12, 355, 363 Amazonia 180, 185
possessive 217 Amazon Basin 284, 293
adjective comparison 188 America (see also Indian) 208, 301, 439, 440
adverb (see also particle, phrase) 23, 39, 49, North America 205, 208, 264
98, 118, 118n3, 118n5, 132, 315n4, 347, 403 South America 16, 264, 439
directional 117, 118n3, 119–121, 132, the Americas 432, 440
188–189 Amharic 349
locational 117, 119–121, 123, 129, 132, Amish 148, 244, 258
188–189 analog 73, 255, 406
adverbial marker 346, 355 analogy (see also change, leveling,
Ælfric 261 regularization) 38, 43–44
affiliation 146, 193, 292 analogical plural formation, analogically
Africa 210, 269, 286, 291–292, 294, 440 formed plurals 85, 214
age (as a variable; see also variation) analysis (as language structure; see also
age-and-proficiency continuum, grammar) 219
age-based proficiency continuum 6, analytic(al) construction 140–141
9, 16, 116, 118n5, 119, 122–123, 124–131 analytic structure 143, 190
* Boldface entries represent chapter titles and chapter section or subsection headings that
feature the subject of the index entry.
Nancy Dorian - 978-90-04-26193-8

450 general index
ancestor, ancestry (see also identity, first-language attrition 11, 133, 272

language) 17, 280, 331, 433 second-language attrition 133
ancestral community 438 Australia, Australians (see also
ancestral culture 232 Aborigine) 15–16, 19, 58n13, 107n6, 210,
ancestral identity 236 215–216, 236, 249, 286, 290, 295, 373, 379,
ancestral language 11, 13, 17, 148, 151, 410, 432, 439–440
178, 181, 199, 227, 236, 241–242, 248, Austro-Hungarian Empire 210
264–266, 269n2, 272–273, 275, 277n10, authenticity 253, 257
280, 285, 287, 299–301, 304–305, 311, authority (see also religion) 18, 182, 225,
329, 331, 350, 376–378, 410, 417, 435, 274–275, 287–288, 375, 382, 384
438, 439 Autonomous Aosta Valley Region 273
ancestral population 230 autonomy 238, 274–275, 284–285, 297
Andes (see Peru) awareness 5, 14, 18, 34, 98, 110, 154, 164, 180,
anomaly 98, 149 212n2, 226, 259, 270n5, 299, 405, 427
anthropologist 3, 13, 22, 270n4, 295, 298, 378, “limits of awareness” 179
384, 388, 406, 412, 415 Ayas Valley 273
apparent time 8, 373, 375 Äy̆iwo 434
aptitude 24, 212, 216 Aztec 206, 254, 264, 268n1
Arabic 207, 247–248, 252, 261–262
Arawakan 298, 400n1 backwardness 12, 149, 241
archaism 181, 184–185 Baniwa 181
archive, archiving 21–22, 419, 420, 425–427 Bará 193
Argentine Chaco 22 Basque, Basques 267, 288
Arizona Tewa 19, 27, 178, 275, 280, 296, Bauchi area (Nigeria) 291
298–301 behavior 24, 51, 56n32, 65, 83, 89, 95, 170,
Arvanitika 10, 258, 260, 374–377, 379 200, 234, 264, 277, 287, 301, 311, 382, 413,
aspect (linguistic) 103, 110 427, 435
progressive 70 language/linguistic behavior 12, 17, 34,
Assamese 284 170, 173, 175, 196, 199–200, 303, 318, 375,
assessment 95, 146, 179–185, 186, 188–190, 433, 440
218, 239, 302–303 respect behavior 276
assimilation (see also pressure) 64, 176, 301, speech behavior 96, 287
303, 438; (phonological) 115 Belfast 259, 395, 415, 426–427
assumption 17–20, 137–38, 143, 178, 206n1, beliefs 17, 162, 225, 270–272, 302–303, 316,
228, 269, 375, 404 412
Assynt 196 Berks County, Pennsylvania 10, 148–149,
attitude (see also change, purism) 146, 169, 149n2, 154, 154n5, 161n5, 170, 172, 230, 383,
205–206, 211–212, 216, 218, 225, 227–230, 386
254, 268, 275, 279, 295–297, 298n2, 301–302, Bhojpuri 376
380, 383, 387, 395, 420–421, 435, 439 bias 46n18, 95, 289, 383, 405–407, 418, 421
conservative 64, 252–253, 256 bilingualism 12, 19–20, 34, 40, 90, 153n4,
negative 20, 168, 175, 226, 229–230, 169, 198–199, 205, 207, 213n5, 218, 272, 275,
276–377, 387 275n9, 277, 284–307
positive 10, 216, 218 home-school 169
puristic 247–249, 252, 258 passive, receptive 213n4, 254, 287
social 158, 163 terminal 40, 61–64
Western 20, 266 bilingual 1–3, 5, 9, 33–34, 36n2, 69, 95, 103,
attrition 1, 146, 213, 215, 215n6, 219–220, 272, 116–117, 138, 148–149, 157, 159, 164–165,
431 170–171, 174, 185, 195, 197–199, 272, 285, 189,
Nancy Dorian - 978-90-04-26193-8

general index 451
311, 327, 329–346, 349–350, 380, 384–385, ceilidh (lexical item) 52, 55;
387, 402, 419 (institution) 163, 313
(fully) fluent 39, 64, 159, 239 Celtic 197, 217, 316n5, 327, 329, 342, 359,
(near-)passive 15, 90, 107n6, 154, 157–160, 361–362, 394–395, 402
162, 164, 170 census 117n2, 242, 299, 331n2, 379, 393
Biloxi 143, 372 Central Asia 235
birth order 10, 149n2 ceremony, ceremonials 168, 194, 208, 276
Boeotia 10 Chadic 291
Bolivia 271, 437 change (see also analogy, economics,
Bòrd na Gàidhlig (the Gaelic Board) 242 mutation, style, vowel)
borrowing 38, 109, 180, 250, 254–255, 261, analogical 82, 82n5
296, 318, 320, 348 attitudinal 421
negative borrowing 115–133 demographic 438
Brazil 194, 208, 298 grammatical 33–65, 67, 138, 187,
Brazilian Amazon 15, 400 399–401
Breton(s) 8–9, 67, 90, 147, 167, 226, 235, 257, internally motivated 115
266–267, 273n8, 288, 372, 377 morphological 73
Britain, Great Britain 4, 169, 175, 207, 215, morphophonemic 139–140
236, 269, 290, 384, 394 mutational 35–36, 40, 51, 61–62
British (language) 169, 238, 242 ongoing 7, 34, 39, 170, 399
broadcast, broadcasting, broadcaster 16, phonological 63, 399
168, 232, 239, 248, 255, 284, 313, 326, 365, quantity (phonology) 71–74, 77, 84, 88,
375 90
Brora 1, 2, 5–7, 9–10, 24, 33–34, 37n3, 37n5, social 438
38n8, 41n13, 42n15, 51n25, 52–53, 64, 68, 88, change in progress 5, 8, 35, 37, 39, 58, 63,
139, 142, 151, 187, 194–195, 197, 217, 316n5, 116–117, 123–124, 142, 186–187, 402
327, 329, 342, 359, 361–362, 394–395, 402 change process, process of change 13,
Bugotu 235 115, 133, 144, 190
Bulgar 265 direction of change 21, 63, 190, 374
Bureau of Indian Affairs 231 rate of change 63–64
Burma 293 Charmey 63
Cherokee 280, 373, 377
California 16, 137, 149, 168, 236, 410, 432, 434, Echota Cherokee 280
439 Chiapanec 16
Master-Apprentice Program 240, 244 Chicago, University of 392
calque 181–183, 226 childhood 10, 34, 107, 119, 131, 149–152, 178,
Cameroon 292 195, 217, 392, 402, 417–418, 420
Canada 167, 169, 208, 210, 237, 284, 412n2 child, children (see learner, speaker) 1–2,
Cape Breton 147, 167, 177 4–5, 9, 13, 38, 94, 112n8, 117, 138, 138n2, 139,
Carnegie, Andrew 227 149n2, 150, 152–153, 159, 170–171, 173–175,
the Carolines (Western Island group, 178, 180–182, 189n5, 195, 207, 209, 229–232,
Micronesia) 276 237, 240, 250–252, 257–258, 258n4,
case (grammar; see also dative, genitive, 259–260, 264, 266, 271, 273, 277n10, 287,
nominative, vocative) 291, 294, 297–299, 301–305, 323, 327, 329,
case distinction 112, 400 331, 337, 338n6, 376, 386, 302, 411, 414, 418,
case system 39, 49–61, 64 420, 428, 430–431, 433–435, 439
caste 175, 290 Chile 271
category: see grammar, morphology Chimbu 293
Cayuga 372 Chinese 210, 432
Nancy Dorian - 978-90-04-26193-8

452 general index
Chitimacha 372, 378 small(-language) 18, 187, 209, 254, 266,

Cholón 16 277–278, 280, 405
Chontal 268 speech 9, 12, 18, 29, 33, 35–36, 85, 114, 128,
church 36n2, 148, 251, 259, 312–313, 385, 404 156–165, 183, 186, 191, 205, 209, 216, 226,
Chuuk State 276 240, 275, 329, 338, 371–390, 402, 405,
Circassians 277 409, 417, 429
class (see social class) community language 11, 153, 191, 239,
closed class 124, 127, 255 345, 395, 438
classroom 11, 206, 211, 218, 226, 272, 304–305, community members 10–11, 14, 191, 205,
321, 430–432, 434 299, 411, 413–414, 416–417, 434
classifier 182, 348 compartmentalization (language) 275, 296
clause 7, 102–104, 106, 198, 108n7, 321, 324, compensation (language)
362–364 compensatory mechanisms 112–13
condition-contrary-to-fact 103 competence 22, 25, 162, 183, 291, 301, 376, 378
(prepositional) relative 100, 171 completeness (linguistic) 10, 18, 137, 409, 425
subordinate 103, 348 complexity (see also morphology) 12, 93, 106,
Clearances, Highland Clearances 207 272, 292, 440
clefting 364–365 comprehension 15, 107n6, 213n6, 218
clergy 266, 351 compromise 169, 176, 207, 224, 237, 247–261,
code 259, 275, 332, 380 374, 419
code choice 329–346 conditional (tense; see also inflection,
code-mixing 275, 373 suffix) 38, 104, 106, 110, 140–143, 171–73,
code-shifting 6 188–189, 214, 346, 355
code switch, code switching 96, 318, 330, confidence 276, 303, 421, 427
333, 335, 337, 341, 344–345 self-confidence 225, 230, 272–274, 279
codification 430–431, 434 confidentiality 427
coinage 254–256, 256n3, 261, 430 conjunction 18–23, 49, 101–102, 106, 108,
colonialism 301 108n7, 109, 110–112, 132n9, 140, 171, 224, 274,
colonial language 292, 297 319, 354–355, 363, 394, 403
Columbia 14 subordinating 214
communication (see also language of wider conquest (see also Normandy) 174, 207, 260,
communication) 87, 119, 137–138, 226, 284, 292
234, 260, 266–267, 303, 375 consent, informed 21, 414–424, 425–427,
communication network 63, 285 428–429, 437
communications 255, 297 conservatism (see also grammar, norm,
community (see also fishing, home, speech, structure) 38–39, 188, 217, 247–248, 251,
tradition) 253–255, 258, 277, 374
bilingual 3, 10–11, 149, 345 cultural 371
closed 432 linguistic 257, 296
ethnic 13–15, 169, 243, 410–411, 434 paradigmatic 174
exile 11, 216 consonant length 77–78, 86n7
fishing, fisherfolk 3, 6, 12, 151, 197, 199, consultants 18, 20, 183, 406
313, 344, 393–394, 396, 402, 404, 428 contact 133, 147–148, 151, 156, 229, 241, 261,
indigenous 13, 430, 434, 436, 440 265, 269n2, 284–286, 288, 290, 303, 312–313,
language 4, 11, 13–14, 20, 137–138, 144, 149, 373, 377, 403, 416, 434
178–192, 220, 230, 248, 254, 261, 274, 277, language contact 1, 19, 66, 91, 93, 115, 154,
287, 292–293, 304–305, 345, 395, 437 272
rural 19, 242 language contact phenomena 131, 254
Nancy Dorian - 978-90-04-26193-8

general index 453
language contact settings 115, 133 material 270

language contact situations 115, 154, traditional 181, 276
205–222, 252, 254 cultural content 179, 236
contact effects 6, 115 cultural context 290, 425, 429, 432, 440
contact with English 115, 241, 318 cultural disinheritance 232
context (see also culture) cultural institutions 230
discourse 179, 376 cultural knowledge 181–182
fieldwork 425–427 cultural legacy 243
revitalization 429–432 cultural material 17, 189
social 1, 244, 396, 399 cultural practices 178
continuum (see also age, proficiency) 82, cultural revitalization 244
84, 224, 250, 376 cultural settings 179
contraction 94, 113, 137, 167, 206, 228, 252 cultural superiority 295–296
control (see also grammar, phonology) cultural tradition 207, 209, 231
38n8, 67, 98, 109, 153–154, 162, 162n6, 163n8, cultural values 210, 285
171–174, 179, 195, 217–218, 265, 277, 286–287, Cupeño 143, 144n7, 168
292, 302, 385, 414, 437 currency 68, 146–151, 154–55, 247, 249, 287,
active, productive 100, 113, 161–162, 216, 300, 431
219, 312–313 language of wider currency (LWC), wider
political 444 currency language 148, 167, 226, 248,
receptive 158, 377 260, 300, 350, 436
social 14 curriculum 232, 239
convergence (see also grammar) 19, 66, 93, Cushitic 349
115, 258–260, 272 Czech, Czechoslovakia 210
conversation (see also interaction, speech
style) 5–7, 17, 69, 97, 128, 160, 160n3, 161, Dakar 294
164, 171, 180, 185, 242, 298n2, 315, 327, Dalmatian 36
329–330, 330n1, 332–335, 337–340, Danes, Danicization 274
343–345, 364, 376, 379–382, 409, 415, 418, Darwinism (social) 270, 272
420, 426–427 data 6, 8, 18, 23, 25, 35, 46n18, 58n33, 68, 75,
(tele)phone 116n1, 187n3, 330, 340, 343 89–91, 118, 129, 131, 138, 144–145, 171, 187,
conversation(al) partner 7, 12, 93, 95–96, 220, 289, 295, 314, 316, 371–390, 392, 396,
105, 112n8, 119, 139, 186n1, 187n2, 190, 315, 402, 410
329, 335, 338, 338n7, 343–344, 399, 409 elicited 119, 122, 126, 314
Coptic Egyptian 264 freely spoken (as opposed to elicited)
Cornish 36, 236, 256–257 119, 122, 126, 314
corpus 21, 24, 111, 128, 129n8, 145, 321n14, 328, database 6–7, 404
332, 363, 372, 440 dative 51–52, 53–59, 60–62
court 210, 267, 288, 292, 312 death (see language)
courtesy 380 bottom-to-top 16
courtesy rules 190 decay 17, 35, 67, 90, 105, 311, 326, 372, 400,
creole, creole language 1, 66, 90, 260, 287, 400n1, 404
376 decline 2, 5, 12–13, 34, 76–77, 80, 83–84, 94,
crofter (see also Gaelic) 195, 351, 396 105, 113, 141, 176, 197, 223, 227, 231, 252, 258,
culture (see also conservatism, tradition) 21, 291, 326, 433
209, 212, 230, 234, 242, 250, 280, 302, 351, deficiency (see also speaker) 96, 269n2,
406, 426 373
ancestral 232 Delaware 378, 381–382
Nancy Dorian - 978-90-04-26193-8

454 general index
delivery 319, 431n1 direct address 118, 122, 128, 129n8, 132, 341,

manner of delivery 137, 142n6, 157 352, 412n3
demography (see also change) 286 direction of change see change
demographic base 210–211 disappearance 12, 19, 61, 93–94, 115, 239, 241,
Dene 275n9, 295 275, 348, 416
Denmark 211, 224, 238 discourse (see also particle) 347, 352, 367,
Dewey 378 429
descendant 5, 14–15, 33–34, 36n2, 91, 116, context 179, 376, 425
156, 158–159, 186, 329, 336, 349–351, 394, device 269
418, 428 effects 349, 352, 364
description (linguistic) 1–2, 13, 17, 21–22, tone 269n2, 364
35–36, 64, 185, 352, 405, 417, 417n1, 437, 440 discrimination 69, 149, 420–421, 438
descriptive linguistics 8, 13 disfluencies 95, 190
descriptive linguists, descriptivists 10, 18, displacement 205, 301
179, 400–401 distance, distancing (see also social
deviation (see also grammar) 12, 37, 43, 53, distance) 36, 63, 252, 438
94, 105, 157, 178, 182, 186n1, 190, 374–375, emotional (self-)distance/distancing
435 217–218
device see grammar, morphology, syntax distancing effect 265, 347, 363
dialect 2, 7, 67–68, 74, 81, 85, 93–94, 98–99, distinctiveness (see also occupation) 2, 200,
117, 142n6, 154, 168, 198–200, 224, 226, 209, 353, 367, 438
268–269, 375, 392, 400n1, 411 ethnic 266, 420
dying 2, 33–65, 68, 90, 113, 149, 154n5, 196 linguistic 33, 199
Breton 90, 372 distinction (language-related; see also
English 171, 226 grammar) 8, 13, 60, 68, 81, 84, 104, 106,
French 267 112, 117, 123, 132, 188–89, 253, 294, 319n12,
German 148, 297, 230, 382 349, 400, 404, 417n11
Gaelic 36, 64, 67–68, 133, 147–148, 156, social distinction (see also status) 403
186, 226, 255, 315, 331, 335, 345, 350, 352, disuse 5, 93, 402
365, 393, 395–398, 401–402, 409, 411, 418 diversity 209, 266, 279, 289–290
local 7, 117, 147–148, 209, 227, 243, documentation 13, 16, 18, 20–25, 36, 199, 236,
385–386, 409, 411, 431, 434 409–443
mainstream 117, 138, 241, 353, 404, 418 Dom, Domari 240, 438
minority 225–226 domain (language-related) 19, 146, 288–89,
prestigious 154n5, 268 326, 352, 376
regional 154, 237, 254, 268, 312, 379, 437 domain-separation 176
unwritten 36n2, 113, 234, 317 dominance (see also language) 19, 68, 90, 95,
dialect geography 395, 401 265, 267, 276, 278, 286, 292, 301, 345, 429
dialect (in)tolerance 14, 226 dominant group 275–276, 272, 278n12,
dialect variation 226 285, 290, 300–301, 303–304
dialect(al) differences 183, 225, 253, 376 domination 207, 261, 286
dialectological features 4–6 Dornoch 156, 194, 335, 357
dialectological survey 4, 411 Dunrobin 156
dialectologist, dialectology 241, 393 Dupaningan Agta 421
dictionary 18, 22, 251, 411, 417, 417n6, 434, Dyirbal 128, 426, 430, 433
436, 439 Young People’s Dyirbal 435
diglossia 312
diglossic relationship 376 East Sepik Province 436
diminutive (see also suffix) 7, 320n13 Eastern Pomo 17, 28, 209, 221
Nancy Dorian - 978-90-04-26193-8

general index 455
Easter Ross 156 endearments 118, 132

Echota (see Cherokee) 280 endogamy 148, 159
economy 272–273, 278, 285 England 265, 289, 416, 440
cash 178 English 3–5
farming 148 American 123n6, 227, 229
economics (see also dominance, pressure) East Sutherland 68, 106, 327
economic base 148, 273, 438 Middle 392, 401
economic change 224, 438 nonstandard 101
economic conditions 207–208, 403 Old 392
economic development 227, 235, 239 Oxford 227, 229
economic factor 175, 235, 302 Standard 16, 68, 103, 376
economic growth 228 erasure 420
economic incentive 174 errors 38, 173, 220
economic opportunity 175, 299 ethics 416
economic success 272, 303 ethical dilemmas, problems 21, 411
Ecuador 271 ethical issues, questions 421, 427
Edinburgh 151, 217, 393, 396, 412 Ethiopia 349
University of 397 ethnicity (see also community, distinctive-
education (see also Gaelic, Welsh) 148, 164, ness, heritage, identity, language) 193–
197, 211, 238–239, 242–243, 254, 277, 284, 196, 199–200, 225, 236, 268, 279–280, 379,
298, 301–303, 312, 327 399, 438
bilingual 230–232, 436 ethnic group 16, 22, 156, 164, 200, 215,
primary 241 235, 237 243, 266, 304, 418–419
secondary 298 ethnic heritage 231
education(al) system 239, 255, 273, 429 ethnic history 230
elder (church) 312 ethnic marker 193–202
elicitation 5–9, 11, 18, 69, 129, 144, 171, 291, ethnic past 232, 378
381–383, 385–387, 398–99 ethnic population 17, 235–37, 242, 418
elimination 118, 289, 434 Europe, Europeans 169, 210, 216, 250, 266,
elite (see also urban) 148, 210, 238, 247–248, 268–270, 274, 278, 278n11, 278n12, 297, 300,
293 366, 432, 439
Embo 1, 2, 5–7, 9–11, 24, 26, 33–39, 41–45, European Community 279
49–54, 56–58, 63–64, 68, 88, 98, 102, 104, European Union 240
112n8, 116, 116n1, 117, 117n2, 118, 118n5, 119, evaluation 137, 143, 158, 163–164, 178–192,
123–124, 127, 129, 131–133, 138–139, 142–143, 237, 260, 275, 378
186–191, 194–195, 197–198, 258, 313, 319, 327, self-evaluation 216
334–335, 337–340, 342, 344, 354, 356–357, eviction 404
364, 394–395, 399–403 evictees 116, 351, 407
Emenyo 275n9, 295–297 evidential 181–183, 349
emigration 146, 238 evidential marker 271
emphasis see suffix evidential property 347
employment 239, 287, 298–299 exile (see also community) 133, 139, 139n4,
enclave 209, 257, 261, 291 144, 149, 149n2, 151, 152n3, 154, 159, 215,
enclaved community/people/population 215n6, 216–218, 371
9, 223, 371 exile/exiled (semi-)speaker 142, 152,
enclavement 275, 301 213–214
endangerment (language) 2–3, 12–17, 235, exploitation 416, 421
243, 270n4, 287, 440 exogamy (see also marriage) 293
endangerment crisis 20, 290 expansion (see also language) 61, 66, 118n3,
endangered language see language 197, 208, 210, 261, 325, 414
Nancy Dorian - 978-90-04-26193-8

456 general index
expert, expertise 13, 21–22, 182, 211, 244, 281, herring fishing/fishery 36n1, 317, 334
378, 388, 398, 437, 439 line fishing 402, 418, 428
expert knowledge 179, 297, 406, 417 fluency 10, 19, 96, 124, 137–138, 157, 162,
exposure (to language) 6, 197, 154, 170, 213, 171–172, 174, 179, 186, 186n1, 217, 236–237,
261, 298, 305, 383 234, 296, 409
expressivity, expressiveness 179, 272 focus marker, focus marking 352–367
expressive bleaching 350–352 formality (see also context) 168, 247, 325,
expressive capacity 270, 349–350 327
expressive power 362–367 formulaic expressions 162, 162n6
extinction (see also language) 33, 35–36, 82, fossil, fossilization (linguistic) 40n10, 93, 98
89, 91, 94, 139, 144, 144n7, 154n5, 167, 207, fossil(ized) expression, form, phrase 41,
209, 211, 213, 220, 270n4, 311–312, 326, 331, 51n24, 248
372, 377, 400, 402 Franco (Francisco) 273, 366
Eyak 66 Franco-Provençal 273
Franks 265
factions, factionalism 24, 256–257 Frederick the Great 210
factors (see grandmother) French 89n9, 169, 210, 237, 260–261,
demographic 302 267–268, 269n2, 270, 270n5, 279, 284, 288,
economic, socioeconomic 175, 299, 302 293–294, 303, 315n4, 345, 378, 432
motivating 277, 366 Cajun 167
Faetar, Faeto 299 Norman 261
family 167–177 Ontario 115
language 143, 298 Pidgin 89n9
nuclear 4, 149, 151, 154 Standard 89n9, 273
Faroese 224–225, 379 French Canadian 167
fate 20, 66–92, 104, 209, 223, 405 frequency (of occurrence, linguistic) 6,
Fatick 294 15–16, 39, 58–61, 63, 70, 73–74, 77–78,
feature see dialect, discourse, grammar, 85–86, 90, 97n1, 98, 101, 106–108, 111, 118,
morphology, salience 128–129, 132–133, 151, 157, 169, 171, 319, 352,
fieldwork, fieldworkers (see also 354, 366, 377
anthropology) 1–6, 8, 23–25, 33, 46n18, friendship 378, 383–384, 406, 415–416
137–138, 142, 144, 159n2, 175, 186–187, friendship networks 148, 159, 384
188n4, 240, 290, 294–295, 314–315, 371, Friulian 207, 377
373, 375–376, 383–384, 386–387, 391–393, Fulnio 194, 208
395–398, 404–407, 409–410, 415, 416–417, function (language) 6, 12, 91, 113, 146, 194,
418, 420–421, 425–429, 435–437 205, 209, 238, 312–314, 326, 353, 375, 433
late-stage 499, 410–414 functional restriction 167
long-term 25, 391, 406 functional segregation 169
team, team-based 21, 427 funding 22, 242, 392–393, 396, 410n1, 421,
telephone 23 415, 433, 435
Finnish, Finland 210, 238 future (tense) 140, 142–143, 172–173, 333, 336,
First Mesa 278, 301 346, 355, 358, 361
fishing (see also industry) 5, 12, 33, 116, 148,
150, 407, 411, 418, 420, 435 Gael 433
fishing community 3, 6, 12, 151, 197, 199, Gaelic (see also Irish)
313, 393–394, 402, 404, 428 Bible, Biblical 36, 95, 117, 313, 324, 335,
fishing village 1, 4–6, 11–12, 58n33, 116, 377
152, 156n1, 172, 241, 335 Common 2
Nancy Dorian - 978-90-04-26193-8

general index 457
“correct” 34, 225, 230, 321 second-ascending 87, 91, 150, 229

crofter 199, 396–397 third 189n5, 209
East Sutherland passim; as section younger 53, 64, 298n2, 331–332
header, 66–91, 116–124 generational seniority 181–183
(East Sutherland) fisherfolk 1, 4, 6–10, genitive 49–51, 51n24, 53–54, 56, 60–62,
12–13, 18, 23–24, 116–117, 133, 190, 196, 259
240, 394, 397, 412–413 genocide 168
fisher Gaelic 199 geography (see also variation) 2, 167, 201,
fisherfolk 116–117 235, 252, 267, 290, 292, 300, 348, 371, 376,
formal 313 385, 400n1, 403
local 1–2, 5, 7, 12, 15, 133, 148, 157–158, 172, dialect 395, 401
213n5, 241–242, 315, 323, 331, 338n7, 377, geographic isolation 33, 36, 93, 261, 277,
381, 384, 393, 397, 399, 420 205, 372
locally authentic 243, 253, 420 geolinguistics (see linguistics)
mainstream 117, 133, 241, 353, 404, 412, German 148, 148n2, 154, 167–169, 207–208,
418 210, 230, 256n3, 258, 260, 287, 288, 293, 347,
non-local 241, 312, 395, 418 348, 377, 380, 382, 385–386, 410n1, 432
Scottish 66–92, 93–94, 116, 147–148, Alemannic Old High 374
156, 167, 174, 195, 207, 211, 223–224, 226, East Middle 374
234n1, 241, 254–255, 269n2, 280, 287, New World 264, 271
304, 311, 318, 320n13, 329–346, 372, 375, Pennsylvania (Pennsylvania Dutch) 10,
377, 380, 386, 391–393, 398, 401, 411, 433 148, 148n1, 149, 153–54, 161n5, 167, 170,
Gaelic Board 242 172, 230, 258, 260, 287, 377, 382–383,
Gaelic League 238–239 385–386, 411n1
Gaelic-medium (primary) education Old World 206, 265
241, 435 Swiss 374
Gaelic-medium schools 241–242, 435 standard 148
Gaeltacht, Gaeltachtaí 228, 232, 239, 243, Germany (see also language) 210, 226, 289
253 gerund 40, 67–68, 69–74, 78–91, 99, 101, 123,
Gahuku 296 123n6, 125, 346, 400
Ge’ez 264 formation 9, 71–74, 78, 80–81, 82–88, 142
gender (language; see also grammar, Gimi 293
noun) 51, 51n25, 52–53, 60–61, 67, 70n3, Glamorgan County 146
85, 106, 112, 115, 259, 324n16, 348, 352, 355, globalization 285, 438
400 Golspie 1–2, 5–7, 24, 33–34, 37n3, 37n5,
generation 4, 35–36, 53, 87, 91, 146–148, 154, 39n8, 41n13, 42n15, 50, 51n25, 52–53, 63–64,
156, 180, 182–183, 189n5, 206, 209, 229, 232, 88, 104, 194–198, 315n3, 316n5, 317n5, 317n6,
238, 248, 252, 274, 277n10, 287, 290, 298, 317n7, 319–321, 323, 325, 327, 329–332,
298n2, 300, 311, 327, 411, 427, 432, 438, 440 340–341, 353, 358–359, 394–396, 402, 409
bilingual 39, 64, 249 government (see also nation, support) 146,
final, last 298, 311 148, 168, 211–212, 228–229, 235, 237–238,
first-ascending 87, 329 242, 279, 284, 297, 302, 304, 429
first-descending 182 government services 211, 232
future 60, 146, 440 grammar (see also change, deviation)
grandparental 377 analytic 49, 62
older, oldest 50, 87–88, 147, 180, 258, 311, synthetic 62
439 grammatical category 49, 49n20, 50, 62,
parental 150, 337 67, 105, 132, 142–143, 190, 226, 352
Nancy Dorian - 978-90-04-26193-8

458 general index
grammatical completeness 18 heritage language 10–11, 14, 21, 303, 350,

grammatical conservatism 186–187, 189 432
grammatical control 11, 113, 162 hierarchy (see also social) 67, 164, 175,
grammatical convergence 19 286–288, 293
grammatical device 118n5, 347–349, Highlands (Scottish; see also Scotland) 1, 33,
351–352 94, 97, 105, 116–117, 123, 123n6, 152, 156, 167,
grammatical deviation 157, 186n1, 435 169, 186, 194, 200, 207, 215, 224, 231–232,
grammatical distinction 8, 13, 319n12, 240–241, 280, 311, 313, 350, 364, 383, 392,
404 394, 397, 403, 406, 411, 433
grammatical features 4, 10, 115–116, 188, Hindi 247–248, 261, 376
241, 259, 348, 403 historical linguistics 36, 64, 138
grammatical form 18, 186, 190 historian 210, 232, 418
grammatical norm 64, 105, 157–158, 163, history (see also acquisition) 10, 17, 19, 64,
405 156, 167, 174–175, 217, 223, 226, 231, 235–236,
grammatical structure 117, 123–124, 241–242, 250, 252, 268–269, 272n7, 278,
128–129, 133, 187–188, 305, 351 278n12, 280, 284, 301, 317–318, 350, 372, 375,
grammatical system 35, 38–39, 49, 62 382, 387, 409, 418, 430, 434, 438
grammatical(ity) judgment 18, 294 oral 3, 413–415, 417–418, 418n12, 421, 428
grandchild 5, 181, 431 Hittite 271
grandfather 151, 181, 373 home (see also transmission) 9, 11, 88, 94,
grandmother, granny 87, 112n8, 150–152, 431 112n8, 139, 139n4, 152–154, 159, 161, 168, 170,
grandmother factor 87 178, 195, 195n2, 209, 214, 215n6, 232, 234,
grandparent 34, 87, 209, 234, 305, 383, 419, 237, 244, 252, 258n4, 259, 265, 270, 274, 276,
431 299, 305, 326, 334, 396, 403, 411, 413, 419,
Greece 10, 209, 258, 375 426, 429, 430–431, 433–435
Greek 168, 206–207, 209, 260, 271, 376, home community 139, 215, 215n6, 412
379 home language 20, 88, 149, 149n2, 168,
Greenland, Greenlanders 237, 274, 285 216, 219, 387, 430
Gros Ventres 372 home village 9–11, 15, 149n2, 213, 217,
group see age, dominance, ethnicity, 403, 418
indigenous group, kin, membership, homogeneity 119, 289, 404
minority, peer, private, resistance, social homogenization 418
group, stigma, subordination Hopi 275, 277–278, 296, 301
Guatemala 406, 415, 422 Horom 291
Guayqueries 194, 209 hostility 176, 196, 230, 235, 278, 434, 437
Guyana 376 household, householder 47, 94, 107, 112n8,
Guyanese Creole 376 138n2, 151–152, 159, 164, 172–174, 216, 312,
Guyanese Bhojpuri 376 332, 334, 380–381, 383, 396, 403, 412, 421,
427
Hamburg (Pennsylvania) 148, 153, 170 Hua 293, 295
Hamp, Eric 209, 259–261, 392 human rights (linguistic) 303
Hapiya 406, 415 Hungarian 19, 208, 210, 377, 379–381
Hawaii, Hawaiian 237, 304 American Hungarian 115
Hebrew 210, 223, 237, 256
Hebrides 232, 317, 375 Icelanders 270
hegemony 292 identity (see also ethnicity, nation) 12, 17, 174,
heterogeneity (linguistic) 267, 288 194, 199–200, 208–209, 218, 223, 231, 234,
heritage 22, 152, 230–232, 235, 237, 240, 243, 240, 242, 259–260, 273, 292, 295, 280, 300,
273, 280, 305 345, 380–381, 388, 415, 428, 432–433, 435
Nancy Dorian - 978-90-04-26193-8

general index 459
ancestral 236 Mexican Indian 174–175, 392

disfavored 264 South American Indian 194, 208
ethnic, ethnolinguistic 193, 195, 198–199, indigenous community/group/people (see
224, 240, 266, 303–304 also language) 22–23, 302, 304, 430, 434,
linguistic 180, 303–304 440
local, quasi-local 381, 386 indigenous-language survival 243
national 242, 266, 287 indigenous-language teacher 244, 430
ideology 4, 15, 17–20, 23, 181, 267, 270, 270n3, Indo-Iranian 438
279–280, 284–285, 288–290, 295–297, 301, Indonesia 437, 439
305 industrialization 146, 228–229, 266–267
European 288–289, 301 industry 302
language, linguistic 185, 191, 264–281, 291 coal 237
local 17, 179–181 fishing 1, 5, 33, 148, 402
puristic 19 herring 36n2
“ideology of contempt” 268–269, 272, publishing, broadcasting 225
280, 288, 301 tourist 273
idiom 197, 215n6, 253–254 infix 80, 348–350
idiosyncrasy 37, 51n25, 56n32, 58, 71, 78, 81, inflection (see also morphology) 8, 60, 66,
85, 105, 107, 394, 409 89n9, 184–185, 249, 253, 319, 320n13
idiosyncratic form, formation 73, 81, 85 1st-person-singular 189
idiosyncratic variation 409 conditional 140, 189
illiteracy 36, 63, 105, 113, 183, 327, 404 plural 320
illiterate society/speech community 138, synthetic 188–189
183 vocative 139
immersion (see also schooling) 304–305, informality 316, 326–327, 387, 393, 428
435 informant 35, 41n13, 43n16, 56n31, 58n33, 69,
experience 407 71, 90, 105n5, 137, 139n3–4, 144n8, 145, 163,
programs 304 196, 207, 375, 378, 383, 385, 405
immigration 208–209 in-group marker/membership/privilege
immigrants 133, 290, 302 168, 199, 435
immigrant language 205 Inka, Inka Empire 264–265, 278n12
imperative in-migration 146, 273, 420
negative 123–124, 124n7, 125, 128–129, innovation 19, 81, 85, 87, 126
131–132, 140, 188–189, 214, 217 innovative forms 21, 181
imperative plural 214 institution (see also culture) 148, 228, 230,
inalienables 41, 41n12, 217 284, 312, 411
inanimates 70n3, 118, 355 academic, educational 175, 267, 424
inauthenticity 237 regional 228
incentive (see also economics) 212, 290, 299 state, national 284
socioeconomic 290, 299 institutional support 148, 287, 302, 410n1
income 149 instruction 234, 239, 241, 251, 318, 332, 382,
low-income livelihood/occupation/subsis- 420, 434
tence mode 116, 241, 438 school instruction 434
incomers 167, 331, 333, 338n6 medium, materials of instruction 230,
Indian (see Bará, Eastern Pomo, Kwak’wala, 250, 259, 298
Maidu, Menomini, Rumsien Ohlone) 22, intactness 137, 142, 144, 186, 186n1, 258, 261
193, 209, 231, 235, 297–298, 300, 378 integrity 10, 258
American, North American 149, 176 structural 94
East Indian 376 intelligentsia 274
Nancy Dorian - 978-90-04-26193-8

460 general index
interaction (see also social interaction) 11, Italian 207, 209, 273, 299

36n2, 69, 88, 107–108, 151, 156, 159, 160–161, Italy 209, 226, 237, 377, 410
161n5, 162, 164, 171, 213n5, 247, 286, 294, Northern 237, 400n1
314n1, 316, 326–327, 330–331, 344, 359, Southern 299
377–378, 382, 404, 427, 433
conversational 128, 332, 338 Japanese 212, 432
cross-generational 159 Jerusalem 240, 438
sender/receiver 113 Jingulu 400n1
verbal 114n9, 160, 162, 327, 330, 344 job 211, 230, 232, 239, 251, 303
interaction networks 159, 161 joke 159, 213n5, 321n14
interference 219–220, 372 Juchitán 174–175
covert 115 juxtaposition 109–110, 112, 182
lexical 219
structural 219 Kachin 293, 295
interjection 321–323, 362 Kayor 294
interlocutor 7, 95, 132, 327, 329, 331–333, Kaurna 236
338–339, 341–342, 344–345, 384, 402 Kazakh 235
interlocutor factor 95 K’emant 349
intermarriage 275, 301 Keres 434
internet 414, 418 kin, kin(s)folk 88, 107–109, 112, 181, 187n2,
intervention 414, 428, 437 189, 214, 276, 299, 329
interventionist agenda 439 kin circle 15, 429
interview, interviewer, interviewing 2, 15, kin group 276
45, 46n18, 58n33, 69, 149, 161, 161n4, 196, kin(ship) network 7, 10–11, 107, 117, 159,
212n3, 255, 291, 314, 314n2, 316n7, 326, 328, 171–172, 303, 316, 331–332, 378, 384, 413
330n1, 355, 362, 379–381, 385–387, 420 kin(ship) tie 5, 119, 159, 181, 413
multispeaker interview 330, 401 kinship system 182
intimacy 413, 416, 426, 430, 433 kinship term 41, 181
intimate setting 431 kinsperson, kinswoman 154, 159, 171, 218
intolerance 19, 226, 240, 266, 278n12 Koasati 277n10
invariance 10, 18, 25 Komunku 295–296
investigation 17, 34, 51, 66–67, 100, 147, 311, Kulere 291
316, 371–373, 402 Kurux 273n8
Ireland 168, 193, 228, 231, 233, 239–240, 259, Kwaio 277
277, 278n11, 375 Kwak’wala 303, 305
Irish, Irish Gaelic 16, 62, 210, 223, 227–229,
231–232, 238–240, 242–243, 252–254, Labov, William 8, 63, 101, 163, 311
259–260, 271, 278, 321n13, 351, 364–365, 375 Labovian correlational sociolinguistics/
irregularity (see also verb) 38n8, 97, 111, studies 4, 399
140–143, 172, 187–189, 219 Ladin 237, 273–274
irregular noun plural 62, 140, 142, 214 lag (linguistic) 193–200
isogloss 37n5 language (see also ancestor, attrition, behav-
isolation (see also geography) 68, 96, 147, ior, colonialism, community, contact,
176, 243, 297 currency, endangerment, heritage, home,
linguistic 64 loss, maintenance, nation, norm, privacy,
social 93, 277, 285 revitalization, schooling, shift, structure,
Israel 237, 277, 283 support, trade, tradition)
Isthmus of Tehuantepec 172 Aboriginal 16, 194, 236, 295, 400n1
Nancy Dorian - 978-90-04-26193-8

general index 461
colonial 292, 297 second (see also acquisition, learner, skills,

creole 66, 90 speaker, teaching) 220, 251
disfavored 147, 155, 298, 429 secret 16
dominant, dominant-group 3, 6, 86, “sleeping” 16, 243
115–133, 167–177, 216, 220, 243, 260, small (see also community, population,
265, 272–273, 278–279, 290, 303–304, setting) 6, 20, 244, 255, 258, 261,
371, 429 264–283, 350, 429, 440
dying 2, 13, 35, 64, 66–68, 91, 137–138, standard, standardized 18, 25, 34, 36n2,
148–149, 210, 219, 231, 311, 326, 376 40, 40n11, 49n20, 50, 56, 60–61, 183, 247,
endangered 3, 8, 17, 19–21, 23–24, 280, 253, 268–270, 269n2, 272, 275, 278, 289,
284–307, 349–350, 409–410, 436, 439 313, 314n1, 353, 359, 374
ethnic 237, 240, 260, 266, 279, 304 state-promoted 168, 270, 287
expanding 1, 115, 178, 206, 234, 238, 265, stigmatized, stigmatizing 5, 240, 345, 412
311, 350 subordinate 93, 266, 268–269, 278n12
extinct 167, 281 superordinate 93, 379
first 1, 94, 117, 239, 252, 259, 301, 402 threatened 2, 17, 154, 167–168, 210, 220,
foreign 198, 206, 211–212, 216, 296, 348, 223, 225–226, 240, 249n1, 260–261, 264,
379 270–271, 274, 349, 375–377
Germanic 260, 392 traditional 249–252, 276, 287, 304, 436
habitual 112n8, 329 unwritten 1, 8, 24, 36n2, 113, 179, 224,
heritage 10–11, 14, 21, 235, 237, 240, 303, 317n9, 405, 436, 438
350, 432 written 17–18, 41n14, 117, 168, 224–226,
immigrant 167, 205 241, 252–253, 256, 271, 317n9, 347, 354,
indigenous 36, 115–134, 175, 210, 238, 364, 405, 409, 429, 431, 436
243–44, 249–250, 269–272, 279, 298n2, language choice 290
430–432, 440 language death 12–17, 19, 66–91, 93–114,
local 12, 15–16, 21, 144, 149, 168, 179, 207, 137–145, 154, 154n5, 167, 174, 219–220, 386,
238, 240, 276, 281, 350, 376, 384, 387, 388, 438
395, 409, 420, 430, 433, 436, 439 language endangerment 2, 3, 12–17, 235, 243,
low-prestige 149, 154–155 270n4, 287, 290, 440
majority(-group) 1, 115, 168, 287, 290, 302, language loyalty 10, 67, 139, 146–147, 149n2,
305, 434 175, 190, 206, 206n1, 207, 216, 376–377
minority 1, 4, 12, 23, 115, 168–69, 176, 205, language maintenance 4, 10, 14, 154,
220, 226–227, 236–237, 248, 252, 269, 205–222, 223–233, 234–246, 273, 275–277,
273n8, 278–279, 284–305, 377, 387, 290–292, 296, 299, 300–02, 303–304,
429–435 417n8, 436
native 16, 19–20, 95, 128, 220, 258, 272, language mixing 17, 180, 260
294–295, 377 language obsolescence 2, 4–6, 8–17, 91, 93,
Native American 236, 437 131, 257, 386, 400n1, 402
non-standard (see speech) language program 432, 434, 435
obsolescent, obsolescing 167–176, 376, language retreat 6, 12, 16, 223, 311–312, 314,
400n1 429
pidgin 66 language shift 7–8, 14, 115–133, 146–155, 178,
prestige 210 187n2, 234–237, 240, 276–277, 280, 287,
private(-sphere) 413, 431, 433 297, 303, 430, 433, 436–437
receding 8, 12–13, 19, 33, 178–191, 237, language suicide 174, 438
240, 243, 409–411, 410n1, 425, 436, language use 1, 3, 66, 158, 169, 299, 367, 425,
438– 440 431
Nancy Dorian - 978-90-04-26193-8

462 general index
language of wider communication linguistics (see also description,

(LWC) 21, 234–235 sociolinguistics) 349, 437
Latin 264, 392 descriptive 8
Latvia 278 historical 36, 64, 138
Lavukaleve 436 structural 411
law 82, 148, 231, 319 geolinguistics 146
learner 11, 16, 21, 107, 133, 158, 162, 241–242, Lisu 293
247, 256–257, 259–260, 304, 351–352, 366, literacy 63, 168, 224, 226, 248, 251, 255, 266,
407, 409 387, 404 430, 434
second language learner, learning 11, passive 36n2, 117, 314n1, 335
218–219, 237, 348, 433, 435 livelihood 33, 116, 230, 241, 418
foreign language 158, 162 loanword 73, 96, 146, 153, 196, 200, 220, 226,
legal system 231, 284 247, 249, 256n3, 258, 291, 318, 319, 319n12,
lenition (consonant mutation) 37, 37n3, 320, 320n13, 323, 325, 341, 372, 396, 403
39–40, 41n13, 42–44, 45n17, 49n20, 49–50, loanword avoidance, avoidance of loan-
52–53, 55–59, 61–62, 100–101, 106, 112, 118, words 23, 260
118n5, 128, 132, 319, 323, 351–352 London 11, 112n8, 151–152, 152n3, 153, 152n3,
obligatory-lenition adjuncts 39–40 159, 215n6
leveling 62, 88, 260 lore (see also proverb) 3, 181, 230, 236, 378,
analogical 34, 62, 140–144, 353 387
Lewis, Isle of 224 loss 94, 113, 142, 241, 272, 346, 350, 434
lexicon 15, 34, 38, 60, 63, 67, 110, 180, 207, of diversity 279
215n6, 243, 249, 254–256, 319, 324–325, 348, of land 265
412, 431–432, 434 of language 2–3, 167, 205–222, 224,
open-class lexicon 110 240–241, 244, 280, 291, 299, 301, 376,
lexical gaps 142n6 405, 435, 438–440
lexical interference 219 of speakers 3, 401
lexical items 4, 8, 15, 24, 105, 157, 247, of structural features 6, 12–13, 19, 38,
259, 291, 314, 362, 393, 395, 397, 409 49n20, 49–51, 54, 61, 62–64, 67, 77–78,
lexical recall 96 81, 85, 93–94, 98, 113, 115–116, 129, 133,
lexical verb 70, 99n2 137, 142–143, 146, 190, 219, 226, 272, 434
lexicostatistics, lexicostatistical test/(word-) of stylistic options/range 67, 327
list 69, 380–381 Louisiana 167, 169
lifeways 230, 297, 421 Lowlands (Scotland) 139, 213, 340
lingua franca 19, 287–288, 291–292, 296–297, loyalty (see also language loyalty) 156, 169,
300 215
linguist 3, 8, 14, 16–23, 63, 114, 186, 188–190, Luther, Martin 374
178–179, 191, 240, 252, 270n4, 272, 295, 298, Luiseño 143–144, 144n7, 168
314n2, 348–351, 366, 378, 384, 388, 391–392, LWC see currency
400n1, 405–406, 410, 436–437, 439–440
Americanist 440 Machair-Chat (Sutherland, Scotland) 196
Australianist 440 Maclean, Sorley 366–367
descriptive 10, 13, 400 Madagasgar 437
“parachute” linguist 21 Maidu 16, 417n7
“salvation” linguist 439 Maine 167
Linguistic Survey of Scotland, Survey of the maintenance see language
Gaelic Dialects of Scotland, Survey 2–4, Malagasy 437
8, 372, 386, 392–396, 398, 403, 405, 409 Malinowski, Bronislaw 397
Nancy Dorian - 978-90-04-26193-8

general index 463
Mandara Mountains 292–293 uniform-state 18

Mandarin Chinese 210 for analogical change/innovation/
Maori 237 restructuring 38, 82n5, 348
Marin Miwok 437 modernity 280–281
markedness 6 modernization 274, 297
marriage (see also intermarriage) 94, 148, Moen 276–277
293, 318, 426 Mohawk 349, 409
exogamous 180, 209, 293, 295, 297–298 monolingual(s) 9, 12, 69, 137, 138n1, 139, 156,
Massachusetts 236 158–159, 171, 187n2, 196, 198–199, 217, 274,
Master-Apprentice Program 240, 244 311, 315, 329–345, 384, 397, 433
media 20, 146, 148, 257, 284, 289 monolingualism 12, 19, 197–199, 285–286,
broadcast media 239 286–288, 289, 301
mass media 299, 302 monostylism 311
medium of instruction 239, 250, 259, 298 Montana Salish 9
membership 15, 91, 158–160, 162, 183, 193, Moray 116, 156
199, 218, 280, 413 morpheme 34, 61, 144n8, 182, 356–357
(in-)group 199, 231 morphology (see also norm,
Mennonites 244 morphophonology, morphosyntax) 34,
Menomini 8, 66, 138, 183–185 38, 49n20, 50, 61–62, 68, 74, 85, 87–91, 99,
merger 93, 105, 110–112, 190, 205, 353 107, 214–215, 319, 319n12, 412
Mexico 12, 16, 67, 175, 207, 254, 372, 379, 430, derivational 87, 90
434 inflectional 87, 319
Mexicano see Nahuatl suffixal 40n10, 85–86, 319n12
Miami 16 morphological category 88
Michigan 412 morphological complexity 66–92, 88, 93
University of 391, 401 morphological device 71–78, 83, 86
Micronesia 76, 286 morphological elements,
migration 146, 228, 273, 284, 371, 420 non-unique 270n5
in-migration 146, 273, 420 morphological feature 85, 180
out-migrants 371 morphological formation 67, 81, 84, 98
military service 312 morphological simplification 9, 88–89,
minister 197–198, 278, 313, 412 91, 93
minority (see also community, dialect, morphological structure 88, 96–99,
language, speaker) 1, 22, 240, 242 382
minority group 217, 284, 290, 301–302 morphological system 87, 253
minority peoples 284–285, 289–290, morphophonology 50, 60, 96–97, 107, 112,
297–300, 303 214–215
minority population 235, 284, 287, 302 morphosyntax 180
missionary 3, 22, 436 mother tongue 34, 105, 168–170, 195, 208
mistake 7, 9, 25, 34–35, 98, 102, 104, 119, 147, motivation 15, 22, 211–213, 219, 382, 393, 430
157, 160n3, 214–215, 327, 338n6, 407, 412 motivating factors 277, 366
Mixteco 392 multilingualism 19–20, 205, 272, 279,
mobility 273, 280 284–308, 440
mobilization 236, 248, 437 “irrational” 291
model 43, 46n18, 58, 73, 272, 275, 289, 305, receding 297–300
313, 314n1, 348, 433 multilinguals 19, 285, 372
puristic 296 multilingual peoples/populations 19–20,
traditional 252, 298 297
Nancy Dorian - 978-90-04-26193-8

464 general index
multilingual region/setting 20, 180, 193, native speaker see speaker

298 Navajo 230–231, 296
mutation 34, 36–38, 38n8, 39–43, 45n17, 49, negation 37–38, 103, 108n7
49n20, 51–53, 54n28, 61–62, 64, 101, 106, neighbor 20, 251, 187, 200, 226, 229, 235, 241,
108, 112, 132, 253, 319, 319n12, 320, 323, 352 244, 248, 259–260, 265, 275, 286, 292, 302,
consonant 40, 62, 100–101, 118n5, 323, 317, 401, 411, 413, 429, 430
351–352, 355 neighboring (ethnic) group/people 20,
final 99, 62, 71–81, 83–84, 86n9, 88, 90 200, 235, 285, 292
initial 34–35, 36–40, 43, 49, 62, 112, 122, neighboring language 248, 260, 265, 292
132n9, 143, 319n12, 321 neighborhood 151, 226, 229, 244, 259, 411,
change-resistant initial mutations 429–430
36–40 the Netherlands 211, 289
mutational change 35–36, 40, 49, 51, 62 networks (see also friendship, kin(ship),
mutational grammar 34, 40, 60, 63 language, marriage) 11, 63, 160, 214, 239,
mutational system 35–36, 38, 38n8, 285, 385–386, 395, 432
39–40, 48–49, 61, 63 interaction 107, 159, 161, 171
mutational variability 49, 51–52 social 148, 239, 307, 395
Mutsun 236 network processes 244, 280
New York City 63, 158, 163
Nahuatl (see also Mexicano) 12, 67, 206, 254, New Zealand 237, 412n2
265, 271, 372, 379, 430 newspaper 16, 148, 289, 420
names 86, 156, 180, 189, 193, 225, 268, 415 Ngas 291, 301–303
given 132, 189n5, 266 Nicaragua 17, 410, 421
personal 118, 128, 129n8, 341 Niger-Kordofanian 291
by-names 189, 193, 338n7, 412n3 Nigeria 291–292, 294
surname, family name 189n5, 225 Plateau State 291
narrative 23, 103, 160, 316, 316n6, 316n8, 317, nobility 266, 334
318–322, 322n7, 322n15, 323–27, 329–346, nominative 50, 50n21, 50n23, 51–56, 56n30,
353, 355, 359, 362–364, 367, 413, 420 57–62
narrative tone 367 norm, norming (see also grammar, rhetoric)
nasalization (consonant mutation) 37, 12, 18, 24, 34, 37, 53, 57–58, 63–64, 71, 73–74,
37n4, 38, 38n7, 41n13, 43–44, 50, 52–53, 77–78, 81, 87–88, 94, 105, 105n5, 138, 169,
55–60, 106, 112, 132n9, 143, 321 174, 198–199, 213–214, 252, 255, 262, 265,
Natchez 373 272, 286, 292, 300, 329, 345, 373–374, 402,
nation, nation state (see also government, 404
setting) 13–14, 19, 22, 168–169, 175, 212, 236, church-Gaelic 404
239, 261, 266, 269, 275, 277–79, 285–290, conservative 53, 58, 68, 71, 73–74, 78, 81,
301, 438 138, 188, 213–214, 259, 373–374, 402
national government 235, 237–238, 302, dominant-language 269n2
429 external, extra-community, nonlocal,
national identity 242, 266, 287 outside 18, 24, 224–225
national language 210–211, 220, 227, 236, fluent-speaker, older fluent speaker
247, 266, 268, 270, 279, 289–290, 297, 302 (o.f.s.) 77, 83, 87–88, 157–158, 258n4
nationalism 210, 266, 271, 277–279, 280, 289, grammatical 64, 157–158, 163, 405
301 linguistic 188, 199
nationality 211, 278 local 199, 224–225, 315, 374, 405, 433
Native American (used adjectivally) 168, narrative 345
230, 298, 406, 437 phonological 64, 158, 163, 405
Native California Network 240, 439 politeness 329–330
Nancy Dorian - 978-90-04-26193-8

general index 465
quotational 345 occupation 116, 164, 194, 241, 411, 413, 438

socioeconomic 438 occupational distinctiveness/
sociolinguistic 10, 158, 160, 162, 218–219, separateness 1, 33
258, 377, 382 Ofo 64
spoken 223, 255 Ojibwa 184, 185
standard-language 404 Oklahoma 378, 381
traditional 248, 435 opportunity (for language use) 9, 96, 139,
Normandy, Normans, Norman period/ 190, 213, 216, 265, 379
Conquest 265, 238, 259 economic 175, 299
Norman French 134 orthography 22, 113, 226–227, 256, 270n5,
Norn 379 317n9
North America (see also Indian) 208, 264, Osage 272n7
439 Otomanguean 16
North Queensland 107n6 Ottoman
North Sea 224 Empire 266, 278
Northern Ireland 259, 278, 415 Turkish, Turks 206–207, 300
Norway, Norwegian 301, 379 outsider 13, 16–17, 168, 179, 216, 224, 228–229,
noun (see also plural, verb) 7, 38, 40–42, 45, 258, 268n1, 281, 304, 313, 380, 385–387,
48–49, 49n20, 50–51, 51n24, 51n25, 53–54, 413–416, 417n8, 419, 428, 436, 439
54n28, 55, 56n30, 57, 60–62, 67, 67n1,
68–75, 77–78, 82–86, 86n7, 87–91, 97, 97n1, Papua New Guinea 284, 286–287, 293,
101, 118n4, 140–142, 157, 172, 214, 249, 255, 417n8, 436, 439
269n2, 294, 320, 320n13, 321, 325, 353, 355, paradigm 18, 40–41, 59
363, 367, 398 paradigmatic conservatism 179, 416
attributive 49, 60, 62 parallel (structural, social, attitudinal) 6,
feminine 51, 51n25 67, 86–87, 93, 100, 115–117, 123, 127, 288, 294,
inanimate 355 347, 363
masculine 7, 49–51, 53, 60–61, 321 parent 4, 5, 10, 34, 94, 112n8, 150–154, 170, 174,
noun class, noun classification, noun class 178, 232, 234, 240, 244, 252, 259, 264, 287,
system 51n25, 249, 294, 434 302–304, 383, 418–419, 430–431
noun plural 67–68, 69–78, 82–87, 88–91, parental generation 150, 377
140, 142, 172, 214 participant-observation 95, 159, 380
nominal gender 7, 11 particle 23, 106, 108, 108n7, 110, 132, 132n9,
nominal structure 324n16 346, 353, 355, 365
nominal system 38–39 adverbial 98
Nova Scotia 147, 167, 304 complement 346
Nsur 291, 301–303 discourse 23
number (grammatical) 43, 48–49, 62, 70, 106 emphasizing 353, 366–367
number, cardinal and ordinal 96–97 enclitic 353
number of speakers 5, 24, 68, 91, 93–94, 113, negative 123–124, 132
129, 217, 238–239, 242, 252, 276, 291 negative imperative 123
numbering of speakers/sources 43n16, 118 negative interrogative 346
preterite 124, 126, 129, 131–132, 132n9,
Oaxaca State 207–208 346, 355
Oberwart 19, 208, 379, 380 relative 106, 111–112, 346, 362
obsolescence 1, 4–6, 8–12, 12–17, 91, 131, 133, vocative 49n20, 118
191, 257, 386, 402 passive 39, 40–49, 62, 90, 100, 103–105,
obsolescence processes 400 140–142, 219, 324, 324n16, 325, 400
Occitanian 267, 288 bith-passive 42–43
Nancy Dorian - 978-90-04-26193-8

466 general index
dol-passive 43–48 verb 38, 70, 106, 118n3, 132

passive structure 46n18, 48, 219 phrasal switch 319–320, 323–325
past tense 103–104, 159, 321, 352 Pictish 156, 242
independent 118, 118n5 pidgin, pidgin language 66, 89n9, 91, 250,
negated, negative 123–124, 126, 129, 250n2, 296
131–132 Pidgin French (see French)
peer 51n25, 56n32, 187, 258n4, 349, 418n12 pidginization 66–67, 89–91
peer group 87, 170, 215 Pike, Kenneth 392
Pennsylvania 10, 115, 148, 161n5, 170, 230, place-names 179, 181, 396, 398
437 Plateau State see Nigeria
Berks County see Pennsylvania German pluperfect (verb form) 103–104, 106
McKeesport 115 plural (see also imperative, inflection, noun,
University of Pennsylvania 437 vocative) 9, 40n11, 41–43, 45n17, 50n22,
Pennsylvania German, Dutch 10, 148, 148n1, 62, 97, 97n1, 141–142, 219–220, 270n5,
149, 153–154, 161n5, 167, 169–170, 172, 230, 319–320, 320n13, 324, 346, 356, 394, 400
258, 260, 287, 377, 382–383, 385–386, 410n1 sibilant 90, 220, 320
performance 15–16, 20, 88–89, 93, 139, 141, zero 78, 86–86, 89–90
143–145, 147, 171, 185, 214, 216, 272, 313–314, plural allomorph 82, 319
332, 378 plural noun 67–68, 71, 73–74, 78, 82–91,
periphery (see also dialect) 117, 206, 261, 140–142, 214
393, 411 plural suffix, suffixal plural 82, 85–86,
periphrasis 70 346
Persian 206–207, 247 policy, policy-making 10, 148n1, 175, 209,
persistence 113, 132, 261, 277–278, 303 231–232, 239, 242–243, 247, 251, 267, 280,
personality 64, 96, 149, 152–54, 217, 367, 284, 288, 289n2, 430, 434
382–383 economic 243
Peru 271 language 175, 239, 243, 246, 280
Peruvian Andes 16 national 284
Amazonian Peru 235 social 243, 287
phone see telephone politeness (see also norm) 329–330
phonology (see also change, polysynthetic language 249
morphophonology) 2, 51, 62, 64, 67, 107, Pomattertitsch 400n1
146, 149, 162n8, 180, 186, 213, 236, 249, 258, pool (of speakers, sources) 4, 7, 137–138, 146,
372, 378, 395, 398 157, 399
phonological class 7, 51, 321 Popoluca 268
phonological inventory 13, 395 population (see also ancestor, enclave,
phonological structure 305 ethnicity, speaker, subordination,
phonological system 2, 393, 403 urbanization) 2, 5, 9, 12–13, 17, 19, 24,
phonological variation 158 33–34, 36, 63, 68, 91, 94–95, 116–117, 117n2,
phrase (see verb) 42, 50, 54n28, 61, 70, 85, 138, 149, 158, 167, 169, 174, 176, 185, 194–195,
97–98, 101, 102n4, 106, 218, 314, 319–321, 195n2, 197–200, 206, 208, 210, 219, 223–224,
325n17, 340, 348, 355, 363, 398, 434 228, 238–239, 242, 244, 254, 260, 265, 268,
adverbial 98 272–273, 275–276, 279, 284–285, 288, 291,
conjunctival 101–102, 106 293, 296–297, 299, 301–302, 304, 331, 331n2,
fixed 107, 377 350–351, 371–372, 375, 377–378, 383, 387,
fossilized 51 394–395, 402–403, 404, 407, 411, 418, 429,
noun 70, 85, 355, 363, 398 432–433
prepositional 42, 50–51, 54n28, 60–62 bilingual 36, 329, 377, 433
prepositional relative 106 fisherfolk 9, 116, 407
Nancy Dorian - 978-90-04-26193-8

general index 467
majority 19, 377 high 264, 378

minority 19–20, 235, 284, 287, 302 higher 154, 269
monolingual 12, 199 low 147, 149, 154–155, 264–265, 378
multilingual 19–20 negative 87, 148–149
rural 403 prestige transfer 268
school 259 preterite (see also particle) 324
small language 244, 279 primers 22, 250, 417
population base 223, 275, 350, 429 privacy 406, 419, 426, 427–429
population mixture 404 private sphere 311–328, 429, 431
population size 63, 228, 285, 287, 404 private sphere language 413, 431
positive reinforcement 105 private-sphere use 311–328, 430, 433
possession (grammatical; see also preposition, productivity 73
relative) 182, 217, 269, 275, 285, 354 proficiency (see also age) 3, 5, 7, 9–11, 15, 21,
possessive adjective 217 68, 68n2, 69, 84, 88–90, 93–94, 116, 118–119,
possessive pronoun 40–41, 42–43, 122–124, 127, 132, 138–139, 141–142, 157–161,
48–49, 62, 64, 324n16, 353, 355 161n5, 162, 164–165, 178–179, 183–184, 186,
possessive pronoun system 64 186n1, 188–191, 213n4, 216–219, 239, 243,
posterity 22, 440 329, 346, 373–374, 402
Potawatomi 184 proficiency continuum 9, 88, 94, 116,
poverty 116, 238, 327, 395, 407 118–119, 122–123, 127, 132, 139, 374
power 23, 146, 156, 168, 206–207, 264–265, proficiency/age-and-proficiency group,
267–268, 272, 276–278, 280–281, 284, 286, groupings 119, 186–188, 189–190
314n2, 348, 362–367, 371 proficiency level 11, 118, 188–190, 346
pragmatism 17, 207–208, 376 proficiency scale 157, 184
Prague 374 program (see also Master-Apprentice,
prayer, praying 17, 95, 216, 312–313, 429 revitalization) 231, 240, 244–245, 278,
precenting, precentor 312 282, 430, 432, 434–435
prefix, prefixation 182, 249, 348–350, 434 bilingual (education) 230, 250–251,
prejudice 196, 225, 418, 420–421 436
preposition (see also phrase) 41, 44–45, 50, immersion 304
53–54, 54n8, 55, 56n30, 56n31, 57– 58, 61, language-maintenance 223
101, 106, 110, 140, 217, 354 school(-based) 251, 432
complex 188–189 teacher-training 251, 430
compound 56 trilingual 273
conjugated/conjugating prepositions, pronoun 7, 67, 188–189, 269, 353, 359, 363
prepositional pronouns 140, 353–54, indefinite 363
363 personal 353–354, 364
double prepositional structure 101 possessive 40–41, 43, 48, 62, 70, 324n16,
preservation 53, 61–62, 86, 91, 124, 133, 187, 353, 355
260 possessive pronoun system 64
pressure 6, 115, 146, 169, 174, 226, 248–249, prepositional 353–354, 363
274, 276–277, 301, 385, 438 subject 48, 48n19, 49, 62, 217
assimilative 169, 176, 304 pronoun object 140, 214, 217
economic 279 pronoun reference 400
social 169, 276 pronoun replacement 7, 70, 70n3, 188
system 115, 133 pronominal forms 354, 362, 366
prestige 93, 146–147, 168, 205–206, 210, 241, pronominal system 39
253, 264, 267–269, 275, 286–288, 296, 350, pronunciation 50n22, 56n32, 181–182, 184,
378 226, 241, 256, 382, 431
Nancy Dorian - 978-90-04-26193-8

468 general index
prospects 3, 20, 238, 242–244, 264, 425, 442 regularization 132n9

small-language 264–283 analogical 219
prosperity 231, 272–274 regularity 132
proverb 54n28 relative (grammar) 106, 171, 173
proverbial lore 95 relative future suffix 355
psychologists 114 relative particle 106, 111–112, 346, 362
public sphere 313–314, 327, 429–432 prepositional (possessive) relative
Pueblo 17, 275, 296 100–101, 106, 166
Pueblo de Cochiti 434 relative (kinsperson) 158–159, 161, 218,
Pulaar 294–295 241, 324, 340, 378, 380, 383, 386, 395, 400,
Pulap, Pulap Islanders, Pulapese 276–277 418–419
purism (see also ideology) 247–263 reliability 372–373, 379
puristic attitudes 247–249, 252, 258 religion (see also register) 88, 168, 194, 209,
purity (linguistic) 196, 198, 258–259, 261, 277 291, 302, 311
religious authority 275, 327
Quapaw 220–221 religious ceremony/ceremonials/ritual
quasi-speaker 14 168, 194, 208, 276, 429
Quebec 237, 261 religious language 312–13, 327
Quechua 206, 265, 271, 278n12 religious life 22, 312, 327, 403
Quichua 431, 434 religious practice 264, 266, 296
questionnaire 1–2, 14, 216, 302, 393 religious register 327
quotation (see also verb) 129, 218 religious usage 312, 325
direct 329–346, 364, 414 relocation 298
remoteness 175, 291
race 24 repertory, repertoire 19, 36, 96, 108, 110, 210,
radio 15, 36n2, 105–168, 212n3, 232, 315n4, 338n6
326, 365, 375 repetition 111, 189n5, 324–325, 325n17, 332,
Radio Eireann 375 416
Rama 17, 410, 421 replacement (see also pronoun) 37n3, 37n4,
rarity 61–62, 93, 105–107, 372, 382 31, 56, 137, 256, 379
Reay Country 196 reported speech 7, 129, 187n2
reclamation 16, 236, 240 research
reconstruction (historical) 35, 138, 256 community-directed 14
recording (see also taping, video) 21, 315–16, field 391, 395
316n7, 355–359, 380–381, 383, 413, 416, 419, flawed 302
425–426 linguistic 179, 421
recording equipment/machinery 413, minority language 4
425 research agenda/goals/purposes/
recording session 316, 320, 326, 328, priorities 23, 58n33, 212, 405
330n1, 344, 415, 429 researcher 1, 4, 8, 13–14, 17, 19, 23–24, 105,
reduction 13, 63, 66–67, 74, 93–94, 113, 115, 169, 240, 291, 295, 301, 311, 345, 374, 384, 391,
137, 142–144, 181–182, 238 401–402, 405–407, 409–411, 414, 416–417,
reduced form 63, 137–138, 144 425–429, 435–437
reduced proficiency 178 endangered language 23, 410
reduced use 13, 138, 144, 159 linguistic 179
redundancy 42, 49, 62, 86 outside 14, 437
reduplication, partial 348 Western 20
register (see also religion) 14, 185, 305, 315n5, research priorities 23
347, 429 resettlement 407
Nancy Dorian - 978-90-04-26193-8

general index 469
residence 33n1, 124, 301, 316, 371, 380, 386, Sanskrit 247, 264

431n1 Sauris 207
residence patterns 4 Scandinavia 208, 237, 265, 398
residential areas 5, 195, 411, 413 scholar 18–19, 22, 174, 196, 231, 239, 302, 373,
resistance 5, 37, 58, 60, 91, 97, 147, 154, 174, 379, 384, 392, 405–407, 410–412, 414–417,
219, 235–237, 249, 256, 304, 307, 401, 438 419
resistant groups 277 scholarship 224, 231
resource 21, 107–108, 235, 257, 268, 273, school, schooling (see also bilingualism,
286–287, 292, 297, 303–304, 350, 363–364, instruction) 20, 44, 117, 148, 148n2, 168,
412, 417, 427 170, 174–75, 197, 206–207, 209, 213, 218,
responsibility 20, 228, 347, 398, 409–424 224, 227–228, 230–231, 239, 242–244, 247,
restoration 236, 238–239, 242–243 250–252, 259, 269n2, 277n10, 280, 287, 297,
restriction 252, 311, 327, 427 304, 350, 373, 403, 429–431, 433–434
functional 29 boarding 298–99, 301
retention 11, 90, 93, 106, 123, 129, 131–133, 140, English-language 148, 149n2, 277n10
147, 188–189, 194, 208, 211–212, 214–218, 229, primary 17, 172, 174, 259
244, 280 secondary 243, 277
retreat (language) 12, 16, 223, 429 immersion schooling 304–305, 435
revalorization 237, 242 school subject 230
revitalization (see also culture) 14, 16–17, school(-based) (language) program 146,
20, 191, 234–246, 247–263, 273, 304–305, 251, 273, 305, 432, 435
425–444 schoolchild(ren) 17, 225, 231–232, 239,
revitalization program 20 432
revival 14, 16, 223, 227, 230–231, 247–263, 436 schoolmate 96, 311
rhetoric 13 School of Scottish Studies 33, 396, 398,
rhetorical norms 259 412–13
rhymes 95, 189, 339n6 schoolteacher see teacher
ridicule 230 Schwyzertütsch 168
Rindre 291 Scotland (see also Linguistic Survey) 68,
rite, ritual 16, 378, 429 123n6, 137, 138, 147, 152, 168, 174–175, 197,
ritual use 16 213, 226, 231–232, 241, 258n4, 304, 329, 335,
ritualistic words 184 350, 391, 393, 398, 412, 419, 433, 435
Romance 207, 265, 392 Highland Scotland 1, 33, 116, 186, 200,
Romani 383 232, 350, 406
Rumsien Ohlone 236 Linguistic Survey of Scotland, Survey of
rural area/district 297, 396 the Gaelic Dialects of Scotland 372,
Russia, Russian 206, 210, 265, 293, 432 386, 392, 396, 409
Scots 116, 242, 255, 336n5
Sahara 286, 292 Scottish Gaelic see Gaelic
Salesians 298, 298n2 Scottish Highlands see Highlands, Highland
salience 144, 364 Scotland
salient feature 77 Scottish Parliament 242
Salina Cruz 174 scripture 312–313
Sami 304 Seereer 294–295
sample 2, 6–8, 35, 41n14, 69, 71, 73, 81, 96, segregation 94, 148, 169
108, 119, 132n9, 133, 142–143, 161, 186–187, self-confidence see confidence
189, 213n4, 214, 215n6, 217, 219, 371–372, 379, self-evaluation see evaluation
381, 395, 399–400, 440 semantics 64, 85, 382
sampling procedure 409 semantic collapse 104
Nancy Dorian - 978-90-04-26193-8

470 general index
semantic complexity 93 younger 119, 171

semantic distinctions 349 sibling set 4, 182
semantic extension 255 Sicily 265
semantic field 18 simplification (see also morphology) 9,
semantic range 41n12, 409 66–67, 88–91, 434
semi-speakers 9–11, 13–14, 38n8, 68–69, 87, simplicity 67, 88, 106, 132
94, 118n5, 119, 129, 133, 137–145, 146–155, singular 41–42, 50n22, 51n24, 69, 70n3,
157–160, 160n3, 161, 161n4, 161n5, 162, 162n6, 72, 78, 97, 97n1, 140–141, 189, 214, 321n13,
164–165, 186, 186n1, 187, 187n2, 188–190, 195, 324n16, 325, 353, 355
213n4, 260, 329, 374 Siouan 64, 220
low-proficiency 10, 15, 157–161, 161n5, 162, skills (language-related; see also speaker
164–165, 186n1 skills) 3, 11, 13, 21, 95, 105, 107, 107n6, 108,
high-proficiency 158–159 111, 135–192, 211, 219, 228, 232, 275, 348, 374,
short-burst 157–158 378–379, 382, 385, 387, 430
semi-speaker status 142 active 107–108, 161, 217
strengths of 107–108 foreign-language 211
Semitic 349 morphological 98
separateness (see also occupation) 260 passive 331
social 1, 199 productive 10, 98, 154, 159–160, 162, 165,
setting (see also culture, intimacy) 219, 329
fieldwork 391, 425 receptive 10, 107, 160–162
home 277, 430 multilingual 296
multilingual 193, 298, 307 second-language 433
nation-state 286–289 verbal 95, 190, 316
public 430 Skye, Isle of Skye 224, 232, 357
school 230, 430 Slovenia 236, 285
small-language 425 society (see also change, homogeneity,
small-village 405 norm) 148, 227, 279, 296, 303, 326, 418,
solidary 431 420, 438
sex (as a variable) 399, 403–404 social context 1, 244, 396, 399
shame 195, 230, 236, 339 social correlates 400
Shetlands, Shetland Islanders 379, 397 social distance 269n2
shift (see age, code, style, utility) 10, 22, 82, social group 24, 160, 163–164, 280, 345, 421
101, 109, 115–134, 156, 207–208, 234–237, social interaction 171, 316
239, 272, 276, 287, 290, 300–302, 304, 349, social intercourse 94, 311
350–352, 380 social liability 395
language shift 7–8, 14, 115–133, 144, social mobility 273
146–155, 174–175, 178, 187n2, 234–237, social organization 24, 266
240, 276–277, 280, 287, 297, 303, 430, social penalties 150, 287
433, 436–437 social pressure 169, 276
reversing language shift (RLS) 235, 430, social standing 199, 265, 271, 395
433 social strata 266, 270, 351
Shiwilu 235 social stratification 18
Shoshoni 67, 376 social structure 10, 209, 278
Siane 293, 295 social unnaturalness 416–418
sibling 4–5, 96, 150–151, 159, 170–173, 187, 215, socialization 91, 149–150, 298
386–387, 396, 403 socioeconomic condition, status 153, 208,
older 5, 149n2, 150, 170, 178 299, 403–404, 429
Nancy Dorian - 978-90-04-26193-8

general index 471
socioeconomic hierarchy/structure 175, 194, non-native 229

209, 290, 299, 438 older fluent (o.f.s., OFS) 6, 68, 68n2, 69,
sociolinguistics 23, 289 71, 73–84, 86–89, 94, 98, 100n3, 102, 104,
correlational 4, 8, 401 119, 121–126, 130–131, 187–190, 213, 258,
solidarity (see also setting) 52, 215, 345, 413, 374, 381
430, 433 second-language 432, 435
Solomon Islands 235, 277, 434, 436 short-burst 190
Somalia 169 terminal 9, 64, 143–144, 376, 379, 387
South America see America, Indian traditional 181, 400
South Asia 440 younger fluent (y.f.s., YFS) 6, 68, 68n2,
South Pacific 176, 235 69, 74–84, 86–88, 94, 100n3, 104,
South Tyrol 273 119–132, 187–190, 215, 217, 374
sources (linguistic) 4–5, 7–8, 11, 16, 18, 21–25, semi-speaker (SS) 9–13, 15, 38n8, 68–69,
101, 116n1, 118–119, 129, 137, 295, 315n3, 327, 87, 94, 106, 107–108, 109–113, 114n9,
336n5, 372–373, 381, 384, 387–388, 393–401, 118n5, 119–132,133, 137–145, 146–155,
403, 405–406, 410, 413–414, 418, 418n12, 419, 157–160, 160n3, 161, 161n4, 161n5, 162,
421, 426–428, 437 162n6, 164–165, 186, 186n1, 187, 187n2,
Soviet Union 207, 210, 235 188–190, 195, 213n4, 329, 374
Spanish 67, 169, 174–175, 208, 210, 254, 265, speaker base 238, 248, 261
270, 273n8, 296, 300, 348–349, 379–380, 432 speaker population 13, 167, 224–225,
Castilian 273 236–37, 248, 311, 378, 433, 438
Los Angeles 115 speaker skills 178–192
speaker (see also competence) speaker status 15, 142, 183, 190, 378–379,
age-atypical 178 432–433
child 9–10, 107, 213, 257 speaker typology 190
deficient 182–183 speech (see also community)
final 400n1, 420 conversational 334, 373
formerly fluent (FFS) 7, 9, 15, 119, good 17, 181, 183
123–124, 129, 130–131 inner 216, 379
fully fluent 1, 4, 9–12, 94–96, 99, 101, non-standard (see also English) 1, 266,
105–106, 118n5, 119, 129, 157, 159–160, 328
170–173, 186–187, 189–190, 213, 218, 239, quoted/reported 7, 129, 187n2, 330, 335,
242, 326, 329–330, 374, 412 341, 345
habitual 12, 434 spontaneous 96, 99, 383
imperfect 7–12, 14–15, 25, 107, 107n6, speech behavior 96, 287
139n4, 142, 147, 149, 154, 170–172, 186, speech form 1, 3, 5, 12–13, 15, 18, 20–21, 23,
190, 213, 213n4, 213n5, 214, 217–220, 345, 68, 93, 113, 133, 154n5, 168, 223, 237, 258,
373–374, 377, 399–400 260, 265–269, 271, 371–377, 379, 385,
local 2–3, 12–13, 117, 225, 241, 313–315, 317, 395, 400, 409–411, 416, 438–439
345, 350, 404, 411, 413–414, 426, 431, 439 speech habit 7, 403
marginal 165 speech island 2, 156, 224
minority-language 11, 226, 304, 430 speech performance 185
native 9, 20, 95, 105, 107, 114, 179, 184, speech style 133, 327, 373
205, 207, 227, 230, 236, 240, 242, 248, speech variety 129, 132, 158, 162, 269, 275,
253–255, 259, 293–295, 305, 348, 277, 354, 371, 380, 411, 418, 431
350–352, 366, 372, 377–381, 406, spouse 52, 88, 96, 229, 231, 396, 428
432–433, 435 Sprachbund 133
Nancy Dorian - 978-90-04-26193-8

472 general index
standing (see also social standing) 168, style (see also factors, speech) 23, 93, 247,
237–238, 242, 264, 272, 274, 284, 386, 421, 251–252, 314–316, 324, 326–327, 329, 332,
436 347, 367, 380
Statistical Account (of Scotland, of careful 321
Sutherlandshire) 197, 199, 207 casual 311, 323–324, 326
status (see also speaker) 15, 61, 119, 149, formal 168, 251, 325
153n4, 170, 174, 195, 223, 229, 237, 240–241, narrative 329
249n1, 267, 269, 271, 278, 284, 290, 302, 332, personal 96
337, 376, 380, 385, 416 style-shifting 382
official (language) 271, 284, 292 stylistic consistency 275
social 269, 287, 327 stylistic option 67, 140, 143
socioeconomic 153n4 stylistic range 314, 316, 326–27
status distinctions 164 stylistic variation 24, 311–328, 380
status factors 302 subjunctive 347–350, 355
status sets 278 subordination (language structure; see also
stereotypes, stereotyping 12, 144, 149, 164, clause) 217
232, 364 subordination (social) 260, 284, 288, 406
stereotyped feature 364 subordinate group/people/population
stigma, stigmatization 116, 149, 236, 241, 169, 265–266, 268–269, 285, 288, 290,
403, 438 300–301
stigmatized groups 5, 164, 215, 218, 230, subordinate language 93, 266, 268–269,
240, 345 278n12
stimulus sentences (see also test sentence) subsistence, sub-subsistence (economic
69, 71, 104, 111–112 mode) 195, 272, 297, 304, 351, 396,
story (see also tradition) 6, 23, 50, 58n33, 438–439
180, 183, 190, 230, 236, 251, 289, 314n2, 316, substitution (of one linguistic element for
318–326, 329, 332–333, 336n5, 339, 343, 350, another) 39, 62, 71, 97, 106, 109–110, 112,
359, 362–364, 382, 399, 420, 427–428 140, 143, 399
story collection 22 success (see also economics) 16, 23, 107,
storybook 417, 421, 426, 436 111, 113, 162, 174, 210, 217–218, 223, 229, 231,
storyteller, storytelling 181–183, 185, 190, 237–238, 244, 247, 255–256, 272, 274–275,
329–230, 330n1, 367, 379, 381 303–305, 373, 383, 385, 425, 435, 439
stratification 18, 268 suffix (see also diminutive, morphology,
structure (see also analysis, deviation, plural) 271n5, 320n13, 349–351, 400, 408
grammar, integrity, interference, lexicon, conditional 355
morphology, noun, passive, phonology, deictic 351, 353
society, surface structure, syntax) emphatic 269n2, 352–367
conservative 6, 19, 96, 101, 132, 188–189 emphatic-deictic 353–354, 359, 363–365,
embedded 171, 173 367
high-frequency 67 idiosyncratic 78
language, linguistic 1, 23, 163, 183, 271 subjunctive 355
matched 116, 123–124, 129–131 3rd-person 348
non-traditional 127 vocative 118n4
socioeconomic 194, 209 suffixation 71–73, 75, 77–86, 86n7, 88–90,
synthetic 143, 219 98–99, 102, 246, 404, 418
traditional 123–124, 127, 129, 131 simple 75, 80, 82–83, 86, 88–90
troublesome 96–105 Sumer, Sumerian 271
unmatched 116, 117–123, 124, 129–132 suppletion 41, 72, 74–75, 78–79, 86, 86n7,
verb(al) 40, 249 87–88, 324
verb complement 100–101, 106 suppletive past-tense forms 124, 568
Nancy Dorian - 978-90-04-26193-8

general index 473
support 20, 22, 131, 148, 162, 169, 235–239, Tasmania 137

241–252, 255, 273, 279–281, 304, 312, 392, teacher, school teacher 182, 225, 227,
417, 417n7, 417n8, 434, 436, 440 231–232, 244, 247–248, 251, 294, 304–305,
community 230, 432 321, 323, 340, 357, 415, 430–431, 435
financial 224, 238, 410n1 foreign-language 211
government(al), official 93, 148, 205, 231, indigenous language 244, 430
238, 429, 433, 438 second-language 14
institutional 148, 287, 302 teaching materials 256, 420, 439
language 235–239, 240–244, 417, 432 teacher-training program, teacher
suppression 232, 236, 238, 273, 286 trainees 251, 257, 430, 437
suprasegmentals 62, 325, 352, 364–365 teaching 15, 225, 230–231, 251–252, 257, 430
suprasegmental features 188, 132 second language 14
surface structure 61, 118, 124, 132, 356 teaching strategies 434
survey 4, 229, 267, 293, 398, 411 telephone (see also conversation) 3, 15,
single-interview 401 116n1, 187, 187n3, 255, 330, 340, 343, 404,
Survey see Linguistic Survey of Scotland 412, 417, 419–420
survival (see also indigenous) 2, 11, 14, 21, telephone fieldwork 23, 420
54, 64, 67–68, 169, 176, 223, 241, 243, 254, television 168, 239, 242, 384
257–258, 270, 273, 279, 281, 301, 303–304, tense 37–38, 100, 103–104, 108, 110, 126, 159,
376, 435, 437, 440 171, 173, 214, 217, 259, 321, 348, 352
Sutherland, East Sutherland passim independent past tense 118n5
Sweden, Swedish 210, 255, 301 negated/negative past tense 123–124,
switch see code, phrase 126, 129, 131–132
Switzerland 168–169 past tense 103–104, 118n5, 123–124, 126,
syncretism 67, 110–12, 254 129, 131–132, 159, 321, 352
syntax (see also morphosyntax) 67, 99, 107, present tense 348
184–85, 214, 249, 324, 412 tense formation 215
syntactic construction 85, 93 test, testing 11, 15, 52–53, 55, 55n29, 56,
syntactic device 36 56n31, 57, 59, 68n2, 69, 82n5, 90, 103–104,
syntactic environment 34–35, 37, 50, 60, 109–111, 115, 138–140, 142, 161–162, 164,
69–71, 105 171–173, 212–217, 219, 253, 255, 302, 326,
syntactic structure 12, 99, 382 380, 384
syntactic variation 163 comprehension 107n6
synthetic form (see also grammar, matched guise 431
polysynthetic language, structure) 140, read-back 11, 59
143, 188–190, 214, 219, 259 translation 69, 102, 108, 158, 187, 187n2,
system see case, grammar, kinship, 219, 399–400
morphology, morphophonology, test phrase, test sentence (see also
mutation, noun, phonology, possessive stimulus sentence) 7, 50, 52, 56,
pronoun, pressure, pronoun, verb, writing 68–69, 71, 74, 78, 82, 82n5, 83, 86, 88, 90,
102, 104, 109, 171, 385
Taiap 287 Texas 277n10
taping, tape recording 3, 23, 59, 116n1, 164, text 16, 41n14, 66, 227, 231, 249–251, 256–257,
314n2, 319, 321n14, 322n15, 330, 359, 415, 289, 316, 317n9, 318, 319n12, 364, 366,
417, 419 382–383, 399, 427–428, 431n1, 434
Tapshin 291, 301–302 textbook 68, 252
Tariana 15, 17, 179–180, 181–183, 187, 189, 191, Thailand, Thai kingdom 278, 300
298, 298n2, 300, 400n1 threat 2, 168, 254, 256, 265, 273, 278, 281,
younger Tariana speakers 15 290, 299
Nancy Dorian - 978-90-04-26193-8

474 general index
threatened language/speech form/speech parent-(to-)child 244, 430

community 17, 154, 168, 210, 220, 223, transmission bridge 237
225–226, 230, 240, 260–261, 270, 274, transmission failure 167–77, 402, 431,
281, 349, 375–377 435
tip (linguistic) 167–177 transmission process 241
Tiwi 249–252 transport, transportation 88, 297, 394
Tlaxcala 430 Trobriand Islanders 397
Tlingit 280, 287 Tsimihety 437
Toba 22 Tucano/Tukano 180–183, 297–298, 300
tolerance 14, 226, 254, 266, 277, 300, 303
Tolowa 17–18, 409–410, 416, 432, 434–435 Ugong 410
tone (see also discourse, narrative) 185, uniformity 18, 67, 158, 254
269n2, 318, 336n5, 348, 362–364, 367, social uniformity 391–408
431n1 United Kingdom 277
topic 8, 13, 66, 95, 311, 314n2, 316, 327–328, United States 207–212, 230, 237, 279, 290
330, 413, 419, 425, 430 Unst 379
tourism 273, 303 Urat 436
trade 285, 287 urbanization (see also variation) 148, 276,
trade language 296 436, 438
tradesmen 33, 351 urban boarding schools 299
tradition (see also culture, language, lore, urban centers 208
norm, society, speaker) 207, 66, 90, 184, urban communities 239
207, 209, 212, 224–225, 228, 231, 241, 252, urban elite 247
256–257, 260, 273, 276, 383, 405, 438 urban origin 431n1
cultural 207, 209, 231 urban population 293, 403
folk 23 urban-rural distinction 294, 431n1
written 224, 252 urban wage-laborers 438
traditional activities, practices 168, 178, usage (see also religion) 5–6, 8, 18, 36,
181, 189n5, 276, 378, 409, 434 38n38, 41, 44, 46–47, 53, 56, 58, 58n33,
traditional grammar 43–44, 60, 98, 63–64, 68n2, 73–75, 78, 83, 85–87, 90, 123,
253–254, 352–353 138, 156, 158, 163, 171, 181, 184, 188, 215n6,
traditional life, lifeways 230, 248, 297, 247, 250, 254–255, 261, 312, 315, 323, 325,
303 344–345, 353, 379, 396, 399, 401–403
traditional material 162n6 conservative 53, 58, 86, 88
traditional stories 6 contemporary, current 58, 353
transcription 21, 42n15, 317n9, 355–356, 382 variable 401–402
transfer 51n25, 83, 115, 117, 268, 351 use (see also private-sphere, syncretism)
translation (see also test) 6–7, 46n18, 53, active 20, 36, 98, 111, 157–58, 314n1, 377
68–69, 97n1, 99, 99n2, 100, 100n3, 101–102, increased 82, 105, 299
102n4, 103–104, 106–107, 139, 206–207, 139, reduced 13, 138, 144
159, 171, 187, 213n5, 214, 225, 250, 330, 345, usefulness 207, 235, 254, 273
355–356, 363, 366, 374, 382–384 user 44, 87, 107n6, 162, 162n6, 226, 239–240,
translation task 6–7, 11, 23, 99, 102–104, 248, 253, 270n3, 321n14, 433
112, 129, 187, 187n2, 379, 385 active 240
translator 7, 187n2, 336n5, 367, 385 nonuser 108
transmission 4, 13, 180, 229–230, 232, 244, utility 190, 205, 208–11, 235, 251, 254, 287, 296
250–251, 287, 311, 398, 417, 425, 430, 436 utility-based shift 234–37
home 242, 304–305, 435 Uto-Aztecan 430
non-transmission 376 Uzbek 235
Nancy Dorian - 978-90-04-26193-8

general index 475
validity 4, 20, 144, 189, 316, 347, 419 verbal enclitic 180

value 20–23, 36n2, 61, 66, 152, 179, 181–185, verbal noun 40, 42, 45, 50, 67n1
191, 207–208, 218, 223–233, 234, 236–237, verbal system 37–39, 61, 271
241, 254, 260, 271n6, 279, 289n1, 296, 304, Vermont 227
377–378, 385–386, 395, 419, 433, 436, 440 vernacular 223, 231, 248, 267, 270, 303, 376,
entertainment 326, 375 405
value(s) (see also culture) 20, 148, 163, video taping 22, 24, 417, 426
183–183, 183–185, 210, 276, 285, 299, 391, Vietnam 89n9
405 Vikings 265
Vancouver Island 303 village (passim; see also fishing, home)
the Vaupés 179, 185, 193, 201, 293, 297–298, village life 117, 405, 418
300 vitality (see also culture) 13, 93, 429
variable (linguistic) 24, 294, 404 ethnolinguistic 302–303
variability (see also mutation) 18, 24, 37, 39, vocabulary 8, 16, 63, 66–67, 96, 151, 184–85,
45, 49–52, 64, 98, 382, 391–408 219, 247, 275, 313, 318, 378–79, 382, 392,
variant 24, 37, 58, 60, 111, 115, 144n8, 163–164, 394
173, 401, 404 commonplace, high-frequency 107, 314
variation (see also age, dialect, phonology, core 69
register, style, syntax) 4, 24, 48, 64, 158, vocabulary retention 216
163–164, 254, 354, 382, 399, 403–404 vocative (see also particle) 49n20, 62,
geographical 4, 24 117–118, 118n4, 118n5, 122–123, 128–129,
idiosyncratic 409 129n8, 131–133, 139, 188–189, 341, 352
individual 4, 17–18, 24, 180, 216 vocative plural 139
personal pattern/personally patterned Voltaire 210
18, 23–24 vowel 24, 37n3, 42, 42n15, 48, 56n30, 60, 75,
regional 163–164 77–78, 181–182, 270n5, 317n9, 352–353
register 315n5 initial 98, 398
social 163 internal vowel change 60–61
style, stylistic 24, 311–328, 380 vowel alternation 71–77, 79–84, 86, 86n7,
urban 4 88, 90
Venezuela 194, 209 vowel shortening 75, 78–79, 81
verb (see also conditional, interaction,
performance, phrase, pluperfect, wage labor, wage-laborers 178, 351, 438
skill) 38–39, 40n10, 41–42, 67, 70, 73, Wales 146, 237, 273–274, 303, 431n1
82–83, 86, 89n9, 99n2, 100, 100n3, 101, 103, Wandala 292
106, 108, 111, 118n5, 123–124, 126, 129, 132, Welsh 237, 273–274, 289, 431
157, 249, 250n2, 259, 271n5, 319n12, 321, Welsh-medium education 273–274,
324–325, 347, 354, 401, 434 431n1
defective 363 Western Mono 17, 409
finite 41–42, 44, 46n18, 140–141, 348 Westernization 276
high(est) frequency 39, 111 Western language ideologies 264–83, 291
irregular 41–42, 123–124, 140, 142, 259 Winnebago 184
quotative 185, 321, 323, 363 Wisconsin 183
dependent verb form 101, 106, 355 Wittenberg 374
verb complement 99–101, 106 Wolof 294–295
verb root, verb stem 83n6, 123–124, workmates 311, 329
140–142, 172, 180 word order 38, 62, 140, 190, 320, 350
Nancy Dorian - 978-90-04-26193-8

476 general index
worth 268, 276, 376, 406, 436 Yiddish 438

writing (see also language) 36n2, 113, 117, 138, Yidiɲ 373
183, 224, 241, 251, 261, 345, 401 Youchigant 375
writing system 14, 113, 225, 237, 271, 439 Young People’s Dyirbal see Dyirbal
writings 19, 22, 113 Yugoslavia 236
written materials 271, 436 Yup’ik 349
written record 143
Zapotec, Isthmus Zapotec 174–175, 208
Yahi 137 Zuni 406, 415
Yap 245, 286
Nancy Dorian - 978-90-04-26193-8


(9789004261938 - Small-Language Fates and Prospects) Small-Language Fates and Prospects

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(9789004261938 - Small-Language Fates and Prospects) Small-Language Fates and Prospects

Uploaded by

Copyright:

Available Formats

Small-Language Fates and Prospects

Nancy Dorian - 978-90-04-26193-8

Alexandra Y. Aikhenvald (Cairns Institute, James Cook University)

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

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

Library of Congress Cataloging-in-Publication Data

Copyright 2014 by Koninklijke Brill nv, Leiden, The Netherlands.

This book is printed on acid-free paper.

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

1 Grammatical Change in a Dying Dialect (1973) 33

5 The Problem of the Semi-Speaker in Language Death (1977) 137

Nancy Dorian - 978-90-04-26193-8

11 Language Loss and Maintenance in Language Contact Situations

17 Stylistic Variation in a Language Restricted to Private-Sphere Use

20 Gathering Language Data in Terminal Speech Communities

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

Chapter 1, ‘Grammatical Change in a Dying Dialect’ was published in Language 49,

Chapter 2, ‘The Fate of Morphological Complexity in Scottish Gaelic Language Death:

Chapter 4, ‘Negative Borrowing in an Indigenous Language Shift to the Dominant

Chapter 6, ‘Language Shift in Community and Individual: The Phenomenon of the

Chapter 8, ‘Abrupt Transmission Failure in Obsolescing Languages: How Sudden the

Nancy Dorian - 978-90-04-26193-8

Chapter 13, ‘The Ambiguous Arithmetic of Language Maintenance and Revitalization’

Chapter 15, ‘Western Language Ideologies and Small-Language Prospects’ was

Chapter 17, ‘Stylistic Variation in a Language Restricted to Private-Sphere Use’ was

Nancy Dorian - 978-90-04-26193-8

Chapter 21, ‘Surprises in Sutherland: Linguistic Variability Amidst Social Uniformity’

Chapter 22, ‘Documentation and Responsibility’ was published in Language &

Nancy Dorian - 978-90-04-26193-8

In American descriptive linguistic practice during the first decades after

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004261938_�02 Nancy Dorian - 978-90-04-26193-8

by peripatetic fieldworkers from the Gaelic division of the Linguistic Survey

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

and accounting for the apparently long-standing prominence of individual

2 From Focus on Dialectological Features to Focus on Features of

Because of my original link to a dialectological survey, documenting geo-

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

continuum be established, with older speakers using relatively conservative

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

4 The Novel Inclusiveness of Language Obsolescence Work

Established field practice in descriptive linguistics favored a certain kind of

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

Once information about the language skills of imperfect minority-language

Nancy Dorian - 978-90-04-26193-8

While identifiably imperfect speakers have a special place in the study of

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

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

to stress the exceptional resilience of these speakers in resisting the language

Nancy Dorian - 978-90-04-26193-8

a reliable gauge of Gaelic speaking ability. One low-proficiency semi-speaker

Nancy Dorian - 978-90-04-26193-8

Nancy Dorian - 978-90-04-26193-8

Reviewing the complex history of Rama revitalization in Nicaragua, Grinevald

© koninklijke brill nv, leiden, ��4 | doi ��.��63/9789004261938_�02 Nancy Dorian - 978-90-04-26193-8

to have on long-term dialect fortunes if each of a number of settlements pre

© koninklijke brill nv, leiden, ��4 | doi ��.��63/9789004261938_�03 Nancy Dorian - 978-90-04-26193-8